ICES WGBAST REPORT Report of the Baltic Salmon and Trout Assessment Working Group (WGBAST) 27 March 4 April 2017.

Transcription

1 ICES WGBAST REPORT 2017 ICES ADVISORY COMMITTEE ICES CM 2017/ACOM:10 Report of the Baltic Salmon and Trout Assessment Working Group (WGBAST) 27 March 4 April 2017 Gdańsk, Poland

2 International Council for the Exploration of the Sea Conseil International pour l Exploration de la Mer H. C. Andersens Boulevard DK-1553 Copenhagen V Denmark Telephone (+45) Telefax (+45) info@ices.dk Recommended format for purposes of citation: ICES Report of the Baltic Salmon and Trout Assessment Working Group (WGBAST), 27 March 4 April 2017, Gdańsk, Poland. ICES CM 2017/ACOM: pp. For permission to reproduce material from this publication, please apply to the General Secretary. The document is a report of an Expert Group under the auspices of the International Council for the Exploration of the Sea and does not necessarily represent the views of the Council International Council for the Exploration of the Sea

3 ICES WGBAST REPORT 2017 i Contents Executive Summary Introduction Terms of Reference Participants Ecosystem considerations Salmon and sea trout in the Baltic ecosystem Ecosystem impacts of fisheries and mixed fisheries overview Data for HELCOM salmon core indicator Response to last year s Technical Minutes Salmon fisheries Description of gears used in salmon fisheries Recreation trolling fishery Assessing recreational salmon fisheries Catches Discards, unreporting and misreporting of catches Discards Fishing effort DCF stock sampling of salmon Tagging data in the Baltic salmon stock assessment Finclipping Estimates of stock and stock group proportions in the Baltic salmon catches based on DNA microsatellite and freshwater age information Management measures influencing the salmon fishery Effects of management measures Other factors influencing the salmon fishery River data on salmon populations Wild salmon populations in Main Basin and Gulf of Bothnia Rivers in assessment unit 1 (Gulf of Bothnia, Subdivision 31) Rivers in assessment unit 2 (Gulf of Bothnia, Subdivision 31) Rivers in assessment unit 3 (Gulf of Bothnia, Subdivision 30) Rivers in assessment unit 4 (Western Main Basin, Subdivisions 25 and 27)... 95

4 ii ICES WGBAST REPORT Rivers in assessment unit 5 (Eastern Main Basin, Subdivisions 26 and 28) Rivers in assessment unit 6 (Gulf of Finland, Subdivision 32) Potential salmon rivers General Potential rivers by country Reared salmon populations Releases Straying rate M74, dioxin and disease outbreaks M74 in Gulf of Bothnia and Bothnian Sea M 74 in Gulf of Finland and Gulf of Riga Disease outbreaks Summary of the information on wild and potential salmon rivers Reference points and assessment of salmon Introduction Historical development of Baltic salmon stocks (assessment units 1 6) Updated submodels Changes in the assessment methods Status of the assessment unit 1 4 stocks and development of fisheries in the Gulf of Bothnia and the Main Basin Abundance trends of wild stocks in Recent development of fisheries, catches, and harvest patterns Stock projection of Baltic salmon stocks in assessment units Assumptions regarding development of fisheries and key biological parameters Results Additional information affecting perception of stock status M74 and disease outbreaks and potential effects on stock development Revision of basic input data Conclusions Benchmark and a road map for development of the stock assessment Benchmark Road map for development of the assessment Needs for improving the use and collection of data for assessment Sea trout Nominal catches Biological catch sampling

5 ICES WGBAST REPORT 2017 iii 5.2 Data collection and methods Monitoring methods Marking and tagging Assessment of recruitment status Methods Data availability Data presentation Trout in Gulf of Bothnia (SD 30 and 31) Trout in Gulf of Finland (SD 32) Trout in Main Basin (SD 22 29) Recruitment status and trends Recruitment status Recruitment trends Reared smolt production Recent management changes and additional information Management changes Additional information Assessment result Future development of model and data improvement Compatibility of the EU-MAP with the data needs for WGBAST Recommendations References Literature Annex 1: List of participants Annex 2: Stock Annex for salmon in SD Annex 3: Recommendations Annex 4: Data for HELCOM salmon core indicator.xlsx Annex 5: Technical minutes from the Salmon Review Group

6 ICES WGBAST REPORT Executive Summary The Baltic Salmon and Trout Assessment Working Group [WGBAST] (Chair: Stefan Palm, Sweden) met in Gdańsk, Poland, 27 March 4 April A total of 27 persons from all Baltic Sea countries attended the meeting (whereof five via correspondence). The group was mandated to assess the status of salmon in Gulf of Bothnia and Main Basin (Subdivision 22 31), Gulf of Finland (Subdivision 32) and sea trout in Subdivision 22 32, and to propose consequent management advices for fisheries in Salmon stocks in Subdivision were assessed using Bayesian methodology, with a stock projection model used for evaluation of the impacts of different catch options. Due to technical problems, the model could only be used with data up to The development in stock status, however, was evaluated using the latest fisheries and river monitoring data (from 2016). Section 2 of the report covers catches and other data on salmon in the sea, and also summarizes information affecting the fisheries and the management of salmon. Section 3 reviews data from salmon rivers and stocking statistics. Status of salmon stocks in the Baltic Sea is evaluated in Section 4. The same section also deals with sampling protocols and data needs. Section 5 presents the status of sea trout stocks. The total salmon catch in 2016 (excluding recent estimates of trolling catches; see below) was the second lowest in the time-series since the 1970s, although the level has been similar in recent years. Efforts in several important commercial fisheries decreased to their lowest recorded. The total share of recreational catches in the sea and rivers continues to increase. In particular, the offshore trolling fishery has developed rapidly. According to preliminary estimates, the total salmon catch may have been individuals larger in recent years than previously known. Most of this extra fishing mortality, so far not accounted for in the assessment, has likely yielded too high estimates of natural (post-smolt) mortality. The natural salmon smolt production has gradually increased in the Gulf of Bothnia rivers, and in the most recent years also in the Main Basin and Gulf of Finland. Continued increase of smolt production is predicted in for most rivers, mainly as a result of good spawning runs in The current (2016) total production in all Baltic Sea rivers is close to 3.5 million wild smolts, corresponding to about 85% of the overall potential smolt production capacity. In addition, about 4.5 million reared salmon smolts were released into the Baltic Sea in An increasing proportion of the assessed river stocks have reached 75% of PSPC with high or very high certainty, especially in the north. At current fishing pressure and natural mortalities, a continued positive status development is predicted. As previously, most weak salmon rivers are located in the Main Basin and Gulf of Finland, but status has improved also among these southern stocks. In particular, wild Estonian (Gulf of Finland) stocks show recovery. M74-related juvenile mortality increased in spring 2016, and is expected to increase further in It is hard to predict how long elevated levels of M74 will persist beyond Recent disease outbreaks (the cause still un-

7 2 ICES WGBAST REPORT 2017 known) in several wild salmon rivers with large numbers of dead spawners is another health-related concern for the future. The exploitation rate of Baltic salmon in the sea fisheries has been reduced to such a low level that most stocks are predicted to recover. However, weak stocks also need longer term stock-specific rebuilding measures, including fisheries restrictions in estuaries and rivers, habitat restoration and removal of migration obstacles. Some positive signs can be seen for sea trout in the Baltic Sea, but many populations are still considered vulnerable. Stocks in the Gulf of Bothnia are particularly weak, although spawner numbers are improving. In general, stock status is higher in the Main Basin and in the southern Gulf of Finland. Exploitation rates in most fisheries that catch sea trout in the Baltic Sea area should be reduced. This also includes fisheries for other species where sea trout is caught as bycatch. In areas where stock status is good, existing fishing restrictions should be maintained in order to retain the present situation.

8 ICES WGBAST REPORT Introduction 1.1 Terms of Reference 2016/2/ACOM10 The Baltic Salmon and Trout Assessment Working Group (WGBAST), chaired by Stefan Palm*, Sweden, will meet in Gdansk, Poland, 27 March 4 April 2017 to: a ) Address relevant points in the Generic ToRs for Regional and Species Working Groups; Material and data relevant for the meeting must be available to the group no later than six weeks prior to the meeting. WGBAST will report by 12 April 2017 for the attention of ACOM. 1.2 Participants Rafał Bernaś Janis Birzaks Poland Latvia Johan Dannewitz (part of meeting) Sweden Piotr Debowski Poland Stanislovas Jonusas (observer, part of meeting) European Commission Anders Kagervall (Skype & by correpondence) Sweden Martin Kesler Vytautas Kesminas Estonia Lithuania Marja-Liisa Koljonen (by correspondence) Finland Antanas Kontautas Adam Lejk Lithuania Poland Egidijus Leliūna (part of meeting) Lithuania Hans Jakob Olesen (part of meeting) Denmark Katarzyna Nadolna-Altyn Tapani Pakarinen Poland Finland Stefan Palm (chair) Sweden Stig Pedersen Wojciech Pelczarski Denmark Poland Christoph Petereit (by correspondence) Germany Henni Pulkkinen (Skype & by correpondence) Finland Atso Romakkaniemi Harry Vincent Strehlow (part of meeting) Stefan Stridsman Finland Germany Sweden Susanne Tärnlund (part of meeting) Sweden

9 4 ICES WGBAST REPORT 2017 Sergey Titov Russia Simon Weltersbach (part of meeting) Germany Rebecca Whitlock (Skype & by correpondence) Sweden 1.3 Ecosystem considerations Salmon and sea trout in the Baltic ecosystem Salmon (Salmo salar) and sea trout (Salmo trutta) are among the top fish predators in the Baltic Sea. Together with European eel (Anguilla Anguilla) and migratory whitefish (Coregonus lavaretus/coregonus maraena) they form the group of keystone diadromous species in the Baltic Sea. On the species level, based on the IUCN criteria, salmon and the sea trout has been categorised as vulnerable (VU) by HELCOM. As a result of precise homing of salmon and sea trout to their natal rivers, each river and even in some cases each river section, may have a genetically unique and demographically largely independent population; thus the conservation of biodiversity requires safeguarding of the genetic variation and integrity of local populations. Likewise, the development and status of single river stocks of salmon and trout needs to be accounted for, to allow an effective resource management. Salmon and sea trout are anadromous, i.e. they hatch in freshwater, spend one to five years in river and after this migrate for a long period to the sea, then return to freshwater to spawn. Therefore, good connectivity between the sea and rivers, as well as in the rivers, is of ultimate importance for the existence of these species. Salmon has the widest migration routes over the Baltic Sea catchment area. As an example, salmon juveniles occupy the headwaters of the River Tornionjoki km upstream from the sea, which is the northernmost point of the Baltic Sea drainage area. After 3 5 years growth in freshwater, juveniles migrate to the sea, at first feeding on insects and other invertebrates and half a year later they shift to feed on herring and sprat in the southwestern part of the Baltic Sea proper. Salmon mature after 1 4 years growth on the feeding grounds, after which they migrate the 2000 km distance back to their natal headwater rivers for spawning. Sea trout basically has the same life cycle as salmon. The most important difference is that most strains do not migrate as far as the salmon. Instead they spend the time in sea in coastal waters where the majority of sea trout from a specific strain stay within a few hundred kilometres from their home river. Some specimens, however, migrate further and in some strains in the Southern Baltic most sea trout seem to migrate longer distances into the open sea. Sea trout spawn and live during the first period of life in smaller streams than salmon. For this reason the connectivity from sea to spawning areas may be even more critical, compared to salmon, due to possible minor barriers in the small streams. In the Baltic Sea area, sea trout are found in a much larger number of streams than salmon. Many of these streams are in lowland areas that are often strongly influenced by human activity. One effect from this is for example elevated siltation, deteriorated spawning possibilities and reduced survival of eggs. At each stage of migration and life cycle, salmon occupies a specific niche which cannot be occupied by any other species in the ecosystem. For instance, salmon juveniles are one of the few species that can utilise fast-flowing freshwater habitats in the large northern rivers. No other fish species was able to replace salmon juveniles in fish

10 ICES WGBAST REPORT production and populate the empty rearing habitats during the deep depression in salmon abundance in the latter half of the 20th century. Salmon is adapted to uniquely utilise and link the low-productive, fast-flowing river habitat, which is a good environment for reproduction, with the pelagic sea habitat, which offers good conditions for fast growth due to the high abundance of prey species (Kulmala et al., 2013). Similarly sea trout has adapted to live in the smaller tributaries and at slightly lower water velocities. All this demonstrates how connectivity between river habitat, coastal transitional zone and open sea is the lifeline for Baltic salmon and sea trout, and how the requirements imposed to biotic and abiotic habitat vary in time and space, depending on the life stage of the species concerned. Today, Baltic salmon reproduce naturally in nearly 40 rivers. In the past, however, the number of rivers with wild Baltic salmon stocks is known to have been considerably higher, i.e. around one hundred rivers. Also the number of rivers with wild sea trout stocks has declined considerably. Damming, habitat destruction, pollution and intensive fishing have been identified as the main causes of the decline. Presently, a majority of the wild salmon originates from rivers located in Sweden, Finland, Latvia and Estonia. Most of the current spawning rivers of wild sea trout stocks are located in Denmark and Sweden. Salmon and sea trout play an important role in maintaining the balance in riverine foodwebs, both by harvesting invertebrate populations and also providing an important food source for other predatory species (Kulmala et al., 2013). The total nutrient transportation between freshwater and sea is nowadays lower than in the past due to damming and other human activities, which have decreased fish abundance, destroyed natural migration and life cycle of salmon in many spawning rivers. Salmon and sea trout turns over gravel in the river bed while spawning. This bioturbation cleans river bed from, for example, organic particles the sedimentation of which is high in the Baltic rivers. Spawning removes also macrophytes and invertebrates from the sediment, which may more easily be fed by river fish. Salmon is a top fish predator in the Baltic Sea that mainly eats sprat and herring (in the south mainly sprat and towards the north increasingly herring). Thus, salmon in one sense refines various micronutrients for use of other top predators like mammals, including humans (Kulmala et al., 2013). Salmon muscle indeed contains plenty of polyunsaturated fatty acids, which are beneficial for human circulatory system. However, being at the top of the food chain salmon unfortunately also accumulates harmful substances, i.e. various environmental toxicants (e.g. dioxins). Salmon and sea trout are frequent prey species of grey seals, especially in the Gulf of Bothnia (e.g. Lundström et al., 2010). The increasing seal population is likely to consume more salmonids, which is expected to impact salmon and sea trout population principally in a similar manner as fishing. The thiamine deficiency syndrome M74 is a reproductive disorder, which causes mortality among yolk-sac fry of Baltic salmon (see Stock Annex and Section 3.4). The development of M74 is caused by a deficiency of thiamine in the salmon eggs that, in turn, is suggested to be coupled to an abundant but unbalanced fish diet with too low concentration of thiamine in relation to fat and energy content (Keinänen et al., 2012). The intake of thiamine for Baltic salmon in relation to energy and fat remains lowest by eating young clupeids, especially young sprat (Keinänen et al., 2012). Total biomass of sprat in the Baltic main basin and salmon growth are positively correlated. Further, variation in the condition factor of pre-spawning salmon is explained by fluctuations in the biomass of sprat (Mikkonen et al., 2011). The high growth rate of salmon seems not as such be the cause, but rather the abundance of prey and its qual-

11 6 ICES WGBAST REPORT 2017 ity are responsible for M74 (Mikkonen et al., 2011). To inhibit M74, great variation in the size of prey stocks utilized by salmon should be avoided. In the Baltic main basin, where sprat reproduce, the size of the sprat stock needs to be under control, and possibly not allowed to exceed that of herring. The safest strategy for attaining this objective would be to ensure a large, stable cod stock (Casini et al., 2009), to prey on the sprat. Alternatively, increased fishing on sprat would have the same inhibiting effect on M74 (Keinänen et al., 2012) Ecosystem impacts of fisheries and mixed fisheries overview In a time-span of about one century, salmon fishing has first moved from rivers and coastal area near the river mouths to the offshore. And again during the last two decades, the balance has shifted back to coastal and river fishing. The expansion of offshore fishing coincided with the expansion of hatchery-rearing and stocking programmes of salmon juveniles for fishing. Stocking volumes have lately somewhat decreased. Catch of sea trout, especially in the coastal gillnet fishery, both as a targeted and as a non-target species poses a problem for the recovery of threatened sea trout stocks in many Baltic Sea areas. Sea trout are also caught as bycatch of some river fishing targeting salmon. Discarding of seal-damaged salmon occurs mainly in the coastal trapnet and gillnet fishery, but also in the offshore longline fishery. Some specimens of seals drown in trapnets. Seal-safe trapnets have been developed, which has lately decreased seal damages, discarding and seal deaths in gear. Salmon and sea trout are caught by several gear types, and in some cases this has raised concerns about the reliability of catch estimates of the TAC controlled salmon fishery vs. the non-controlled sea trout via misreporting of salmon as sea trout. This may skew species specific estimates of fishing pressure and undermine effectiveness of management measures Data for HELCOM salmon core indicator The core indicator used by HELCOM for evaluation of salmon stock status is based on the comparison of assessed smolt production versus assessed potential smolt production capacity on the assessment unit (AU) level. To facilitate data transfer, AUspecific smolt production estimates needed for the HELCOM indicator are presented in Annex 4, where AU 1 2 have been combined to better match the division used for HELCOM assessment units. For AU 5 6, the production estimates are summed up from estimates presented in Tables and Response to last year s Technical Minutes The Technical Minutes of last year s report were entirely focused on a special request by the EU on potential management measures for salmon in the Gulf of Finland (see ICES 2016, Section 1.4). Therefore the present report does not contain the usual responses from the WG on how comments/criticism from reviewers have been handled with respect to this year s assessment.

12 ICES WGBAST REPORT Salmon fisheries 2.1 Description of gears used in salmon fisheries A description of gears used in different Baltic salmon fisheries, both commercial and recreational, is given in the Stock Annex (Annex 3). Extensive descriptions of gears used, as well as historical gear development in the fisheries, are also available in ICES (2003). This year the WG held a special session during the meeting on needs for more and higher quality data from the developing recreational salmon fisheries in both marine and fresh waters. This report therefore contains two new sections with that focus. Below comes first a description of the growing trolling fishery, including an updated time-series of trolling catches from a recent expert elicitation (Section 2.1.1). After that follows a summary of more general aspects on needs and complexities associated with data collection from recreational fisheries (Section 2.1.2) Recreation trolling fishery Recreational trolling is an increasingly common and popular fishing method to catch salmonids in the Baltic Sea. Trolling is not only practiced by private anglers but also by professional guiding operators. The recreational salmon fishery, including the trolling sector, supports an industry that provides jobs involved in manufacture, sale or provision of tackle, boats, professional guide services, hotels, restaurants and more. Recent survey estimates from Germany revealed that trolling anglers spend on average 3500 annually (Kaiser, 2016). The name trolling originated from the verb to troll, describing a fishing practice of slowly dragging a lure or bait from a moving boat. Thereby recreational fishermen troll a number of fishing lines, baited with lures or natural bait through the water. Fishing lines are spread horizontally with help of planer boards and vertically using downriggers. Common trolling speeds vary from 1.5 to 3 knots. Small boats used for trolling vary between 3 and 8 meters. Fishing grounds are usually over deeper water, and boats may venture more than 20 nautical miles off-shore. Therefore, weather conditions have a strong impact on the trolling effort, and bad weather conditions may prevent trolling boats to leave their homeports periodically during the trolling season. The trolling season varies between the different sea areas and depends on the feeding and spawning migration of salmon and/or seasonal closures. In the western Baltic and the Main Basin it typically starts in late fall and ends the middle of May. In the Åland Sea and Gulf of Bothnia the trolling season starts at the end of May and ends in late summer. Recreational salmon trolling has been practiced in the Baltic Sea for more than 30 years, however, catch data from individual countries is still incomplete or missing. One reason for this is that trolling data collection is often not included or sufficiently covered in national recreational catch sampling schemes. The magnitude of this fishery also varies between countries, and while in some countries trolling effort has levelled off (e.g. Sweden) it is developing in others (e.g. Poland and Lithuania). To account for this source of fishing mortality and to facilitate the inclusion of trolling data in the Baltic salmon stock assessment, a time-series comprising both retained and released components was developed as part of the recent benchmark (WKBaltSalmon, 2017; see Section 4.6). National experts were asked to reconstruct time-series of the number of retained and released salmon caught in the recreational trolling fishery,

13 8 ICES WGBAST REPORT 2017 starting from 1987, by using quantitative data from surveys (if available) and/or qualitative inquiry of stakeholders (e.g. experienced trolling fishers, local authorities, guiding operators and angler associations). In addition to providing a mode number of retained and released salmon for each year and area, national experts were also asked to provide a minimum and maximum value (similar to a 95% probability interval) to provide a semi-quantitive measure of uncertainty. National guestimates were asked to cover the three main areas with feeding or spawning migrating salmon (i.e. SD 22 28, SD and SD 32). The triangular probability distributions per year and area collected from national experts (min-mode-max) were combined into joint medians (with 90% probability limits) using the same transformation as applied to similar expert estimates of discarding and unreporting (Section 2.2, ICES, 2016, Annex 4). The total number of retained salmon includes an assumed post-release mortality rate of 25% for troll-caught released salmon. As no post-release mortality estimates for trollcaught Atlantic or Baltic salmon in marine waters exist, the 25% mortality rate was derived from a review of studies dealing with troll-caught Pacific salmon (Parker et al., 1959; Butler and Loeffel, 1972; Wertheimer, 1988; Wertheimer et al., 1989; Gjernes et al., 1993; Orsi et al., 1993). The resulting joint expert estimates for trolling catches in SD 22 28, SD and SD 32 over time are depicted in Figure This preliminary evaluation suggested that the true recreational catch at sea has potentially been salmon larger than assumed in the assessment period Exploratory model runs were planned to evaluate the response of the assessment and the sensitivity to the stock size. Unfortunately, the model did not perform (Section 4.1) and no scenario could therefore be run including these updated recreational trolling catches. This will instead be assessed by intersessional work. The group points out, that according to the nature of the Baltic salmon assessment model, the underestimated trolling catches are currently most likely included as a component of natural mortality Assessing recreational salmon fisheries Commercial and recreational fisheries coexist and exploit the same stock. In the past 20 years, commercial salmon catches in the Baltic Sea have declined by nearly 80%, while recreational salmon catches remain high and/or are increasing (both freshwater and marine). In contrast to commercial catch data, which rely on mandatory reporting, recreational catch data rely on estimates provided by recreational fishing surveys. While many freshwater catches are fairly well covered, either on the level of individual rivers (reporting systems, e.g. by sport fishing clubs) or in larger national surveys with a focus on recreational freshwater fishers (e.g. Finland, Sweden), available data on marine catches is patchy and for most countries missing completely. European Member States (MS) are obliged to annually collect marine recreational fishery data of salmon in the Baltic Sea since 2002 (EC) No 1639/2001. In 2016, the EU multiannual plan was prolonged, specifying that MS are obliged to collect numbers and weight or length for caught and released catch components of salmon and sea trout (including in freshwater) (EU) 2016/1251. There are three main notable challenges associated with recreational fisheries data collection: (1) there is no central registration of recreational fishers, (2) recreational catches are not documented, and (3) recreational fishers often fish in remote and hard to access areas. As a result, recreational fishing surveys are complex and difficult to conduct, often requiring a number of different surveys.

14 ICES WGBAST REPORT The main drivers for the collection of recreational fishery data include: collecting recreational fishing mortality for inclusion in stock assessment, designing effective controls of recreational fishing and monitoring outcomes, estimating economic value and social benefits to local communities, developing long-term management plans, and supporting the delivery of environmental and marine spatial planning legislation (ICES, 2015b). The type of recreational fishery data needed involves information on the characteristics of the different types of recreational fisheries in a region, the size compositions for retained and released fish, and the numbers of fish retained and released per individual fishing trip. WGBAST recognizes the need for developing the evidence base to support decision making and scientific advice. It is now for the MS to set up national recreational fishery data collection schemes that provide robust and accurate estimates, especially for the marine recreational salmon fishery (i.e. salmon trolling fishery). To estimate total catches and releases, the following information is usually needed (ICES, 2015b): The total number of recreational fishers, boats, number of fishing trips or other measure of participation or fishing effort, generally estimated from a national survey. Catch per unit of effort (or catch per person or per boat depending on the type of survey) recorded for a representative sample of fishers, boats or trips, etc. for example from on-site surveys of individual anglers or completion of catch diaries or vessel logbooks. Data are needed for the retained (harvested) catch as well as for released fish if total fishery removals are to be estimated using data on post-release mortality. Demographic and avidity (frequency of fishing) data where these are needed to re-weight samples to be more representative of the population and improve accuracy. Biological data on catches size or age composition are required both for caught and released components if catch at size or age is needed for an assessment model. Direct on-site measurements of fish length are known to be more accurate than self-reported data. To estimate economic value of recreational fisheries direct expenditure data by spend categories is also needed. This information should be collected alongside existing recreational fisheries surveys as the costs are not significantly greater. Collection of data on an annual basis is preferable, as imputations for missing years introduce uncertainty and there are strong indications that the spatial and seasonal variability of fishing effort and catches is highly dynamic. Moreover, historical evidence shows that recreational fisheries may become more or less important over time, thus there is a need for time series data to show trends. The most cost-effective way to conduct recreational fishing surveys is having a licence system in place where licence holders can be contacted. Lithuania even requires mandatory catch reporting allowing for a census of recreational catch data. If no national registry is available, a screening survey is required sampling from a broad coverage frame like residential households to obtain total numbers of recreational fishers. This is usually done by means of off-site surveys (telephone, mail, online). On-site surveys like access point intercept or roving creel surveys are conducted to obtain cpue data. Visual surveys such as aerial or camera surveys are conducted to estimate effort. A

15 10 ICES WGBAST REPORT 2017 combination of several survey methods is usually required to estimate recreational catch and effort. 2.2 Catches Commercial catch statistics provided for ICES WGBAST are mainly based on logbooks and/or sales notes. Non-commercial catches are mainly estimated by questionnaires or special issues. Detailed information on catch statistics in the Baltic Sea (in total and by country) is given in the Stock Annex (Annex 3). The presented catch tables include both commercial and non-commercial fisheries from sea, coast and rivers. Discards and unreported catches are not included in nominal catches, but are presented separately. Estimation procedures for discards and unreported catches are described in the Stock Annex (Annex 3). More detailed information on discards and unreporting on a country-by-country level is given in Section 2.3. The catches in weight from by country, including separate columns for noncommercial catches, discards and unreported catches from 1994 and onwards, are presented in Table The catches in numbers are presented in Table 2.2.2, where also the share of discards and unreported catches from 1994 and onwards are presented in separate columns. Catches by area and country in tonnes are presented in Table 2.2.3, and by subdivision (SD) in Table Nominal catches in numbers by country from sea, coast and rivers are presented in Table Values on discards and unreported catches (Tables and ) are calculated using conversion factors (see Section 2.3 and Annex 3) and are reported in terms of most likely value and 90% probability interval (PI). An overview of management areas and rivers is presented in the Stock Annex (Annex 3). The recreational (non-commercial) catches in numbers by country are presented in Table There has been a decline of the total nominal catches in the Baltic Sea, starting from tonnes in 1990 and decreasing to 900 tonnes in Since then, the catches increased somewhat in In the last two years the total nominal catch again decreased, and in 2016 it was 940 tonnes (Table 2.2.1). Catches by type of gear in percent (weight) are presented in Figure Due to the total driftnet ban being enforced in 2008, the percentage of the total catches by driftnet has been zero since that year. Over time, the proportion of the coastal catch (mostly trapnets) has gradually increased, and in 2016 it was 51.2% out of all nominal catches (70.8% of the commercial). The proportion of non-commercial catches has grown in relation to the total nominal catches. In 1994, non-commercial catches were 10% of the total nominal catches in tons. In 2016 this share reached 40%. The proportion of the non-commercial part of the total catches (including river catches) from 2004 and onwards is illustrated in Figure Below follows more detailed catch information, country by country: Denmark: The Danish salmon fishery is an open sea fishery. The commercial fishery uses longlines and takes place in the cold months (December March), when the water temperature is below 10 C and garfish are not active. In 2016 the commercial catch was salmon (51.1 tonnes). The recreational fishery is mainly trolling, but some household fishing also takes place around Bornholm. Furthermore a recreational fishery with fixed gears occurs. It is estimated that the recreational catch in 2016 was 8000 salmon (approximately 40 tonnes), resulting in a total catch of salmon. Almost all catches, including the recreational fishery, were caught in ICES SD It is likely

16 ICES WGBAST REPORT that the effort in the commercial fishery has decreased in recent years due to the increasing number of seals in the waters close to Bornholm. Estonia: There is no Estonian fishery targeting salmon particularly. In the coastal fishery, salmon is a bycatch and the main targeted species are sprat, flounder and perch. The share of salmon in the total coastal catch is less than 1%. In total, the Estonian salmon catch in 2016 was 11 tonnes, which is 22% more than in the previous year. A vast majority of the salmon is caught from the Gulf of Finland (SD 32). In that region there are about 570 commercial fishermen, and besides of that up to monthly gillnet licences are distributed annually to non-commercial fishermen. Commercial fishermen take about 68% of all caught salmon; a vast majority (88%) are caught in gillnets and the rest in trapnets. About 75% of the annual salmon catch is taken in September November, and during this time of the year nearly all caught salmon are spawners. Finland: In 2016 in total, Finnish fishermen caught salmon (396 tonnes) in the Baltic Sea, which was about 3% less than in The commercial catch was salmon (190 tonnes). The recreational catch (including river catches) was salmon (11% increase from 2015). A decrease in the catches occurred mainly in the commercial fishery. Since 2012, there has been no commercial fishery for salmon in the Southern Baltic Sea by Finnish vessels. All of the commercial catch was taken in the coastal fishery, mainly by trapnets, and the catch here decreased about 19% from Finland had an extra quota of salmon swapped from Latvia. Commercial catch data from year 2016 are preliminary. River catches (recreational) were salmon (130 tonnes), an increase with 19% compared to in Catch estimates of the recreational fishery in sea were highly uncertain, particularly in the Gulf of Finland (see confidence intervals of catch estimates in Table 2.2.6). The estimates of recreational salmon catches in the sea for years are based on results of the Finnish Recreational Fishing 2014 survey. The methodology used for estimating these catches is described at River catches in Tornionjoki and Simojoki have been estimated by annual surveys, and by interviews and voluntary riverside catch statistics in other rivers. In the Gulf of Finland (ICES SD 32) the Finnish commercial catch in 2016 was salmon (42 tonnes), and the recreational catch including river catches was salmon (16 tonnes). Almost all of the recreational catch was caught in river Kymijoki. Trapnets caught 99% of the commercial salmon catch in this area. In all, 44 fishermen fished salmon with 135 trapnets with an effort of trapnet days which was 5% less than in There was only sporadic longline off-shore fishing for salmon in the area (47 salmon, 257 kg). Germany: The total reported commercial catch of salmon was 8.1 tonnes (1 616 individuals) in 2016 (SD 22 25), which was 20% less compared to in Currently there is very little German commercial fishery that directly targets salmon, and a considerable part of the salmon catch is caught as bycatch in other fisheries (particularly trawl and gillnet fisheries). However, in recent years an extensive trolling fishery has been observed. In 2016, a national marine recreational fisheries survey was initiated (including salmon and sea trout). A total of 37 on-site samplings were conducted and 237 trolling boats targeting salmon were interviewed. The total number of landed salmon was estimated to be (SE ± 369) salmon in the two harbours during the study pe-

17 12 ICES WGBAST REPORT 2017 riod. There are no data available on freshwater catches, but commercial and recreational freshwater catches are most likely insignificant as there are no rivers with a significant salmon spawning and fishery along the German Baltic coast. Latvia: The total catch in 2016 was salmon (5.9 tonnes), which is on the same level as in Included in the total catch are both recreational and commercial coastal catches (1 700 salmon, 4.7 tonnes). In 2015, one Latvian vessel reported 137 salmon caught by longlines. An additional 153 salmon were caught in the rivers, mainly in brood stock fisheries. In the river Daugava all fishing with gillnets is closed since 2005, and only use of trapnets is allowed in the river. Lithuania: In 2016 Lithuanian commercial fishermen caught 344 salmon (1.3 tonnes), which is about 30% more than in Most of the salmon were caught in the Curonian Lagoon. Additionally, 31 salmon were caught in the rivers for artificial reproduction (brood stock fishery) and five for scientific purposes. In addition, salmon were taken in recreational sea (trolling) and river fisheries (Table 2.2.6). Poland: Overall, the off-shore and coastal catch was salmon, 3% less than in The river catch was only 122 fish, and the main reason for this low number was seal predation. Most of the Polish salmon catch was taken from SD 26. The reported river catch originated mostly from Vistula River (90%) and Pomeranian rivers. The river catch was taken exclusively for brood stock purposes. Russia: There is no specific fishery targeting salmon in Russia, but salmon (and sea trout) can be caught as bycatch in the coastal fishery (by trapnets and gillnets) where the main targeted species are herring, sprat, smelt, perch, pikeperch, etc. In Russia there are no official statistics on bycatches. In 2016, Russian fishermen caught 419 salmon (1.7 tonnes) from the Baltic Sea, whereas 16 salmons were caught as bycatch in the Gulf of Finland. In the rivers, 403 spawners (1.6 tonnes) were caught during brood stock fishing (223 in Neva River, 164 in Narva River and 16 in Luga River). Sweden: The total Swedish salmon catch in 2016 was 395 tonnes (close to salmon) which is approximately the same weight as in This is the lowest recorded annual catch for the period (Table 2.2.1). The total catch in the coastal fisheries increased from 190 tonnes in 2015 to 197 tonnes in At the same time, the river catch decreased to 182 tonnes in 2016 (compared to 192 tonnes in 2015). In 2016, the commercial sea fisheries caught salmon (197 tonnes) which was 5% more than in The vast majority of salmon were caught in trapnets along the coast in the Gulf of Bothnia (SD 30 31). In addition, there was a commercial riverine catch of salmon (64 tonnes), whereof 53 tonnes were taken in River Luleälven (SD 31) and the rest (11 tonnes) in River Ljusnan (SD 30). In Luleälven and Ljusnan large numbers of reared smolts are released annually and no wild production exists. These commercial catches taken in river mouths (fresh water) are not counted against the Swedish TAC. In the Swedish recreational fisheries in 2016, about 10% of the estimated salmon catch was caught at sea, and about 90% were taken in rivers. Based on a similar estimation as in 2015, a recreational off-shore trolling catch (landed) of 16.4 tonnes (1 603 salmon) was assumed also for In the wild and reared rivers a total of close to salmon were landed (118 tonnes, not including the above mentioned commercial river catches). Out of this total landed river catch, brood stock fisheries in reared rivers took a significant part (24%). The remaining catch was divided between rod-fishing (58%) and various types of net fishing (18%). Note finally that Swedish salmon anglers are practising catch & release to an increasing extent; about salmon taken in rivers were released back in 2016.

18 ICES WGBAST REPORT A new study on recreational fisheries in SD was conducted in 2015 (Hasselborg, 2016). One main results of that study was that the catches in the recreational coastal trapnet fishery in the Bothnia Bay have decreased considerably in recent years. One main reason is a new legal requirement that it is mandatory to have a commercial fishing licence if you want to sell your catch. Hence, many of the former recreational fishermen are now fishing commercially. Catches from the Swedish trolling fishery in 2015 were also estimated in the same year. Compared to the last previous study of Swedish trolling (performed in 2011) a significant drop in landed fish was estimated. The main reason is a national regulation (since 2013) that only allows reared salmon with removed adipose fin (mandatory in the Swedish hatchery programme) to be landed in the Swedish trolling fishery that mainly takes place in the Baltic Main Basin. Distribution of catches by countries in comparison with the TAC Until 1992 the TAC was given in tonnes, but from 1993 the TAC has been given in numbers. The commercial landings in numbers (excluding river catches) compared to TAC, by fishing nations and areas in , are given in Table Unreported catches and discards are not included in the utilisation of the TAC, but total catches of salmon including unreported catches and discards are presented in percentage of TAC in Figure In 2016, 73% of the TAC in SD was utilised. Total TAC was individuals (according to COUNCIL REGULATION (EU) No 2015/2072 of 17 November, 2015; amending Regulations (EU) No 1221/2014 and (EU) 2015/104). In the Gulf of Finland, 56% of the EC TAC of individuals was utilised. The Russian catch was just 16 salmon. It should be noted, that occasionally there can be some quota swapping between countries, which may result (or not) in exceeded original national TACs. In 2016 such an exchange took place in Finland, where 3400 salmon were exchanged from Latvia, and in Sweden, where all together 2442 salmon were exchanged with Denmark and Germany. The total TAC for salmon was allocated and utilized in the following manner in 2016: Subdivision (SD) Subdivision (SD) 32 Contracting Quota Sea/Coast Catch Quota Catch Utilized (%) party (nos.) (nos.) Utilized (%) (nos.) (nos.) Denmark Estonia Finland* ) Germany Latvia Lithuania Poland Sweden Total EU Russia 1) TOTAL ) No international agreed quota between Russia and EC. * with use of swapped quota.

19 14 ICES WGBAST REPORT 2017 The major part of the salmon catch in the Baltic Sea was caught by professional fishermen, either with longlines in the off-shore areas or by trapnets and gillnets in the coastal areas. The catches in the recreational fishery using commercial gear-types are for self-consumption. These catches are usually not reported through the official channels and therefore the figures have to be estimated. Table and Figure give estimates of the magnitude of this fishery, and it appears that non-commercial fisheries constitute a considerable and growing part of the total catch of salmon. In 2016, noncommercial catches (in numbers from coast, sea and river) constituted approximately 40% (in tons) of the total reported salmon catches. In addition, new information shows that the recreational salmon catch has been even higher than previously recognised; in recent years, total Baltic trolling catches appear to have been severely underestimated (Section 2.1.1; Figure 2.1.1). 2.3 Discards, unreporting and misreporting of catches Estimation procedures for discards and unreported catches for years and misreported catches for years are described in the Stock Annex (Annex 3). In general, data for years on discards and unreporting of salmon from different fisheries in the Baltic Sea are incomplete and fragmentary. For years , the estimates for discards and unreporting have been computed with a new method based on updated expert evaluations, as described below. The new estimation method was applied for the first time by WGBAST in Coefficient factors for unreporting and discarding by country and fisheries were updated for fishing years during the IBPSalmon in autumn 2012 (ICES 2012b), and in addition for years (see WGBAST reports ). Expert evaluations have been provided from Poland, Denmark, Sweden and Finland for all relevant fisheries of each country, respectively. These four countries cover the main salmon fisheries, and together they have caught more than 95% of the total Baltic salmon catch in the last several years. Parameter values for the elicited priors and pooled (average) probability distributions for different conversion factors (by country and year period) are given in the Table In WGBAST the average conversion factors were calculated for all parameters separately for years before and after 2008, because of the change in relative weight between the fisheries in 2008 due to ban of driftnet fishing. In addition, Finland and Sweden banned salmon off-shore fishing in the Main Basin in 2013, which further changed the relative weight between the fleets. Therefore the relevant conversion factors were computed separately for fishing years from 2013 and onwards. Since WGBAST 2015, the average conversion factors for certain parameters have not been used in computations, since they were considered to give a too biased estimate for certain fisheries and fleets. For example, the average share of seal damaged salmon in the off-shore fishery based on Swedish, Danish and Finnish data was considered to give too high estimates for discarded seal damaged salmon in the Polish off-shore fishery before year The average values of the following parameters were seen inapplicable and consequently abandoned: (i) share of unreported catch in off-shore fisheries, (ii) share of unreported catch in coastal fisheries, (iii) share of discarded seal damaged salmon in longline fisheries, (iv) share of discarded seal damaged salmon in driftnet fisheries and (v) share of discarded seal damaged salmon in trapnet fisheries. The average values of these parameters were removed from the Table Instead of average values, a minimum available observed value of the parameter concerned was

20 ICES WGBAST REPORT used for the countries and fisheries, where neither data nor expert evaluation was available. As a result of this change, the total estimates of discarding and unreporting before year 2012 changed to some extent in WGBAST 2015, compared to the corresponding estimates in the reports. The expert evaluation of unreporting rate in the Polish fisheries in 2014 was significantly shifted downwards in the WGBAST 2015, and has since then been considered low. The rest of parameters still have the same values. Apart from the parameters listed above, the average values were used for German, Lithuanian, Latvian, Estonian and Russian fisheries, as country specific expert evaluations for coefficient factors were missing for those countries. However, the catches of these countries represent less than 5% of the total catch of Baltic salmon. Details on the transformation method of parameters of expert elicited triangular probability distributions into parameters of lognormal distributions is presented in Annex 4 in ICES (2016). Assumptions in estimation of unreported catch and discards: In the estimation of unreported catch in the Polish salmon fishery, it was assumed that the same rate of unreporting prevails in misreported as in reported catch. In the estimation of seal damages and discarded undersized salmon in all fisheries, the unreporting (and misreporting in the Polish off-shore fishery) was counted into the total catch, i.e. similar rates were assumed for unreported catch components as for the reported catch. In the Finnish salmon fisheries, seal damaged catch is derived from log book records. These catches were raised by the relevant unreporting rates, i.e. the same unreporting rate was assumed for the seal damaged catch as for the unharmed catch. For seal damaged catch in the Swedish salmon fisheries, the same assumption is due. Here, though, the official statistics do not contain a complete quantitative measure of seal damaged catch, and instead the seal damaged catch is estimated (see below for more details). Estimated unreported catch and discarding for the whole Baltic Sea are presented in Tables and Comparison of estimated unreporting and discard between the year period and shows that the main difference is in the order of magnitude of estimates in discards. This is mainly as a result of updated expert opinions and partly the adoption of new computing model which was realized at the IBPSalmon in 2012 and in WGBAST reports since Main part of the discards is seal damaged salmon and it occurs in the costal trapnet and gillnet fishery but also in the off-shore longline fishery (Table ). Small amounts of undersized salmon were estimated to be discarded in the off-shore longline fishery. Since 2015 there has been landing obligation in longline fishery, however, it has not fully implemented since little reporting of such landings has occurred. Country specific estimates for discards and unreporting are presented in the Table The estimates are uncertain and should be considered as a rather rough magnitude of discards and unreporting Discards In 2016, approximately 8290 salmon were discarded due to seal damages (Table 2.3.2). Slightly more than half of these discards (4450 salmon) took place in the Danish and Polish LLD fishery in the southern Baltic Sea (Table 2.3.2). Estimates were based on the observed proportion of seal damaged catch in subsamples that has been extrapolated to the total catch. In this calculation also the potential misreporting and unreporting

21 16 ICES WGBAST REPORT 2017 was included in the total catch. In the Danish LLD fishery approximately 15% (5% 30%) of the catch was seal damaged, and in the Polish LLD fishery in SD 26 about 25% (5% 65%). Representativeness of these estimates is unknown to the WG. Amounts of seal damaged catches in the Main Basin have increased to significant rates in the last few years, and monitoring will be needed to attain reliably estimates of the seal damages in salmon fisheries. Seal damages started to escalate gradually from 1993, but since the introduction of seal safe trapnets the catch losses in coastal fisheries have levelled off. In 2016, approximately more than one third of the total seal damaged discards (668 salmon in the Gulf of Bothnia and 1923 salmon in the Gulf of Finland) took place in Finnish coastal trapnet fisheries. In Finland, data on seal damages are based on logbook records. In Sweden, the level of seal damages is estimated based on data from a voluntary logbook system and available data on seal interaction in the official statistics, for which an additional expert assessment has been made. The reported amounts of seal damaged salmon should, however, be regarded as a minimum estimate. The reporting rate of the seal damaged catch is assumed to be the same as for the undamaged catch in the coastal fishery. In 2016, seal damages in the Finnish trapnet fisheries covered about 7% of the total salmon catch. The last estimates from Sweden are from 2011, and the same rate of damage is assumed also for the fishing years Dead discards of undersized salmon in 2016 were estimated to 1548 salmon in the whole Baltic Sea (Table 2.3.2). Proportions of undersized salmon in the catches of different fisheries are mainly based on the sampling data (Table 2.3.1) and are considered to be rather accurate. Mortality estimates of the discarded undersized salmon released back to the sea are based on expert opinions. Mortality of the undersized salmon released from longline hooks is presently assumed to be high (around 80%), but few studies have been carried out on this issue and the true rate is uncertain. In the trapnet fishery the post-release mortality is assumed to be lower (around 40%), but again the true rate is uncertain. Both the experimental design and the settings to study these mortalities are very challenging, but such empirical studies are needed in order to get better estimates of the survival rate of discarded salmon. Post-smolts and adult salmon are frequently caught as bycatch in pelagic commercial trawling for sprat (mostly for supplying fish for production of fish meal and oil), but are probably often unreported in logbooks because the relative amount of salmon in these catches is low and can be identified only during unloading (ICES, 2011). Estimates of these potential removals, however, are so uncertain because of insufficient data that they are not taken into account in the present assessment. Besides, there is no estimate on the potential unreporting of bycatch of legally sized salmon in the pelagic trawl fishery. Only the reported catch from the trawls is accounted for in the catch data, even though it has been very low over the years Misreporting of salmon as trout in the Polish salmon fishery In the WGBAST 2014, the Polish misreporting was recomputed for years This was due to that the WG received new data on the catch compositions in the Polish longline fishery. These data were collected by the Polish Marine Fisheries Research Institute in DCF sampling trips on the Polish longline vessels which operated off-shore in SDs 25 and 26 in the years (Table 2.3.3). The data are available in the ICES Regional Database (RDB).

22 ICES WGBAST REPORT According to Polish experts the sampling in 2010 represented only 0.5% of the total number of days at sea, and the proportion of sampled trips has been low also in Even though there is a clear under-representativeness in the sampling of the total fishery, the observed proportion of salmon in the catches of the sampled trips is consistent and with little variation; none of the observations has indicated a substantial proportion of sea trout in the catch. These data suggest that the Polish longline fishery is an almost true salmon fishing with only few sea trout in the catches (approximately 0 3%; Table 2.3.3). The catch compositions in these data correspond to catch compositions that have been observed in catches of other countries vessels fishing in the same area (ICES, 2012). Based on the given data, a 97% proportion of salmon was assumed in the total Polish longline catch for fishing years and for coming years. This is a conservative estimate, and it excludes potential misreporting in the Polish coastal fishery. Misreporting estimates for earlier fishing years ( ) were based on the assumption that cpue in the Polish off-shore fisheries (driftnet and longline) corresponded to 75% of the cpue of other countries fleets in the corresponding fishery and the same area (see e.g. ICES, 2012). The total catch of the Polish off-shore fishery decreased significantly until 2015, but increased again in The total estimated misreporting in 2016 was salmon, about four times more than estimated for 2015 (Table 2.2.2). This increase was mainly due to an increase of longline and gillnet fishing effort off-shore. The Polish reported 2016 catch in the longline fishery was 2224 salmon and sea trout. In addition 6905 sea trout were reported as caught by gillnets in the off-shore. Misreporting in the coastal gillnet fishery was not estimated, even though a potential misreporting could take place there too. However, recent Polish sampling data present very low proportions of salmon in coastal catches (annually maximum 5%). In 2014, Poland presented Regulation (EC) No 1223/2013, which was based on results of extensive EC and national controls (45% coverage of all landings) and gave a decrease by 23% of Polish TAC in Such a percentage as an official factor was proposed to WGBAST for using in the assessment of misreporting. A similar document (Commission Implementing Regulation (EU) No 871/2014), decreasing the Polish TAC by 3.2% in 2014, was presented to the WG in 2015, which make up proof of better control and lower level of potential misreporting of salmon in Poland. However, WGBAST did not use these sources of information in their assessment. The present misreporting estimates should be considered as rough. The WG would benefit from Polish contribution in providing more data or relevant reports that would support the estimation of misreporting rates in off-shore and coastal salmon fisheries. Poland should make sure that the whole catch is counted during each of the EU-MAP sampling trips, and that the planned number of trips will be carried out with an appropriate areal and temporal coverage. Comments provided by Poland on Misreporting of salmon as trout in the off-shore fishery Poland sustains most of its explanations and comments given in its statements in WGBAST Report 2012 and Data on proportion of salmon and sea trout in Polish LLD fishery, considered for use in WGBAST 2015 and further in , based on Polish DCF sampling, have only indicative not quantitative value. DCF salmon/sea trout sampling is aimed for collecting biological measurements, not for assessing share of catch, because it is not only based on sampling at sea but also on sampling on land, where catch is already separated and

23 18 ICES WGBAST REPORT 2017 do not necessarily reflect the real share of both species in a single catch. Moreover, EU_MAP sampling at sea has very low representativeness concerning days at sea comparing to yearly amount of days at sea of the whole Polish salmon LLD fleet (e.g. 0.2% in 2007, 0% in 2008, and 0.5% in 2010). Thus it cannot be considered as a reliable source for obtaining the share of salmon and sea trout respectively in the catches. Information on unreporting and misreporting by country Below follows more detailed information on discards and mis- and unreporting of catches, country by country: Denmark has no information in the official statistics from which it is possible to estimate discard percentages, even if this should be available from the DCF data sampling. Since the quota for salmon in recent years has not been fully utilized, it seems unlikely (however uncertain) that there are unreported catches in the commercial salmon fishery. At three trips with on-board observers in February 2016, 2.6% of the catch was below 60 cm. On the other hand, restrictions in possibilities for marketing larger salmon due to restrictions from dioxin contents could result in unreported catches. Damage by seals to caught salmon has been reported from fishermen to be reaching 40 50%. However, in two trips in December 2015 with inspector on board, the proportion seal damaged salmon was approximately 4%, and in two other trips in February 2016 the proportion was 0.8%. There are no records of misreporting of salmon as other species (e.g. sea trout). Recently seals have been observed to attack salmon being hooked in the trolling fishery. Anglers have also observed seal damaged salmon in their catches. The bycatch of salmon in other fisheries has been observed to be quite low. Observers from the DTU-Aqua participated in the Baltic herring and sprat fishery in the winter 2007/2008 for about 50 days, and bycatches of only a few salmon were observed in this fishery. In Estonia, seal damage is a serious problem in the coastal gillnet fishery where salmon and sea trout are caught. Information from fishermen shows that damages by seals have increased over time. A quantitative assessment of these damages is not available, however, as fishermen in most cases do not present claims for gear compensation. In Finland, the reported discards of seal damages were 2591 salmon (13 tonnes) in This was about 20% less than in the previous year. Seals caused severe damages to all fisheries, mainly in SD 32 where seal damages comprised 11% of the total commercial catch in the area. In Main Basin and Gulf of Bothnia discards of seal damaged salmon were 1923 fish (9 tonnes) comprising about 7.8% of the total commercial catch in SDs In the Gulf of Finland, discards of the seal damaged salmon were 668 fish (4 tonnes) which is 35% less than in In Germany there is no detailed information about potential discards in the commercial fishery, although anecdotal information indicate that discard rates seem to be low. Concerning the low catches, it is unlikely that there is a discard problem in the commercial fishery, but there could be some unknown unreported salmon catches. There are no data available on predation by seals. However, German commercial fishermen reported increased predation rates on salmon longline catches around the Island of Bornholm, which led to the cessation of the directed salmon fishery by German vessels in Also in the recreational trolling fishery, anglers have started to report predation by seals during fish retrieval. Misreporting may be an issue concerning salmon catches, whereby salmon may be reported as sea trout. This could either occur deliberately, since sea trout catches are generally not limited by quota, or unintentionally

24 ICES WGBAST REPORT through species misidentification. In 2013, the federal state authority of Mecklenburg- Western Pomerania initiated a pilot study to investigate the level of misreporting. Within the remits of this study fin samples were genetically analysed. Preliminary results show that misidentification occurred in 30 40% of the commercially landed salmon and sea trout but that the misidentification was evenly distributed between the species, thus indicating no directed misreporting. In Latvia the direct catch losses of salmon due to seal damages has increased significantly from In the most affected area, the southern part of the Gulf of Riga, the percentage of salmon damaged by seal in the coastal fishery increased from 5% in 2002 to 40% in 2003, and to 60% in In recent years, the number of seals has continued to increase. Due to this, the salmon fisheries in late autumn in the coastal waters of Latvia have become economically unfavourable. This holds especially for the gillnet fishery. Experimental fishing with a seal-safe gear (produced in Sweden) was unsuccessful. The gear was too fragile for fishing in the open Latvian coastal waters, with a dominating SW-NW wind direction. In Lithuania information on discards, misreporting and unreporting is not available. In Poland, sampling of 2015 and 2016 longline catches resulted in a total of 2% undersized fish. No undersized salmon were reported in the gillnet fishery. There are no data available on unreported catches. A rapidly increasing amount of seal damages has been observed in recent years, both in off-shore and coastal fisheries in SDs 25 and 26 (Gulf of Gdansk area). No damages have been reported so far in SD 24. Preliminary national data from 2013 indicate that the share of seal damaged fish in separate catches was on average 25% (minimum 5%, maximum 65%). In 2016, losses of 721 salmon and 846 sea trout were recorded in logbooks. The seal colony at the Vistula River mouth has grown up to 300 individuals that are hunting on neighbouring fishing grounds. Since two years, the large amount of seals in the area has almost completely stopped the previous salmon and sea-trout fishery in the lowest few kilometres of the Vistula. In the last few years no catch was reported, and in autumn no spawners of sea trout or salmon could be collected in the Vistula river mouth due to seal attacks. In the past, Vistula used to be the best places for sea trout fishing and for collecting live spawners. In Russia information on discard, misreporting and unreporting is not available. However, unofficial information indicates presence of significant poaching of salmon and sea trout, both in the coastal area and in rivers. In Sweden earlier estimates on discards, misreporting and unreporting are still the best available (see WGBAST 2015 report). Further analysis is needed to evaluate the quality of these estimates and the data on which they are based. As described above, discards due to seal damages have been estimated by expert assessment, using data from a voluntary logbook system and available data on seal interaction in the official statistics. Seal interaction includes seal damaged gears and/or seal damaged fish. Currently, in the official statistics it is possible to report if you have had seal interaction on a fishing trip. These records form the basis for the system that handles seal damage compensation from the government to commercial fishermen. It is also possible, but not mandatory, to report seal damaged salmon in the official statistics, and in 2016 only a total of 51 seal damaged salmon were reported (compared to 343 in 2015). Besides, information on salmon discarded out of other reasons should be included in the official statistics. In 2016, 1237 salmon were reported as discarded (compared to 744 in 2015).

25 20 ICES WGBAST REPORT Fishing effort The total fishing effort by gears in the Main Basin, and in the three main assessment areas for the coastal commercial salmon fishery (AU 1 3), excluding Gulf of Finland, is presented in Table This table includes Baltic salmon fishing at sea and along the coasts in The coastal fishing effort on stocks of AU 1 (Table 2.4.1) refers to the total Finnish coastal fishing effort and partly to the Swedish effort in SD 31. The coastal fishing effort on stocks of AU 2 refers to the Finnish coastal fishing effort in SD 30 and partly to the Swedish coastal fishing effort in SD 31. The coastal fishing effort on stocks of AU 3 refers to the Finnish and Swedish coastal fishing effort in SD 30. Because sea trout in Poland are fished with the same gear type as salmon, effort from the Polish fishery targeting sea trout was included in the table before Development over time in fishing effort for the off-shore fishery is presented in Figure 2.4.1, and for the coastal fishery in Figure The fishing effort is expressed in number of gear days (number of fishing days times the number of gears). The total effort in the longline fishery in 2016 decreased to hookdays, compared to in 2015 and in 2009 (Figure 2.4.1). The effort in the 2016 trapnet fishery was 12% lower than the effort reported in 2015 (Figure 2.4.2). An overview of the number of fishing vessels engaged in the off-shore fishery for salmon during the last 17 years in SD is presented in Table Note that 2016 is the third year in a row, when fishing effort in the Danish off-shore fishery has not been presented. In 2016, 53 Polish vessels were engaged in the off-shore fishery, a number rather similar to in 2015 (44 vessels). Catch per unit of effort (cpue) on a country-by-country level for off-shore driftnetting (until 2008) and longlining in SD is presented in Table The total effort and catch in the Finnish off-shore fishery in Gulf of Finland (mainly longlining) has been too low since 2010, to allow any conclusion regarding the development in cpue. For the Finnish trapnet fishery in the Gulf of Finland, cpue was 0.8 salmon per gear and day in both 2015 and 2016 (Table 2.4.4). Regarding Swedish cpue, further analysis is needed to evaluate especially the quality of past effort data (off-shore fisheries) in official catch statistics. 2.5 DCF stock sampling of salmon All EU Baltic sea countries are obliged to follow the Data Collection Framework (DCF). The national data collection programmes under the DCF include different fisheries regions (off-shore, coastal, river), different fisheries (e.g. commercial, angling, brood stock), and different origin (wild or reared fish). General information on the structure of the data collection in different fisheries, including length of time-series, is presented in the Stock Annex (Annex 3). The sample collection includes individual information (length, weight, sex, etc.) and scales for age and/or genetic analysis. An overview of DCF samples (biological sampling) collected in 2016 is given below:

26 ICES WGBAST REPORT Time period Number of sampled fish by SD Country (month number) Fishery Gear(s) Total Denmark 2 Off-shore Longline Finland 5 9 Coastal Trapnet Off-shore longline River Latvia 3, 8 11 River Trapnet Estonia 1 12 Coastal Gillnet Lithuania 8 10 River Electrofishing Poland 2 12 Off-shore Longline Russia 9 11 River Gillnet+Trapnet Sweden 5 7 Coastal Trapnet River Various Germany* 0 0 Total * not counted as EU DCF. Denmark: There were 235 scale samples collected from Danish landings in All collected samples were used for DNA analyses and scale reading. The analyses were conducted in Finland (Luke, Helsinki). Estonia: Since 2005 Estonia follows the EU sampling programme. Sampling takes place occasionally and is carried out by fishermen. In 2016, 35 samples were collected. Finland: Catch sampling in 2016 brought in 2383 salmon scale samples from the Finnish commercial salmon fisheries and 849 samples from the recreational river fisheries. The samples represented different fisheries in terms of time and space. The whole pool of samples was resampled by stratification according to the real catches. The final amount of analysed samples was optimally adjusted to meet the quality criteria of DCF. Finally, the total numbers of samples were analysed by scale reading and part of these were also analysed using DNA-markers (microsatellites). Germany: In the past, only catch statistics have been assembled. No biological data have been collected in the commercial salmon fishery, as the reported catches are very low making it difficult to collect commercial samples. In 2016 no biological sampling was performed. Latvia: None of Latvia s vessels have been engaged in salmon off-shore fisheries since In the coastal fisheries targeting salmon, biological sampling in 2016 was carried out from June to November. In total 200 salmon were sampled in coastal fisheries. Lithuania: Since 2005 sampling in Lithuania has followed the EU Minimum programme. Lithuanian fishermen have not been carrying out specialized salmon fishing. In 2016, a total of 31 samples were collected in rivers. Poland: Sampling was conducted on landed fish from off-shore catches. According to DCF, total number of sampled fish should be 500 for the whole Polish salmon fishery, but in fact 958 fish were sampled for age, length and weight in Age was estimated based on scale readings. Data collection was conducted in Subdivisions 25 and 26, and

27 22 ICES WGBAST REPORT 2017 covered mainly longline fishery but also some pelagic trawl catches. Part of the samples of salmon scales (n=400) was sent to Finland (Luke, Helsinki) for genetic analyses. Russia: Since Russia is not an EU member state, the country is not obliged to follow DCF. There is no other biological sampling programme running in Russia. However in 2016, 383 salmon collected in the river fishery were aged, and lengths and weights were also recorded. Sweden: The Swedish DCF National Programme (NP) includes sampling in coastal trapnet fisheries targeting salmon. Métier sampling (concurrent sampling at sea) was planned for in total 12 trips (four trips in SD 30 and eight in SD 31) during the fishing season 2016, and nine trips were performed (four trips in SD 30 and five in SD 31). Also, stock sampling is performed by contracted fishermen throughout the whole season. In 2016, 538 salmon were sampled for age analysis (285 salmon in SD 30 and 253 in SD 31). The sampled fish are aged by scale reading, and at the same time, it is determined if the fish is of wild or reared origin. The data from the commercial sampling have been uploaded to ICES Regional Data Base FishFrame (RDB Fish- Frame). Biological sampling from salmon (adults, smolts) caught in rivers is also included in the Swedish DCF NP. New EU legislation: The Commission Implementing Decision (EU) 2016/1251 of 12 July 2016, adopting a multiannual Union programme for the collection, management and use of data in the fisheries and aquaculture sectors for the period , affects also salmonid fish data collection. According to the regulation biological data should be collected from stocks caught by Union commercial fisheries in Union and outside Union waters, and also by recreational fisheries in Union waters. In addition data on anadromous and catadromous species (including salmon) caught by commercial fisheries during the freshwater part of their lifecycle, irrespective of the way these fisheries are undertaken, must be collected. Growth of salmon The average weight of salmon by age group increased around year 1990, simultaneously with an increase in sprat abundance (Figure 2.5.1). Despite some annual variation, the level of growth has been rather stable since that. In 2016, catch samples indicate a slight increase in mean weights by age which is potentially a result of the strong 2014 year class of sprat and also Baltic herring. Despite that salmon shares feeding areas with cod in the southern Baltic Main Basin, there is no clear reduction in the growth rate of salmon as has been observed for cod in the last few years. The estimated post-smolt survival decreased strongly from the mid-1990s until 2005 (Figure ) but this cannot be recognised in the growth data. Mortality mechanisms seem to affect salmon population in such a way that survived individuals grow approximately as large in periods of high mortality as in periods of low mortality. 2.6 Tagging data in the Baltic salmon stock assessment Tagging data (mainly Carlin tags; Table 2.6.1) have been used within the Baltic salmon assessment to estimate population parameters as well as the exploitation rates by different fisheries (see Annex 3 for more detailed information). For various reasons, the number of tag returns has become very sparse after 2009, and tag return data have not been used in the assessment after fishing year As tagging data used in the model are based on external tags, it is vital that fishermen find and report tags. However, earlier reports (summarised in e.g. ICES, 2014) indicate an obvious unreporting of tags.

28 ICES WGBAST REPORT As the tag return data influence e.g. the annual post smolt survival estimates, which is a key parameter in the Baltic salmon assessment, there is a need to supplement or replace the sparse tagging data in the near future. The WGBAST 2010 (ICES, 2010) dealt with potential measures to improve and supplement the tagging data. These consist of alternative tagging methods and supplementary catch sample data. Also, inclusion of smolt tagging in the EU DCF has been suggested. In 2010, the WG also noted the need of a comprehensive study to explore potential tagging systems before a change over to a new system in the Baltic Sea area can be considered. Salmon smolts are still tagged for other monitoring purposes (Table 2.6.1). The total number of Carlin tagged reared salmon released in the Baltic Sea in 2016 was (Table 2.6.2), which was 41% less than in Carlin tagged salmon smolts were released by Finland, Sweden and Latvia. As alternative methods, T-bar anchor tags are also used for tagging of smolts in Finland and Estonia, and in Sweden PIT-tags were used in several wild (index) and reared rivers (Table 2.7.2). No salmon tagging with Carlin tags was conducted in Poland 2016, because of low recapture rate in previous years. As mentioned above tag return rates show decreasing trends, as illustrated in Figures and for tagged salmon released in Gulf of Bothnia and Gulf of Finland, respectively. The return rate of 1-year old Carlin tagged salmon smolts in the Gulf of Finland in Estonian experiments varied around 0.2% in years There were no returns of tags in 2006, but in the following year the recapture rate exceeded 0.8%. Because of the low recapture rate and changes in stocking practices, no 1-year-old salmon smolts have been Carlin tagged in Estonia since The mean recapture rate of 2-year-olds in Estonian experiments for years was 0.7% and varied around 0.1% in years (Figure 2.6.2). Since 2015, only T-bar anchor tags are used in Estonian experiments for tagging of salmon smolts. The mean recapture rate was around 0.1% for years In the recapture rate of Finnish Carlin tagged reared salmon smolts released in the Gulf of Finland and Gulf of Bothnia was close to zero (Figures and 2.6.2). A similarly low recapture rate has been seen for Polish tags in recent years (Figure 2.6.3). 2.7 Finclipping Finclipping makes it possible to distinguish between reared and wild salmon in catches. Such information has been used to e.g. estimate proportion of wild and reared salmon in different mixed-stock fisheries. However, since only a part of the Baltic salmon smolt released are presently finclipped (Table 2.7.1), this type of information is not directly utilised in the WGBAST assessment model. Since 2005 it has been mandatory in Sweden to finclip all released salmon. All reared Estonian and Latvian salmon smolts released in 2016 were also finclipped. In Poland, all types of tagging were stopped in 2013 and 2014, because of national veterinarian s objections. In 2015 tagging was again permitted, but finclipping of Polish smolts was not continued until in From 2017 and onwards all salmon released in Finland will be finclipped (except releases for enhancement purposes). Salmon smolts released in Russia, Lithuania, Poland, Germany and Denmark in 2016 were not finclipped. In 2016 the total number of finclipped young salmon released was , an increase of 1% compared to in Out of this total, were parr and smolts. Compared to in 2015, the number of finclipped smolts increased with about 3% whereas the number of finclipped parr decreased with about 35%. Most finclippings (in numbers) were carried out in SD 30 31, but part of the finclipped fish were also

29 24 ICES WGBAST REPORT 2017 released in SD (Table 2.7.2). In Finland about 22% of the released smolts and 86% of 1-year old parr were finclipped in Additionally, in Sweden and Estonia a total of parr and smolts were finclipped and stocked in 2016 (Tables and 2.7.2). 2.8 Estimates of stock and stock group proportions in the Baltic salmon catches based on DNA microsatellite and freshwater age information Combined DNA- and smolt-age-data has been used to estimate stock and stock group proportions of Atlantic salmon catches in the Baltic Sea with a Bayesian method since year 2000 (Pella and Masuda, 2001; Koljonen, 2006). For the 2016 catch, the same genotype baseline data were used as for the 2015 catch (Table 2.8.1, Figure 2.8.1). The current baseline stock dataset includes information on 17 DNA microsatellite loci assayed in samples from 39 Baltic salmon stocks from six countries, totalling 4453 individuals (Table 2.8.1). In all, 1275 samples were analysed from 2016 salmon catches, and DNA results was received from 1260 fish. Because smolt-age information was used for stock proportion estimation, the fish in the catch samples were divided into two smolt age classes according to the smolt age information from scale reading: younger,1 2-year old smolts, and older, 3 5 year old smolts. As all released hatchery smolts are younger than three years, salmon in catch samples with a smolt age of older than two years originated presumably, or a priori, from any of the wild stocks, whereas individuals with a smolt age of one or two years might have originated either from a wild or a reared stock. This same assumption is used in scale reading as well, when defining wild and reared fish. Correspondingly, smolt age distributions were needed for all baseline stocks in addition to genetic data (Table 2.8.5). The smolt age distributions for Tornionjoki wild, Kalixälven, Råneälven and Simojoki were updated to correspond to the mean distribution of smolt year classes from 2013 to 2015, of which the catches of adult salmon in 2016 were mainly composed. For the other stocks an average of smolt ages over the years was used (Table 2.8.5). Results In 2016, trends in salmon stock proportion changes over time differed somewhat in the three studied sea areas (four samples). A marked increase in the proportion of wild stocks could be seen in the Bothnian Bay Swedish catches and in the Main Basin catches (Table 2.8.2). Whereas the share of most northern wild stocks in assessment unit (AU) 1 have remained stable, the more southern Swedish wild stocks in AU2 had increased in their relative abundance, which could be seen especially in those catches (Table ). These more southern wild stocks don t occur in Finnish Bothnian Bay catches, and the stock composition in Finnish catches had thus remained very similar from 2015 to A partly different trend could be seen in the Åland Sea catches, as an increase in the proportion of reared salmon was observed there. In particular, an increase for Swedish hatchery salmon could be seen in the Åland Sea catches, possibly because several of these reared stocks don t migrate so far north in the Bothnian Bay. In the Main Basin catches, the share of wild salmon stocks increased from about 76% in 2015 to 81% in 2016 (Table ). The survival from smolts to maturing adults has generally improved in recent years, partly because of released fishing pressure, especially in the Main Basin area. Therefore, an improvement in sea survival for both wild and hatchery fish, paralleled by a

30 ICES WGBAST REPORT gradual increase in wild smolt production, seems as a possible reason for the observed increases of stock proportions in different areas. In Åland Sea, there is no temporal fishing regulation of the coastal fishery, which is targeting bypassing spawning migrating salmon of the year. Spawners do not gather at river mouths here, like in the Bothnian Bay area. The proportions of wild salmon stocks have been regularly high in this early season Finnish fishery. The proportion of wild stocks reached its peak in 2011, with 92% (88 95%) (Table 2.8.2, Figure 2.8.2). In 2016 the proportion of wild Bothnian Bay stocks was significantly lower 81%, (76 86%). The proportion of Finnish reared stocks has remained stable and at a quite low level (6 7%) during last five to seven years in these Åland Sea catches. The proportion of Swedish reared stocks, however, continued to increase in 2016 from 6% (1 5%) in 2015 to 11% (7 15%). This increase came from Luleälven, Ångermanälven and especially from Indalsälven salmon (Table 2.8.3). The most abundant salmon stocks contributing to the Åland Sea catches over the years , have correspondingly been mainly wild stocks: Tornionjoki wild (34%), Kalixälven wild (23%), Tornionjoki hatchery (6%), Iijoki hatchery (6%), Vindelälven wild (5%), Byskeälven wild (4%), Simojoki wild (3%), and Luleälven hatchery (3%)(Table 2.8.3). In the Bothnian Bay, Finnish-Swedish coastal catches were analysed both separately and together as a pooled sample. The stock proportions of Finnish and Swedish catches differ considerably, and it is presently unclear to what extent stock and stock group proportion estimates for the pooled sample are representative for the total coastal Swedish-Finnish fishery, as they can be regarded correct only if sample sizes both countries are the same and they represent the fishing of the representative country (see below). The Finnish coastal catch has usually in practice only comprised wild salmon and Finnish reared stocks. However, in the years 2015 and 2016 a small proportion (3 4%) of Swedish hatchery fish also has occurred in these catches (Table ). In 2016 the proportion of wild stocks was about the same average level (70%; 64 75%) as in 2015, and the pooled share of reared stocks remained at the long-term annual average level, about one third of the local catch, from which 26% (21 32%) came from Finnish released fish (Table 2.8.2). The increase of hatchery fish in the catches from the 2014 year level was a result of increased share of stocks from rivers Iijoki (9%), and especially Oulujoki stocks (17%) in 2016 (Table ). The Swedish hatchery fish came from the Luleälven stock. In the analysed Swedish salmon catch, the proportion of wild stocks increased from the lowest ever 64% (57 70%) in 2015 to 77% (71 82%) in Correspondingly, the share of reared salmon of both Finnish (from 17% to 8%) and Swedish (from 19% to 14%) origin decreased in these Swedish salmon catches (Table ). The increase of wild stocks did not come from the Tornionjoki and Kalixälven as their common share was the same 36%, but from several more southern Swedish wild stocks, Piteälven, Byskeälven, Kågeälven, Sävarån, and Lögdeälven had increased their proportions (Table ). During the years, the overall proportion of wild salmon has systematically been higher in the Swedish (mean 82%) than in the Finnish (mean 70%) coastal catches. In the pooled sample of 2016, the proportion of wild salmon increased markedly from 67% (62 72%) in 2015 to 74% (70 77%) in 2016 (Table 2.8.2, Figure 2.8.2). This 74% is reflecting the overall average level of wild stock component of both Finnish and Swedish catches. In the pooled catch the proportion of Finnish reared catch decreased slightly from the last year high level of 22% to 17% in 2016, and the corresponding decrease for the Swedish reared proportion was only from 11% to 9%.

31 26 ICES WGBAST REPORT 2017 The analysed Finnish salmon Bothnian Bay catch sample is assumed to be representative of the total Finnish coastal catch of the area, as it is resampled from the total salmon catch scale sample distribution. The Swedish sample, however, is not necessarily as representative for the total coastal catch. Before 2013, the coastal Swedish Bothnian Bay sample comprised a mixture of salmon from three traps spread along the coast, whereof two were located close outside Kågeälven and the potential river Moälven. In 2013, the latter two traps were replaced by others located in the same coastal areas but more distant from any river mouth. Hence, comparisons over the years of results for Swedish coastal catch samples before 2013 (and pooled SE-FI samples) are not straightforward. Furthermore, in 2013 and 2014 a detailed MSA-survey of stock composition in the Swedish coastal fishery was carried out (Östergren et al., 2015; Whitlock et al., in prep.). These results revealed that stock compositions were stable across years but differed markedly geographically, with local catches being dominated by salmon originating from the nearest (wild or reared) river. Only in areas more far from any salmon river, a higher number of stocks in more even proportions occurred. In the Main Basin the feeding Gulf of Bothnia salmon are assumed to be mixing relatively randomly, although some migration differences may occur among river stocks. However, the Gulf of Finland salmon stocks are not representatively occurring in these Main Basin catches (they partly stay for feeding in the Gulf of Finland). Over half of the pooled international salmon catch sample has regularly originated from wild Bothnian Bay salmon stocks since 2006 (62 79%; Table 2.8.2) (Figure 2.8.2). Some variation has occurred, however, and a strong increasing trend in wild stock proportion can be seen from 2012, from 63% to until 79% in In 2016 the share of wild stocks reached its new historical maximum level, 79% (74 84%). This increasing trend was similar to that observed in Swedish Bothnian Bay catches, but opposite to the trend in the Åland Sea catches. Partly different stock components are targeted in these two fisheries. In the Main Basin, a wider variety of year classes is represented and in early spring time all fish are still there, as the spawners of the year have not yet left. Bothnian Bay and Åland Sea catches, on the other hand, are composed of maturing salmon caught during their spawning migration. Earlier Carlin tagging studies have also indicated that the wild and reared component have slightly different migration patterns, and the migration distance of the reared fish has been reported to be shorter on average. Temporal regulations of fisheries in various parts of the Bothnian Bay may also cause fisheries to target partly different stock components; in the Åland Sea area all stocks are targeted, whereas more northern coastal fisheries (with a later fishing start date) tends to let wild stocks escape the fishery more than reared stocks, as wild stocks migrate earlier than hatchery stocks. In the analysed Main Basin catches the share of Swedish reared (Gulf of Bothnia) stocks continued to be higher (14%, 10 19%) than that of the Finnish reared stocks 3% (1 6%), which was especially low in The reason for this clear difference in relative proportions of reared salmon from different countries is unclear, but may be partly expected as Swedish releases of smolts into the Gulf of Bothnia in recent years have been ca % higher than the Finnish releases. The Finnish reared stocks may also have a tendency for a shorter feeding migration. When analysing the total Main Basin catches from divided into wild and reared salmon from different assessment units (Table 2.8.4; Figure 2.8.3), the changes from 2015 to 2016 were clearest in relation to the share of wild AU2 salmon stocks. The proportion of wild fish from the AU1 (Tornionjoki, Simojoki, Kalixälven and Råne) has remained at exactly the same quite high level, 58%, from 2014 to However an

32 ICES WGBAST REPORT increase from 15% to 21% in the share of the more southern wild Swedish salmon stocks in AU2, could be seen. From those stocks especially Piteälven, Byskeälven, Kågeälven and Vindelälven increased their share in the Main Basin catches. The same stocks were also more common in the 2016 Swedish Bothnian Bay catches, and they thus show increased abundance in relation to other contributing stocks. 2.9 Management measures influencing the salmon fishery Detailed information on international regulatory measures is presented in the Stock Annex (Annex 3). National regulatory measures are updated quite often, sometimes on a yearly basis, and therefore they are presented below and not in the Stock Annex. National regulatory measures In Denmark no new national regulation measures were implemented in According to regulations for the period (BEK nr 212 af 01/03/ Bekendtgørelse om regulering af fiskeriet i ) the following rules must be followed: all commercial vessels fishing salmon needs to be registered as salmon fishing boats and have a specific permission for the fishery; one vessel is permitted a maximum of 10% of the TAC; discard is not allowed; vessels with a catch of ten or more salmon must notice the Fisheries Inspection before entering the harbour. All salmon and sea trout streams with outlets wider than 2 m are protected by closed areas within 500 m from the mouth throughout the year; otherwise the closure period is four months at the time of spawning run. Estuaries are usually protected by a more extended zone. Gillnetting is not permitted within 100 m of the watermark. A closed period for salmon and sea trout has been established from November to 15 January in freshwater. In the sea this only applies for sexually mature fish. In Estonia no new national regulation measures were implemented in There is an all-year-round closed area of 1000 m radius at the river mouths of the present or potential salmon spawning rivers Purtse, Kunda, Selja, Loobu, Valgejõgi, Jägala, Pirita, Keila, and Vasalemma, and at the river mouths of the sea trout spawning rivers Punapea, Õngu, and Pidula. Since 2011 the closed area for fishing around the river mouth was extended from 1000 to 1500 m for the time period 1 September 31 October for rivers Kunda, Selja, Loobu, Valgejõe, Pirita, Keila, Vääna, Vasalemma and Purtse. In rivers Selja, Valgejõgi, Pirita, Vääna and Purtse, recreational fishery for salmon and sea trout is banned from 15 October to 15 November. The autumn in 2015 was exceptionally dry and the fishing was closed until 25 November. In the case of the most important Estonian sea trout spawning rivers (Pada, Toolse, Vainupea, Mustoja, Altja, Võsu, Pudisoo, Loo, Vääna, Vihterpadu, Nõva, Riguldi, Kolga, Rannametsa, Vanajõgi, Jämaja) a closed area of 500 m is established from 15 August 1 December. In the case of smaller sea trout spawning streams, an area of 200 m radius around the river mouths is closed from 1 September 30 November. Apart from lamprey fishing, no commercial fishery in salmon and sea trout spawning rivers is permitted. In most of the rivers angling with natural bait is prohibited. Besides, only licensed sport fishing is permitted. A closed period for salmon and sea trout sport fishing is established in the rivers Narva, Purtse, Kunda, Selja Loobu, Valgejõgi, Jägala, Pirita, Keila, Vasalemma, and Pärnu from 1 September 30 November, and in other rivers from 1 September 31 October. Exceptions in sport fishing closure are allowed by decree of the Minister of Environment in rivers with a reared (River Narva) or mixed salmon stock (rivers Purtse, Selja, Valgejõgi, Jägala, Pirita and Vääna). Below of dams

33 28 ICES WGBAST REPORT 2017 and waterfalls all kind of fishing is prohibited at a distance of 100 m. In the River Pärnu, below Sindi dam, this distance is 500 m. In Finland no changes in the regulation for coastal fisheries have occurred since In the Main Basin, commercial salmon off-shore fishing is forbidden for Finnish vessels from year In the Gulf of Bothnia salmon fishing is forbidden from the beginning of April to the end of the following dates in four zones: Bothnian Sea (59 30 N N) 16 June, Quark (62 30 N 64 N) 21 June, southern Bothnian Bay (64 00 N N) 26 June and northern Bothnian Bay (65 30 N >) 1 July. Commercial fisherman may, however, start fishing for salmon by two trapnets one week before these dates. In three weeks from the opening, five trapnets per fisherman are allowed. After this, for another three weeks eight trapnets at maximum are allowed per fisherman. Non-professional fisherman may start fishing salmon two weeks after the opening of the commercial fishery by one trapnet at maximum (and only in private water areas). In the terminal fishing area of River Kemi, the salmon fishing may start on 11th June. In the area outside the estuary of River Simojoki salmon fishing may start on 16th July, and outside the estuary of river Tornionjoki on June 25th. In 2017 an individual catch quota system will be implemented in the Finnish salmon fishery. In addition, the time regulation will probably be changed so that limited salmon fishing will be allowed in the early summer. This would mean that commercial fishers could possibly start fishing in May by one trapnet, using only a certain part of their quota. In Germany all EC rules apply to German EEZ waters, whereas some additional measures (e.g. fixed protected areas) are in force within territorial waters. There are two German federal states bordering the German Baltic coast: Schleswig-Holstein (SH) and Mecklenburg Western-Pomerania (MWP). Recreational and commercial (coastal) fishing is under the jurisdiction of the German federal states. Consequently, marine fishing is managed by the federal states with different legislation. Legal regulations for the commercial and recreational salmon fishery of the two German federal states bordering the German Baltic coast cover the following: minimum landing size of 60 cm, closed areas; closed seasons: 1 October 31 December in SH (only coloured fish) and 15 September 14 January in MWP; recreational fishery: a bag limit of 3 salmonids only in MWP. In Latvia no new fisheries regulations were implemented in In the Gulf of Riga salmon driftnet and longline fishing is not permitted. In coastal waters salmon fishing is prohibited from 1 October 15 November. Salmon fishing in coastal waters has been restricted indirectly, by limiting the number of gears in the fishing season. In rivers with natural reproduction of salmon, all angling and fishing for salmon and sea trout are prohibited, with the exception of licensed angling of sea trout and salmon that is permitted in rivers Salaca and Venta in spring time. The daily bag limit is one sea trout or salmon. From 2003, all fisheries by gillnets are prohibited all year round in a 3 km zone around the River Salaca outlet. Fisheries restriction zones were enlarged around the rivers Gauja and Venta from 1 to 2 km in In rivers Daugava and Bullupe (that connects rivers Lielupe and Daugava), angling and commercial fishing of salmon was allowed from However, gillnets are not allowed for fisheries in these rivers. Thus, the salmon resources are not utilised or utilised by illegal fisheries. In Lithuania no new fisheries regulations were implemented in The commercial fishery is regulated during salmon and sea-trout migration in the Klaipėda strait and the Curonian lagoon. Fishing is prohibited the whole year round in the Klaipėda strait; from the northern breakwater to the northern border of the 15th fishing bay. From 1

34 ICES WGBAST REPORT September 31 October, during the salmon and sea trout migration, fishing with nets is prohibited in the eastern stretch of Curonian lagoon between Klaipėda and Skirvytė, at a 2 km distance from the eastern shore. In the period 15 September 31 October, recreational fishing is prohibited within a 0.5 km radius from Šventoji and Rėkstyne river mouths, and from the southern and northern breakwaters of Klaipėda Strait. During the same period commercial fishing is prohibited within a 0.5 km radius from Šventoji River mouth, and 3 km from the Curonian lagoon and Baltic Sea confluence. From October 1st to December 31st all types of fishing are prohibited in 161 streams, because of brown trout and sea trout spawning. In larger rivers such as Neris and Šventoji (twelve rivers in total) special protect zones have been selected where schooling of salmon and sea trout occurs. In these selected zones, licensed fishing is only permitted from September 16th to October 15th. From October 16th to December 31st any kind of fishing is prohibited in these areas. From January 1st, licensed salmon and sea trout kelt fishing is permitted in the Minija, Veiviržas, Skirvytė, Jūra, Atmata, Nemunas, Neris, Dubysa, Siesartis and Šventoji river. Licence fishing is allowed from January 1st to October 1st in designated stretches of the listed rivers. The minimum landing size of salmon is 60 cm for commercial fishery and 65 cm for angling. In Poland, in addition to EC measures, seasonal closures and fixed protected areas are in force within territorial waters managed by Regional Fisheries Inspectorates. Fishing for salmon and sea trout is not allowed between 15 September and 15 November in a 4-mile belt of the coastal zone. Since 2005 the commercial fisheries for salmon/sea trout in rivers is based on new implemented rules. Fisheries opportunities were sold in 2005 by the state on a tender basis, where the bidder had to submit a fishing ten-year operational plan including restocking. New rules were recently introduced for recreational salmon fishing in Polish EEZ (Law on Fisheries of 19 December 2014). These include: catch quotas (per day/per angler), minimum size limits (TL.), protection periods and areas for fish species, minimum distance between anglers. Rod fishing (coastal fishing, boat/belly boat fishing, and organized cruises on board fishing vessels) and spearfishing is allowed. Recreational fishing with nets is not allowed. A new system of obtaining fishing licences is established. Currently, proof of bank transfer with a specified personal data is needed for legal fishing. The licence can be issued for a period of one week, one month or one year. In case of rivers Ina, Reda, Parseta and Słupia the anglers must release all fish that has been caught. In Russia no changes in the national regulations were implemented in the period The international fishery rules are extended to the coastline. In all rivers, and within one nautical mile of their mouths, fishing and angling for salmon is prohibited during all year, except fishing for breeding purposes for hatcheries. In Sweden, as in recent years, the bulk of the salmon quota for the commercial fisheries in 2016 was allocated to the coastal fishery, as the Swedish off-shore fishery targeting salmon and sea trout was phased out in Management measures for salmon include a minimum landing size of 60 cm. Starting dates of the Swedish commercial coastal fishing season in 2016 were the same as in South of latitude 62º55 N, commercial coastal fishing in 2016 was allowed from 1 April. North of latitude 62º55 N the season started 17 June. Exemptions from this seasonal regulation of the salmon fishery was allowed by the local county board to professional fishermen in the area north of latitude 62º55 N up to the border between the counties Västerbotten and Norrbotten, so that the fishery could start 12 June with a maximum of two trapnets per fisherman. Status and development of weak salmon populations should be taken into consideration in decisions about eventual exemptions. The aim of the early summer ban in the

35 30 ICES WGBAST REPORT 2017 coastal fishery is to ensure that a part of the migrating population of adult salmon ascend rivers before the fishing season starts. The total Swedish quota in 2016 ( salmon) was divided between subdivisions with the aim of increasing the exploitation of reared salmon and reducing the exploitation of weak wild salmon populations. In SD 31 the regional quota was set to salmon. Among these, not more than were allowed to be not finclipped (wild) salmon. When the limit of not finclipped salmon was reached, only fishing for finclipped (reared) salmon in terminal fishing areas outside reared rivers, and in the restriction area outside River Umeälven, was allowed. In SD 30 the regional quota was set to salmon. In SD the regional quota was set to 350 salmon because of the higher expected proportion of salmon from weaker populations in these catches (as compared to SD 30 and 31). The remaining 1020 salmon were not allocated to any specific area, but were reserved for bycatch in fisheries targeting other species and as a buffer if catches in any area would exceed the quota. Summarizing the 2016 commercial fishing season, catches of salmon exceeded the regional quotas in both SD 31 and 22 29, and the total Swedish commercial salmon catch exceeded the Swedish salmon quota by approximately 2500 salmon. The overdraft was solved by swapping quotas with other EU Member States. Recreational fisheries in both the sea and in rivers are managed through national regulations. Coastal recreational fisheries with trapnets in the counties of Norrbotten, Västerbotten and part of Västernorrland were allowed from 1 July until the quota of salmon within the commercial fishery was fulfilled. In SD31, the regional salmon quota was fulfilled already in end of June and the fishery was terminated 30 June. Hence, no recreational salmon fishing with trapnets could be conducted in SD31 in In SD30, the commercial salmon fishery continued longer but was stopped 18 July due to a fulfilled quota. Since 2013, Swedish off-shore trolling fisheries (that mainly takes place in the Main Basin) are only allowed to land salmon without an adipose fin (i.e. finclipped reared salmon). In all rivers there is a general bag limit of one salmon and one sea trout per fisherman and day. Also fishing periods are regulated on a national level. In Gulf of Bothnian wild rivers, for example, angling for salmon is forbidden from 1 September until 31 December, and in some rivers angling is also forbidden between 1 May and 18 June. In addition to national regulations, local fishing and management organizations may stipulate more restrictive fishing regulations. The management of fisheries in River Torneälven, including the coastal area directly outside the river mouth, is handled through an agreement between Sweden and Finland. The Swedish-Finnish agreement includes, for example, a specified time period in which the commercial coastal fishery in the river mouth is allowed to start. Regulations targeting the river fishery are also handled in the agreement. Deviations from the agreed fishing regulations in this area are negotiated and decided upon on an annual basis by the Swedish Agency for Marine and Water Management (according to a Government commission from the Swedish Ministry of Enterprise and Innovation) and the Finnish Ministry of Agriculture and Forestry. In order to improve the situation for the weak sea trout stocks in SD 31, a number of changes have been implemented in recent years. The minimum size for landed sea trout was raised from 40 to 50 cm in the sea from 1 July Furthermore, a ban of fishing with nets in areas with a depth of less than 3 meters during the period 1 April 10 June and 1 October 31 December was implemented in order to decrease the bycatch

36 ICES WGBAST REPORT of trout in fisheries targeting other species (e.g. whitefish). Further restrictions for rivers in Bothnian Bay (SD 31) were adopted in 2013, including shortening of the autumn period for fishing with two weeks, restrictions of catch size (window size of cm), and landing of only one trout per fisherman and day. In River Torneälven, trout fishing is completely banned since Effects of management measures International regulatory measures Minimum landing size No change in the measures has occurred since An evaluation of the effects of the minimum landing size and minimum hook size was provided by ICES (2000). However, the changes in the regulatory measures in the EC waters (Council Regulation (EC) 2187/2005) might have changed the situation compared to the years before enforcement of this regulation. The minimum landing size in the Baltic salmon fishery is 60 cm, except for on the Finnish side of SD 31, where it has been decreased from 60 cm to 50 cm. An evaluation of this change was provided in ICES (2007). The minimum hook size for longlining in EC Baltic Sea waters is 19 mm. Longlines do not have the same pronounced size selectivity as for driftnets, thus the minimum landing size in the off-shore fishery is important. Landing obligation Discarding refers to the practice of returning unwanted catch, dead or alive, back to the sea. During autumn 2014, the European Commission decided to introduce a discard ban for commercial fisheries, covering all species under TACs including salmon (Commission Delegated Regulation (EU) No 1396/2014 of 20 October 2014, establishing a discard plan in the Baltic Sea). The aim of this landing obligation is to stop the wasteful practice of discarding, promote the development of more selective fishing gears and to increase the quality of catch data. The landing obligation will be implemented gradually between 2015 and 2019, and salmon fisheries in the Baltic Sea are subjected to the new rules from 1 January Under the landing obligation, all salmon caught in commercial fisheries targeting salmon should be landed, reported and counted against the quotas. The landing obligation also covers salmon caught as bycatch in commercial fisheries targeting other species. The impact of the landing obligation on salmon fisheries varies depending on geographical location and gear type used. In the longline fishery in the Main Basin, the share of undersized (unwanted) salmon has been estimated to be around 2 3 percent (Table 2.3.1), which means that the landing obligation has only had minor effects on this fishery. For countries that have applied for an exemption, the coastal fisheries using salmon traps (and a few other gears) are temporally exempted from the discard ban. The exemptions are valid as long as the discard plan in the Baltic CFP region is valid, i.e. until 31 December Sweden has applied for and received a temporally exemption from the discard ban for salmon and cod caught in traps (and a few other gears) in the Baltic. The exemption applies as long as the current Baltic discard plan is valid, which is until 31st of December At the moment, possibilities to extend this exemption from 2018 and onwards are evaluated. Salmon caught with trapnets are considered to have comparably high survival possibilities after release back into the sea. The knowledge about long-term survival rate of salmon after release from trapnets is, however, limited and further investigations are

37 32 ICES WGBAST REPORT 2017 urgently needed. Therefore, when the countries apply for a new exemption, the Commission will require potential new information on survival rate. Pilot studies aimed at exploring the survival of discarded salmon in coastal fisheries, and studies where new technical solutions are tested to make the gears more selective and/or less harmful for the fish, have recently been performed or are initiated/planned. A future discard ban that involves also trapnet fisheries would probably affect the coastal exploitation pattern of both salmon and other species. The estimated share of undersized salmon in coastal fisheries with traps is low (1 5%), so a discard ban will not have any major impacts on the total amount of salmon caught. However, the possibility of releasing wild salmon back into the sea as a measure to steer the exploitation towards reared (finclipped) salmon will disappear. Also, under a discard ban, trapnet fisheries targeting other species may have to be more strongly regulated than today if salmon are taken as bycatch (and must be counted against the quota), but this effect may be overcome by the development of selective gears that minimizes the bycatch of salmon. Summer closure and maximum hook number Since 2013 only Danish and Polish longline fleets operate in SD The previous summer closure for this fishery had a small effect, since longlining with a high cpue is possible only during the winter months (from November/December to February or possibly March). The rule concerning a maximum number of hooks per vessel (previously 2000) has also been dropped from the EC Council regulation. This measure might contribute to an increased fishing effort by longlining. As longline fishery is very labour intense, it is not possible to increase the number of hooks so much. In addition some of the boats involved in longline fishery are small and they do not have capacity to use more than 2000 hooks. TAC An evaluation of the TAC regulation can be found in the Stock Annex (Annex 3). Driftnet ban In the northern feeding areas, Bothnian Sea (SD 30) and Gulf of Finland (SD 32), offshore fishing with longlines would be theoretically possible with small boats and a small crew (1 2), but seals and a busy ship traffic practically prevent such fishing in these areas. Besides, this fishery was banned Sweden and Finland in The present off-shore fishing of salmon (currently only Denmark and Poland) takes place in the most southern part of the Baltic Main Basin. Previously important fishing took place also in the northern Baltic Sea at the Gotland Deep, and in the Bothnian Sea and Gulf of Finland. Fishermen have reported that densities of feeding salmon have been low in northern areas, and therefore they have switched to more southern fishing areas where catches are higher. The reason for the appearance of feeding salmon mostly in the areas of Bornholm deep and Gdansk deep is unknown. The share of undersized salmon in catches is most likely to be larger in off-shore longline fishery than in the past driftnet fishery, mostly due to higher selectivity of driftnets. In the Danish off-shore fishing in undersized salmon in longline catches varied between 1.7% and 20.3% (mean 11.5%), whereas in the driftnet catch, the mean percentage of salmon smaller than 60 cm was 3% (ICES, 2003b). Likewise, in Polish catch samples from the Main Basin longline fishery in , the proportion of undersized salmon was 1.5 4%. In fact, small salmon in longline catches is not a new finding, although small salmon have often been classified as sea trout. According to Järvi (1938),

38 ICES WGBAST REPORT for example, Polish salmon catches from the 1930s could be dominated by small salmon (post-smolts with an average weight of about 0.5 kg). Also Alm (1954) discussed the catches of small salmon with longlines in the Baltic Sea, and suggested that this fishery should be prohibited in winter (December March) because of the high proportion of post smolts in the catches during that time of the year. In summary, catch of undersized salmon in the present longline fishery may be noticeably, although additional information is needed on how it varies in time and space. Furthermore, the survival rate of undersized salmon that have been released from hook and put back to sea is not much known. However, Polish data from indicate that 20 30% of undersized released fish was alive. Without information on how large proportion of released salmon that actually survives, it will be impossible to gauge the effect of this type of discard with respect to stock assessment and in terms of reduced catch options (i.e. by not catching the fish later in life, when it has grown larger). Therefore studies on survival would be important. In addition on-board sampling is important to obtain further data on discards of undersized salmon. Delayed opening of the coastal salmon fishery ICES (2007) concluded that the delayed opening of the coastal salmon fishery is an effective measure for saving a proportion of the spawning run from being harvested. However, the run timing varies between years, which mean that with multi-annually fixed opening dates, the saved proportion of the spawning run is highly variable. This regulatory measure results in a higher harvest rate of late-migrating than early-migrating salmon (ICES 2007). As older fish and females dominate in the early part of the spawning run, a late opening of the fishery saves the most valuable part of the run. Also, it can be possible to direct the fisheries towards either wild or reared fish by changing the opening date, since a difference in the migration timing has been observed depending on origin (wild salmon tend to arrive somewhat earlier than reared) Other factors influencing the salmon fishery The incitement to fish salmon as an alternative to other species is likely to be influenced by a number of factors, such as the possibilities for selling the fish, problems with damage to the catches from seals, the market price for salmon compared to other species and possibilities to fish on other species. Dioxin The maximum level of dioxin and dioxin-like PCB set for the flesh from salmon will be 8 pg WHO-PCDD/F-PCBTEQ (COMMISSION REGULATION (EC) No 1881/2006 and 1256/2011). Overall levels of dioxin and related substances tend to increase with size (sea age) of the salmon, but varies also in different parts of the fish flesh with fat contents (Persson et al., 2007). In general, the levels found are above the maximum EU level. Sweden, Finland and Latvia have derogations in the regulation allowing domestic use of the salmon, if dietary advice is given to the public. The derogations are not timelimited. Export of salmon to other EU countries is not permitted. In Sweden the latest published report show elevated levels of dioxin in Baltic salmon caught along the coast (Fohgelberg and Wretling, 2015). In Denmark salmon above 5.5 kg (gutted weight) were not permitted to be marketed within the EU. However, from 9 February 2009 it has been allowed to land and sell (to countries outside the EU) all size groups of salmon. In March 2011 deep skinned salmon were analysed and a general decrease in

39 34 ICES WGBAST REPORT 2017 the dioxin content was observed. The latest results (2013) shown high levels of dioxins, comparable to levels from While there is no information available from Germany, Polish samples of salmon were examined in 2005, 2006 and again in The results from these have not resulted in any marketing restrictions.

40 ICES WGBAST REPORT Table Nominal catches, discards (incl. seal damaged salmon) and unreported catches of Baltic Salmon in tonnes round fresh weight, from sea, coast and river by country in in sub-division The estimation method for discards and unreported catches are different for years and ( 95% PI = probability interval) Year Reported catches by country Reported catches Non commercial catch. Discard Total unreported catches 3) Total catches Denmark Es tonia Finland Germany Latvia Lithuania Poland 1) Russia Sweden USSR total included in tot. catch. median 90% PI median 90% PI median 90% PI na na na 13 na na na na na na na na na 17 na na na na na na na na na 20 na na na na na na na na na 10 na na na na na na na na na 7 na na na na na na na na na 6 na na na na na na na na na 4 na na na na na na na na na 4 na na na na na na na na na 22 na na na na na na )

41 36 ICES WGBAST REPORT All data from includes sub-divisions 24-32, while it is more uncertain in which years sub-divisions are included. The catches in sub-divisions are normally less than one ton. From 1995 data includes sub-divisions Catches from the recreational fishery are included in reported catches as follows: Finland from 1980, Sweden from 1988, Denmark from Other countries have no or very low recreational catches. Danish, Finnish, German, Polish and Swedish catches are converted from gutted to round fresh weight w by multiplying by 1.1. Estonian, Latvian, Lithuanian and Russian catches before 1981 are summarized as USSR catches. Estonian, Latvian, Lithuanian and Russian catches are reported as whole fresh weight. Sea trout are included in the sea catches in the order of 3 % for Denmark (before 1983), 3% for Estonia, Germany, Latvia, Lithuania, Russia, and about 5% for Poland (before 1997). Estimated non-reported coastal catches in Sub-division 25 has from 1993 been included in the Swedish statistics. Danish coastal catches are non-profesional trolling catches. 1) Polish reported catches are recalculated for assessment purposes (see Section 5) 2) In 1993 fishermen from the Faroe Islands caught 16 tonnes, which are included in total Danish catches. 3) Including both unreporting for all countries and the estimated additional Polish catch

42 ICES WGBAST REPORT Table Nominal catches, discards (incl. seal damaged salmon) and unreported catches of Baltic Salmon in numbers from sea, coast and river by country in Sub-divisions The estimation method for discards and unreported catches are different for years and ( 90% PI = probability interval) Year Country reported Discard Estimated Polish Total unreported catches 2) Total catches Denmark Es tonia Finland Germany Latvia Lithuania Poland Russia Sweden total median 90% PI misreported catch median 90% PI median 90% PI ) All data from , includes sub-divisions 24-32, while it is more uncertain in which years sub-divisions are included. The catches in sub-divisions are normally less than one tonnes. From 1995 data includes sub-divisions Catches from the recreational fishery are included in reported catches as follows: Finland from 1980, Sweden from 1988, Denmark from Other countries have no, or very low recreational catches. 1) In 1993 Fishermen from the Faroe Islands caught 3200 individuals, which is included in the total Danish catches. 2) Including both unreporting for all countries and the estimated additional Polish catch

43 38 ICES WGBAST REPORT 2017 Table Nominal catches of Baltic Salmon in tonnes round fresh weight, from sea, coast and river by country and region in S=sea, C=coast, R=river. Main Basin (Sub-divisions 22-29) Year Denmark Finland Germany Poland Sweden USSR Total S S+C S S S R S C+R S C+R GT Main Basin (Sub-divisions 22-29) Year Denmark Estonia Finland Germany Latvia Lithuania Poland Russia Sweden Total S C S C S C R S S C R S C R S C R S C R S C R S C R GT * na na 45 na na * na na 38 na na * na na 76 na na * na na 72 na na * na na 162 na na * na na 137 na na * na na 267 na na * na na 93 na na * na na 80 na na * na na 195 na na * na na 77 na na * na na 170 na na * na na 191 na na * na na 184 na na * na na na * na na na na * na na na

44 ICES WGBAST REPORT Table Continued Gulf of Bothnia Main Basin+Gulf of (Sub-divisions 30-31) Bothnia (Sub-divs. Year Denmark Finland Sweden Total 22-31) Total S S S+C C S C R S C R GT S C+R GT Gulf of Bothnia Main Basin + Gulf of ( Sub-divisions 30-31) Bothnia (Sub-divisions Year Finland Sweden Total 22-31) Total S C R S C R S C R GT S C R GT )

45 40 ICES WGBAST REPORT 2017 Table Continued Gulf of Finland (Sub-division 32) Sub-division Year Finland USSR Total S S+C C S C+R S C+R GT Gulf of Finland (Sub-division 32) Year Estonia Finland Russia Total Sub-division Total S C R S C R C R S C R GT S C R GT ) All data from , includes sub-divisions 24-32, while it is more uncertain in which years sub-divisions are included. The catches in sub-divisions are normally less than one tonnes. From 1995 data includes Catches from the recreational fishery are included as follow s: Finland from 1980, Sw eden from 1988, Denmark from Other countries have no, or very low recreational catches. Danish, Finnish, German, Polish and Sw edish catches are converted from gutted to round fresh w eight w by multiplying by 1.1. Estonian, Latvian, Lithuanian and Russian catches before 1981 are summarized as USSR catches. Estonian, Latvian, Lithuanian and Russian catches are reported as hole fresh w eight. Sea trout are included in the sea catches in the order of 3 % for Denmark (before 1983), 3% for Estonia, Germany, Latvia, Lithuania, Russia, and about 5% for Poland (before 1997). Estonian sea catches in Sub-division 32 in include a small quantity of coastal catches. Estimated non-reported coastal catches in Sub-division 25 has from 1993 been included in the Sw edish statistics. Danish coast catches are non-profesional trolling catches. 1) In 1993 fishermen from the Faroe Islands caught 16 tonnes, which are included in total Danish catches.

46 ICES WGBAST REPORT Table Nominal catches of Baltic Salmon in numbers, from sea, coast and river by country and region in S=sea, C=coast, R=river. Main Basin (Sub-divisions 22-29) Year Denmark Estonia Finland Germany Latvia Lithuania Poland Russia Sweden Main Basin (sub-divisions 22-29) Total S C S C S C R S S C R S C R S C R S C S C R S C R GT

47 42 ICES WGBAST REPORT 2017 Gulf of Bothnia ( Sub-divisions 30-31) Main Basin + Gulf of Bothnia Year Finland Sweden Total (Sub-divisions 22-31) Total S C R S C R S C R GT S C R GT

48 ICES WGBAST REPORT Gulf of Finland (Sub-division 32) Year Estonia Finland Russia Total S C R S C R C 1) R S C R GT S C R GT Data from the recreational fishery are included in Swedish and Finnish data. Recreational fishery are included in Danish data from Other countries have no, or very low recreational catches. In 1996 sea trout catches are included in the Polish catches in the order of 5%. 1) Russian coastal catches have in earlier reports been recorded as sea catches. Sub-divisions Total

49 44 ICES WGBAST REPORT 2017 Table Nominal catches of Baltic salmon in tonnes round fresh weight and numbers from sea, coast and river, by country and sub-divisions in Subdivisions (SD 200=SD 22-29; SD 300=SD 30-31). S=sea, C=coast, R=river, W=weight (tonnes), N=number of fish SD Fishery - DE DK EE FI LT LV PL RU SE Grand Total 200 S W N S W 1 1 N S W N C W 0 0 N S W N C W N R W N S W N C W N R W N C W 0 0 N 6 6 R W 0 0 N C W N R W 1 1 N S W N C W N R W 0 0 N S W 0 0 N C W N R W N S W 0 0 N 3 3 C W N R W N S W 0 0 N C W N R W N C W N Total S+C+R W N Total 32 S+C+R W N S W N C W Grand Total N R W N S+C+R W N

50 ICES WGBAST REPORT Table Non-commercial catches of Baltic Salmon in numbers from sea, coast and river by country in in sub-divisions and sub-division 32. (S = Sea, C = Coast, PI = probability interval). Sub-divisions Year Denmark Estonia Finland Germany Latvia Lithuania Poland Russia Sweden S+C River Grand S+C S+C S+C (95% PI) River S+C S+C River S+C River S+C River S+C River S+C River Total Total Total 1997 na na na na na na na na na 0 na na na 0 na na na 9040 (±6370) 5100 na na na na na na 0 na na na na 9040 (±6370) 400 na na (±5490) 4150 na na na (±5490) 3750 na na 3640 (±1070) 3925 na na 3640 (±1070) 4525 na na (±7300) 5950 na na (±7300) 6725 na na 6180 (±3710) 2640 na na 6180 (±3710) 3590 na (±4380) na na 9090 (±4380) 7934 na na 3270 (±3600) 4910 na na 3270 (±3600) 5475 na na 3090 (±2830) na (±2830) na (±5450) na (±5450) na (±5450) na

51 46 ICES WGBAST REPORT 2017 Sub-division 32 Sub-division Year Estonia Finland Russia S+C River Grand S+C River GT S+C River S+C (95% PI) River S+C River Total Total Total Total Total 1997 na na na na na na na na na 5150 (±3630) 5100 na na (±3630) na (±5780) na (±5780) na 2550 (±750) na 2550 (±750) na 3090 (±1430) (±1430) (±110) (±110) (±350) (±350) (±400) (±400) (±3170) (±3170) (±3270) (±3270) (±3270)

52 ICES WGBAST REPORT Table Nominal catches (commercial) of Baltic Salmon in numbers from sea and coast, excluding river catches, by country in and in comparison with TAC. comparison with TAC.Subdivisions Years include also sea catch of the recreational fishery in Sweden and Finland. Baltic Main Basin and Gulf of Bothnia (Sub-divisions 22-31) Year Fishing Nation Total TOTAL Landing of Denmark Estonia Finland Germany Latvia Lithuania Poland Russia Sweden TAC TAC (in %) ,

53 48 ICES WGBAST REPORT 2017 Table Continued Gulf of Finland (Sub-division 32) Year Fishing Nation Total EC Landing of Estonia Finland TAC TAC (in %) Russia All data from , includes sub-divisions 24-32, while it is more uncertain in which years sub-divisions are included Russia are not included in the TAC in Sub-division 31. The catches in sub-divisions are normally less than one tonnes. From 1995 data includes sub-divisions Estonia: Offshore catches reported by numbers, coastal catches converted from weight. Catches from the recreational fishery are included as follows: Finland from 1980, Sweden from 1988, and Denmark from Other countries have no, or very low recreational catches. Estimated non-reported coastal catches in sub-division 25, have from 1993 been included in the Swedish catches. Sea trout are included in the sea catches in the order of 5% for Poland before ) In 1993 Polish, Russian and Faroe Islands numbers are converted from weight. 2) In 1993 Fishermen from Faroe Islands caught 3100 salmons included in the total Danish catches. 3) In 1998 German numbers are converted from weight.

54 ICES WGBAST REPORT Table Summary of the uncertainty associated to fisheries data series according to the expert opinions from different countries backed by data (D) or based on subjective expert estimation (EE). The conversion factors (mean) are proportions and can be multiplied with the nominal catch data in order to obtain estimates for unreported catches and discards, which altogether sum up to the total catches. Driftnet fishing has been closed from Finland and Sweden have had no off-shore fishing for salmon after Parameter Country Year Source min mode max mean SD DK EE FI EE Share of unreported catch in offshore fishery EE PL 2014 EE EE SE EE EE FI EE EE PL Share of unreported catch in coastal fishery EE SE EE SE EE SE EE FI Share of unreported catch in river fishery PL EE EE SE EE Average share of unreported catch in river fishery Others DK D, EE D, EE FI D, EE Share of discarded undersized salmon in longline fishery D PL D SE D, EE Average share of discarded undersized salmon in longline fisheryothers DK EE Mortality of discarded undersized salmon in longline fishery FI EE SE EE PL D, EE Average mortality of discarded undersized salmon in longline fis Others Share of discarded undersized salmon in driftnet fishery DK EE, D FI D Average share of discarded undersized salmon in driftnet fisheryothers Mortality of discarded undersized salmon in driftnet fishery DK EE, D FI EE Average mortality of discarded undersized salmon in driftnet fishothers Share of discarded undersized salmon in trapnet fishery FI EE SE EE, D Average share of discarded undersized salmon in trapnet fishery Others Mortality of discarded undersized salmon in trapnet fishery FI EE, D SE EE, D Average mortality of discarded undersized salmon in trapnet fishothers FI D D SE EE, D EE, D Share of discarded sealdamaged salmon in longline fishery EE DK EE, D EE D PL D EE, D Share of discarded sealdamaged salmon in driftnet fishery Share of discarded sealdamaged salmon in trapnet fishery DK FI EE, D D FI SE D EE, D

55 50 ICES WGBAST REPORT 2017 Table Estimated number of discarded undersized salmon and discarded seal damaged salmon by management unit in Estimates of discarded undersized salmon are proportional to nominal cathes by the conversion factors (see Table 2.3.1). Estimates of seal damages age based partly on the logbook records (Finland and Sweden) and partly to the estimates proportional to nominal catches by conversion factors. Estimates should be considered as a magnitude of discards. Discard undersized Discard seal damaged Management unit Year Driftnet Longline Trapnet Other gears Driftnet Longline Trapnet Other gears Total Disc_GND Disc_LLD Disc_TN Disc_OT Seal_GND Seal_LLD Seal_TN Seal_OT SD * * Disc_GND Disc_LLD Disc_TN Disc_OT Seal_GND Seal_LLD Seal_TN Seal_OT SD * * * There is a discard ban since year 2015, but data from reports of caught undersized fish are not available

56 ICES WGBAST REPORT Table Number salmon and sea trout in the catch of sampled Polish long-line vessels in (SAL=salmon and TRS=sea trout). SamplingType Year Month Trip_id SAL TRS % SAL Sea sampling % % % % % % % 2009 Total % % % % % % 2010 Total % % % % % % % 2011 Total % % % % % 2012 Total % % % % % % % % % % 2013 Total % 2014 Total % 2015 Total % 2016 Total % % % % % Sea sampling Total % Market sampling % 2009 Total % % 2010 Total % Market sampling Total % Grand Total %

57 52 ICES WGBAST REPORT 2017 Table Estimated number of seal damaged salmon, dead discard of undersized salmon, unreported salmon in sea and river fisheries and misreported salmon by country and management unit in Estimates should be condered as order of magnitude. Denmark Estonia Finland Germany Latvia Lithuania Poland Russia Sweden Seal damage Discard Unreported sea Seal damage Discard Unreported sea Unreported river Seal damage Discard Unreported sea Unreported river Seal damage Discard Unreported sea Seal damage Discard Unreported sea Unreported river SD SD Seal damage Discard Unreported sea Unreported river Seal damage Discard Unreported sea Unreported river Misreporte d Seal damage Discard Unreported sea Unreported river Seal damage Discard Unreported sea Unreported river

58 ICES WGBAST REPORT Table Fishing efforts of Baltic salmon fisheries at sea and at the coast in in subdivision (excluding Gulf of Finland). The fishing efforts are expressed in number of geardays (number of fishing days times the number of gear) per year. The yearly reported total offshore effort refers to the sum of the effort in the second half of the given year and the first half of the next coming year (e.g., effort in second half of effort in first half of 1988 = effort reported in 1987, etc.). The coastal fishing effort on stocks of assessment unit 1 (AU 1) refers to the total Finnish coastal fishing effort and partly to the Swedish effort in subdivision (SD) 31. The coastal fishing effort on stocks of AU 2 refers to the Finnish coastal fishing effort in SD 30, and partly to the Swedish coastal fishing effort in SD 31. The coastal fishing effort on stocks of AU 3 refers to the Finnish and Swedish coastal fishing effort in SD 30. AU 1 AU 2 AU 3 Offshore Offshore Commercial Commercial Commercial Commercial Commercial Commercial Commercial Year driftnet longline coastal coastal coastal coastal coastal coastal coastal driftnet trapnet other gear trapnet other gear trapnet other gear

59 54 ICES WGBAST REPORT 2017 Table Number of fishing vessels in the offshore fishery for salmon by country and area from Number of fishing days divided in 4 groups, 1-9 fishing days, fishing days, fishing days and more than 40 fishing days (from 2001 also and > 80 days, total 6 groups). Sub-divisions and Sub-division 32. Year Area Country Effort in days per ship > Total Number of fishing vessels 1999 Sub-divisions Denmark Estonia na na Finland Germany na na na na na Latvia Lithuania na na na na na Poland Russia Sweden Total Sub-div. 32 Finland Sub-divs Total Year Area Country Effort in days per ship > Total Number of fishing vessels 2000 Sub-divisions Denmark Estonia Finland Germany na na na na na Latvia Lithuania na na na na na Poland Russia na na na na na Sweden Total Sub-div. 32 Estonia Finland Sub-divs Total Year Area Country Effort in days per ship >80 days Total Number of fishing vessels 2001 Sub-divisions Denmark Estonia Finland Germany na na na na na na na Latvia Lithuania na na na na na na na Poland Russia na na na na na na na Sweden Total Sub-div. 32 Finland Sub-divs Total

60 ICES WGBAST REPORT Continued Year Area Country Effort in days per ship >80 days Total Number of fishing vessels 2002 Sub-divisions Denmark Estonia Finland na na na na na na 0 Germany na na na na na na na Latvia Lithuania na na na na na na 0 Poland na na na na na na 50 Russia na na na na na na 0 Sweden Total Sub-div. 32 Finland Sub-divs Total Year Area Country Effort in days per ship >80 days Total Number of fishing vessels 2003 Sub-divisions Denmark Finland Germany na na na na na na na Latvia Lithuania na na na na na na 0 Poland Russia na na na na na na 0 Sweden Total Sub-div. 32 Estonia Finland Sub-divs Total Year Area Country Effort in days per ship >80 days Total Number of fishing vessels 2004 Sub-divisions Denmark Finland Germany n.a. n.a. n.a. n.a. n.a. n.a. n.a. Latvia Lithuania Poland Russia na na na na na na n.a. Sweden Total Sub-div. 32 Estonia Finland Sub-divs Total

61 56 ICES WGBAST REPORT 2017 Continued Year Area Country Effort in days per ship >80 days Total Number of fishing vessels 2005 Sub-divisions Denmark Finland Germany na na na na na na na Latvia Lithuania Poland Russia na na na na na na na Sweden Total Sub-div. 32 Estonia na na na na na na na Finland Sub-divs Total Year Area Country Effort in days per ship >80 days Total Number of fishing vessels 2006 Sub-divisions Denmark Finland Germany na na na na na na na Latvia Lithuania Poland na na na na na na na Russia na na na na na na na Sweden Total Sub-div. 32 Estonia na na na na na na na Finland Sub-divs Total Year Area Country Effort in days per ship >80 days Total Number of fishing vessels 2007 Sub-divisions Denmark Finland Germany na na na na na na na Latvia Lithuania Poland na na na na na na na Russia na na na na na na na Sweden Total Sub-div. 32 Estonia na na na na na na na Finland Sub-divs Total

62 ICES WGBAST REPORT Continued Year Area Country Effort in days per ship >80 days Total Number of fishing vessels 2008 Sub-divisions Denmark Finland Germany na na na na na na na Latvia Lithuania Poland Russia na na na na na na na Sweden Total Sub-div. 32 Estonia na na na na na na na Finland Sub-divs Total Continued Year Area Country Effort in days per ship >80 days Total Number of fishing vessels 2009 Sub-divisions Denmark Finland Germany na na na na na na na Latvia Lithuania Poland Russia na na na na na na na Sweden Total Sub-div. 32 Estonia na na na na na na na Finland Sub-divs Total Continued Year Area Country Effort in days per ship >80 days Total Number of fishing vessels 2010 Sub-divisions Denmark Finland Germany Latvia Lithuania Poland Russia Sweden Total Sub-div. 32 Estonia Finland Sub-divs Total

63 58 ICES WGBAST REPORT 2017 Continued Year Area Country Effort in days per ship >80 days Total Number of fishing vessels 2011 Sub-divisions Denmark Finland Germany Latvia Lithuania Poland Russia Sweden Continued Total Sub-div. 32 Estonia Finland Sub-divs Total Year Area Country Effort in days per ship >80 days Total Number of fishing vessels 2012 Sub-divisions Denmark Finland Germany Latvia Lithuania Poland Russia Sweden Total Sub-div. 32 Estonia Finland Sub-divs Total Continued Year Area Country Effort in days per ship >80 days Total Number of fishing vessels 2013 Sub-divisions Denmark Finland Germany Latvia Lithuania Poland Russia Sweden Total Sub-div. 32 Estonia Finland Sub-divs Total

64 ICES WGBAST REPORT Continued Year Area Country Effort in days per ship >80 days Total Number of fishing vessels 2014 Sub-divisions Denmark n.a. n.a. n.a. n.a. n.a. n.a. n.a Finland Germany Latvia Lithuania Poland Russia Sweden Total Sub-div. 32 Estonia Finland Sub-divs Total Year Area Country Effort in days per ship >80 days Total Number of fishing vessels 2015 Sub-divisions Denmark n.a. n.a. n.a. n.a. n.a. n.a. n.a Finland Germany Latvia Lithuania Poland Russia Sweden Total Sub-div. 32 Estonia Finland Sub-divs Total Year Area Country Effort in days per ship >80 days Total Number of fishing vessels 2016 Sub-divisions Denmark n.a. n.a. n.a. n.a. n.a. n.a. n.a Finland Germany Latvia Lithuania Poland Russia Sweden Total Sub-div. 32 Estonia Finland Sub-divs Total

65 60 ICES WGBAST REPORT 2017 Table Catch per unit effort (CPUE), expressed as number of salmon caught per 100 nets and per 1,000 hooks, by fishing season in the Danish, Estonian, Finnish, Latvian, Russian and Swedish offshore fisheries in the Main Basin, in the Gulf of Bothnia, and in the Gulf of Finland from 1980/1981 (Denmark from 1983/84) to Fishing Denmark season Sub-divisions Sub-divisions Driftnet Longline Driftnet Longline 1983/ / na / na / na / / / / / / / / / / / / / / N.A. 0.0 N.A N.A 0.0 N.A.

66 ICES WGBAST REPORT Fishing Finland season Sub-divisions Sub-divisions Sub-division 32 Driftnet Longline Driftnet Longline Driftnet Longline 1980/ na / na / na / na / na / na / na / na / na / na / na / na / na / na / / / / / / /

67 62 ICES WGBAST REPORT 2017 Continued. Fishing Estonia Latvia season Sub-divisions Sub-divisions and 28 Russia Sub-division Sweden Sub-divisions Driftnet Driftnet Driftnet Longline Driftnet Longline Driftnet Longline 1980/1981 na na na 0.0 na na 1981/1982 na na na 0.0 na na 1982/1983 na na na 0.0 na na 1983/1984 na na na 0.0 na na 1984/1985 na na na 0.0 na na 1985/1986 na na na /1987 na na na /1988 na na na /1989 na na na /1990 na na na /1991 na na na /1992 na na na / na / na / na / na / na / na / / /2001 na na na na na na na na 2004 na na na na na na na na na All data from 1980/ /1994 includes sub-divisions 24-32, while it is more uncertain which years sub-divisions are included. The catches in sub-division are normally less than one ton. From 1995 data includes sub-divisions Estonian data from sub-div has earlier been given as sub-div

68 ICES WGBAST REPORT Table Trapnet effort and catch per unit of effort in number of salmon caught in trapnets in the Finnish fisheries in Subdivision 32 (number of salmon per trapnetdays). Effort CPUE

69 64 ICES WGBAST REPORT 2017 Table Number of tagged hatchery-reared and wild salmon smolts released in assessment units 1, 2 or 3 and used in the salmon assessment (data not updated since 2012) RELEASE YEAR Reared salmon stocked in rivers without natural reproduction Reared salmon stocked in rivers with natural reproduction Wild salmon AU1 AU2 AU3 AU1 AU2 AU3 AU Table Number of Carlin-tagged salmon released into the Baltic Sea in Country Total Denmark 0 Estonia 0 Finland 4,744 4,744 Sweden 7,800 7,800 Poland 0 Russia 0 Lithuania 0 Germany 0 Latvia 1,000 1,000 Total , ,800 4, ,544

70 ICES WGBAST REPORT Table Releases of adipose fin clipped salmon in the Baltic Sea and the number of adipose fin clipped salmon registered in Latvian (sub-divisions 26 and 28) offshore catches. Releases of adipose fin clipped Latvian offshore catches salmon, Sub-divs Sub-divs. 26 and 28 Year Parr Smolt Adipose fin Sample clipped salmon N in % , , , ,149 69, , , , , , , , , , , , , , , , , ,165, , , , , , , ,626, , , , , ,004, , ,284, , , , , , , , , ,500 2,124, ,714 1,753, ,126, ,984 2,450, ,731 2,325, ,123 2,084, ,496 2,341, ,094 1,971, ,200 1,768, ,670 2,038, ,361 2,690, ,113 2,777,

71 66 ICES WGBAST REPORT 2017 Table Adipose finclipped salmon released in the Baltic Sea area in 2016 (and clipped or unclipped tagged using other methods). Country Species Stock Age Number River Sub- Other tagging parr smolt division Estonia salmon Daugava 1 yr 11,100 Pärnu 28 salmon Daugava 2 yr 10,300 Pärnu T-bar salmon Daugava 1 yr 4,700 Pärnu 28 salmon Kunda 1 yr 9,600 Purtse 32 salmon Kunda 2 s 5,400 Purtse 32 salmon Kunda 1 s Purtse 32 salmon Kunda 1 yr 4,000 Purtse 32 salmon Kunda 2 yr 5,500 Selja T-bar salmon Kunda 1 yr 400 4,900 Selja 32 salmon Kunda 1 yr 1,300 3,800 Loobu 32 salmon Kunda 2 yr 5,400 Loobu T-bar salmon Kunda 1 yr 900 4,000 Valgejõgi 32 salmon Kunda 2 yr 5,100 Valgejõgi T-bar salmon Kunda 2 yr 5,100 Jägala T-bar salmon Kunda 2 yr 5,400 Pirita T-bar Russia salmon 1 yr 0 Vammeljoki T-bar (Finnish tags) Finland salmon Tornionjoki 2 yr 17,163 Aurajoki 29 salmon Tornionjoki 2 yr 6,000 Eurajoki 30 salmon Tornionjoki 1 yr 7,500 Kokemäenjoki 30 salmon Tornionjoki 2 yr 91,067 Kokemäenjoki 30 salmon Iijoki 1 yr 62,400 Kiiminkijoki ARS salmon Iijoki 2 yr 13,296 Kiiminkijoki ARS salmon Iijoki 2 yr 8,419 Iijoki Carlin, ARS salmon Tornionjoki 2 yr 35,000 Kemijoki T-bar, ARS salmon Tornionjoki/Iijoki 2 yr 22,658 Kemijoki ARS salmon Iijoki 2 yr 0 Kemijoki Carlin, ARS salmon Oulujoki 2 yr 0 Oulujoki T-bar, 35 Arrow-anch salmon Tornionjoki/wild smolts 0 Tornionjoki Carlin salmon Neva 2 yr 93,509 Kymijoki T-bar, ARS salmon Neva 2 yr 5,125 Karjaanjoki ARS Sweden salmon Luleälven 1 yr 252,831 Luleälven 31 salmon Luleälven 2 yr 362,221 Luleälven PIT-tag salmon Skellefteälven 1 yr 122,853 Skellefteälven 31 salmon Skellefteälven 1 yr 6,000 Gideälven 30 salmon Umeälven 2 yr 60 Sävarån 31 salmon Umeälven 1 yr 28,717 Umeälven PIT-tag salmon Umeälven 2 yr 92,006 Umeälven PIT-tag salmon Rickleån wild smolts 0 Rickleån PIT-tag salmon Lögdeälven wild smolts 0 Lögdeälven PIT-tag salmon Ångermanälven 1 yr 189,881 Ångermanälven Carlin salmon Ångermanälven 2 yr 41,925 Ångermanälven Carlin salmon Indalsälven 1 yr 320,936 Indalsälven 30 salmon Ljusnan 1 yr 165,658 Ljusnan 30 salmon Testeboån wild smolts 0 Testeboån PIT-tag salmon Dalälven 1 yr 6,513 Dalälven 30 salmon Dalälven 1 yr 203,712 Dalälven Carlin pelvic fin clipped + 57 telemetry salmon Dalälven 2 yr 3,035 Dalälven 30 salmon Dalälven 1 yr 12,000 Stockholms ström 27 salmon Mörrumsån wild smolts 0 Mörrumsån PIT-tag Latvia salmon Daugava 1 yr 599,710 Daugava Carlin salmon Venta 1 yr 13,800 Venta 28 Total salmon 93,113 2,777,782 *) ARS = alizarin red otholith marking

72 ICES WGBAST REPORT Table List of Baltic salmon stocks included in the genetic stock proportion estimation of catches. Stock Sampling year Propagation N 1 Tornionjoki, W 2011 Wild Tornionjoki, H 2006, 2013 Hatchery Simojoki 2006, 2009, 2010 Wild Iijoki 2006, 2013 Hatchery Oulujoki 2009, 2013 Hatchery Kalixälven 2012 Wild Råneälven 2003, 2011 Wild Luleälven 2014 Hatchery 90 9 Piteälven 2012 Wild Åbyälven 2003, 2005 Wild Byskeälven 2003 Wild Kågeälven 2009 Wild Skellefteälven 2006, 2014 Hatchery Rickleå 2012, 2013 Wild Säverån 2011 Wild Vindelälven 2003 Wild Umeälven 2006, 2014 Hatchery Öreälven 2003, 2012 Wild Lögdeälven 1995, 2003, 2012 Wild Ångermanälven 2006, 2014 Hatchery Indalsälven 2006, 2013 Hatchery Ljungan 2003, 2014 Wild Ljusnan 2013 Hatchery Testeboån 2014 Wild Dalälven 2006, 2014 Hatchery Emån 2003, 2013 Wild Mörrumsån 2010, 2011, 2012 Wild Neva, Fi 2006 Hatchery Neva, Rus 1995 Hatchery Luga 2003, 2011 Wild, Hatchery Narva 2009 Hatchery Kunda 2009, 2013 Wild, Hatchery Keila 2013 Wild Vasalemma 2013 Wild Salaca 2007, 2008 Wild Gauja 1998 Hatchery Daugava 2011 Hatchery Venta 1996 Wild Neumunas Hatchery 166 Total 4453

73 68 ICES WGBAST REPORT 2017 Table Medians and probability intervals of stock group proportion estimates (%) in Atlantic salmon catch samples from based on microsatellite (DNA) and smolt age data. Proportions of wild salmon estimated by scale reading for the same samples are given for comparison (with range for cases where prop. wild has been calculated with/without fish with missing data). D Danish, F Finnish, L Latvian, P Polish, and S Swedish catches. Gulf of Bothnia, WILD 2.5 % 97.5 % G. of Bothnia, HATC, FIN 2.5 % 97.5 % G. of Bothnia, HATC, SWE 2.5 % 97.5 % Others 2.5 % 97.5 % Sample size Scale reading - wild % 1. Åland Sea 2016 F F F F F F F F F F F F F F F F Mea n Bothnian Bay Finnish catch 2016 F F F F F F F F

74 ICES WGBAST REPORT Table Continued. 3. Bothnian Bay Swedish catch 2016 S S S S S S S S Bothnian Bay together 2016 FS FS FS FS FS FS FS FS FS FS FS Mean

75 70 ICES WGBAST REPORT 2017 Table Continued, Main Basin. GULF OF BOTHNIA, WILD 2.5 % 97.5 % G. OF BOTHNIA, HATCHERY, FIN 2.5 % 97.5 % G. OF BOTHNIA, HATCHERY, SWE 2.5 % 97.5 % GULF OF F INLAND, WILD 2.5 % 97.5 % GULF OF FINLAND, HATCHERY 2.5 % 97.5 % WESTERN MAIN B., WILD, SWE 2.5 % 97.5 % EASTERN MAIN BASIN 2.5 % 97.5 % SAMPLE SIZE SCALE READING - WILD % 3. Main Basin 2016 DP DP DP DP DFP S DFP S DFP S FP P FPS DFL PS Mean

76 ICES WGBAST REPORT Table Medians of individual river-stock proportion estimates in Atlantic salmon catches from the Gulf of Bothnia. Finnish and Swedish Bothnian Bay catches were analysed both pooled and separately. D for Danish, F Finnish, L Latvian, P Polish, and S for Swedish catch. Tornionj.Wild Tornionj. Hatch. Simojoki, W Iijoki, H Oulujoki, H Kalixälven, W Råne, W Luleälven, H Piteälven, W Åbyälven, W Byskeälven, W Kågeälven, W Skellefteälven, H Åland Sea, Finnnish catch 2016 F F F F F F F F F F F F F F F F Mean Ricleå, W Sävarån, W Vindelälven, W Umeälven, H Öreälven, W Lögde, W Ångermanälven, H Indalsälven, H Ljungan, W Ljusnan, H Testeboån, W Dalälven, H Neva-FI, H Sample size Bothnian Bay, pooled Finnish-Swedish catch 2016 FS FS FS FS FS FS FS FS FS FS FS Mean

77 72 ICES WGBAST REPORT 2017 Table (Continued.) Individual river stock proportions in the Finnish and Swedish Bothnian Bay catches. Tornionj.Wild Tornionj. Hatch. Simojoki, W Iijoki, H Oulujoki, H Kalixälven, W Råne, W Luleälven, H Piteälven, W Åbyälven, W Byskeälven, W Kågeälven, W Skellefteälven, H Ricleå, W Sävarån, W Vindelälven, W Umeälven, H Öreälven, W Lögde, W Ångermanälven, H Indalsälven, H Ljungan, W Ljusnan, H Dalälven, H Sample size Bothnian Bay, Finnish catch 2016 F F F F Mean Bothnian Bay, Swedish catch 2016 S S S S Mean

78 ICES WGBAST REPORT Table (Continued.) Medians of individual river-stock proportion estimates in Atlantic salmon catches from the Baltic Main Basin, for salmon stocks which did occur in the catches. D for Danish, F Finnish, L Latvian, P Polish, and S for Swedish catch.

79 74 ICES WGBAST REPORT 2017 Table Atlantic salmon catches from the Baltic Main Basin by assessment units, from 2006 to Year Estimate AU1 Wild; Fin, Swe AU1 Hatchery, Fin AU2 Wild, Swe AU2 Hatchery, Swe AU3 Wild, Swe AU3 Hatchery, Swe AU4 Wild, Swe AU6 Wild, Est, Rus AU6 Hat., Est, Fin, Rus AU5 Wild, Lat, Lit AU5 Hatchery, Lat, Pol 2016 Median % % Median % % Median % % Median % % Median % % Median % % Median % % Median % % Median % % Median % % Median % % Overall mean

80 ICES WGBAST REPORT NO ASSESSMENT UNIT RIVERS COUNTRY 1 AU1 Wild Simojoki, Tornionjoki; W, Kalixälven, Råne 4 Fin, Swe 2 AU1 Hatchery Tornionjoki; H, Iijoki, Oulujoki 3 Fin 3 AU2 Wild Pite, Åby, Byske, Kåge, Ricleå, Säverån, Vindel, Öre, Lögde 9 Swe 4 AU2 Hatchery Luleälven, Skellefteälven, Umeälven 3 Swe 5 AU3 Wild Ljungan, Testeboån 2 Swe 6 AU3 Hatchery Ångerman, Indals, Ljusnan, Dalälven 4 Swe 7 AU4 Wild Emån, Mörrumsån 2 Swe 8 AU6 Wild Luga, Kunda, Keila, Vasalemma 4 Est, Rus 9 AU6 Hatchery Neva-FI, Neva-RUS, Narva 3 Est, Fi, Rus 10 AU5 Wild Salaca, Gauja, Venta, Neumunas 4 Lat, Lit 11 AU5 Hatchery Daugava 1 Lat

81 76 ICES WGBAST REPORT 2017 Table (or Appendix 1). Prior proportion of 1 2 year old smolts in the Atlantic salmon baseline stocks used for Baltic salmon catch composition analysis for No Stock Smolt age 2.5% Prior median % 97.5% Data from years 1 Tornionjoki, W 1-2 years 4,3 5,6 7, Tornionjoki, H 1-2 years 99,8 100,0 100,0 All 3 Simojoki 1-2 years 42,3 52,2 61, Iijoki 1-2 years 99,8 100,0 100,0 All 5 Oulujoki 1-2 years 99,8 100,0 100,0 All 6 Kalixälven 1-2 years 3,9 5,7 8, Råneälven 1-2 years 2,7 6,2 11, Luleälven 1-2 years 99,8 100,0 100,0 All 9 Piteälven 1-2 years 16,8 20,1 23,8 All 10 Åbyälven 1-2 years 22,0 29,9 39,6 All 11 Byskeälven 1-2 years 22,5 30,2 39,4 All 12 Kågeälven 1-2 years 21,2 30,3 40,4 All 13 Skellefteälven 1-2 years 99,8 100,0 100,0 All 14 Rickleå 1-2 years 19,4 25,0 31,4 All 15 Säverån 1-2 years 19,1 25,0 31,6 All 16 Vindelälven 1-2 years 31,1 36,9 43,8 All 17 Umeälven 1-2 years 99,8 100,0 100,0 All 18 Öreälven 1-2 years 14,4 21,4 29,7 All 19 Lögdeälven 1-2 years 21,6 29,3 38,5 All 20 Ångermanälven 1-2 years 99,8 100,0 100,0 All 21 Indalsälven 1-2 years 99,8 100,0 100,0 All 22 Ljungan 1-2 years 27,9 37,7 46,9 All 23 Ljusnan 1-2 years 99,8 100,0 100,0 All 24 Testeboån 1-2 years 28,2 37,2 46,7 All 25 Dalälven 1-2 years 99,8 100,0 100,0 All 26 Emån 1-2 years 92,3 97,1 99,4 All 27 Mörrumsån 1-2 years 92,9 97,1 99,3 All 28 Neva, Fi 1-2 years 99,8 100,0 100,0 All 29 Neva, Rus 1-2 years 85,7 90,0 93,1 All 30 Luga 1-2 years 92,9 96,0 98,2 All 31 Narva 1-2 years 99,8 100,0 100,0 All 32 Kunda 1-2 years 97,9 99,0 99,6 All 33 Keila 1-2 years 97,8 99,0 99,6 All 34 Vasalemma 1-2 years 97,9 99,0 99,7 All 35 Salaca 1-2 years 97,9 99,0 99,7 All 36 Gauja 1-2 years 99,8 100,0 100,0 All 37 Daugava 1-2 years 99,8 100,0 100,0 All 38 Venta 1-2 years 99,8 100,0 100,0 All 39 Neumunas 1-2 years 99,8 100,0 100,0 All

82 ICES WGBAST REPORT Main Basin (SD 22-28) Åland Sea and Gulf of Bothnia (SD 29-31) Gulf of Finland (SD 32) Figure Combined expert estimates of total trolling catches (dead fish) for Baltic salmon, (medians with 90% p.i.). Red lines show previous estimates of trolling catches. Note the different scales.

83 78 ICES WGBAST REPORT % 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Blank TN OT LLS LLD GNS GND AN Figure Proportion of catch of Baltic salmon by weight in different types of gear Variables: GND=driftnet, AN=angling, GNS=gillnet, LLD=longline, OT=other, TN=trapnet. Blank=unidentified gear. 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Commercial and non-commercial catches of salmon Non-commercial catch Commercial catch Figure Commercial and non-commercial catches in percent (weight) in in Subdivisions (total catches from sea, coast and rivers).

84 ICES WGBAST REPORT Figure Catches of salmon in % of TAC in For years ( for Gulf of Finland) it is not possible to divide the total reported catch into commercial and recreational catches. Estimates of discards and unreported catches are presented separately in Table

85 80 ICES WGBAST REPORT Offshore driftnet fisheries Offshore longline fisheries Effort Year Figure Fishing effort in Main Basin offshore fisheries (x 1000 geardays) Coastal driftnet fisheries Coastal trapnet fisheries Coastal gillnet fisheries 800 Effort Year Figure Effort in Main Basin and Gulf of Bothnia coastal fisheries (x 1000 geardays).

86 ICES WGBAST REPORT A.1 A.2 A.3 A Weight (kg) Year Figure Mean weight of spawners in the Gulf of Bothnia by year. Values in from catch statistics in the Rivers Oulu and Torne. Values in are from Swedish tagging records and in from the Finnish catch sampling data. Weights of A.4 salmon based on small material in some years in % 18% 16% 14% GoB GoF 12% 10% 8% 6% 4% 2% 0% Year Figure Return rates of Finnish Carling tagged reared salmon released in Gulf of Bothnia and Gulf of Finland in (updated in March 2017 but no returns from 2013, cohorts).

87 82 ICES WGBAST REPORT 2017 Figure Recapture rate (%) of two-year-old Estonian Carlin tagged salmon in the Gulf of Finland. Carlin tagged from and T-bar anchor tags since Figure Return rates for salmon in in Poland.

88 ICES WGBAST REPORT Genetic distances between Atlantic salmon baseline stock samples in Tornionj. Wild 37 Tornionj. Hatch. 28 Kalixälven Simojoki 83 Iijoki 62 Råne 81 Oulujoki Kågeälven 48 Åbyälven 100 Byskeälven Säverån 97 Öreälven 99 Lögde Luleälven 91 Piteälven Skellefteälven 81 Ricleå Vindelälven 100 Umälven 82 Ångermanälven Indalsälven 98 Ljungan Ljusnan 47 Testeboån 99 Dalälven Emån 100 Mörrumsån Neva, Fin 100 Neva, Rus Kunda 54 Keila13 25 Vasalemma Luga 29 Narva 31 Daugava 28 Gauja 36 Salaca 13 Venta 71 Neumunas Figure Neighbour joining dendrogram (based on Nei s pairwise DA genetic distances) depicting genetic relationships among Atlantic salmon baseline samples for catch analysis. Numbers represent percentage support values based on 1000 bootstraps.

89 84 ICES WGBAST REPORT 2017 Åland Sea 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Bothnian Bay 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 06FS 07FS 08FS 09FS 10FS 11FS 12FS 13FS 14FS 15FS 16FS Main Basin 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 06DFLPS07FPS 08P 09FP 10DFPS11DFPS12DFPS 13DP 14DP 15DP 16DP Gulf of Bothnia, wild G. of Bothnia, hatchery, FIN G. of Bothnia, hatchery, SWE Gulf of Finland Main Basin Figure The proportion of Atlantic salmon stock groups in salmon catches of three Baltic Sea areas.

90 ICES WGBAST REPORT Figure Proportions of Atlantic salmon assessment units in the Baltic Sea Main Basin catches in international catch samples over the years

91 86 ICES WGBAST REPORT River data on salmon populations The Baltic salmon (and sea trout) rivers are divided into four main categories: wild, mixed, reared and potential. Details on how rivers are classified into these categories are given in the Stock Annex (Annex 3). 3.1 Wild salmon populations in Main Basin and Gulf of Bothnia Current wild salmon rivers in Main Basin and Gulf of Bothnia are listed per country and assessment unit in the Stock Annex (Annex 3) Rivers in assessment unit 1 (Gulf of Bothnia, Subdivision 31) River catches and fishery During the past centuries and even during the early 1900s, assessment unit (AU) 1 river catches were generally on a much higher level than during the late 1900s, as illustrated by catch statistics from Tornionjoki (Figure ). During the 1980s, river catches were the lowest ever recorded: only kg/year in Simojoki, and some tonnes/year in Tornionjoki and Kalixälven, indicating that the escapement to the spawning grounds was very low (Table , Figure ). In river catches increased and peaked in 1997, when they were 4, 74 and 10 tonnes in Simojoki, Tornionjoki and Kalixälven, respectively. Thereafter the catches again decreased to 25% 60% of that of 1997, until two new prominent rises; first in 2008 and second in Exceptional circumstances (warm and low vs. high and cool river water) may have affected fishing success in some years, but the catches generally reflect trends in the abundance of salmon in spawning runs (Table , Figure ). In 2012, the catch in Tornionjoki was three times higher than in 2011 and exceeded for the first time 100 tonnes since the beginning of the time-series of annual catch statistics (Table ). In 2014, catch rose to 147 tonnes. Catch somewhat decreased in 2015 (131 tonnes), but rose again in 2016 and reached a new record, 161 tonnes (Table ). Catch levels similar to those observed in were observed in the early 20th century (Figure ). Salmon catch in Simojoki did not rise much in , which is partly due to the low fishing effort. However, in 2014 and 2015 there was a clear increase in the catch and the rising trend continued to 2016, when the catch was 1800 kilos (Table ). The catches in Kalixälven has decreased and do not correspond to the registered number of salmon that passed the fishway in latest years. A special kind of fishing from boat (rod fishing by rowing) dominates in salmon fishing in Tornionjoki. This fishing also occurs in Kalixälven, but there it is not as dominating as in Tornionjoki. Cpue of this fishery in Tornionjoki has increased tens of times since the late 1980s (Table ), apparently reflecting the parallel increase in the abundance of spawners in the river. The cpue has been high (over 1000 grams/fishing day) in 1997, 2008 and , when the total river catches were also peaking. Annual changes in cpue and in total river catch follow each other rather closely. In Råneälven the local administration has since 2014 utilized a seasonal catch bag limit regulation of maximum of three salmon per person and season. Both obligatory tagging of killed fish and a digital catch reporting system has been utilized to aid in enforcement. Sport fishing catches in the river were 66 salmon in 2015, with 50 of those being released, and 46 salmon in 2016 with 37 being released.

92 ICES WGBAST REPORT Spawning runs and their composition In Kalixälven monitoring of migrating fish has been performed in the Jokkfall fishway since The fishway is located ca. 100 km from the sea, and not all salmon ascending river is counted. Until 1997 fish passage was monitored by manual controls, and from 1998 an electronic, infrared fishcounter, Riverwatcher (Vaki Aquaculture System Ltd, Iceland) has been used. Registration of species has been carried out during the whole migration seasons Every species passing both up- and downstream is identified using video recording. Totally six species (salmon, trout, whitefish, grayling, bream, and ide) have been registered during In 2001 and 2002 over 8000 salmon passed the Kalixälven fishway. During the years the run in the fishway was over 6000 individuals. The run in 2011 was the lowest in the past ten years. In 2012 the run increased to the same level as in 2001 and 2002, but with the difference that number of multi-sea winter salmon was the highest recorded. In 2013 the run increased to the highest level observed when more than salmon passed the fishway. The runs in have been at the same level. No reared salmon (adipose finclipped) was registered 2016 and 2015, and in previous years the registrations of adipose finclipped salmon has been very few. A hydroacoustic split-beam technique was employed in to count the spawning run in Simojoki. It seems evident that these counts covered only a fraction of the total run, as there are irregularities in the river bottom at the counting site, allowing salmon to pass without being recorded. Since 2008, the split-beam technique has been replaced by an echo sounder called DIDSON (Dual frequency IDentification SONar). According to the monitoring results, seasonal run size has ranged from less than 1000 up to more than 5000 fish (Table ). Spawning runs gradually increased from 2004 to , but dropped in In 2012 the run increased fourfold from the previous year (to about 3000) and was about as high also in 2013 and 2015 (about 3100). In 2014 the run peaked to 3800 salmon, and in 2016 the run was record-high with 5400 salmon counted (Table ). A lot of back-and-forth movement of salmon has been detected in Simojoki, which erodes the accuracy of the counts. There have also been problems connected to the differentiation of species. The spawning runs into Tornionjoki have also been monitored using the DIDSON technique since The observed seasonal run size has ranged from (year 2010) to (year 2014) salmon (Table ). The run size in 2016 ( salmon) was almost as high as in the record year 2014 ( salmon). The counting site is located about 100 km upstream from the river mouth. Therefore, salmon which are either caught below the site or stay to spawn below the site must be assessed and added into the hydroacoustic count in order to get an estimate of the total run size into the river (Lilja et al., 2010). Also, according to auxiliary studies, a small fraction of the spawners pass the counting site via the fast-flowing mid-channel without being detected by sonars. In 2016, the total amount of spawners entering the river mouth probably was somewhere between individuals. By subtracting the river catch from this number, the 2016 spawning population in the Tornionjoki is estimated to be 16 24% smaller ( spawners). Grilse account for a minority (7 21%) of the annual spawning runs. In the spawning run into Råneälven has been monitored with an ultrasound camera called SIMSONAR. The technique is similar to that used in Tornionjoki and Simojoki. The counting site is located about 35 km upstream from the river mouth, and the counts represent the total run of salmon. The total counted salmon runs in 2014, 2015 and 2016 was 3756, 1004 and 1454 individuals, respectively.

93 88 ICES WGBAST REPORT 2017 About catch samples have been collected from the Tornionjoki salmon fishery since the mid-1970s. Table shows sample size, sea age composition, sex composition and proportion of reared fish (identified either by the absence of adipose fin or by scale reading) of the data for the given time periods. Caught fish have generally become older, and the proportion of repeat spawners has increased in parallel with a decreasing sea fishing pressure (see Chapter 4). The strong spawning runs into Tornionjoki in were a result of fish from several smolt cohorts. In these years the proportion of females has been fairly stable (61 69% of total biomass, except in 2015 when the proportion dropped to 55%). The proportion of repeat spawners decreased in to 6 8%, after the record high level (14%) observed in Recently very few, if any, reared salmon has been observed in the catch samples and some of them are probably strayers from nearby Finnish compensatory releases (with intact adipose fins). Parr densities and smolt trapping The lowest parr densities were observed in the mid-1980s (Table , Figures and ). During the 1990s, densities increased in a cyclic pattern with two jumps. The second, higher jump started in Between these jumps there was a collapse in densities around the mid-1990s, when the highest M74 mortality was observed. Average parr densities are nowadays 5 60 times higher than in the mid-1980s. Since the turn of the millennium, annual parr densities have varied 2 6 fold. In Simojoki, some years with higher-than-earlier densities of 0+ parr have been observed recently, but annual variation has been large and densities of older parr have not increased in this river. In the other rivers, however, the densities have continued to increase rather steadily. In some years, like in 2003, high densities of parr hatched in spite of relatively low preceding river catches (indicating low spawner abundance) in Simojoki, Tornionjoki and Kalixälven. Similarly, high densities of 0+ parr were observed in Tornionjoki in 2008 and 2011, although river catches in the preceding years were not among the highest. Possible reasons for this inconsistency include exceptionally warm and low summer-time river water, which might have affected fishing success in the river and even the measurements of parr densities. In years 2006, 2013 and 2014 conditions for electrofishing were favourable because of very low river water levels, whereas they were the opposite in 2004 and These kinds of changes in electrofishing conditions may have affected the results, and one must therefore be somewhat cautious when interpreting the data obtained. In Simojoki the mean density of one-summer old parr increased by about 50% from 2015 to 2016 (Table ). The 2016 density of 0+ parr is the fourth highest in the timeseries. The uppermost sites lacked 0+ parr. The density of older parr doubled from 2015, and this density is the fifth highest in the time-series. In Tornionjoki the densities of 0+ parr in 2014 and 2015 were clearly higher than in any earlier year in the timeseries. In 2016, several flood peaks due to heavy rains prevented electrofishing on the lower and on some of the middle and upper sections of the river system. Therefore, only 50 of the 80 regular sampling sites could be electrofished, and the results are not well comparable to earlier years results. The average density of 0+ parr on the sampled sites was the fourth highest in the time-series, about 15 ind./100 m 2 lower than in The average density of older parr was the third highest density in the time-series, only slightly lower than in In Kalixälven and Råneälven the densities of 0+ parr increased in 2013, even if river catches and the result from the Kalixälven fishway indicated a decreased spawner abundance in In 2014, the Kalixälven densities of 0+ parr stayed at the same level

94 ICES WGBAST REPORT as in 2013, despite the high number of salmon registered in the fish ladder in the preceding year. The density of 0+ in 2016 decreased with more than 50% compared to 2015, whereas the older parr density stayed at same level as in recent years. Also in Råneälven the 2016 density of 0+ parr decreased with half compared to in the three latest years, whereas the older parr density stayed at the same level as in previous years. Smolt production has been monitored in Simojoki and Tornionjoki by annual partial smolt trapping and mark recapture experiments (see Annex 3 for methodology) since 1977 and 1987, respectively (Table ). A hierarchical linear regression analysis (below referred to as River model ) has been applied to combine the information from electrofishing and smolt trapping results, to obtain updated estimates of the wild smolt production in the rivers, including years when high water flow has prevented complete smolt trapping. With a 1 3 year time-lag (needed for parr to transform to smolts) wild smolt runs have followed changes in wild parr densities. In the late 1980s, the annual estimated wild smolt run was only some thousands in Simojoki and less than in Tornionjoki (Table ). The first increase in the production occurred in the early 1990s, and a second, higher jump occurred in the turn of the millennium. Since then, smolt runs have not increased in Simojoki, while in Tornionjoki the runs have continued to increase especially during the last ten years. Since the turn of the millennium, annual estimated runs of wild smolt have exceeded and smolts with high certainty in Simojoki and Tornionjoki, respectively. Since 2008, estimates of wild smolt runs have exceeded one million smolts in the Tornionjoki. In 2016, successful smolt trapping was carried out both in Tornionjoki and Simojoki. In Tornionjoki, the estimated number of smolts in 2016 based on smolt trapping was about 3 million smolts (Table ), but this estimate is very uncertain (95% PI: million). The river model with the newest data (up to 2016) estimates the updated smolt run for Tornionjoki to about 1.8 million (95% PI million) and predicts about 2.5 million smolts for 2017 and The smolt trapping estimate of the 2016 smolt run size for Simojoki was smolts (95% PI: , Table ). Here the river model with the newest data does not change the median value of the smolt run estimate for 2016, but just slightly decreases the associated uncertainty (posterior median , 95% PI ). The river model predicts an increase to approximately smolts/year for 2017 and 2018 in Simojoki Rivers in assessment unit 2 (Gulf of Bothnia, Subdivision 31) River catches and fishery The 2016 catches in Piteälven and Åbyälven stayed at the same low level as in previous years. Catches in Byskeälven during the 1980s varied between kg, but in the beginning of the 1990s they increased noticeably (Table ). The highest catch occurred in 1996 (4788 kg) after which the catches have been lower but with much annual variation. In 2011, for example, the catch decreased with 40% (to 870 kg) compared to in 2010, whereas in 2012 the catch increased three times compared with in the previous year (Table ). In 2016, the Byskeälven river catch was at the same level as in the three last years, not corresponding to the increase in number of salmon (highest recorded) that passed the fishway in the same year. In Kågeälven (wild river since 2014) the sport fishery was regulated in 2012 by the local administration to become 100% catch and release, with all fish released to be registered in the obligatory reporting system. Numbers of caught and released salmon in the river fishery have been 35, 24, 67, 92 and 61 in years , respectively.

95 90 ICES WGBAST REPORT 2017 In Sävarån the catches have been very low in recent years, and in 2016 only 13 salmon where caught and released. The catches in Ume/Vindelälven decreased from 185 salmon in 2015 to 90 (70 wild and 20 reared) salmon in In Öreälven the 2016 catch decreased to 200 salmon (whereof 400 released), compared to 450 in In Lögdeälven the catch in 2016 was 107 salmon (whereof 28 released), compared to 125 in Spawning runs and their composition In almost all AU 2 rivers the upstream migration in fish ladders is counted by electronic, infrared fish counters, Riverwatcher (Vaki Aquaculture System Ltd, Iceland). In Piteälven a power plant station (the only one in the river) equipped with a fishway was built in the end of the 1960s, about 40 km from the river mouth. In 1992 the power plant company built a new ladder and in 1998 they installed an electronic fish counter. In 2001 a camera was also installed for the detection of species. The run counted is the entire run, as no salmon spawning areas exist downstream the fish ladder. The total run 2012 increased to 1418 salmon which was three times higher than the two earlier years, and the run has stayed at that level until In the two last years the run has increased, and in 2016 the run was the highest ever recorded with 2009 salmon passing (Table , Figure ). Low water level has no effect on the possibility for salmon and sea trout to enter the fishway, but very high water can temporary stop and delay the migration. In Piteälven an ultrasound camera (Simsonar) was operated by the County Administrative Board of Norrbotten in 2015 and 2016, one kilometre below the hydropower tailrace. One purpose of the survey was to evaluate the total number of ascending salmon and sea trout below the power plant in relation to the number of fish passing the fishway. Another aim was to identify potential delays in the migration. The results from 2015 revealed a delay that varied between 3 5 weeks, and the total run was 22% lower in the fishway than counted with the ultrasound camera. The delay accrues both in the tailrace and at the spillgate close to the outlet of the fishway. At the end of the ultrasound study, close to the beginning of the spawning period, downstream migration of fish was recorded with the ultrasound camera, indicating that fish were leaving the area below the entrance to the fishway. During earlier spawning runs no downstream migration occurred in the fishway, except for a few kelt in the beginning of the migration run. Due to a broken computer storage disc, the collected data from the survey in 2016 has yet not been fully analysed. In Åbyälven a power plant station (the only one in the river) equipped with a fishway is located 30 km from the river mouth. In year 2000 the power plant company installed an electronic fish counter (Riverwatcher), and in 2009 a camera was added to allow identification of species. The number of salmon passing the fish ladder only represents a small part of the entire run. The counted run in 2016 was almost double as high compared to the previous year (150 salmon passed the fishway; Table , Figure ). Low water levels can cause shut downs of the power plant which makes it almost impossible for fish to enter the fishway. In the fall of 2013, the power company filled pits with stones and concrete to prevent fish from getting caught in the pits when closing the spillgates. However, a test with water spill in the former river bed, to study the undertaken measurers to steer salmon when spill occurs, attracted more fish into the former river bed than was counted in the fish ladder during the total season (approximately compared to 113). This strongly indicates that there are problems for salmon to find the ladder, or problems within the ladder causing fallbacks. A project has been started to find measures for optimising up- and downstream passage, and in 2017 a new entrance to the fishway will be installed.

96 ICES WGBAST REPORT In Byskeälven a new fishway was built in 2000 on the opposite side to the old ladder at Fällfors, about 40 kilometres from the sea. The waterfall is a partial obstacle for salmon, and the number of fish counted only represents a part of the entire run. In 2000 also an electronic fish counter (Riverwatcher) was installed in the new ladder, whereas a Poro counter (camera) was installed in the old ladder. The latter was replaced with a Riverwatcher (VAKI) in The water level in the natural waterfall at Fällfors affects the possibilities for salmon to pass the fall without being counted. In 2008 the number of salmon registered increased to 3409, the highest level since The counted run in 2009 decreased almost by half, to 1976 salmon, and in 2010 and 2011 the number of salmon remained at the same low level as in However, since 2012 the number of ascending fish has again increased. The 2016 run was the highest recorded so far with 7280 salmon counted (Table , Figure ). In Rickleån the power plant company built four fishways in 2002 at the three hydro power stations located in Robertsfors, some 15 kilometres upstream from the river mouth. Fish migration is controlled using an electronic counter placed in the uppermost fishway. Before the fishways were constructed, salmon migration had been impossible closed since the early 1900s (over 100 years) when the first power plant station was built. The counted number of fish is just a small part of the entire run. No salmon was counted in the fishways in compared to one, 6, 7 and 5 in In 2014 the number increased to 27 salmon, which is the highest number recorded until present. In 2016 a total of 17 salmon passed the fishways. The water level does not affect the migration in the four fishways except when it being extremely low; then the migration can decline or even stop. The fish ladder in Ume/Vindelälven is located at Stornorrfors, about 25 km from the sea. The original ladder built in 1960 was replaced by a new one in 2010, opened before the start of the migration season. The new fish ladder has a length of ca. 300 meter and is one of the longest in Europe. It is constructed to also serve as a passage gate for downstream migrating fish. A future aim is also to monitor migration of smolts and kelts passing through the fishway. No spawning areas exist below the ladder, and all spawners must pass the fishway to find their spawning areas located in the main tributary Vindelälven, merging with Umeälven ca. 10 km upstream the dam at Stornorrfors. The salmon run in Ume/Vindelälven is much affected by yearly fluctuations in the amount of water in the old riverbed leading up to the fishway, which influences the possibilities for salmon and sea trout to find their way upstream. As compensation for hydropower exploitation of Umeälven, where no salmon production remains, hatchery reared smolts are released annually at Stornorrfors (Section 3.3). From the early 1970s the total salmon run into Ume/Vindelälven has been divided into wild and reared (absence of adipose fin) salmon. In 2012 the run of wild salmon increased to 8058 which was 65% higher compared to 2011, and in 2013 the run was the highest recorded ( wild salmon passed the fish ladder). In 2016 a total of 9134 salmon passed the fishway (Table and Figure ), whereof 6442 were wild and 2692 reared. The new ladder in Ume/Vindelälven has been operating for six years. Some construction modifications were carried out in 2013 that may have resulted in improved possibilities for the fish to force former thresholds in the river section below the fishway and to detect the entrance of the fish ladder. In further tests and modifications were carried out in the first pool section to create stronger water velocity into the diffuser to attract salmon into the fishway. Measures to assist the fish migration are also planned in the lowermost rapid section, Baggböleforsen, to reduce its steepness.

97 92 ICES WGBAST REPORT 2017 In Öreälven the control of ascending fish ended in 2000 (Table ). The reason was high water levels that destroyed the part of the dam where the fish trap was located. Parr densities and smolt trapping Electrofishing surveys have been done with the same kind of equipment (Lugab), a portable motor and transformer, until in 2011 when it was exchange to the model ELT60IIHI from Hans Grassl. During the time-series, same groups of people have made most of the electrofishing in Swedish rivers in AU 1 4. In the beginning of the monitoring surveys the average size of the sites was around m², especially in AU 1 and 2. The reason for using large sizes of sites was to increase the possibility to catch parr. From 2003 and onwards, however, the sizes of sites in AUs 1 and 2 have been reduced to about m² due to higher parr densities. In the summers of 2006, 2013 and 2014 conditions for electrofishing were extraordinary because of very low water levels, i.e. opposite to the conditions prevailing in For the electrofishing carried out in 2009, 2010, 2012 and 2015, the water levels were normal, but in 2011 and 2016 high water levels due to rain prevented surveys in several rivers. The densities of salmon parr in electrofishing surveys in AU 2 rivers in the Gulf of Bothnia (ICES SD 31) are shown in Table and in Figures and In Piteälven no consistent electrofishing surveys were made during the 1990s. In 2002, 2006, 2007, 2008 and 2010 surveys were carried out. The density of 0+ parr has been rather low in most of the years (Table ). No surveys were done 2011 and 2012 due to high water levels. In 2014 the densities of 0+ parr was the highest recorded (12 parr/100 m²). In 2016 the average density increased compared to in the previous year. The density of older parr has also been low, varying between 4 9 parr/100 m² the latest four years. In Åbyälven the mean densities of 0+ parr in were about 3 parr/100 m². In 1999 the densities of 0+parr increased to 17 parr/100 m², about five times higher than earlier. In 2016 the densities increased to the so far highest recorded level (37 parr/100 m²). The densities of older parr have been stable in the latest six years with a mean of 13 parr/100 m² (Table ). In Byskeälven the mean densities of 0+ parr in were about 5 parr/100 m². In the densities increased to about 11 parr/100 m², and in 1999 and 2000 the 0+ parr densities increased further (they were about 70% higher than in ). During the 2000s, the densities have been on rather high levels with a few exceptions, and in 2016 the densities of 0+ increased to the so far highest recorded level (43 parr/100 m²). The densities of older parr have remained rather stable during later years with a mean around 20 parr/100 m² (Table ). In Kågeälven the last releases of reared salmon parr were made in 2004 which means that the wild-born 0+ observed in 2013 were mainly offspring of spawners which themselves were also wild-born. Stable occurrence of 0+ parr in recent years indicates that the population has become self-sustaining, although the 0+ density in 2016 decreased for the third year in a row (Table ). Spawning occurs in the whole river stretch. In Rickleån the mean density of 0+ parr were only about 0.5 parr/100 m² in , whereas since 1998 the mean density has been around 3.7 parr/100 m². The mean 0+ density decreased slightly in 2016, compared to in 2015 when the so far highest density was recorded (14.6 parr/100 m²). In Table are also average densities from extended electrofishing surveys in Rickleån presented, that include sites in the upper part of the river that recently was colonized (for more details see Section in ICES, 2015). Since some years, mean densities that include the extended electrofishing surveys are used as input in the river model (see Stock Annex).

98 ICES WGBAST REPORT In Sävarån the mean densities of 0+ parr in were about 1.4 parr/100 m². In 1996 the densities increased to 10.3 parr/100 m² and in 2000 to 12.8 parr/100 m². No electrofishing was made in 2001 and The 0+ density in 2015 was the so far highest recorded (45 parr/100 m²) followed by the highest one for older parr in 2016 (34 parr/100 m²), whereas the 0+ density in 2016 decreased to 32 parr/100 m² (Table ). In Ume/Vindelälven mean densities of 0+ parr in were only about 0.8 parr/100 m². In 1997 the densities increased to 17.2 parr/100 m². During the 2000s, densities have fluctuated a lot from about 5 25 parr/100 m². No surveys were carried out in 2011 due to high water level. In 2014, the density of 0+ parr increased to the so far highest recorded (39 parr/100 m²) followed by a decrease in 2015 with almost 50%. In 2016 the mean density of 0+ parr declined to a very low level (about one parr/100 m²), a value not seen since the peak years of the M74-disease in the early 1990s. The reason for the low density seems to be linked to the low number of salmon females passing the fishway in Stornorrfors in 2015 (Table ) and the amount of ascending spawning fish suffering of fungus in the same year. The establishment of fungus weakened the fish and resulted in high mortality, which was observed in the fishway and at the intake grid to the hydro power station. In addition the M74-frequency increased in the spawning year 2015 (Section 3.4) These factors combined probably led to a low egg deposition in autumn 2015, seen as the very low densities of one summer old salmon parr in In Table also average densities from extended electrofishing surveys in Vindelälven are shown, that also include sites from upper parts in the river that recently have been colonized (see Section in ICES, 2015). Since some years, mean densities that include the extended electrofishing surveys are used as input in the river model (see Stock Annex). In Öreälven mean densities of 0+ parr in were very low, just about 0.5 parr/100 m². The 0+ parr densities increased during the early 2000s, and then stayed on levels of 3 10 parr/100 m² until in 2015 when the density increased three times compared with earlier to the highest value recorded so far (21.6 parr/100 m²). In 2016 the 0+ densities were nearly halved compared to in the previous year (Table ). In Table also average densities from extended electrofishing surveys in Öreälven are shown, including sites from upper parts of the river that recently have been colonized (see Section 4.2.2). In coming years, these mean densities that include also the extended electrofishing surveys will be used as input in the river model (see Stock Annex). In Lögdeälven mean densities of 0+ parr in were about 1.4 parr/100 m². In 1998 the densities increased to 13.7 parr/100 m². Densities during the 2000s have fluctuated between three and almost 15 parr/100m². The highest value so far was observed in 2014 and stayed at same level at 2015 and 2016 (Table ). In Table also average densities from extended electrofishing surveys in Lögdeälven are shown that include sites from upper parts of the river that recently have been colonized (see Section 4.2.2). In coming years, these mean densities that include also the extended electrofishing surveys will be used as input in the river model (see Stock Annex). In Rickleån smolts of salmon and sea trout have been counted during their downstream migration in using a Rotary-Screw-trap and mark recapture experiments. The trap was positioned close to the river mouth. In 2014 a total of 434 salmon smolts were caught. The calculated recapture rate for tagged salmon was 20.3%, which was used to estimate the total smolt production to 2149 (Table ). Because of many breaks in drifting the screw-trap in 2015, no reliable estimate of the smolt production could be carried out. The trap was in drift the whole season 2016 and the estimated run was about 4000 smolts (Table ).

99 94 ICES WGBAST REPORT 2017 In Sävarån smolts of salmon and sea trout were caught on their downstream migration using Rotary-Screw-traps from 2005 to The trap was positioned 15 km upstream from the mouth of the river. On average ca. 470 wild salmon smolts were caught annually in from mid-may to mid-june. Smolts were measured for length and weight, scale samples were taken for age determination and genetic analyses. The dominating age group among caught smolts was three years. The proportion of recaptured tagged fish in the trap varied between 4 31% during the trapping years. Estimates of total smolt production are presented in Table No trapping of smolts was carried out in as the smolt trap was used in Rickleån (to get basic data from this river for increasing precision in analyses of its stock status). In Vindelälven a smolt fykenet for catching smolts, similar to the one used in Tornionjoki, was operated between 2009 and The entire smolt production area is located upstream of the trapping site. Around 2500 salmon smolts on average were caught annually from 2009 to 2015, and the proportion of recaptured tagged fish in the varied between %. In 2009 the trap was operating from end of May to beginning of July, and smolts were likely caught during the whole time period with a peak in mid-june. In 2010 a pronounced spring flood caused problems to set up the fykenet and a considerable part of the smolt run was missed. In 2011, an episode late during the season with very high water flow again prevented smolt trapping. Although the break was rather short (six days) a very high smolt catch the day immediately before the break indicated presence of a significant "peak" that was missed. In , again several episodes of high water flow resulted in repeated breaks, and for those years it seems difficult to even produce a crude guess of the proportion of the total smolt run that was missed. Although direct smolt estimates from the mark-recapture experiments with the fykenet have not been possible to produce, due to the above mentioned interruptions in the function of the trap, estimates have still been possible to calculate based on data for returning 1SW adults (grilse). Since 2010 all captured smolts have been marked using PIT-tags. VAKI counters and PIT-antennas in the Ume/Vindelälven fishway record all marked and unmarked wild returning spawners. Assuming a common smolt-to-adult survival rate for marked and unmarked grilse, the size of a given smolt cohort is thus possible to indirectly estimate (see Table ). Since 2016, the Vindelälven smolt trapping has been moved to a newly built permanent smolt trap within the fish ladder at Stornorrfors (hydropower dam that must be passed by down-migrating smolts) just a few kilometres downstream the former trapping site. In a smolt wheel was operated in Lögdeälven, close to the river mouth. The number of caught salmon smolts were 299 (2015) and 463 (2016), with 11% and 10% of the marked smolts being recaptured. In 2015, the trap had to be closed before the migration was finished, and the total smolt run for this year was therefore likely underestimated. In 2016, however, the whole run was monitored, yielding an estimate of about 5200 smolts (Table ) Rivers in assessment unit 3 (Gulf of Bothnia, Subdivision 30) River catches and fishery In Ljungan the salmon angling catch was only 14 salmon in 2016, compared to the average annual catch of 81 for the period (with 145, 228 and 161 salmon caught in 2013, 2014 and 2015, respectively). In general, the catches have increased since the early 2000s ( ) when 18, two and one salmon, respectively, were caught.

100 ICES WGBAST REPORT In Testeboån (wild river since 2013) landing of salmon is not allowed. Parr densities and smolt trapping Average densities of 0+ salmon in Ljungan varied markedly from 3 to 45 parr/100 m² between 1990 and 2008, but without any clear trend (Table and Figure ). However, in 2012, 2014 and 2015 (especially) parr densities have showed signs of increase, although more years of data are necessary before any clear conclusions can be drawn. The density decreased in 2016 to parr/100 m². It should be noted that electrofishing data from Ljungan are missing in several years due to high water levels in late autumn, making electrofishing impossible. For example, the relatively high value for 2012 only mirrors data from one electrofishing site (Table ) as the other sites could not be fished due to high water level. In Testeboån the latest releases of reared salmon (fry) were made in 2006 which means that the wild-born 0+ parr observed at electrofishing from 2012 and onwards most likely were offspring of salmon which themselves were wild-born. Fairly stable levels of 0+ parr densities in recent years, except for in 2008 when 0+ parr were absent due to a very poor spawning run in 2007, indicates that the population is self-sustaining (Table ). The densities of 0+ parr decreased in 2014 compared to the four previous years, but again increased in 2015 and 2016 displayed the so far highest recorded density (about 28 parr/100 m²). In Testeboån, which received status as a wild salmon river by ICES/WGBAST in 2013, smolt trapping using a smolt wheel has taken place since In 2015, Testeboån was equipped with permanent facilities for counting of both smolts and ascending adults. Hence, in the future it may be possible to also utilize Testeboån as a full index river. Annual estimates of the total smolt runs in have varied in the range from about 2000 to 4300 smolts Rivers in assessment unit 4 (Western Main Basin, Subdivisions 25 and 27) River catches and fishery In Emån the 2016 catch was eight salmon, compared to 49 in In 2014 only two salmon were caught. No salmon was reported as caught and retained in 2012 or The retained catches in were 12, 9, one, 15, 5 and 3 salmon, whereas earlier in , the catch was 143, 83 and 89, respectively. During the last years anglers have increasingly applied catch and release, and the sport fishery in Emån is nowadays basically a no-kill fishing. This is likely an important reason for the decreasing catches. In Mörrumsån the retained salmon catches have varied between 145 and 536 salmon during the last five years. In 2015 the catch increased to 831 salmon, but in 2016 only 94 salmon were retained. Also in Mörrumsån anglers have increasingly applied catch and release in the latest years. This is likely the reason for the declining landed catches in recent years. Parr densities and smolt trapping For Emån densities of parr in electrofishing surveys below the first partial obstacle are displayed in Table and Figures and The densities of 0+ parr have varied between parr/100 m² during , with a mean density of 43. The highest 0+ density so far occurred in The density of 0+ parr was 53 parr/100 m² in 2016, which is just over the mean value for earlier years in the time-series. The densities

101 96 ICES WGBAST REPORT 2017 of older parr have varied from 1 10 parr/100 m² during the period with a mean value of 5 parr/100 m² during recent years. The smolt production estimates in River Emån have been a problem in the current assessment model used by WGBAST; the current production appears very low compared to the production capacity. In 2007 an overview of the conditions in the river concluded that probably the difficulties for particularly salmon spawners, and to a minor extent also sea trout, to ascend fishways may give rise to low production of juveniles above the fishways. Electrofishing sites in these upstream areas do therefore normally show low juvenile abundance. On the other hand, there is a highly successful sea trout and salmon fishery in the lower part of the river (at Em) and this fishery have not shown signs of a lower abundance of either species. On the contrary, salmon seems to have increased in abundance. Monitoring of salmon migration in one fishway during also suggested that very few salmon could reach some of the upstream potential spawning areas. In recent years, however, some actions have been undertaken to improve the conditions for fish migration in the river. In 2006 the lowermost dam (at Emsfors) was opened permanently, and since then increased electrofishing densities for salmon have been recorded at the closest upstream electrofishing site. Activities are also ongoing to facilitate upand downstream migration at the second dam counted from the sea, above which significant habitat areas regarded suitable for salmon reproduction are located. In order to get a quantitative estimate of the smolt run in the river, smolt traps were operated in the river Emån in 2007 and The primary purpose was to get an overview of the smolt production in the river. Two smolt wheels were installed within 200 m from the river mouth. In 2008 the smolt traps were operating through most of the smolt migration period. Almost the entire catch of salmon and sea trout smolts in the traps were utilized for mark recapture estimation, and the trap efficiency was estimated to 6.1%. The estimated salmon smolt run in 2008 was 3473 smolts (95% confidence interval: ). A considerable emigration of Salmo sp. fry (species not identified more precisely) in the length interval mm occurred in 2007 and 2008, indicating that this migration can be a common phenomenon. It was not possible to estimate the catch efficiency for small fry, but it is certainly much lower than for smolts. Assuming that the trap efficiency for fry is half that of salmon smolts, or 3%, the estimated number of fry emigrating from the river would be in the order of However, the actual numbers might be much higher if the trap efficiency is even lower. This kind of mass emigration has not been observed in any of the other Swedish rivers where smolt wheels have been operating (Mörrumsån, Testeboån, Lögdeälven, Sävarån and Rickleån). In Mörrumsån the densities of parr in electrofishing surveys are shown in Table and Figures and In the period the densities of 0+ varied between parr/100 m². The highest average density was observed in The densities decreased in 2006 and 2007, but 2008 the densities of 0 + parr again increased to 102 parr/100 m². In 2011 the densities of 0+ decreased to 36 parr/100m², the lowest value since the mid-1990s. In 2016 the densities decreased somewhat compared to the mean value for the latest years. One probable reason for the lower densities in 2007 and 2011 was high water levels, as only part of the survey sites were possible to electrofish. However, it should also be noted that the number of ascending salmon counted in the preceding autumns (2006 and 2010) was the lowest recorded at Marieberg power plant, ca. 13 kilometres from the sea, since a VAKI counter was installed in the fish ladder 2002.

102 ICES WGBAST REPORT In river Mörrumsån, hybrids between salmon and trout have been found during the electrofishing. In the proportion of hybrids was high, up to over 50% in some sampling sites. After that, the occurrence of hybrids has varied. In 1995 and 1996 it was only some percent of the total catch. In 2005 the density of 0+ hybrids was 14 parr/100 m² which is higher than in the three years before. The amount of hybrids has decreased during In 2016 the densities of hybrids was 2.3 parr/100 m². In 2004 two new fish ladders were built at the power plant station about 20 km from the river mouth which opened up about 9 km of suitable habitat for salmon, including about ha of production area. In , a smolt wheel has been operated in Mörrumsån, ca. 12 km upstream from the river mouth. About 55% of the total production area for salmonids is located upstream the trap. A main reason for choosing this upstream location was that ascending adults are counted in a nearby fishway close to the smolt trap site, which should allow comparisons among numbers of ascending spawners and smolts from the upper part of Mörrumsån. So far, however, only preliminary numbers of ascending adult spawners exist; to obtain such reliable estimates further work will be needed that accounts for (i) a relatively large share of missing or unclear species identifications (due to absent or low quality camera images from the fishway) and (ii) the fact that a rather high proportion of salmon trout hybrids exists in the river (Palm et al., 2013). In , the estimated smolt production in the upstream parts of the river was lower than expected (ca ). As a comparison, Lindroth (1977) performed smolt trapping in at a site close to the one presently used and estimated the average annual salmon smolt production to (range ). However, since 2013, the smolt production in the monitored upper reaches of Mörrumsån has increased. In 2013 it was estimated to ca , and in 2014 it was estimated to be the highest recorded so far (ca ). In 2015 the estimated smolt production decreased to ca , but in 2016 it again increased to ca Mörrumsån is located quite close to Emån. When the Mörrumsån smolt trap was operated in in 2009, a total of 35 Salmo sp. fry were caught in the trap, and if we assume that the trap has half the catching efficiency for them compared to smolt the total number would be around 1200 fry. This is only about 1% of the number of fry estimated to emigrate from Emån in previous studies Rivers in assessment unit 5 (Eastern Main Basin, Subdivisions 26 and 28) Estonian rivers The River Pärnu flows into the Gulf of Riga and is the only Estonian salmon river in the Main Basin. The first obstacle for salmon migrating in the river is the Sindi dam, located 14 km from the river mouth. The fish ladder at the dam is not effective due to its small size and the location of the entrance. Electrofishing surveys on the spawning and nursery ground below the dam have been performed since 1996; the number of parr/100 m² has been very low during the whole period (Table and Figure ). No parr were found in 2003, 2004, 2007, 2008, 2010 and In 2016 the 0+ parr density below Sindi dam was 0.6 parr/100 m². The habitat quality below the dam is poor and that is the main cause for the low parr density. Since 2013, electrofishing is carried out also upstream from the Sindi dam. Above the dam salmon parr are found in only some years, and the densities have been very low. No 0+ parr were found above the dam in Spawning areas above the dam are relatively good, and low parr density is associated with poor accessibility. Future plans to restore the Pärnu salmon population include reconstruction of the Sindi dam, and a juvenile release programme was

103 98 ICES WGBAST REPORT 2017 initiated in The first juvenile salmon was released in 2013 and the Pärnu River must therefore be considered as mixed. Latvian rivers There are ten wild salmon rivers in Latvia, mainly flowing into the Gulf of Riga. Some rivers have been annually stocked with hatchery reared parr and smolts, with the result that salmon populations in these rivers consist of a mixture of wild and reared fish. In 2016, 28 sites in eleven rivers were sampled by electrofishing. The salmon parr densities are presented in Table and Figure The wild salmon population in river Salaca has been monitored by smolt trapping since 1964 and by parr electrofishing since From 2000, no releases of artificially reared salmon have been carried out in river Salaca. In 2016, eleven sites were electrofished in the river and its tributaries. All sites in the main river held 0+ age salmon parr. Salmon 0+ parr also occurred in the tributaries Jaunupe, Svētupe and Korģe. The average density of 0+ salmon was 67.7 parr/100 m², whereas the density of 1+ and older parr was 5.5/100 m². The smolt trap in the river Salaca was in operation between April 10 and May 18, In total 3571 salmon and 1369 sea trout smolts were caught; 750 of them were marked using streamer tags for total smolt run estimation. The smolt trap catch efficiency was 9.5%. Thus, in total salmon and sea trout smolts were estimated to have migrated from the river Salaca in In the river Venta, wild salmon parr was only found below the Rumba waterfall in The number of 0+ parr decreased compared to 2015; from 16.7 to 3.8 parr/100 m². As in 2015, older parr were found in low densities. In river Gauja wild salmon parr production was very low in 2016, and no 0+ parr was found in the tributary Amata. Wild salmon were found in the small Gulf of the Riga rivers Vitrupe, Aģe and Pēterupe. Parr age structures in these rivers testify that salmon reproduction does not occur in every year. Only 0+ parr in low densities were caught in the Main Basin rivers Tebra (Saka system) and Užava. No wild salmon parr was caught in river Irbe in Lithuanian rivers Lithuanian salmon rivers are typical lowland ones, and many of them are tributaries in the Nemunas river system. They are mainly sandy, gravely rivers flowing in the heights of upper and lower Lithuania. Nevertheless, salmonids inhabit more than 180 rivers. In total, sea trout is spawning in 76 rivers whereas Baltic salmon spawns in rivers. Leaning on historical data and today s situation, Lithuanian salmon rivers can be divided into the same groups as for Baltic salmon in general (see Annex 3): i.e. 1) rivers inhabited by wild salmon; 2) rivers inhabited by artificially reared salmon; 3) rivers inhabited by a mixed (wild and reared) salmon population; 4) potential rivers, i.e. where salmon occurs only occasionally; 5) rivers where salmon has gone extinct (Kesminas et al., 2003). There are 12 rivers in Lithuania inhabited by salmon populations of different abundance and with different status. Purely natural salmon population are inhabiting Žeimena and its tributaries Mera and Saria. In the index river Žeimena there is only natural population; this river has never been stocked with artificially reared salmonids. Mixed populations, i.e. natural and reared, are found in the rivers Neris, Šventoji, Vilnia, B. Šventoji, Dubysa, Siesartis, Širvinta, and Vokė. Reared populations occur in

104 ICES WGBAST REPORT the Virinta, Jūra, and Minija rivers and some smaller tributaries. In these rivers, salmon releases are done regularly. Electrofishing is the main monitoring method for evaluation of occurrence and densities of 0+ and older salmon parr. Monitoring covers all main salmon rivers (including all potential rivers). In 2016, salmon parr were found in Žeimena, Mera (tributary of Žeimena) and Neris. Parr densities in Lithuanian rivers are presented in Table and Figures and The abundance of salmon parr depends on hydrological conditions, spawning success, and protection of spawning grounds. In 2016 the average density of salmon 0+ parr in the index river Žeimena decreased to 3.3 parr/100 m²; and only 0.4 >0+ parr were caught. Salmon parr existed at all six electrofishing sites, and in one site in the Mera tributary. These density values are near the long-term average. A low parr density was also registered in the Neris River in Wild salmon parr were caught at nine out of eleven sites with densities varying between 0 5 parr/100 m 2. The mean density was relatively low during the study season compared to in recent years (Table ); the abundance of 0+ parr in the Neris River decreased to 1.52 parr/100 m 2 and >0+ juveniles amounted to 0.3 parr/100 m 2. A main reasons behind the low parr densities in southeastern Lithuanian rivers might be attributed to climatic condition in summer 2015, which were characterized by low levels of water, caused by drought and high temperatures, especially during July August. Efforts to increase the areal of suitable salmon habitats in Lithuania have been successful with salmon restored in the Šventoji, Siesartis, Vilnia, Vokė and Dubysa rivers. Salmon also spawn in lower reaches of many small rivers: Mera, Kena, Musė, Širvinta, Virinta, Dūkšta, Žalesa, Saria. Salmon parr densities strongly depends on climatic and hydrological conditions, and water temperature, as indicated by the 2010 year data. All these factors influence salmon parr and cause natural changes in their abundance. A salmon restocking program in Lithuania was carried out , and monitoring and stocking is still ongoing. There are many measures implemented every year to augment salmon populations, including artificial rearing, construction of fishways, protection of spawning grounds, stock monitoring and scientific projects. Despite measures taken, data from monitoring show that the smolt production in the Nemunas basin is increasing very slowly. A notable increase in production has only been observed in recent years; during the period estimated total smolt production increased substantially from to individuals. However, adverse ecological conditions in 2010 significantly decreased parr abundances in many important salmon rivers. As a consequence, the estimated smolt production in 2011 decreased to 6656 individuals. In 2012 the estimated smolt production again increased to about smolts, followed by a new decrease in to around In 2016 the total production was , a level similar to that estimated in 2015 ( individuals). Salmon smolt production estimates for single rivers increased notably in some cases (Neris, Žeimena, Šventoji and Minija) and slightly decreased in others (Siesartis, Vilnia, Vokė, Virinta, B. Šventoji and Dubysa). According to survey data, 2016 was not an exceptional year; it was characterized by parr densities corresponding to average values of multiannual results from the monitoring programme. However, water temperature was substantially above the average in the last few years and water levels were low. The correlation between salmon juvenile density and water temperature during July, the hottest month of the year, was investigated in two rivers characterized by different thermal regimes; Neris (r = -0,530, p =0,035) and Žeimena (r = -0,555, p =0,021). It was

105 100 ICES WGBAST REPORT 2017 found that during a period of several years, water temperatures in July varied within a range of a few degrees (19.1 C on average). However, in 2010 it reached 22.6 C which could have had a lethal impact on some of the weaker juveniles in the river. During that year also the parr density was estimated to be the lowest in Žeimena; only 0.2 parr/100 m². Average temperature during July in Neris is 20.9 CA. Temperatures above stress level (>22 C) were seen six times during a period of 16 years, in 2001, 2002, 2006, 2010, 2012 and The result of this study demonstrated that the thermal regime is very important for salmon production in Lithuanian salmon. Other concerns include pollution, and that rivers are of lowland type with scarce parr rearing habitats. Finally, quite high mortality rates are expected due to predation; predator densities are significantly higher in comparison to typical salmon rivers in the northern Baltica Rivers in assessment unit 6 (Gulf of Finland, Subdivision 32) All three wild salmon populations in the Gulf of Finland area are located in Estonia: Kunda, Keila and Vasalemma. These rivers are small and their potential production is small. In addition there is natural reproduction supported with regular releases in ten other rivers: Kymijoki, Gladyshevka, Luga, Purtse, Selja, Loobu, Valgejõgi, Jägala, Pirita and Vääna. In these mixed rivers natural reproduction is variable, and enhancement releases have been carried out in (Table ). The salmon in rivers Narva, Neva and Vantaanjoki are of reared origin. Status of wild and mixed AU 6 populations Parr density in the wild river Keila started to increase significantly in 2005, and in 2013 the 0+ parr density reached its highest reported density (157.1 parr/100 m²). In 2014 the density of older parr also increased to approximately 48.9 parr/100 m², the highest level so far recorded. Since then total density (0+ and older) has remained above 120 parr/100 m². Therefore it can be stated that the river Keila population is in a good state and with a clear positive trend (Figure ). The parr densities have been varying in river Kunda and a positive trend is only evident in the past two years (Table ). In comparison, the river Vasalemma is in a more precarious state, although some stronger year classes have occurred. The average 0+ density in 2015 decreased to 2.1 parr/100 m². In 2016 the 0+ density increased to 18.2 parr/100m², which is slightly above the long-term average. The most important change in the 1990s was the occurrence of salmon spawning in the Estonian mixed rivers Selja, Valgejõgi and Jägala, after many years without natural reproduction. In 2006 wild salmon parr was found also in rivers Purtse and Vääna. Since then a low and varying wild reproduction has occurred in all these mixed rivers (Table ). In the period parr densities increased to relatively high levels in these rivers. However, in 2016 parr densities were again low. In 2016 the Kotka dam in river Valgejõgi broke, and it will not be rebuilt. Thus, in autumn 2016 salmon was able to ascend to potential spawning areas that before were not accessible, and a considerable increase in salmon abundance may be expected in coming years. In 2011 and in 2012 the parr density was low in Finnish mixed river Kymijoki because of exceptional flow conditions. Since then the parr density has increased considerably, and in 2015 parr density increased to its long-term maximum (112.7 parr/100 m²). In 2016 the parr density again decreased to 33.7 parr/100 m², which is near the long-term average. Salmon releases are carried out annually in Valgejõgi (since 1996), in Selja (since 1997), in Jägala and Pirita (since 1998), in Loobu (since 2002) and in Purtse (since 2005). According to the rearing programme by Estonian Ministry of Environment (for the period

106 ICES WGBAST REPORT ) releases will be continued in these rivers. Salmon used for stocking in late 1990s originated from spawners caught in the river Narva and Selja brood stock fisheries. In addition, salmon from the Neva strain was imported as eyed eggs from a Finnish hatchery in In brood fish were again caught from river Narva. A captive brood stock based on salmon from wild river Kunda was established in 2007 at Polula Fish Rearing Centre, and all salmon releases in Estonia (SD 32) are nowadays based on that stock. In river Vääna releases were carried out from 1999 to The stocking was stopped due to the high risk of returning adults straying into the neighbouring Keila, considered to be a wild salmon river. On the northern side of the Gulf of Finland all wild salmon populations in Finland were lost in the 1950s due to gradual establishment of paper mill industry and by closing the river Kymijoki by dams. The geographically nearest available strain, Neva salmon, was imported from Russia in the late 1970s and releases into the rivers Kymijoki and Vantaanjoki started in The River Kymijoki is mainly used for hydroelectric production and pulp industries. The quality of water, however, has improved significantly since early 1980s. Reproduction areas exist on the lowest 40 kilometres of the river. Ascending spawners originate mainly from hatchery-reared smolt releases. Natural production has been estimated to vary between 7000 and smolts in the last fifteen years ( smolts expected in 2017). Along with the gradual increase in natural smolt production, the releases have decreased in the last few years. The released number of smolts (on average per year, ), however, is still clearly larger than the natural smolt production (on average per year, ). The brood stock of salmon is held in hatcheries, and it has frequently been partially renewed by ascending spawners. An inventory of rearing habitats in the river Kymijoki suggests 75 ha of smolt production area in the eastern branches of the river, between the sea and Myllykoski (40 km from the river outlet). Out of this total, about 15 ha of the rapids are situated in the lower reaches with no obstacles for migration, whereas about 60 ha are located beyond dams. The potential smolt production was assessed on the basis of assumed parr density (maximum >1 parr/1 m²) and smolt age distribution (1 3 years). The annual mean potential was assessed to be 1,34 smolts per ha, yielding a total potential of the river of about smolts per year. From this potential, annually about smolts could be produced in the lower reaches and smolts in the upper reaches of the river (Table ). The water conditions in winter time influence the hatching success in productions areas below the lowest dams. Higher water levels in mild and rainy winters in 2005 and 2015 (especially) were followed by higher parr densities. In general the parr densities have been on a moderate level, but some improvement have occurred since 2005 (Table ). Despite rainy autumns, most of the nursery areas in the lower part dry out because of water regulation between the power plants. Good quality habitats are located above the lowest power plants, but presently spawners can only access those areas via two branches where the dams are equipped with fish ways. The fish ladders in the Langinkoski branch do not function well and salmon can ascend the dam only in rainy summers when the discharge is high. Because of higher outflow, usually most of the spawning salmon ascend to the Korkeakoski branch, where a fish pass at the hydropower station was finished in So far the smolt production areas beyond the dams are only partially utilized. The new fish pass is expected to allow access of a much higher number of spawners to the better spawning and rearing habitats above the dams. If the fish pass will work well, it will increase the natural smolt production of

107 102 ICES WGBAST REPORT 2017 the river significantly. However, in the autumn 2016 only ten salmon ascended through the new fish pass, even though a much higher number of spawners were observed below the dam. In the river Vantaanjoki, electrofishing surveys in have shown only sporadic occurrence of salmon parr and only at a few sites. In Russia, Luga and Gladyshevka are the only rivers with wild Baltic salmon reproduction. In Luga the salmon population is supported by large and long-term releases. The released smolts are based on ascending Luga and Narva river spawners, as well as on a brood stock of mixed origin. In the mixed River Luga, a smolt trapping survey has been conducted since The natural production has been estimated to vary from about 2000 to 8000 smolts in different years. There has been some increase in the wild smolt production during the last years; about 6700 wild smolts were estimated in 2010 compared to 4000 smolts in 2009 and 3000 smolts in In 2015 the smolt trapping indicated some decrease (5300 wild smolts). The total potential smolt production of the river has been assessed to be about smolts, and the current wild reproduction is thus very far from its maximum level. The main reason for this poor situation in believed to be intensive poaching in the river. 3.2 Potential salmon rivers General The definition of a potential salmon river is a river with potential for establishment of natural reproduction of salmon (ICES, 2000). Current status of restoration programmes in Baltic Sea potential salmon rivers is presented in Table Releases of salmon fry, parr and smolt have resulted in natural reproduction in some rivers (see Table ). Reproduction and occurrence of wild salmon parr has in some potential rivers occurred for at least one salmon generation. Before any of these rivers may be transferred to the wild salmon river category, the Working Group needs more information of river-specific stock status Potential rivers by country Finland Nine potential salmon rivers are listed in Table Out of these three rivers Kuivajoki, Kiiminkijoki and Pyhäjoki were selected to be included in the Finnish Salmon Action Plan (SAP) programme. These SAP rivers are all located in AU 1 (Subdivision 31). Densities of wild salmon parr in electrofishing surveys in the SAP rivers are presented in Table Hatchery reared parr and smolts have been stocked annually in the rivers since the 1990s. Due to poor success of stock rebuilding to date, especially in the Pyhäjoki and Kuivajoki, the monitoring activities and stocking volumes have been decreased. Current activities include regular salmon releases only in Kiiminkijoki. In 2015, smolts and newly-hatched fry (of Iijoki river origin) were stocked. Electrofishing is currently conducted irregularly in the Kiiminkijoki. In the average densities of wild 0+ (one-summer old) parr in the river have ranged between individuals/100 m 2 (Table ). In the rivers Kuivajoki and Pyhäjoki, the observed densities in ranged from and parr/100 m 2, respectively. The poor success of stock rebuilding in the rivers is probably due to a combination of high fishing pressure and insufficient quality of both water and physical habitats. Also,

108 ICES WGBAST REPORT the flow is temporally low. All this together make the survival rate and reproductive success of salmon low. Small-scale natural reproduction has also been observed in rivers Merikarvianjoki and Harjunpäänjoki (tributary of Kokemäenjoki at the Bothnian Sea, Subdivision 30), and in the Vantaanjoki at the Gulf of Finland (Subdivision 32). The density of wild salmon parr in the lower reaches of the Kymijoki (Subdivision 32) has in recent years ranged from 16 to 113 parr/100 m 2, with the highest density so far in In all the above named rivers, salmon smolts are released annually, and there is angling of salmon and sea trout. Lately, plans have emerged for building up fish ladders and rebuilding migratory fish stocks in the large, former Finnish salmon rivers. Projects are underway to study the preconditions for these activities in rivers Kemijoki, Iijoki, Oulujoki and Kymijoki. For instance, salmon have been caught from at the mouths of Iijoki and Kemijoki, and they have been tagged with radio transmitters, transported and then released to upstream reproduction areas. The in-river behaviour of these salmon was monitored until the spawning time. Furthermore, downstream migration and survival of smolts through dams have been studied in these rivers. Finally, some electrofishing sites have been sampled during in the rivers Perhonjoki, Kalajoki and Siikajoki. No wild salmon parr were found, however. River Pyhäjoki has not been sampled by electrofishing during these years. Although there is no wild reproduction in these rivers at the moment, it does not mean that they are not potential, according to the definition commonly agreed in WGBAST. Sweden Three potential Swedish salmon rivers are listed in Table : Moälven, Alsterån and Helgeån. Densities of wild salmon parr in electrofishing surveys in Alsterån are presented in Table In recent years the former potential rivers Testeboån (2013) and Kågeälven (2014) did receive status as wild rivers by WGBAST. Restoration efforts are ongoing at the regional local level in several of the remaining potential Swedish salmon rivers. However, so far recent stocking activities and/or too low natural production have prevented them from having their status upgraded. Until next year (2018), the intention is to review and potentially update the list of Swedish potential salmon rivers. Lithuania Two potential Lithuanian salmon rivers, Sventoji and Minija/Veivirzas, are listed in Table River Venta, previously included as potential salmon river, is not listed this year. In May 2016, a total of salmon smolts were released into five rivers: Neris, Šventoji (Neris basin), Dubysa, Jūra and Merkys. In addition, a total of salmon fry were released into the Neris basin (rivers Neris, Vilnia, Muse, Vokė, Dūkšta and Kena); into the Šventoji basin (Šventoji, Širvinta, Siesartis, Virinta); into the Dubysa basin, into the Minija basin, into the Jūra basin and 4000 into the Šventoji basin (B. sea). It has been observed that restocking efficiency in smaller rivers is much greater than in larger ones, which is in line with a survey that indicated that in larger rivers mortality of juveniles is greater. Electrofishing densities of wild salmon parr in potential Lithuanian rivers are presented in Table Salmon densities in 2016 were at average levels in some larger tributaries of Neris and Šventoji. Salmon parr densities in Šventoji basin decreased

109 104 ICES WGBAST REPORT times, compared to the last year. Average total parr density in this river were close to the mean levels for the whole study period and reached 4.8 parr/100 m 2 ( and 1.7 >0+/100 m 2 ). However, in Siesartis, a tributary of Šventoji, the average total density of salmon juveniles increased to 9.1 parr/100 m 2 ( and 3.2 >0+/100 m 2 ). The density in 2016 also increased 1.9 times in Virinta, to 4.65 parr/100 m 2 ( and 1.0 >0+/100 m 2 ). In Vilnia and Vokė the densities of juvenile salmonids increased in 2016, compared to in the previous year, and was sufficiently high with 18.1 parr/100 m 2 ( and 3.2 >0+/100 m 2 ) in Vilnia and 4.45 parr/100 m 2 in Vokė. Densities in other potential salmon rivers in 2016 were lower: 0.41 parr/100 m 2 in B. Šventoji, and 0.75 parr/100 m 2 in Dubysa. In Minija the density of salmon parr was also lower with 1.4 parr/100 m 2. Poland Restoration programmes for salmon in seven potential Polish rivers (Table ) were started in 1994, based on releases of hatchery reared Daugava salmon. To date, however, there is no good evidence of successful re-establishment of any self-sustaining salmon population. In 2016 the total number of released hatchery reared fry and alevins was , and the number of released smolts was Since at least 2011, salmon spawners have been observed in the Vistula river system, but there are still no data on wild progeny. In total smolts were released into this river system in 2016 (Subdivision 26). Salmon spawning has been observed in the Drawa River (the Odra R. system) for some years, but the number of redds have stayed on a low level (not higher than ten per year). Until present, there is only one evidence of a few wild salmon progeny born in the river (result from spawning in 2013). In 2016, a total of 3000 smolts were released in the Odra river system (Subdivision 24). In 2016, a total of smolts and fry and alevins were released into Pomeranian rivers (Subdivision 25). In almost all these Pomeranian rivers, ascending and spent adult salmon have been observed and caught by anglers, but so far wild parr has only been found in the Slupia River. Densities of wild salmon parr in that river from electrofishing surveys are presented in Table Russia One potential Russian salmon river is listed in Table The Gladyshevka River was selected as a potential river for the Russian Salmon Action Plan. Stocking of salmon with hatchery reared (Neva origin) young salmon is ongoing in this river. Since 2001, a total or nearly salmon parr and smolts has been released in the river. In addition, about 2000 of one-year old salmon tagged by T-bar tags were released in Densities of wild salmon parr from electrofishing surveys in Gladyshevka are presented in Table Since 2004, wild salmon parr have occurred in the river. In 2013 parr densities varied from parr/100 m 2 (average of 6 parr/100 m 2 ) in the different rapids; in 2014 from parr/100 m 2 (average about 2 parr/100 m 2 ). In 2015 densities increased significantly up to 26,7 40,0 parr/100 m 2 ; average 0+ parr density was 23.2 parr/100 m 2 (about 70% of the total number of juveniles). No electrofishing surveys were carried out in 2016 due to high water level. Estonia No potential salmon rivers have been pointed out in Estonia.

110 ICES WGBAST REPORT Latvia No potential salmon rivers have been pointed out in Latvia. However, rivers Liela Jugla and Maza Jugla (in lower part of the river Daugava system) are regularly stocked by one summer salmon and sea trout parr, and electrofishing and habitat mapping are carried out. Germany No potential salmon rivers have been pointed out in Germany. So far, no German rivers with outlet into the Baltic Sea exist with a known (former) wild salmon population, and there is potentially no significant natural German salmon smolt production. However, in 2014 and 2015 a few salmon were caught during spawning migration in the river Warnow (W. Loch, pers. comm.). Denmark No potential salmon rivers have been pointed out in Denmark. 3.3 Reared salmon populations Releases The total number of salmon smolts released in reared rivers around the Baltic Sea in 2016 is presented in Table In AU 1 5 (Subdivisions 22 31) it was about 3.9 million, with an additional 0.6 million in AU 6 (Subdivision 32), making a grand total of 4.5 million smolts released in Releases of younger life stages (eggs, alevins, fry, parr) are presented in Table These releases have in many areas consisted of hatchery surplus, often carried out at poor rearing habitats. In such cases mortality among parr is high and releases correspond only to small amounts of smolts. On the other hand, when releases have taken place in potential, mixed or wild salmon rivers with good rearing habitats, they have had a true contribution to the smolt production. When comparing the total annual number of releases (of younger life stages) in the latest two years, the number is higher in AU 1, AU 5 and AU 6 in 2016 than in 2015, and lower in AU 2 and AU 3 (no releases in AU 4 since in 2012). However, in the longer term, the releases of younger life stages have decreased in the majority of the assessment units, with the exception of AU 5 where the observed trend is not as evident. Roughly, these releases is expected to produce less than smolts in the next few years. However, the data available to the working group do not allow distinction between rivers and release categories (age stages) and therefore the corresponding number of smolts expected from releases of younger life stages has not been possible to estimate properly. The yield from salmon smolt releases has decreased in the Baltic Sea during the last years, according to results from ongoing national tagging studies (Figures ). Possible explanations for lower catches include decreased offshore fishing and strong regulations in the coastal fishery. Initially, no substantial surplus of fish was observed in the rivers where compensatory releases were carried out, which most likely was due to decreased post-smolt survival. In recent years, however, the amount of salmon returning to reared rivers has increased, in some cases even dramatically.

111 106 ICES WGBAST REPORT 2017 The wild smolt production has increased considerably since the mid-1990s. Catch samples from the years indicate that the proportion of reared salmon has decreased and is presently well below 50 percent in most Baltic Sea fisheries (Table and Figure 2.8.1). Releases country by country The releases in Sweden are regulated through water court decisions. Due to that the reared (and wild) stocks were severely affected by the M74-syndrome in the early 1990s, the number of Swedish compensatory released salmon smolts in 1995 were only percent of the intended amount. However, already in 1996 the releases increased to the level set in the water court decisions. From that year and onwards the releases have been kept at the intended level each year. In 2016, a total of 1.8 million salmon smolts were released in AU 2, AU 3 and AU 4. The releases in AU 4 are minor and amounts to less than one percent of the total Swedish releases (Table 3.3.1). The number of one-year-old salmon smolts released in Sweden has increased, especially in the most southern rivers. From 2007 to 2016 the share of one-year old smolts has increased from 23% to 62% of the total Swedish smolt releases. The development reflects a combination of high-energy feed and longer growth seasons due to early springs and warm and long autumns. Many brood-stock traps in Swedish reared rivers were previously operated with equal intensity throughout the fishing season. The catch could therefore be considered as a relative index of escapement. However, only in two rivers (Skellefteälven and Ljusnan) has the brood-stock trap been operated with equal intensity until present (until 2015 in Skellefteälven). The reduced fishing intensity in most rivers with smolt releases reflects the increasing abundance of returning adults during the last ten years. Brood-stock fishing at low intensity during the migrating season is nowadays sufficient to get the amount of spawners (eggs) needed to fulfil terms in court decisions. Only in one river (Skellefteälven) was the brood-stock trap operated with equal intensity until in The reduced fishing intensity in the rivers with smolt releases reflects the increased abundance of returning adults during the last ten years. In Finland, the production of smolts is based on brood stocks reared from eggs and kept in hatcheries. The number of captive spawners is high enough to secure the whole smolt production. A partial renewal of the brood stocks has been regarded necessary in order to avoid inbreeding, and is consequently enforced occasionally by brood-stock fishing. In 2016, the total releases in AU 1 and AU 3 were 1.2 million smolts and smolts in AU 6 (Table 3.3.1). When the Finnish compensatory release programmes were enforced in the early 1980s, the total annual salmon smolt releases were about 2 million in total, whereof 1.5 million released in AU 1 and AU 3, and 0.5 million in AU 6. In recent years the releases have gradually been reduced. As in Sweden, the reared stocks in Finland have been affected by M74 over the years. In Russia there are annual releases in AU 6; in 2016 a total of reared smolts were stocked. Occasionally Lithuania also makes annual releases of a lower number of smolts in AU 5; in 2016 a total of smolts were released (Table 3.3.1). In Estonia a rearing programme using the Neva salmon stock was started in Eggs were collected from the reared Narva stock and the mixed Selja stock. In the late 1990s eggs were also imported from Finland. A captive stock based on spawners from river Kunda was established in One hatchery is at present engaged in salmon rearing. In 2016, the total annual smolt production was smolts released in AU 6, and another smolts released in AU 5 (Table 3.3.1). This was the first year since 1993 that Estonia released smolts in AU 5.

112 ICES WGBAST REPORT In Latvia the artificial reproduction is based on sea-run wild and hatchery origin salmon brood stock. The brood-stock fishery is carried out in the coastal waters of the Gulf of Riga in October November, as well as in the rivers Daugava and Venta. The mortality of yolk-sac fry has been low, indicating that M74 might be absent in this region. In 2016, the annual smolt production in Latvian hatcheries was about 0.6 million (Table 3.3.1). This is below the average number of releases during the latest decade. Earlier, from 1987 and onwards, the annual Latvian releases ranged up to 1.1 million smolts in several years. In Poland the last wild salmon population became extinct in the mid-1980s. A restoration programme was started in 1984, when eyed eggs of Daugava salmon were imported from Latvia. Import of eggs continued until In eggs for rearing purposes were collected from a salmon brood stock kept in sea cages located in the Bay of Puck. In subsequent years eggs have been collected from returning spawners caught in Polish rivers, besides from spawners reared in the Miastko hatchery. Spawners are caught mainly in the Wieprza River and in the mouth of Wisla River, but also from rivers Drweca, Parseta, Rega and Slupia. The yearly production amounts to million eggs. Stocking material (smolts, one-year old parr and one-summer old parr) are reared in five hatcheries. In 2016 the total smolt production was 259 thousand released in AU 5 (Table 3.3.1). This is slightly higher than in 2015, but lower than in the two previous years ( ) when approximately 0.4 million smolts were released. Starting from 1994, the annual releases have fluctuated between and 0.5 million smolts. In Germany no regular release programme exists in the Baltic region, as there are no known natural salmon populations. Consequently, there were no official releases of salmon in rivers with outlet into the Baltic Sea in However, a few irregular releases have been reported recently and in the past (e.g. river Trave and river Warnow). There is a controversy regarding the historic existence of wild Baltic salmon populations in German rivers. Until 2005, a rearing programme was run in Denmark in a hatchery on the Island of Bornholm using the river Mörrumsån stock (in AU 4). In the last year, tagged smolts were released (Table 3.3.1). The year before (2004), a total of smolts were released, and one experiment on artificial imprinting and another on establishment of a Terminal Fishery was performed. No more releases have been planned Straying rate Observations on straying rates of released salmon vary between areas. The level of straying is evidently dependent on several factors. For example, in Finland rearing of smolts is based on brood stocks kept in hatcheries, whereas in Sweden it is based on annual brood-stock fishing ( sea-ranching ). These differences in rearing practices may also influence straying rates. Strayers are often observed in the lower stretches of the rivers into which they have strayed. This may indicate that not all strayers necessary enter the spawning grounds and contribute to spawning, but instead that a proportion of them may only temporally visit the wrong river. This also implies that the place and time of collecting observations about straying is expected to influence the obtained estimates about straying rates. More information is needed to study these aspects of straying. According to scale analysis of catch samples collected from the Tornionjoki river fishery in , only eight salmon out of a total of 4364 analysed were identified as potential strayers from releases in other Baltic rivers. This indicates that about 0.2% of the salmon run into Tornionjoki were from other (reared) rivers, which corresponds to

113 108 ICES WGBAST REPORT 2017 about 100 strayers per year, if one assumes an average spawning run into Tornionjoki of about salmon. Tag recapture data of compensatory releases in the Finnish Bothnian Bay indicate that the straying rate of these reared fish to other rivers is 3 4%. From all these releases, however, strayers were found only among the Tornionjoki hatchery strain stocked into the mouth of Kemijoki, and all these strayers were observed in the Tornionjoki. Using these tag recaptures to calculate the amount of strayers in the Tornionjoki, assuming no strayers from the Swedish releases, there would be annually about 200 strayers in the Tornionjoki spawning run (corresponding to 0.4% straying into the river, again assuming a spawning run of about salmon). In Sweden, tag recoveries indicate that the average straying rate of reared salmon into other rivers has been % on average, but for some releases the straying rate has been as high as 10 30%. Highest straying rate of tagged salmon is often observed in reared rivers with annual releases, due to a high total exploitation rate from the commercial, recreational and brood-stock collection, and probably also because broodstock fisheries are carried out close to river mouths. 3.4 M74, dioxin and disease outbreaks M74 in Gulf of Bothnia and Bothnian Sea When calculated from all Swedish and Finnish data, the proportion of salmon females whose offspring displayed increased mortality in 2016 was on average 19%, compared to 2% in the preceding year (Table 3.4.1). Hence, the M74 syndrome returned to the Gulf of Bothnia salmon rivers after having remained very low in four preceding years. The thiamine (vitamin B1) concentrations in unfertilized eggs (reproductive period 2015/2016) from females in River Simojoki decreased approximately to half of that in the preceding year (Figure 3.4.1) when no M74-related mortality was observed in the Finnish M74 monitoring data (Table 3.4.2). In Swedish rivers the proportion of offspring groups with increased M74-like mortality varied from 4 to 39 in 2016, compared to 0 7% in 2015 (Table 3.4.3; Börjeson, 2016). In spring 2017 the incidence of M74 is expected to increase considerably: the average free thiamine concentration in eggs from salmon that ascended the River Simojoki in autumn 2016 had still decreased compared to autumn 2015 (Figure 3.4.1). For the second year thiamine was also measured in unfertilized eggs of salmon ascended 2016 in the Swedish Rivers Ume/Vindelälven and Dalälven, where mean free thiamine concentrations were lower than in the previous year (Figure 3.4.2). Moreover, there were wiggling females (i.e. with an uncoordinated swimming behaviour) among spawners of River Simojoki and especially in River Dalälven. The river-specific prognosis for the proportions of offspring groups in spring 2017 suffering from thiamine deficiency causing increased mortality varies from 30% up to 60 70%, depending whether wiggling females, not used for egg stripping, are included or not (Table 3.4.1). Furthermore, in the Swedish rivers Dalälven and Luleälven forced hatching of eggs in warmer water is performed for prognosis purposes, and the preliminary results for 2017 also indicate that the M74 prevalence will increase further compared to in hatching year The M74 frequencies in Table predominantly represent the percentage of females in a hatchery with a recorded increase in offspring mortality. In the rivers Simojoki, Tornionjoki, and Kemijoki, however, mortality estimates are based on both the proportion of females affected by M74 and the mean percentage yolk-sac fry mortality (Table 3.4.2). In Finnish estimates, annual M74 figures are based on female-specific experimental incubations in which M74 symptom-related mortality is ascertained by observations of yolk-sac fry (until the reproductive period 2010/2011) and/or comparing

114 ICES WGBAST REPORT mortalities with thiamine concentration of eggs (from 1994/1995 and onwards). Three estimates are presented: (1) the average yolk-sac fry mortality, (2) the proportion of females with offspring affected by M74, and (3) the proportion of those females whose offspring have all died (Keinänen et al., 2000; 2008; 2014; Vuorinen et al., 2014a). Mean annual yolk-sac fry mortalities and proportions of M74 females correlate significantly, but the M74 frequency has usually been somewhat higher than the offspring M74 mortality, especially in years when many offspring groups with mild M74 occur, i.e. when only a proportion of yolk-sac fry die. In years when the M74 syndrome is moderate in most offspring groups, the difference between the proportion of M74 females and mean yolk-sac fry mortality can exceed 20 percentage units (Keinänen et al., 2008). In contrast, Swedish data are based only on the proportion of females whose offspring display increased mortality (Table 3.4.3). In the hatching years the M74 syndrome resulted in a high mortality of salmon yolk-sac fry with an M74 frequency (i.e. the proportion of the females whose offspring were affected) over 50% in most Swedish and Finnish rivers (Table 3.4.1). Since then the incidence of M74 has on average decreased. However, it has varied greatly even between successive years so that some years (e.g. 1999, 2002, and ) have displayed elevated mortalities compared to other (e.g. 1998, and ) with low or non-existent mortalities. In 2012 the incidence of M74 could be considered as non-existent for the first time since the large outbreak in the 1990s. Earlier there was a tendency that estimates of M74-mortality were higher in Finland than in Sweden, but this difference seems to have disappeared in the years when the mortality has been low (Figure 3.4.3). The difference may be due to the fact that in Finland all females caught for M74 monitoring have been included, whereas in Sweden females that have displayed uncoordinated swimming (wigglers) have been excluded from incubation. Wiggling females are known to inevitably produce offspring that would all die from M74. The proportion of wiggling females was high in the early and mid-1990s (Fiskhälsan, 2007). Trends and annual fluctuations in average proportions of M74-affected females have been very similar in Swedish and Finnish rivers (Figure 3.4.3). However, in some recent years M74 has been insignificant or absent in the Finnish M74 monitoring, whereas rather high M74 frequencies have been reported from some Swedish rivers. It seems that those Swedish results may result from technical failures or too high or variable water temperatures, as reported by Börjeson (2013). In the Finnish M74 monitoring, but not in Sweden before 2015/2016, the mortality and female proportion figures for M74 incidence are ascertained by measuring the thiamine concentration of eggs (Figure 3.4.1). In the Finnish M74 data, the annual M74 incidence among the monitored Bothnian Bay rivers has been very similar. Therefore it is relevant to express the annual M74 mortality and proportion of M74 females as an average of all individual monitored salmon females (and respective offspring groups) that ascended those rivers (Keinänen et al., 2014). However, there may be some differences between salmon populations from rivers in the Bothnian Bay and in the Bothnian Sea, if migration routes and feeding grounds during the whole feeding migration differ. This would also explain different mortalities among offspring of salmon from the River Mörrum, from where the smolts descend directly into the Baltic Proper. Apart from observations in hatcheries and experimental incubations, effects of the M74-syndrome was also observed as decreased parr densities in some of the wild salmon populations in and also in the years 1995 and 1996, despite a high number of spawners (Karlström, 1999; Romakkaniemi et al., 2003; 2014). In the Swedish wild river Ume/Vindelälven in the Gulf of Bothnia, an estimate of the egg deposition is available together with an estimate of the parr densities derived from these

115 110 ICES WGBAST REPORT 2017 brood-year classes. It shows that the densities of 0+ parr were low in the years when the incidence of M74 was high, while parr densities were better correlated to the egg deposition in years when the incidence of M74 was low ( and ). Statistics from the Swedish River Dalälven collected during 14 years ( ) show that females (n = 1866) affected by M74 have a lower average weight than non-affected fish (Börjeson, 2011) and in also 3% lower condition factor (Börjeson, 2015). The reason for the weight difference is not known. It could be that affected M74 fish are younger than healthy females and contrary to older salmon have fed only on smaller and younger prey fish, or that they have grown less due to the nutritional conditions. Backman (2004) found that in wild salmon that ascended earlier and were larger had somewhat lower offspring M74 mortalities than fish that ascended later and were smaller. The same relationship was not found among reared salmon. In intra-annual comparisons among two sea-year salmon, only in some years with low M74 incidence, a negative correlation between the weight or size of females and yolksac fry mortality was found. On the contrary, a large size (weight or length) or high condition factor of mature or pre-spawning female salmon was related to high yolksac fry mortality in years of relatively high M74 incidence (Mikkonen et al., 2011). Although the high condition factor (CF >1.05) of pre-spawning salmon predicted high M74-related mortality, the high growth rate of salmon appeared not as such to be the cause of M74, but rather the abundance of prey and its quality (Mikkonen et al., 2011). Most salmon feed in the Baltic Proper, but reared salmon smolts have been found to remain to feed more often than wild salmon in the Bothnian Sea (Jutila et al., 2003), where salmon growth has generally been slower than in the Baltic Proper (Salminen et al., 1994; Niva, 2001; Keinänen et al., 2012). Evidently, because cod (Gadus morhua) compete with salmon for food in the Baltic Sea (Larsson, 1984), the annual growth rate and the condition factor of pre-spawning salmon were both inversely related to the size of the cod stock (Mikkonen et al., 2011). From the various stock factors of sprat (Sprattus sprattus) and Baltic herring (Clupea harengus membras) in the southern Baltic Proper, the biomass of sprat had the strongest positive relationships with the growth rate and condition factor of pre-spawning salmon, and the total prey biomass with yolk-sac fry mortality (Backman, 2004; Mikkonen et al., 2011). However, sprat was the dominant prey species of salmon in that feeding area in years of high M74 incidence, and already earlier had M74 been shown to be statistically well correlated with parameters describing the sprat stock (Karlsson et al., 1999). The M74 syndrome has unquestionably been linked to a low concentration of thiamine in salmon eggs (Lundström et al., 1999; Vuorinen and Keinänen, 1999; Koski et al., 2001), although some other relationships have also been found. However, yolk-sac fry suffering from M74 can be restored in hatchery to a healthy condition by treatment with thiamine (Koski et al., 1999). A pale egg colour in M74 eggs (Börjeson et al., 1999; Keinänen et al., 2000) is a result of a low concentration of carotenoids, especially astaxanthine having antioxidant property (Lundström et al., 1999; Pettersson and Lignell, 1999; Vuorinen and Keinänen, 1999). An increase in the concentrations of particular organochlorines in salmon spawners ascending the River Simojoki, coincidentally with the outbreak of M74 at the start of the 1990s, was concluded to have resulted from enhanced feeding on sprat in which the concentrations of these organochlorines were high in younger age groups with the greatest fat content (Vuorinen et al., 2002). Bioaccumulation of specifically these organochlorines, coplanar PCBs, was most distinctly affected by the fat content of the prey and predator fishes (Vuorinen et al., 2012).

116 ICES WGBAST REPORT The fat concentration of sprat is nearly twice that of herring, and it is highest in the youngest sprat. The percentage of lipid also varies more in sprat than in herring (Keinänen et al., 2012). The average thiamine concentration in sprat and herring (of the size preferred by salmon as prey) sampled in different seasons and years are quite similar (Keinänen et al., 2012; 2017), although in autumn samples it was lower in sprat than in herring (Vuorinen et al., 2002). However, in both prey species the thiamine concentration by several times exceeded the nutritional guidelines on growth of salmon. The thiamine concentration changed curvilinearly with the age of both sprat and herring, being lowest in the youngest age groups [and also in the oldest herring of length >19 cm, and hence not often included as salmon prey (Hansson et al., 2001; Vuorinen et al., 2014b)] and greatest at 6 10 years in sprat and 3 7 years in herring (Keinänen et al., 2012). As thiamine has a central role in the energy metabolism, its nutritional requirement is determined by the energy density of the diet, which means the fat content of prey fish. Thus, ample abundance of (fat) young sprat as food for salmon increases the requirement of thiamine. Contrary to demand, the thiamine content per unit fat and energy in the diet of salmon has been least during years and in areas where recruitment and biomass of sprat have been high (Mikkonen et al., 2011; Keinänen et al., 2012). An abundance of dietary lipid increases the content of unsaturated fatty acids, especially DHA, in the diet of salmon (Keinänen et al., 2017). These are susceptible to peroxidation and increase oxidative stress. Because of lipid peroxidation and the antioxidant property of thiamine, the thiamine reserves are further depleted at an increasing rate (see Keinänen et al., 2012; 2017) during the long spawning migration followed by a long pre-spawning fasting period of salmon (Ikonen, 2006). Diminished body stores do not allow adequate deposition of thiamine into developing oocytes; the development of offspring cannot be sustained until the end of the yolk-sac period, when fry start external feeding. Because M74 is induced by the ample but unbalanced food resources for salmon (primarily sprat), the incidence of the M74 syndrome may be reduced and even prevented. The safest strategy for attaining this objective would be to ensure a large and stable cod stock in the Baltic Sea (Casini et al., 2009) to prey on the sprat, and possibly also by managing the sprat fishery in years when the cod stock is weak (Mikkonen et al., 2011; Keinänen et al., 2012). Evidently, as a consequence of strengthening of the cod stock and flattening out of the sprat stock (ICES, 2012) the incidence of M74 decreased and was virtually non-existent in However, more young sprat than for many years were observed in connection to annual surveys in the autumn of 2015, indicating a strong 2014 year class of sprat in the Baltic Proper (J. Raitaniemi, oral info). Apparently, an abrupt increase in the abundance of young sprat appeared as low thiamine concentrations in unfertilized eggs of salmon ascended the Gulf of Bothnia rivers in the autumn The thiamine concentrations were even lower in salmon ascended in autumn Thus, M74 returned in the hatching spring 2016 after several good years, and the M74 mortality is expected to increase further in the hatching spring In the Stock Annex (Annex 3, Chapter CA.1.6) a description is given of a Bayesian hierarchical model applied to the Gulf of Bothnian (GoB) monitoring data (Tables and 3.4.3) of M74 occurrence from rivers in Finland and Sweden, to obtain annual estimates of the M74-derived yolk-sac fry mortality. This information is needed to fully assess the effects of M74 on the reproductive success of spawners. Besides annual estimates of M74 mortality in the rivers, where mortality has been recorded, the model provides annual estimates of the mortality for any GoB river, in which no monitoring has been carried out (see Table and Figure in ICES 2016). Most of the wild stocks, including all smaller wild rivers in the GoB, belong to this group. The results demonstrate that in some years the actual M74 mortality among offspring has been

117 112 ICES WGBAST REPORT 2017 lower than the proportion of M74 females indicated, which apparently is related (see above) to mildness of the syndrome, i.e. to partial mortalities in offspring groups M 74 in Gulf of Finland and Gulf of Riga In the River Kymijoki in the Gulf of Finland the incidence of M74 has in many years been lower than in the Rivers Simojoki and Tornionjoki (Table 3.4.1; Keinänen et al., 2008; 2014). However, in the reproductive period 1997/1998, for example, when M74 mortalities among salmon yolk-sac fry of the Gulf of Bothnia rivers was temporarily low, the situation was vice versa; evidently this reflected variation in sprat abundance between the main feeding areas, i.e., the Baltic Proper and the Gulf of Finland. The long-term tendency has, however, been roughly similar. The River Kymijoki of the Gulf of Finland, with introduced salmon originating from the Neva stock, was included in the Finnish M74 monitoring programme from the year 1995, but no data for the years and exist because of problems in salmon collection for monitoring. Therefore the latest mortality data from the R. Kymijoki are from spring 2007 (Table 3.4.1). However, in autumn 2013 a few Kymijoki salmon females were caught for renewing of the brood stock. Based on relatively high thiamine concentrations in unfertilized eggs (mean 3.2 ± 1.1 nmol/g, N = 5) of all five salmon, M74 mortalities in spring 2014 were unlikely. In Estonia, M74 has been observed in hatcheries in some years during the period , but the mortality has not exceeded 15%. A small number of spawners is collected for brood stock from river Kunda since 2013, and no fry mortality has been observed. However, in 2016 the eggs from one female (out of four) displayed mortality after hatching. This recent observation indicates that M74 may occur also in Gulf of Finland. There is no evidence to suggest that M74 occur in Latvian salmon populations. In the main hatchery Tome, the mortality from hatching until the start of feeding varied in the range of 2 10% in the years In addition, parr densities in Latvian river Salaca did not decrease during the period in the 1990s when salmon reproduction in the Gulf of Bothnia was negatively influenced by M74 (Table ). Before ascending the river, salmon from Daugava and Salaca feed in the Gulf of Riga, where the main prey species of salmon was herring during the years (Karlsson et al., 1999; Hansson et al., 2001). Although sprat was the dominant prey species in the Baltic Proper during that time period, the salmon diet in the Gulf of Riga did not include sprat. Furthermore, in contrast to salmon feeding in the Baltic Proper or the Bothnian Sea, the proportion of other prey species, such as sand eel (Ammodytes spp.), perch (Perca fluviatilis), smelt (Osmerus eperlanus) and cod, was considerable in the Gulf of Riga (Karlsson et al., 1999; Hansson et al., 2001). Salmon in River Daugava moreover ascended later than salmon in Gulf of Bothnia rivers (Karlsson et al., 1999). Dioxin In Sweden, the National Food Agency is responsible for sampling, analysis and dietary recommendations regarding dioxin in fish. In the latest published report, the results indicate elevated levels of dioxin in Baltic salmon caught along the coast (Fohgelberg and Wretling 2015). The Swedish control program is set up in accordance with EU regulation 589/2014. Limits are set out in EU Regulation 1881/2006 with updates in EU Regulation 1259/2011. Sweden has an exception to the limits of dioxin when it comes to salmon and a few other fish species in the Baltic Sea and in the lakes Vänern and Vättern. Also Finland has an exemption to the EC regulation 1259/2011 which allows selling of Baltic salmon and sea trout in the domestic market. No export of wild-caught salmon or sea trout has taken place.

118 ICES WGBAST REPORT In Denmark the following restrictions for marketing of salmon (and sea trout) were enforced from December 5th, 2016: Salmon 5.5 kg gutted weight caught in ICES SD must be trimmed (deep skinned) before marketing. In the same SD s salmon weighing > 5.5 kg and < 7.9 kg can be marketed, if both trimmed and the ventral part of the fish is removed. Each batch of salmon > 2.0 kg caught in ICES SD must also be analyzed for dioxin before marketing. Dioxin levels found in samples taken in 2006 and 2013 were comparable, while samples from 2011 contained slightly lower concentrations of dioxin Disease outbreaks Since 2014 an increasing number of reports from fishermen and local administrators of dying or dead salmon have come from Swedish and Finnish salmon rivers, spanning from Tornionjoki to Mörrumsån. The affected salmon have displayed various degrees of skin damages, from milder erythemas and bleedings, to UDN-like (Ulcerative Dermal Necrosis) lesions and more severe ulcers and traumatic wounds that are typically followed by secondary fungal infections causing death (SVA 2017). To some extent also other fish species, such as sea trout, whitefish and grayling have been reported with the same symptoms. So far the disease prevalence has varied considerably between both rivers and years. In some rivers there so far are no reports of elevated levels of dead salmon. The most severe disease outbreaks occurred in Tornionjoki ( ), Kalixälven (2015), Ume/Vindelälven ( ), Ljungan (2016) and Mörrumsån ( ). In several cases the number of dead salmon (and other species) has been considerable, although quantitative estimates of total death rates are missing. However, in e.g. Mörrumsån, it has been noted that following a year with disease very few overwintered spawners (kelts) appear to remain the following spring according to river catches. In 2015 and 2016 the Swedish National veterinary institute (SVA) and the Finnish food safety authority (Evira) have conducted investigations aimed at identifying the cause of the salmon disease. Analyses of salmon from Tornionjoki in 2015 showed that some of the sampled fish displayed UDN-like symptoms. Cultivation for virus and bacteria in 2016 did not provide conclusive answers, although in some cases bacteria associated with skin lesions were identified. Next generation sequencing indicated presence of herpes- and iridoviruses in individuals with erythemas. These viruses may cause skin lesions, but these findings need to be investigated further (SVA and Evira, in prep.). Although it appears likely that the disease outbreaks in Swedish and Finnish salmon rivers during recent years have a common cause, this still remains to be proved. Notably, since 2007 severe disease problems have also occurred in all Polish Pomeranian sea trout rivers. Similar to in northern salmon rivers, the disease problems in Poland have been variable over time with peaks and drops (Bartel et al., 2009). Further, the affected sea trout display UDN-like skin damages followed by fungal infections, high mortality and lack of kelts. Also salmon and grayling with symptoms have been observed. Polish veterinary studies have been performed, but so far without any clear conclusions. Potential consequences of salmon disease outbreaks for the future development of wild stocks, and how this mortality may be monitored and handled in stock assessment are briefly discussed in Section

119 114 ICES WGBAST REPORT Summary of the information on wild and potential salmon rivers Wild smolt production in relation to the smolt production capacity is one of the ultimate measures of management success. Among the wild rivers flowing into the Gulf of Bothnia and the Main Basin (assessment units 1 5), smolt abundance is measured directly in the index rivers Simojoki and Tornionjoki/Torneälven (AU 1), Sävarån (AU 2, replaced with Rickleån since 2013), Vindelälven (AU 2), Mörrumsån (AU 4) and in the Latvian river Salaca (AU 5). In addition, a few years of smolts counting has also been performed in Lögdeälven (AU 2), Testeboån (AU 3) and Emån (AU 4) (Sections ) and counting in additional rivers may take place in the future. The smolt abundance model (Annex 3), which utilises all available juvenile abundance data, is a rigorous tool for formal assessment of current smolt production. Differences in the status of wild stocks have become more apparent in recent years, not only in terms of the level of smolt production in relation to potential production, but also in terms of trends in various indices of abundance. These differences are particularly clear when comparing different regions: most Gulf of Bothnia (AU 1 3) rivers have shown increases in abundance while many of the Main Basin (AU 4 5) rivers have shown either decreasing or stable abundance. Rivers in the Gulf of Bothnia (assessment units 1 3) The parr production in the hatching years of was as low as in the 1980s (Tables , and , and Figures , , , and ), although the spawning runs were apparently larger (Tables , , and Figures , ). In those years, the M74 syndrome caused high mortality (Table and Figure 3.4.1), which decreased parr production considerably. In the hatching years , parr densities increased to higher levels, about five to ten times higher than in the earlier years. These strong year classes resulted from large spawning runs in and a simultaneous decrease in the level of M74. The large parr year classes hatching in resulted in increased smolt runs in 2000 and 2001 (Table ). In spite of some reduction in parr densities during , parr densities and subsequent smolt runs stayed on elevated levels compared to the situation in the mid- 1990s. In 2003, densities of one-summer old parr increased in some rivers back to the peak level observed around 1998, while no similar increase was observed in other rivers. From , densities of one-summer old parr showed a yearly increase in most of the rivers, but in 2007 the densities of one summer old parr again decreased. Despite the relative high spawning run in 2009 the densities of one summer old parr in 2010 decreased substantially in most rivers, compared to the densities in The densities of one summer old parr in 2012 stayed at the same level as in 2011, or even increased, despite the relatively weak 2011 spawning run. The increased spawning run in 2012 did not substantially increase the densities of one summer old parr in 2013, whereas the increased spawning runs in 2013 and 2014 resulted in elevated densities of one summer old parr. Catch statistics and fishway counts also indicate some differences among rivers in the development in number of ascending spawners. To some extent, these differences may reflect problems with fish passages in certain rivers. For example, a survey in 2015 of the efficiency of the fishway in Piteälven indicated a large delay in the spawning run and loss of salmon and trout that didn t pass the fishway located below the spawning areas. Similar observations have also been identified in Åbyälven (Section 3.1.2). There has been pronounced annual variation in the indices of wild reproduction of salmon both between and within rivers. Variation in abundance indices might partly

120 ICES WGBAST REPORT be explained to extreme summer conditions in the rivers during some years, e.g. in and in 2006, which might have affected river catches and the fish migration in some fishways. Counted number of salmon in 2007 increased with about 50% compared to The additional increase in fishway counts in 2008 is in agreement with increased river catches, which more than doubled in 2008 compared to 2007 and were almost as high as in the highest recorded years (1996 and 1997). The spawner counts in 2010 and 2011 in combination with information on river catches indicated weak spawning runs in those years. The large increased spawning run in Tornionjoki in 2012, 2013, 2014 and 2016, as compared to 2011, resulted in increased total river catches with 40 70% compared to the two previous years. Most data from the Gulf of Bothnia rivers indicate an increasing trend in salmon production. Rivers in AU 1 have shown the most positive development, while stocks in the small rivers in AUs 2 3 have yet not showed a similarly positive development. These small rivers are located on the Swedish coast close to the Quark area (northern Bothnian Sea, southern Bothnian Bay). The recent period of low M74-levels has most likely affected the wild production positively. In 2016, however, levels of offspring mortality again raised, and preliminary data from two Swedish hatcheries and one Finnish indicate that M74-mortality among the offspring hatching in 2017 will increase further; preliminary results from Simojoki, Iijoki, Ume/Vindelälven and Dalälven indicate that offspring mortality for those rivers may be reaching around 50%. Disease outbreaks seen in recent years in several rivers is another mortality factor that may have a negative impact on future stock development (Sections 3.4 and 4.4.1). Rivers in the Main Basin (assessment units 4 5) The status of the Swedish AU 4 salmon populations in rivers Mörrumsån and Emån in the Main Basin differ, but they both show a similar slight negative trend in parr densities (Table and Figures and ). The outbreak of M74 mortality in the early 1990s might have decreased smolt production in mid-1990s, after reaching the historical highest parr densities in Mörrumsån at the turn of the 1980s and 1990s. In Emån, the smolt production has for long been far below the required level, which is most likely dependent on insufficient numbers of spawners so far finding their way to reproduction areas further upstream in the river system. The present status of the population in Mörrumsån is higher than in Emån, but at present uncertain. Whereas average electrofishing densities have not increased since the mid-1990s, smolt trapping results for the production in the upper part of Mörrumsån shows a positive trend in the period (Section 3.1.4). Not until recently updated production capacity priors for Mörrumsån and Emån (ICES, 2015) and smolt estimates from a revised southern river model (developed as part of the recent benchmark WKBaltSalmon 2017, in prep.) have been included together into the full life-history model, stock status in these rivers will be possible to assess accurately (Section 4.4). High disease related mortality among spawners in Mörrumsån (but yet not in Emån) in recent years is another factor that also may affect the future stock development (Sections 3.4 and 4.4.1). Among rivers in AU 5, the Pärnu River exhibit the most precarious state: no parr at all were found in the river in In the densities increased slightly, but in 2007, 2008, 2010 and 2011 again no parr were found. Reproduction occurred in 2008, 2011 and 2012 resulting in low densities of parr in 2009 and (Table , Figure ). There has been remarkable annual variation in parr densities, both within and between rivers in AU 5. Since 1997, parr densities in the river Salaca in Latvia have been on relatively high levels (Table , Figure ), but in 2010 and 2011 the densities decreased to the lowest observed level since the mid-1990s. In 2015

121 116 ICES WGBAST REPORT 2017 the density increased to the highest observed so far, but in 2016 densities decreased again. In river Gauja, parr density levels have been very low since In 2014, the 0+ parr density increased to a slightly higher level. It seems that in some of the smaller AU 5 salmon rivers (Saka, Pēterupe and Vitrupe) reproduction occurs only occasionally; as an example, salmon 0+ parr densities in 2015 increased to the so far highest observed in these small rivers, except in Pēterupe, whereas in 2016 the 0+ densities were again very low or even zero. Although only relatively short time-series of parr and smolt abundance are available from Lithuanian salmon rivers, the latest monitoring results (Table ) indicate somewhat similar variation in juvenile production as seen in Latvian rivers. The observed parr densities are very low in relation to observed parr densities in most other Baltic rivers. This illustrates the poor state of several wild salmon stocks in AU 5. These stocks might be in a higher risk of extinction than any of the stocks in AU 1 3 (Gulf of Bothnia). In Lithuania, measures have been carried out since 1998 to increase salmon populations (Section 3.1.5). Implementation of measures has stabilized the salmon populations in Lithuanian rivers, and the production is increasing very slowly. Pollution also affects the salmon rivers. Another important factor in Lithuanian rivers, which are of lowland type, is lack of suitable habitats for salmon parr and high summer temperatures. Besides regulation of fisheries, many of the salmon rivers in the Main Basin (AU 4 5) may need habitat restoration and re-establish connectivity, which aim at stabilizing and improving natural reproduction. For instance, in the Pärnu River, the Sindi dam prevents access to over 90% of the potential reproduction areas. In the rivers Mörrumsån and Emån, new fish passes have increased significantly the available reproduction area to salmon. Rivers in assessment unit 6 (Gulf of Finland, Subdivision 32) The 0+ parr density in Estonian wild rivers Kunda and Keila remained high in In Vasalemma the 0+ parr density increased, approaching the average level in the river. The status of river Keila is considered to be good, whereas improvement has been modest in rivers Vasalemma and Kunda. Because of highly variable annual parr densities in Vasalemma and Kunda, the status of these wild populations still must be considered uncertain. In the Estonian mixed rivers Purtse, Selja, Loobu, Valgejõgi, Jägala, Pirita and Vääna, wild parr densities mostly decreased in However in the preceding three years ( ) parr density stayed above the long-term average in all of these rivers. The clearest positive trend can be seen in Selja, Valgejõgi, Loobu and Pirita. However, because of the high fluctuations in recruitment, the status of these populations remains uncertain. To safeguard these stocks additional regulatory measures were enforced in 2011 (see Section 2.9) and positive effect of these measures can be seen as increases in wild parr densities and as a relatively satisfactory amount of ascending spawners to R. Pirita in recent years ( ). In Russia, wild salmon reproduction occurs in rivers Luga and Gladyshevka. The status of both these stocks is considered very uncertain. However, high densities of 0+ salmon parr occurred in Gladyshevka in 2015, whereas there was no monitoring in Since 2003, there is no information that suggests wild reproduction in river Neva. In Finland, the wild production in the mixed river Kymijoki has increased during the last ten years. However, the present natural reproduction in the lower part of the river has still remained below the river potential (Section 3.1.6).

122 ICES WGBAST REPORT Total natural smolt production in Estonian, Finnish, and Russian rivers in the Gulf of Finland area was estimated to about in In 2016, the estimated wild AU 6 smolt production slightly increased to about It is estimated that the wild smolt production will remain at a fairly similar level also in 2017 (ca smolts). The AU 6 smolt releases since year 2000 has been on a stable level. The exception was the year 2011 when releases were reduced to almost half (Table 3.3.1). The reduction in Russian smolt releases was caused by exceptionally warm climatic conditions in the summer 2010 causing high parr mortality in hatcheries.

123 118 ICES WGBAST REPORT 2017 Table Salmon catches (in kilos) in four rivers of the Subdivision 31, and the catch per unit of effort (cpue) of the Finnish salmon rod fishing in the river Tornionjoki/Torneälven. Simojoki Kalixälven Byskeälven Tornionjoki/ Torneälven (au 1) (au1) (au1) (au2) Finnish Swedish Total CPUE catch, kilo catch, kilo catch, kilo catch, kilo catch, kilo catch, kilo grams/day ) ) 561 2) /736 3) ) Ban of salmon fishing 1994 in Kalixälven and Byskeälven and the Swedish tributaries of Torneälven. 2) Calculated on the basis of a fishing questionnaire similar to years before ) Calculated on the basis of a new kind of fishing questionnaire, which is addressed to fishermen, who have bought a salmon rod fishing license. 4) 5 tonnes of illegal/unreported catch has included in total estimate.

124 ICES WGBAST REPORT Table Numbers of wild salmon (MSW=MultiSeaWinter) in fish ladders and hydroacoustic counting in the rivers of the assessment units 1 and 2 (Subdivisions 30 31, Gulf of Bothnia). Year Number of salmon Simojoki (au 1) Tornionjoki (au 1) Kalixälven (au 1) Råneälven (au 1) Piteälven (au 2) Åbyälven (au 2) Byskeälven (au 2) Rickleån (au 2) Ume/Vindelälven (au 2) Öreälven (au 2) MSW Total MSW Total MSW Total Total MSW Total MSW Total MSW Total Total MSW Females Total Total no control no control no control no control no control no control n/a n/a n/a n/a n/a na 9134 Simojoki: Hydroacoustic counting near the river mouth, started Tornionjoki: Hydroacoustic counting 100 km upstream from the sea, started Kalixälven: Fishcounting in the fishway is a part of the run. No control during Råneälven: Hydroacoustic counting 40 km upstream from the sea, started Piteälven: New fishway built Fishcounting is the entire run. Åbyälven: New fishway built in Fishcounting is only part of the total run. Byskeälven: New fishway built Fishcounting is part of the total run. Rickleån: New fishways built Fishcounting is part of the total run. Umeälven/Vindelälven: Fishcounting in the fishway is the entire run. Öreälven: Fishcounting in the trap is part of the run. The trap was destroyed by high water levels in 2000.

125 120 ICES WGBAST REPORT 2017 Table The age and sex composition of ascending salmon caught by the Finnish river fishery in the River Tornionjoki since the mid-1970s. Year(s) N:o of samples A1 (Grilse) 9% 53% 35% 7% 20% 8% 9% 9% 7% 8% 19% 6% A2 60% 31% 38% 59% 50% 53% 42% 44% 48% 44% 36% 76% A3 29% 13% 24% 28% 26% 31% 41% 40% 38% 36% 38% 11% A4 2% 2% 3% 4% 3% 6% 7% 5% 4% 8% 6% 5% >A4 0% 1% <1 % 2% 2% 2% 1% 2% 3% 5% 1% 1% Females, proportion of biomass About 45 % 49% 75% 71% 65% 67% 63% 61% 64% 69% 55% 67% Proportion of repeat spawners 2% 2% 2% 6% 6% 8% 9% 7% 9% 14% 6% 8% Proportion of reared origin 7% 46 %* 18% 15% 9% 1% 0.0% 0.6% 0.4% 0.0% 0.3% 0.3% * An unusually large part of these salmon were not fin-clipped but analysed as reared on the basis of scales (probably strayers). A bulk of these was caught in 1989 as grilse.

126 ICES WGBAST REPORT Table Densities and occurrence of wild salmon parr in electrofishing surveys in the rivers of the assessment unit 1 (Subdivision 31). River Number of parr/100 m² by age group 2+ & older >0+ (sum of two previous columns) Sites with 0+ parr Number of sampling year (%) sites Notes Simojoki % 14 No age data of older parr available % 14 No age data of older parr available % 16 No age data of older parr available % 16 No age data of older parr available % 16 No age data of older parr available % 22 No age data of older parr available % % % % No sampling because of flood % % % % 29 No age data of older parr available % % 17 Flood; only a part of sites were fished % % % % % % 19 Flood; only a part of sites were fished % 27 Flood; only a part of sites were fished % % % % % % % % % % % 35 Tornionjoki % % % % % 37 Flood; only a part of sites were fished % % % % % % % % % % % % 60 Flood; only a part of sites were fished % % % % % % % % % % % % 50 Flood; only a part of sites were fished.

127 122 ICES WGBAST REPORT 2017 Table continues Number of parr/100 m2 by age group River 2+ & older >0+ (sum of two previous columns) Sites with 0+ parr (%) Number of sampling sites year Notes Kalixälven % % % % % % % 11 Flood; only a part of sites were fished % % % % % % 9 Flood; only a part of sites were fished % % % % % % % % % % % % % 9 Flood; only a part of sites were fished % % % % % 30 Råneälven % % % % % % 1 Flood; only a part of sites were fished % % % % % No sampling because of flood % % % % % % % % % % % % 14

128 ICES WGBAST REPORT Table Estimated number of smolt by smolt trapping in the rivers Simojoki and Tornionjoki (assessment unit 1), and Sävarån, Ume/Vindelälven, Rickleån and Lögdeälven (assessment unit 2). The coefficient of variation (CV) of the trapping estimates has been derived from the mark-recapture model (Mäntyniemi and Romakkaniemi 2002) for the last years of the time-series. In the Ume/Vindelälven, however, another technique has been applied, in which smolts are tagged during the smolt run and recaptures has been monitored from adults ascending the year 1 2 years later. The ratio of smolts stocked as parr/wild smolts in trap catch is available in some years even though total run estimate cannot be provided (e.g., in the cases of too low trap catches). The number of stocked smolts is based on stocking statistics. Tornionjoki (AU 1) Simojoki (AU 1) Sävarån (AU 2) Ume/Vindelälven (AU 2) Rickleån (AU 2) Lögdeälv (AU 2) Smolt trapping, original CV of estimate Ratio of smolts stocked as parr/wild smolts in catch Number of stocked reared smolts (point estimate) Smolt trapping, original CV of estimate Ratio of smolts stocked as parr/wild smolts in catch Number of stocked reared smolts (point estimate) estimate estimate estimate estimate estimate estimate 1977 n/a 29,000 n/a n/a n/a n/a 1978 n/a 67,000 n/a n/a n/a n/a 1979 n/a 12,000 n/a n/a n/a n/a 1980 n/a 14,000 n/a n/a n/a n/a 1981 n/a 15,000 n/a n/a n/a n/a 1982 n/a n/a n/a n/a n/a n/a 1983 n/a n/a n/a n/a n/a n/a 1984 n/a 19,000 n/a n/a n/a n/a 1985 n/a 13,000 n/a n/a n/a n/a 1986 n/a 2,200 n/a n/a n/a n/a ,000 *) ,129 1, ,800 n/a n/a n/a n/a , ,300 1, ,700 n/a n/a n/a n/a 1989 n/a ,829 12, ,841 n/a n/a n/a n/a , ,545 12, ,100 n/a n/a n/a n/a , ,344 7, ,916 n/a n/a n/a n/a 1992 n/a ,000 17, ,389 n/a n/a n/a n/a , ,342 9, ,087 n/a n/a n/a n/a , ,317 12, ,862 n/a n/a n/a n/a 1995 n/a ,986 1, ,580 n/a n/a n/a n/a , ,858 1, ,153 n/a n/a n/a n/a ,000 **) 20,004 2, ,939 n/a n/a n/a n/a , ,033 9, ,942 n/a n/a n/a n/a ,000 17% ,771 8, ,815 n/a n/a n/a n/a ,000 39% ,339 57, ,100 n/a n/a n/a n/a ,000 33% ,000 47, ,111 n/a n/a n/a n/a ,000 12% ,998 53, ,300 n/a n/a n/a n/a ,000 43% ,032 63, ,912 n/a n/a n/a n/a ,000 33% ,000 29, ,900 n/a n/a n/a n/a ,000 25% ,000 17,500 28% ,800 3,800 15% n/a n/a n/a ,250,000 35% ,814 29,400 35% ,000 12% n/a n/a n/a ,000 48% ,458 23,200 20% ,000 3,100 18% n/a n/a n/a ,490,000 37% ,442 42,800 29% ,000 4,570 18% n/a n/a n/a ,090,000 42% ,490 22,700 29% ,000 1,900 49% n/a n/a n/a 2010 n/a ,965 29,700 28% ,240 1,820 32% 193,800 21% n/a n/a ,990,000 27% ,048 36,700 13% ,643 28% 210,000 14% n/a n/a 2012 n/a ,437 19,300 37% n/a 352,900 19% n/a n/a 2013 n/a ,300 37,000 11% ,548 31% 302,600 25% n/a n/a 2014 n/a ,800 36,600 19% n/a 180,600 13% 2,149 16% n/a ,032,000 47% n/a n/a 186,000 13% n/a n/a ,914,000 27% ,900 7% n/a n/a 3,961 15% 5,211 22% Smolt trapping, original CV of estimate Smolt trapping, original CV of estimate Smolt trapping, original CV of estimate Smolt trapping, original CV of estimate *) trap was not in use the whole period; value has been adjusted according to assumed proportion of run outside trapping period. **) Most of the reared parr released in 1995 were non-adipose fin clipped and they left the river mainly in Because the wild and reared production has been distinguished on the basis of adipose fin, the wild production in 1997 is overestimated. This was considered when the production number used by WG was estimated.

129 124 ICES WGBAST REPORT 2017 Table Densities and occurrence of wild salmon parr in electrofishing surveys in the rivers of the assessment unit 2 (Subdivisions 30 31). Detailed information on the age structure of older parr (>0+) is available only from the Åbyälven and Byskeälven. River Number of parr/100 m² by age group 2+ & older >0+ (sum of two previous columns) Sites with 0+ parr Number of sampling year (%) sites Notes Piteälven No sampling 1992 No sampling No sampling 1996 No sampling No sampling because of flood No sampling 2000 No sampling 2001 No sampling No sampling 2004 No sampling 2005 No sampling % % % No sampling % No sampling because of flood No sampling because of flood % % % % 4 Åbyälven % % % % % % % % % % % % No sampling because of flood % % % % % % % % % % % % % % % % , % , % 10

130 ICES WGBAST REPORT Table Continued. Table continues Number of parr/100 m² by age group >0+ (sum of River year & older two previous columns) Sites with 0+ parr Number of sampling (%) sites Notes Byskeälven % No sampling 1988 No sampling % % % % % % % % % No sampling because of flood % % No sampling because of flood % % % % % % % % % No sampling because of flood % % % % % 15

131 126 ICES WGBAST REPORT 2017 Table Continued. Kågeälven % % % % % α 50% α 0% α 0% No sampling No sampling No sampling No sampling α 58% α 30% α 33% α 54% α 46% α 44% α 58% α 82% % % % % % % % % % % 10 Rickleån * 0+ * > % % % % % % % % % % % % % No sampling because of flood % % % % 7/* % % 7/* % 7/* % % % % 7/* % 7/* % 7/* , % 7/* , % 7/*8

132 ICES WGBAST REPORT Table Continued. River Number of parr/100 m² by age group >0+ (sum of & older two previous columns) Sites with 0+ parr (%) Number of sampling sites year Notes Sävarån % % % % % % % % % % % % No sampling because of flood % , % No sampling because of flood % % % % % % % % % % % % 9 Ume/Vindelälven * 0+ * > % % % % % % % % % 6 Flood; only a part of sites were fished % % % % % % % % 19/* % 19/* % 19/* % 19/* % 19/* No sampling because of flood % 19/* % 19/* % 18/* % 19/* % 19/*23

133 128 ICES WGBAST REPORT 2017 Table Continued. Öreälven * 0+ * > % % % % % % % % % % % % No sampling because of flood % % % % % % % % 10/* % 10/* % % 10/* % 9/* % 10/* % 10/* % 10/*13 table continues on next page Table Continued. River Number of parr/100 m2 by age group >0+ (sum of & older two previous columns) Sites with 0+ parr (%) Number of sampling sites year Notes Lögdeälven * 0+ * > % % % 8/* % % % % % % % % % No sampling because of flood % % % % % % % % 8/* % 8/* % % 8/* % 8/* % 8/* % 8/* % 8/*11 *) Average densities from extended electrofishing surveys also including areas and sites in the upper parts of the river which have recently been colonized by salmon. These average densities are used as input in the river model (see stock ann α) stocked and wild parr. Not possible to distinguish stocked parr from wild.

134 ICES WGBAST REPORT Table Densities and occurrence of wild salmon parr in electrofishing surveys in the assessment unit 3 (Subdivisions 30). Detailed information on the age structure of older parr (>0+) is not available. Number of parr/100 m² by age group Number of River year & older >0+ Sites with 0+ parr (%) sampling sites Notes Ljungan % % % % % % No sampling because of flood % % No sampling because of flood % % % % No sampling because of flood % % 3 Flood; only a part of sites were fished No sampling because of flood No sampling because of flood No sampling because of flood Only one site fished because of flood 2013 No sampling because of flood % % % 9 Testeboån n/a n/a n/a n/a n/a n/a n/a % % 11 n/a = reared parr, which are stocked, are not marked; natural parr densities can be monitored only from 0+ parr

135 130 ICES WGBAST REPORT 2017 Table Densities of wild salmon parr in electrofishing surveys in the rivers of the assessment unit 4 (Subdivisions 25 26, Baltic Main Basin). Number of parr/100 m² Number of Number of parr/100 m 2 Number of River by age group sampling River by age group sampling year 0+ >0+ sites year 0+ >0+ sites Mörrumsån Emån * * no sampling because of flood * * Flood, only a part of sites were fished.

136 ICES WGBAST REPORT Table Densities of wild salmon parr in electrofishing surveys in the Latvian and Estonian wild salmon rivers of the assessment unit 5 (Gulf of Riga. Subdivisions 28). River Number of parr/100 m 2 by age group Number of sampling year 0+ >0+ sites Pärnu Salaca

137 132 ICES WGBAST REPORT 2017 Table Continued. Gauja 2003 <1 < < ² < < Venta Amata* < * ²) tributaries to Gauja *) reard fish

138 ICES WGBAST REPORT Table Densities of salmon parr in electrofishing surveys in rivers in Lithauanian of the assessment unit 5 (Baltic Main Basin). River Number of parr/100 m 2 by age group Number of sampling year 0+ >0+ sites Neris Žeimena

139 134 ICES WGBAST REPORT 2017 Table Continued. Mera Saria n/a n/a n/a n/a n/a n/a 2011 n/a n/a n/a n/a 2014 n/a n/a n/a n/a

140 ICES WGBAST REPORT Table Estonian wild and mixed salmon rivers in the Gulf of Finland. River Wild or mixed Water Flow m³/s First quality 1) obstacle km Undetected parr cohorts Production of >0+ parr mean min Purtse mixed IV (since 2006) Kunda wild III Selja mixed V Loobu mixed II Valgejõgi mixed IV Jagala mixed II Pirita mixed V Vaana mixed V Keila wild V Vasalemma wild II ) Classification of EU Water Framework Directive Table Densities of salmon parr rivers with only wild salmon populations, Subdivision 32. River Year Number of parr/100m2 Number of sites River Year Number of parr/100m2 Number of sites and older and older Kunda Vasalemma * * Keila *) = no electrofishing *) = no electrofishing.

141 136 ICES WGBAST REPORT 2017 Table Densities of wild salmon parr in rivers where supportive releases are carried out, Subdivision 32. River Year Number of parr/100m2 Number of sites River Year Number of parr/100m2 Number of sites and older and older Purtse Valgejõ Selja Jägala

142 ICES WGBAST REPORT Table Continued. Loobu Pirita * * * * Kymijoki NA NA NA NA NA NA NA 5 Vääna NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA 5 *) = no electrofishing

143 138 ICES WGBAST REPORT 2017 Table Current status of reintroduction programme in Baltic Sea potential salmon rivers. Potential production estimates are uncertain and currently being re-evaluated. River Description of river Restoration programme Results of restoration Country ICES subdivision Old salmon river Cause salmon population extinction of Potential production areas (ha) Potential smolt production (num.) Officially selected for reintroduction Programme initiated Measures Releases Origin of population Parr and smolt production from releases Spawne rs in the river Wild parr production Wild smolt production Moälven SE 31 yes 3, no yes c,l 2 Byskeälven yes yes >0 >0 Alsterån SE 27 yes 2, no no c,g,l 4 ** ** yes >0 >0 Helgeån SE 25 yes 2, no yes c,e,m 2 Mörrumsån yes yes >0 >0 Kuivajoki FI 31 yes 1, yes yes b,c,f 2 Simojoki yes yes yes 0 Kiiminkijoki FI 31 yes 1, yes yes b,c,d,f 2 Iijoki yes yes yes >0 Siikajoki FI 31 yes 1,2, no yes b,g,m 1.4 m ixed yes * 0 0 Pyhäjoki FI 31 yes 1,2, yes yes b,c,d,f,m 2 Tornionjoki/Oulojo yes yes yes 0 Kalajoki FI 31 yes 1,2, no yes b,e, m 1,4 no * 0 0 Perhonjoki FI 31 yes 1,2, no yes b,f 2 Tornionjoki/Oulojo yes * 0 0 Merikarvianjoki FI 30 yes 1,2, no yes b,c,e 2 Neva yes yes >0 * Vantaanjoki FI 32 no? no yes b,c,f,m 2 Neva yes yes 0 0 Kymijoki FI 32 yes 2,3, no yes b,c,m 2 Neva yes yes yes Sventoji LI 26 yes 2, yes yes m,c 2 Nemunas yes yes Minija/Veivirzas LI 26 yes * * yes yes c 2 Nemunas no no 0 0 Wisla/Drweca PL 26 yes 1,2,3,4 * * yes yes b,l,m 2 Daugava yes yes * * Slupia PL 25 yes 1,2,3,4 * * yes yes b,l,m 2 Daugava yes yes yes * Wieprza PL 25 yes 1,2,3,4 * * yes yes b,m 2 Daugava yes yes * * Parseta PL 25 yes 1,2,4 * * yes yes b,n 2 Daugava yes yes * * Rega PL 25 yes 1,2,3,4 * * yes yes b 2 Daugava yes yes * * Odra/Notec/Drawa PL 24 yes 1,2,4 * * yes yes b 2 Daugava yes yes * * Reda PL 24 yes 1,2,3,4 * * yes yes b 2 Daugava yes yes * * Gladyshevka RU 32 yes 1,2, no yes a,g,k,n 2 Narva, Neva yes yes yes >0 Cause of salmon popul. extinction Measures Releases 1 Overexploitation Fisheries 1 Has been carried out, now finished 2 Habitat degradation a Total ban of salmon fishery in the river and river mouth 2 Going on 3 Dam building b Seasonal or areal regulation of salmon fishery 3 Planned 4 Pollution c Limited recreational salmon fishery in river mouth or river 4 Not planned d Professional salmon fishery allowed in river mouth or/and river * No data Habitat restoration Dam removal ** Not applicable e partial i planned f completed j completed g planned k not needed h not needed

144 ICES WGBAST REPORT Table Densities of wild salmon parr in electrofishing surveys in potential rivers. Country Assess- Sub-div River Number of parr /100 m² Number of ment and year sampling unit 0+ >0+ sites Sweden 4 27 Alsterån no sampling no sampling Finland 1 31 Kuivajoki n/a n/a n/a n/a n/a n/a n/a n/a n/a no sampling 2009 no sampling 2010 no sampling 2011 no sampling 2012 no sampling 2013 no sampling 2014 no sampling 2015 no sampling 2016 no sampling Finland 1 31 Kiiminkijoki n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a no sampling n/a n/a n/a no sampling 2016 no sampling

145 140 ICES WGBAST REPORT 2017 Table Continued. Finland 1 30 Pyhäjoki n/a n/a n/a n/a n/a n/a n/a n/a n/a no sampling no sampling 2013 no sampling 2014 no sampling 2015 no sampling 2016 no sampling Russia 6 32 Gladyshevka no sampling 2012 no sampling no sampling table continues next page * = stocked and wild parr. Not possible to distinguish stocked parr from wild. n/a = reared parr, which are stocked, are not marked; natural parr densities can be monitored only from 0+ parr

146 ICES WGBAST REPORT Table Continued. Contry Assess- Sub-div Number of parr/100 m 2 Number of ment River by age group sampling unit year 0+ >0+ sites Lithuania 5 26 Šventoji Lithuania 5 26 Siesartis Lithuania 5 26 Virinta no sampling

147 142 ICES WGBAST REPORT 2017 Table Continued. Lithuania 5 26 Širvinta no sampling Lithuania 5 26 Vilnia Lithuania 5 26 Vokė no sampling

148 ICES WGBAST REPORT Table Continued. Lithuania 5 26 B. Šventoji no sampling Lithuania 5 26 Dubysa Lithuania 5 26 Minija

149 144 ICES WGBAST REPORT 2017 Table Salmon smolt releases by country and assessment units in the Baltic Sea (x1000) in Year Assessment unit Country Age Finland 2yr yr Total Sweden 1yr yr Total Finland 1yr yr yr Sweden 1yr yr Total Denmark 1yr yr EU 1yr yr Sweden 1yr yr Total Estonia 1yr yr 10 Poland 1yr yr Latvia 1yr yr Lithuania 1yr Total Assessment units 1-5 Total Estonia 1yr yr Finland 1yr yr yr 12 3 Russia 1yr yr Total Grand Total

150 ICES WGBAST REPORT Table Releases of salmon eggs, alevin, fry and parr to the Baltic Sea rivers by assessment unit in Assessment unit age eyed egg alevin fry 1s parr 1yr parr 2s parr 2yr parr year

151 146 ICES WGBAST REPORT 2017 Table Continued

152 ICES WGBAST REPORT Table The M74 frequency (in %) as a proportion of M74 females (partial or total offspring M74 mortality) or the mean offspring M74-mortality (see annotation 2) of sea-run female spawners, belonging to populations of Baltic salmon, in hatching years The data originate from hatcheries and from laboratory monitoring. Prognosis (min max) for 2017 is based on the free thiamine concentration of unfertilized eggs of autumn 2016 spawners and, moreover, on the number of wiggling females. River Sub-d Simojoki (2) *a Tornionjoki(2) Kemijoki Iijoki Luleälven Skellefteälven Ume/Vindelälven Ångermanälven Indalsälven Ljungan Ljusnan Dalälven *b Mörrumsån Neva/Åland (2) Neva/Kymijoki (2) Mean River Simojoki and Tornionjoki Mean River Luleälven, Indalsälven, Dalälven Mean total ) All estimates known to be based on material from less than 20 females in italics. 2) The estimates in the rivers Simojoki, Tornionjoki/Torne älv and Kymijoki are since 1992, 1994 and 1995, respectively, given as the proportion of females (%) with offspring affected by M74 and before that as the mean yolk-sac-fry mortality (%). *a = with wigglers included; without wigglers 31 41%. *b = with wigglers included; withou twigglers 49 69%

153 148 ICES WGBAST REPORT 2017 Table Summary of M74 data for Atlantic salmon (Salmo salar) stocks of the Rivers Simojoki, Tornionjoki and Kemijoki (hatching years ), indicating the total average yolk-sac fry mortality among offspring of sampled females (%), the percentage of sampled females with offspring that display M74 symptoms (%) and the percentage of sampled females with 100% mortality among offspring (%). Data from less than 20 females are given in italics. NA = not available. Total average yolk-sac fry Proportion of females with Proportion of females mortality among offspring (%) offspring affected by M74 (%) without surviving offspring (%) Simojoki Tornionjoki Kemijoki Simojoki Tornionjoki Kemijoki Simojoki Tornionjoki Kemijoki NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA 47 NA NA 74 NA NA 86 NA 50 NA NA 91 NA 50 NA NA 3 NA 3 NA NA 6 3 NA 6 0 NA NA 0 0 NA 0 0 NA NA 0 0 NA 0 0 NA NA NA 0 NA NA 0 NA NA NA NA 0 NA NA 0 NA NA NA NA 4 NA NA 4 NA NA

154 ICES WGBAST REPORT Table Summary of M74 data for nine different Atlantic salmon stocks (hatching years ), in terms of the number of females sampled with offspring affected by the M74 syndrome in comparison to the total number of females sampled from each stock. Wiggling females (in river Dalälven) not included in number of M74 females. Luleälven Skellefteälven Ume/Vindelälven Ångermanälven Indalsälven Ljungan Ljusnan Dalälven Mörrumsån M74 Total M74 Total M74 Total M74 Total M74 Total M74 Total M74 Total M74 Total M74 Total 1985 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA 6 38 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA

155 150 ICES WGBAST REPORT a) tonnes b) Illegal/unreported Sweden Finland tonnes Figure Total river catches in the River Tornionjoki (assessment unit 1). a) Comparison of the periods from 1600 to present (range of annual catches). b) From 1974 to present. Swedish catch estimates are provided from 1980 onwards.

156 ICES WGBAST REPORT Assessment unit Kg Simojoki Tornionjoki Kalixälven Year Figure Salmon catch in the rivers Simojoki, Tornionjoki (Finnish and Swedish combined) and Kalixälven, Gulf of Bothnia, assessment unit 1, Ban of salmon fishing 1994 in the river Kalixälven.

157 152 ICES WGBAST REPORT Assessment unit 1 and Number of wild salmon Kalixälven (au 1) Piteälven (au 2) Åbyälven (au 2) Byskeälven (au 2) Ume/Vindelälven (au 2) Year Figure Total wild salmon run in fish way in rivers in assessment unit 1 and 2, in

158 ICES WGBAST REPORT Assessment unit 1 50 Number of 0+ parr/100m² Simojoki Tornionjoki Kalixälven Råneälven Year Figure Densities of 0+ parr in rivers in Gulf of Bothnia (Subdivision 31), assessment unit 1, in

159 154 ICES WGBAST REPORT Assessment unit 1 40 Number of >0+ parr/100m² Simojoki Tornionjoki Kalixälven Råneälven Year Figure Densities of >0+ parr in rivers in Gulf of Bothnia (Subdivision 31), assessment unit 1, in

160 ICES WGBAST REPORT Assessment unit Number of 0+ parr/100m² Åbyälven Byskeälven Kågeälven Rickleån Sävarån Ume/Vindelälven Öreälven Lögdeälven Year Figure Densities of 0+ parr in rivers in Gulf of Bothnia (Subdivision 31), assessment unit 2, in

161 156 ICES WGBAST REPORT Assessment unit Number of >0+ parr/100m² Piteälven Åbyälven Byskeälven Kågeälven Rickleån Sävarån Ume/Vindelälven Öreälven Lögdeälven Year Figure Densities of >0+ parr in riveres in Gulf of Bothnia (Subdivision 31), assessment unit 2, in

162 ICES WGBAST REPORT Assessment unit 3 Number of parr/100m² Ljungan 0+ Ljungan >0+ Testeboån 0+ Testeboån > Year Figure Densites of parr in Ljungan and Testeboån in the Gulf of Bothnia (Subdivision 30), assessment unit 3, in

163 158 ICES WGBAST REPORT Assessment unit Emån Mörrumsån 250 Number of 0+ parr/100m² Year Figure Densities of 0+ parr in rivers in the Main Basin (Subdivision 25 27), assessment unit 4, in

164 ICES WGBAST REPORT Assessment unit 4 Emån 75 Mörrumsån Number of >0+ parr/100m² Year Figure Densities of >0+ parr in rivers in the Main Basin (Subdivision 25 27), assessment unit 4, in

165 160 ICES WGBAST REPORT Assessment unit 5 8 Number of parr/100m² > Year Figure Densities of parr in the river Pärnu Main Basin (Subdivision 22 29) assessment unit 5, in

166 ICES WGBAST REPORT Assessment unit Number of parr/100m² > Year Figure Densites of parr in the river Salaca Main Basin (Subdivision 22 29) assessment unit 5, in

167 162 ICES WGBAST REPORT Assessment unit 5 8 Number of parr/100m² 6 4 Neris Žeimena Mera Saria Year Figure Densites of 0+ parr in Lithuanian rivers in Main Basin (Subdivision 22 29) assessment unit 5, in

168 ICES WGBAST REPORT Assessment unit Number of parr/100m² Neris Žeimena Mera Saria Year Figure Densities of >0+parr in Lithuanian rivers in Main Basin (Subdivision 22 29) assessment unit 5, in

169 164 ICES WGBAST REPORT Kunda Keila Vasalemma Numer of 0+ parr/100m 2 Figure Densities of 0+ (one-summer old) salmon parr in the three wild Estonian salmon rivers.

170 ICES WGBAST REPORT Pirita 60 Selja 50 Vääna Loobu Number of 0+ parr/100m Jägala Valgejõgi Purtse Figure Densities of 0+ (one-summer old) salmon parr in seven Estonian salmon rivers were suportive releases are carried out.

171 166 ICES WGBAST REPORT Eggs, free thiamine 100 Free thiamine, nmol/g Yolk-sac fry mortality Mean yolk-sac fry mortality, % 0 0 Hatching year (number of females) Figure Relationship between the mean yolk-sac fry mortality (± SE) of River Simojoki salmon and the free thiamine concentration (± SE) in their unfertilized eggs (Vuorinen et al., unpublished data).

172 ICES WGBAST REPORT Figure Concentration of free thiamine in unfertilized eggs of salmon from the rivers Simojoki, Umeå and Dalälven in Data includes wigglers, one in Simojoki, none in Umeå and 16 in Dalälven, for which an estimated thiamine concentration (0.130 nmol g-1) was set. Box depicts range of 25 75%, horizontal line median, diamond mean, whiskers confidence level of 5 95% and stars minimum and maximum observations.

173 168 ICES WGBAST REPORT Sweden Finland 60 % Figure Proportion of M74 positive females in Swedish and Finnish hatcheries.

174 ICES WGBAST REPORT Reference points and assessment of salmon 4.1 Introduction In this chapter, assessment of stock status and information relevant for advising Baltic salmon fishing possibilities in 2018 are summarized (Sections 4.2 to 4.5). The chapter also briefly reviews recent steps taken in developing the assessment methodology as well as planned next steps (Sections and 4.6.2). Finally, relevant issues for salmon data collection under the EU MAP are briefly discussed (Section 4.7). In WGBAST, an analytical assessment of the AU 1 4 salmon stocks has been carried out annually using the Full Life-History Model (FLHM), and the WinBUGS software as programming platform. Initiation of the Markov Chain Monte Carlo (MCMC) sampling has often been difficult and has required a lot of trials, e.g. with various sets of initial values, before being successful. In 2017, trials to run the FLHM with data up to 2016 (i.e. all available data) were started in late February because obtaining results from the run requires 4 5 weeks. Unfortunately, the FLHM with data up to 2016 did not run in spite of many trials. Therefore no results from this model were available for the group by the time of the meeting. Because of the lack of model results based on data from 2016, the WG decided to follow last year s approach (ICES, 2016) to compare the 2016 data and results of stock monitoring with the corresponding model predictions for 2016 based on the latest successfully run model. This comparison allows for a qualitative evaluation of how well the predictions made by FLHM are realised in 2016, thus providing an indication of the feasibility to base advice with catch options for 2018 on the latest successful model run. A few months before the 2017 WG meeting, the FLHM with data up to 2015 (i.e. one more year than used in the last previous model from 2015) was successfully run for the purposes of benchmark work (WKBaltSalmon 2017; Section 4.6.1). The results of this model are here used to compare model predictions with the 2016 monitoring results as described above. Also, to illustrate how one year of missing data may affect model results, we here include several comparisons of the two latest models (the results of the above model with data up to 2015 vs. WGBAST 2015 model with data up to 2014). 4.2 Historical development of Baltic salmon stocks (assessment units 1 6) Updated submodels The river model provides input about smolt production into the life cycle model by analysing all the juvenile survey data from the rivers in AUs 1 3. For rivers in AUs 4 6, other methods are used to estimate smolt production (see Stock Annex, Annex 3, Section C.1.5). As indicated above, this year the FLHM was fed with the river model using juvenile data up to the year The results of this river model run are reported and discussed in the last year s assessment (ICES, 2016, Section 4.2.1). A model for M74 mortality provides input about mortality due to M74 into the life cycle model by analysing all data on incidence of M74 in the stocks (see Stock Annex, Section C.1.6). Similar to the river model, data about M74 was used up to 2015 and also these results are reported in last year s WG report (ICES, 2016, Section 4.2.1).

175 170 ICES WGBAST REPORT Changes in the assessment methods The only change made to the assessment model (versus the previous version in 2015) was the addition of a constant (1) to the parameters of the Beta observation model for the proportion of wild salmon in offshore catches. This was done to improve the numerical stability of the model and is expected to have negligible impact on results. Minor changes were also made to the scenarios (forward projection code) to correct inconsistencies with the population dynamics equations identified during the recent benchmark (WKBaltSalmon 2017; Section 4.6). These were as follows: seal mortality during the coastal fishery was removed for AU4 stocks (Mörrumsån and Emån), and three months of natural mortality were subtracted for AU4 smolts and fish on the spawning migration. These changes were made to check whether lower levels of natural mortality in the assessment model versus the projections may have been responsible for the lack of recovery of AU4 stocks in scenarios with no fishing mortality (e.g. Figure , ICES, 2015): it appears that this was the case (Figure ). However, further changes are needed to both codes in next year, to make natural mortalities applied for Emån and Mörrumsån consistent with other stocks. It is worth re-iterating here that owing to technical problems running an updated version of the model (Section 4.1), this year s results are subject to the same caveats about interpretation as in In particular, updating of priors for PSPC for Mörrumsån and Emån, but not the corresponding smolt production estimates, may cause bias in estimates of status for those stocks. Likewise, recently updated PSPC-priors for rivers Piteälven, Lögdeälven and Öreälven are not included in the model, which makes the presently assessed status of these stocks uncertain (Section 4.4.2). Owing to strongly increasing recreational trolling effort that is currently unaccounted for in the assessment, a preliminary evaluation of the relative recreational catch volume was made, suggesting that the true recreational catch at sea has potentially been salmon larger in recent years than the catch used in the assessment during (Sections 2.1.1, 4.2.5). This catch volume was found to constitute less than a 20% increase relative to the removal corresponding to ICES advice for fishing year 2017 (i.e. difference in removal between scenarios 1 and 2; Section 4.3.1) so that additional scenarios were not considered Status of the assessment unit 1 4 stocks and development of fisheries in the Gulf of Bothnia and the Main Basin The full life-history model (FLHM) was run for about iterations, the first 6000 of which were discarded as burn-in. The chain was thinned with an interval of ten after the burn-in. Another chain was run in parallel for the models and convergence was checked using the Gelman-Rubin diagnostic and by visual inspection of trace plots. The final sample was thinned further to result in a sample of 1000 iterations for each parameter on which the modelling results are based. Whenever possible, the medians and 90% probability intervals (PI s) are shown as statistics of the resulting probability distributions. As in last years assessments, high autocorrelation was found in the MCMC samples of the PSPC estimates for Tornionjoki/Torneälven, and to lesser extent also for Kalixälven and Ume/Vindelälven, as well as in the adult natural mortality estimate. Caution must therefore be taken in the interpretation of these results. The results indicate a decreasing long-term trend in the post-smolt survival until mid-2000, after which survival has somewhat improved (Figure ). The lowest survival (median estimate about 10% among wild and 2 3% among reared smolts)

176 ICES WGBAST REPORT was estimated for salmon that smolted in years and Thereafter the survival has increased to 13 20% for wild smolts and 3 8% for reared smolts (ranges of median estimates in ). Survival improved especially among salmon that smolted in The current survival is slightly lower than in the early 2000s, and less than half of the estimated survival level prevailing in the early and mid-1990s. It should be noted, however, that since the post-smolt survival is one of the few parameters in the model that vary from year to year, these estimates can also contain variation that actually takes place in some other stage of the life cycle (in natural or fishery induced survival). The adult natural annual survival of wild salmon (median 86%, PI 79 90%) is estimated to be higher than that of reared salmon (median 79%, PI 71 89%). The adult survival estimates of wild and reared salmon have been reasonably similar among assessments, but in the last full assessment (ICES, 2015) adult survival was estimated to be higher than in this newest assessment, especially among wild salmon. Maturation of 1-sea winter salmon (grilse) has in most years been around 20% and 20 30% among wild and reared individuals, respectively, i.e. close to the average level of the whole time-series (Figure ). Among 2-sea winter salmon maturation is estimated to have been mostly 30 45% and 40 60% for wild and reared salmon, respectively. Salmon of both origins have had rather similar maturation rates among 3-sea winter individuals (50 60% and 50 70% for wild and reared, respectively) and the maturation rates of 4-sea winter are even more similar. The estimated maturation rates of 4-sea winter are on average lower but more uncertain than those of 3-sea winter salmon. This is against intuition but might be an artefact due to the inconsistency between the model assumptions (no repeat spawners, all fish mature at latest after five sea winters) and the biology of salmon (some repeat spawners exist and some salmon have a longer lifespan than five years at sea). The maturation rates for reared salmon were higher-than average during the period Maturation rate has been almost as high also around , even among wild salmon. One more fairly conspicuous feature in the time-series is the low maturation rate of wild salmon during , following two cold winters. The full life-history model allows estimation of steepness of the stock recruit relationship (Table ) and the PSPC (Table ) for different salmon stocks. Figure gives an indication of river-specific stock recruit dynamics. The blue clouds in the figure panels indicate posterior probability distributions of all the historical estimates of yearly egg deposition and corresponding smolt abundance (the density of the cloud indicates the probability). Curves added in the figure panels are draws from the posterior distribution of the Beverton Holt stock recruit function. Adding the latest information about spawner and smolt abundance together with the latest changes in the model structure has resulted in some changes in posterior probability distributions of the PSPCs as compared to in last year (Figure , Table ). Generally, PCPC estimates have somewhat decreased in several rivers and increased in a few rivers, compared to the 2015 assessment. The largest update was in the PSPC of Kalixälven (18% decrease in median value), Simojoki (15% decrease in median), Råneälven (12% decrease in median) and Ume/Vindelälven (11% decrease in median). The summed estimate of PSPCs over the AU 1 4 rivers decreased by 10% from the 2015 estimate (to a median value of 3.48 million smolts; Table ). The AU specific total PSPC estimates changed from the last year s assessment by maximum 12% (AU 1). The PSPC estimates of AUs 5 and 6 are not updated in this year s assessment.

177 172 ICES WGBAST REPORT 2017 The total PSPC estimate of AUs 1 6 (median 4.01 million) is about smolts lower than the corresponding estimate from the last full assessment. There are reasons to believe that PSPC is underestimated in several rivers of AU s 1 3, but overestimated in AU 4 rivers, mainly because of low-quality priors set for the PSPCs and/or estimates of the historic smolt production in several rivers (see Sections and 4.6). Work to correct these problems is underway, and at the same time also the parametrization of stock recruit dynamics in the assessment model will be improved according to recent benchmark results (WKBaltSalmon 2017; Section 4.6.1). These changes in the methodology and the background information are likely to revise PSPC estimates in near future, thus the current estimates must be regarded transitory. There are also other and even more difficult problems connected to estimation of PSPCs. Walters and Korman (2001) have pointed out that for depleted stocks, when the spawning stocks increase rapidly after long periods of low abundance this may result in locally intense competition within those reproduction areas that are primarily being used. The patchy habitat use may impose local density-dependent effects, which may diminish in the longer run (after several generations) once spawners have dispersed to fully re-establish the natural or most productive habitats (Walters and Korman, 2001). If this phenomenon is valid for Baltic salmon populations, our analysis utilising the recently available stock recruit information underestimates long-term (full) carrying capacity of the Baltic rivers. Since the mid-1990s the status of many wild salmon populations in the Baltic Sea has improved, and the total wild production has increased from less than 0.5 to over three million smolts (Figure , Table ). However, there are significant regional differences in trends in smolt production. For the wild salmon stocks of AUs 1 2, the very fast recovery of smolt production indicates high productivity of these rivers. The only wild stock in AU 3 evaluated in the assessment model so far (Ljungan) has also recovered, but the estimates of both the current and the potential smolt production of this river are highly uncertain. AU 4 stocks have improved since the drop occurred in the latter half of the last decade, and the smolt abundance is now back on the level prevailing around the turn of the millennium. Most of the AU5 stocks have shown a decreasing trend in smolt abundance until the last two years during which a there has been a notable increase in the abundance. The AU6 stocks show improvements similar to the AU3 stocks (see Section 4.2.4). Smolt production in AUs 1 3 rivers is expected to increase beyond 2016, which is a reflection of the recent increase in the number of spawners. By comparing the posterior smolt production (Table ) against the posterior PSPC, it is possible to evaluate current (year 2016) status of the stocks in terms of their probability to reach 50% or 75% of the PSPC (Figures and , Table ). Table contains also AU 5 6 stocks and the two new wild rivers Kågeälven (AU 2) and Testeboån (AU 3). Testeboån and AU 5 6 stocks have not been analytically derived, but for those rivers expert judgments are used to classify their current status. The perception about the overall stock status (amount of rivers in different status classes) has become much more positive compared to earlier years assessments. Within AU 1, all of the stocks are very likely to have reached at least 50% of their PSPCs, and all except one (Simojoki) are likely or very likely to have reached 75% of their PSPCs. The assessed status of the nine AU 2 stocks also has changed from last year; upwards in six rivers (Åbyälven, Byskeälven, Sävarån, Ume/Vindelälven,

178 ICES WGBAST REPORT Öreälven and Lögdeälven) and downwards in none of the rivers. In AU 3, the updated status of Ljungan indicates that it has very likely reached 50% and has likely reached 75% of its PSPC, while the expert judgment about Testeboån indicates that it has unlikely reached even the 50% target. In AU 4, the status for Emån remains low (unlikely to have reached even 50% of PSPC) and the status of Mörrumsån remains high (very likely to have reached even 75% of PSPC), similar to the previous assessment. Note, however, that the current status estimates for Piteälven, Lögdeälven, Öreälven, Emån and Mörrumsån are particularly uncertain, due to ongoing work with updated PSPC and/or smolt priors for these rivers (Sections and 4.6). Status has been assessed to vary considerably among AU 5 6 stocks. Among the 12 AU 5 stocks, the status has been assessed better in five stocks and worse in one stock, compared to the last assessment. Five and three stocks are assessed to likely or very likely reach 50% and 75% of their PSPCs, respectively. However, five stocks are unlikely to have reached even 50% of their PSPCs. Also the AU 6 stocks are generally assessed to be in better status compared to the last assessment. Among the twelve AU 6 stocks, the status was assessed to be higher in five stocks and lower in two stocks, compared to the last previous assessment. Five and three stocks are assessed to likely or very likely reach 50% and 75% of their PSPCs, respectively. However, four stocks are unlikely to have reached even 50% of their PSPCs. Out of the 41 assessed stocks (Table ), 23 (56%) are likely or very likely to have reached 50% of PSPC, and 18 (44%) are likely or very likely to also have reached 75% of PSPC. Generally, the probability to reach targets is highest for the largest northern stocks. However, the probabilities vary also among the northern stocks. Twelve stocks are considered unlikely to have reached even 50% of PSPC, i.e. they are considered to be weak. Most of these weak stocks (nine out of 12) are located in AUs 5 6. While most of the AUs 1 2 stocks show strong indications of recovery over the years, the stocks in AUs 4 5 have shown little or no recovery. Stocks in rivers situated between these areas (i.e. AU 3 and AU 6 stocks) have mostly shown modest indications of recovery (Figures , and Section 4.2.4). The model captures quite well the overall historic fluctuation of catches in various fisheries (Figure ). Also the 2016 catches, not used to fit the model, are in fair accordance with the model predictions. However, the offshore catches from the early and mid-2000s become underestimated. This happens because of higher than average catch per unit of effort in those years in offshore longlining and driftnetting, and on the lack of flexibility for such variation in the current model structure. The same gear specific catchability is assumed over the time-series meaning that model predicted catch is proportional to effort and cannot adjust to data, indicating very high catch per unit of effort observed in individual years. The model fit to the coastal catches is good, although there is some tendency to slightly underestimate catches in the beginning and the end of the time-series. Concerning river catches, the model does not fully capture the high catches of the years The model is fitted to the proportion of wild and reared salmon (separately for ages 2SW and 3SW) in the offshore catch. The posterior estimates of wild vs. reared proportions follow closely the observed proportions (Figure ). The model tends to somewhat underestimate the proportion of wild salmon for the smolt years , the data of which are not used in the model fitting. An increasing trend in the number of spawners is seen in most AU 1 4 rivers (Figure ). Spawner abundance has increased particularly for the years In Simojoki, the very high estimates of spawners around the turn of the millennium

179 174 ICES WGBAST REPORT 2017 reflect intense stocking of hatchery-reared parr and smolts in the river during the late 1990s. The model captures trends seen in fish ladder counts, even short-term variation in rivers where the data is not used for model fitting (e.g. Byskeälven). It should be noted, though, that fish ladder counts themselves are usually not ideal indices of spawner abundance. Annual variation in river conditions often affect the success of fish to pass through ladders. Furthermore, counters are often located some distance from the sea, sometimes above considerable spawning areas (e.g. in Kalixälven and Byskeälven). Therefore modelled predictions of spawner numbers are usually expected to be (much) higher than the observed counts. For Ume/Vindelälven, however, no spawning areas exist below the counter, and the fish counts are therefore good approximations of the total amounts of fish having reached the spawning grounds. In this river also the model based spawner estimates follow closely these observations (Figure ). In Kalixälven, the development in spawner abundance estimated by the model is more optimistic than the development observed in the fish ladder counts. However, the drop from 2009 to and a drastic increase in 2012 observed in spawner counts are well captured by the model. The improvement is probably a consequence of fitting the model to spawner counts in combination with assuming annually varying maturation rates. From 2013 to 2014 the spawner abundance markedly increased further in some, but decreased in other, rivers of the AUs 1 2 (e.g. Tornionjoki vs. Kalixälven). In several rivers a marked increase in counts was observed again in 2016, i.e. the year which could not be used in the model fitting (Section 4.1). This latest jump is not well captured by the model, which usually estimates (predicts) no marked change in spawner abundance from 2015 to It is difficult to explain this contrasting development in spawning runs observed even between neighbouring rivers. Tornionjoki salmon is by far the highest in abundance among Baltic stocks, and even a modest under/overestimation of the abundance at relative scale in this stock may thus result in differences of up to tens of thousands of spawners. In spite of fluctuations, there was a long-term decreasing trend in the harvest rate of longlines (and driftnets during the years driftnetting was allowed) (Figures a and ). After the ban of driftnets in 2008 the longlining harvest rate increased rapidly and reached its all-time high around In the harvest rate of longlines was almost as high as the combined harvest rate of longlines and driftnets in (Figure ). Thereafter, the harvest rate of longlining quickly dropped until 2013, and thereafter it has stayed on about the same level (close to 10% per year for MSW salmon). The harvest rate of coastal trap netting dropped in the late 1990s. Since mid-2000s the coastal harvest rate has started to decrease again, and after 2010 the decrease has accelerated (Figure b). Since 2012 this is at least partly a result of early fishing closures due to filling up of quotas. Estimates of harvest rates in the rivers are inaccurate and lack apparent trends since the mid-1990s (Figure c). River-specific data indicate that there can be substantial variation in the harvest rate between rivers (see Chapter 3), which is not taken into account in the model Abundance trends of wild stocks in In this section, current trends in wild salmon stocks (based on most recent data) are reviewed and compared with model predictions lacking the most recent (2016) data. The river model is used to assess smolt abundances in the AU 1 3 rivers (this model is normally used to provide input for the full life-history model, see Section and Stock Annex). Smolt abundances in the AU 4 6 rivers are based on electrofishing

180 ICES WGBAST REPORT data combined with expert opinions, similar to the approach taken in the earlier years. Available spawner counts from 2016 have also been compared with modelbased predictions, following a procedure reported in e.g. ICES (2012). M74 mortality Figures 3.4.1, and show the incidence of M74 mortality until the hatching year 2016, as well as an indication about the M74 mortality for the hatching year 2017 (based on thiamine contents in eggs spawned in 2016). The mortality has decreased since the mid-2000 and was almost non-existing in However, in spring 2016 the M74-mortality increased, and it is predicted to increase further in It is not possible to accurately predict the level of M74-mortality for 2017, and it is also impossible to predict if the increase will persist beyond The abundance of prey and its quality leading to thiamine deficiency are responsible of M74 (Mikkonen et al., 2011). The intake of thiamine for Baltic salmon in relation to energy and fat remains lowest by eating young clupeids, especially young sprat (Keinänen et al., 2012). The latest strong year class of sprat hatched in 2014, which coincides well with the recent increase in the M74 mortality: the salmon spawners in 2015 and 2016 have been on their feeding migration at a time with high young sprat abundance. According to the most recent (preliminary) data provided by the Baltic International Fish Survey WG (ICES WGBIFS), no indications exist that the 2015 and 2016 year classes of sprat are very strong. If so, then the M74-mortality may increase only moderately and only for a few years period. Information about the strength of the most recent sprat year classes is not, however, fully reliable. Also, there are no predictions available about the strength of future (2017 and onwards) sprat year classes. Abundance trends in AUs 1 4 The overall AU specific smolt abundances predicted by the updated river model follows relatively closely the prediction made last year by the full life-history model (Figure ). Both models predict an increase in overall smolt abundance in AUs 1 3 in the near future ( ). For AUs 1 and 3, the river model with updated information indicates a somewhat larger increase than the FLHM. In AU 1 this difference is substantial, even in terms of absolute abundance. River-specific information (Table , see also Chapter 3) shows a positive development of parr densities for almost all the wild stocks in AUs 1 3 (e.g. Figures ). In AU 4, smolt abundance can be predicted by juvenile data only one year ahead. Smolt production in this AU is expected to decrease somewhat from 2016 to 2017 (Figure ). FLHM predicts a smaller decrease, but according to it the decrease will continue to It is important to note, however, that the time-series of smolt production estimates as well as the priors for the PSPC of the AU 4 rivers (Emån, Mörrumsån) are under full revision, which will most likely result in changes both to the predicted development of smolt production as well as to the status assessment (see Section 4.4.2). Spawner count data are available from several AU 1 2 rivers. The comparison of a spawning run index computed from such counts (number of salmon in relation to in 2009, averaged across rivers) with an identically computed index for model estimates shows that the development in observed and predicted spawning run sizes are generally in good agreement (Figures and ). A tendency can be seen, however, for the FLHM to underestimate the spawning run of 2016, i.e. the latest year with fish counting observations which is not included in the model (Section 4.1).

181 176 ICES WGBAST REPORT 2017 According to the FLHM results, the latest smolt runs (year 2016) in most AU 1 4 rivers have likely or very likely reached 75% of PSPC (ICES 2015; Table and Figure ). This is in accordance with the predictions based on the earlier full assessment (ICES, 2015, Figure a d). Note that PSPC in three Swedish AU 2 rivers, Piteälven, Lögdeälven and Öreälven, have recently been revised and are expected to be underestimated by the current assessment model, which may introduce biases in their current status assessment (likely overestimation, see Section 4.4.2). On the other hand, the salmon stocks also in these rivers are currently showing a recovery trend. The assessment predicts a period with somewhat lower spawning runs for , which obviously reflects the relatively small spawning runs one generation earlier ( ). Consequently, a somewhat lower smolt production is predicted for (Figure a d). Stock projections this far into the future are, however, very uncertain: variation in natural and fishing mortality after smolting, as well as variation in reproductive success may easily modify the expected stock fluctuation in the future. Abundance trends in AUs 5 6 Smolt production in relation to PSPC in AU 5 stocks have previously (from 2011 to 2014) been on a low level in almost every wild and mixed river (Table , Figures and ). Total estimated production of wild smolts in AU 5 has varied a lot, and largest numbers of wild smolts were produced in and in (Figure ). Notable increases in smolt production occurred in , and the overall smolt production in relation to PSPC was near 50%. Based on parr density data the smolt production is expected to slightly decrease in 2017 (Table , Figure ). The longest positive development in AU 5 can be seen in the river Nemunas. Based on redd counting studies done in Nemunas river basin, the smolt production is expected to remain at present level in In spite of the positive trend, however, the observed smolt production in the Nemunas in relation to PSPC has remained far below 50%. In river Pärnu the smolt production level is almost zero. This river is a large watercourse with several tributaries, and many of them have been subject to long-term restoration efforts (restoring connectivity, restocking, etc. see Section 3.1.5). River Salaca in AU 5 has the most extensive timeseries of monitoring data (Section 3.1.5). The smolt production increased from the early to late 1990s to a high level (over 75% of PSPC). However, the smolt production started to decrease again in the 2000s, and was in its lowest level around (Figure ).In 2016 the smolt production in relation to PSPC again exceeded 75%, and based on parr density data it is likely to slightly decrease in 2017 (Table ). Smolt production in AU 6 stocks shows a positive trend in most of the rivers, but with large annual variation (Figures ). In spite of the overall positive development, a majority of the AU 6 mixed rivers are still not likely to be over 50% of their PSPCs. Among the three wild Estonian stocks (Figures and ) smolt production in relation to PSPC in 2015 and 2016 is around or above 75%. The increase in smolt production has been the highest in river Keila; in this river parr densities have in some years even exceeded the previously estimated PSPC, and for this reason it has been necessary to revise (increase) the PSPC estimate on several occasions. A positive trend in smolt production can also be seen in river Kunda, although the estimated smolt production shows high variation. The last period of low smolt production in river Kunda occurred in (Figure , Table ), which may reflect

182 ICES WGBAST REPORT the low number of spawners originating from the weak earlier year classes that occurred in in this river. Smolt production in river Vasalemma has been very variable; however, strong year classes are more frequent. For the small Estonian mixed-stocks the trend has also been positive in recent years (Figure ). However, current PSPC in some of these rivers is severely limited by migration barriers, and there is also a lot of annual variation in these small populations. Smolt production in relation to PSPC in 2017 is expected to remain just below 50% in most rivers. Smolt production in river Pirita is expected to remain above 50% of PSPC. In the Finnish mixed river Kymijoki smolt production remained far below the 50% level in the past. However, a strong year class occurred in As a result smolt production is expected to increase above 75% in 2017 (Figure ). The smolt production estimates of Kymijoki are very uncertain, however, and so is also the PSPC estimate which is presently assumed to be around (HEL- COM, 2011; 2011b). The most conservative estimate of the PSPC of Kymijoki has been applied in Figure In Russian river Luga, wild smolt production has remained below 10% of PSPC, despite large-scale annual smolt releases using salmon of local origin (Figure ) Recent development of fisheries, catches, and harvest patterns In this section, most recent trends in fisheries (based on most recent data) are reviewed and compared with the model predictions lacking the most recent (2016) data. Total salmon catch in 2016 was one of the lowest in the time-series since 1970; it was somewhat higher than in 2015 and slightly lower than in (Tables and 2.2.2). The reported commercial catch (in numbers of salmon) was the lowest in history, but recreational catch in sea and especially misreporting increased from 2016 (Tables and 2.2.6). Also efforts in various commercial fisheries remained at their lowest recorded level (Figures and 2.4.2). In this year s assessment the last data year is 2015 (see Section 4.1 for explanation). The realized development of fisheries in 2016 was in accordance with model predicted catch for that year in future scenarios 1-4 (reported catch , cf. model predictions in Figure ). The long-term trend of increasing recreational catches in relation to the commercial ones continues (Figure 2.2.2). In addition, much of the catch statistics for recreational fisheries are largely incomplete and partly missing, and the real recreational catch is likely to be significantly larger than the figure used in the present assessment input data. As part of the recent benchmark for Baltic salmon (WKBaltSalmon 2017) an expert elicitation was carried out to estimate updated total salmon trolling catches. This preliminary evaluation suggested that the true recreational catch at sea has potentially been salmon larger than assumed in the assessment for period (Section 2.1.1). This underestimation of the recreational sea catch has likely resulted in underestimated natural (mainly post-smolt) survival rates in recent years. Despite that the total catch in recent years is likely underestimated, and various uncertainties related to the development and intensity of recreational fishing remain, it can be concluded that the overall development is in accordance with what has been assumed in the assessment. Although the true magnitude of catch levels is probably slightly higher than in the future scenarios (Section 4.3), the likely overestimated post-smolt mortality is expected to largely counterweight this bias. Therefore, the current stock projections still serves as a workable basis for fisheries advice.

183 178 ICES WGBAST REPORT 2017 The reduced harvest rate in the Main Basin likely has had a positive effect also on the status of the AU 6 wild stocks (see Chapter 3). Most Estonian stocked parr and all stocked smolts have been adipose finclipped and the share of adipose finclipped salmon in Estonian coastal fishery is monitored by gathering catch samples. If the relative production of wild and reared smolt is compared with the share of finclipped fish in the Estonian coastal catch samples (Figure ) the share of finclipped fish is clearly smaller than expected. This indicates that reared fish have very low postsmolt survival and that wild fish are harvested in significant numbers. However, the origin of the wild fish in these catch samples has not been studied. 4.3 Stock projection of Baltic salmon stocks in assessment units Assumptions regarding development of fisheries and key biological parameters Table provides a summary of the assumptions on which the stock projections are based. The basis has been kept as similar to the last full assessment (ICES, 2015) as feasible, in order to allow for a review of how the new information (2015 data) is affecting projections. Fishing scenarios The base case scenario, scenario 1, for future fishing (2018 and onwards) matches the commercial catches adviced by ICES for , i.e. the median commercial removal would equal to salmon. Scenarios 2 and 3 correspond to 20% decrease and 20% increase from the scenario 1, respectively. Scenario 4a equals to F=0.1 harvest rule applied for (total) commercial removals, whereas in scenario 4b the F=0.1 harvest rule contains both commercial and recreational fisheries. Finally, scenario 5 illustrates stock development in case all fishing both at sea and in rivers was closed. Fisheries in 2016 are set to follow closely the realised catches (but which are not included in the model fitting as pointed out in Section 4.1). Fisheries in the interim year (2017) follow the scenarios, except for longline fishing during the first months of the year, which is estimated based on the effort observed during the corresponding months of The scenarios were computed by searching an effort that resulted in the desired median catches. As the scenarios are defined in terms of future fishing effort, the predicted catches have probability distributions according to the estimated population abundance, age-specific catchabilities and the assumed fishing effort. Scenarios 1 3 and 4a-b assume the same fishing pattern (division of effort between fishing grounds) as realized in Figure a b shows the harvest rates prevailing in the scenarios. In all scenarios it is also assumed that the commercial wanted catch reported covers 56% of the total sea fishing mortality, whereas 44% of this mortality consists of discards, misreported, unreported, and recreational sea fisheries. This corresponds to the situation assessed to prevail in 2016 (Figure ). According to the expert evaluation, unreporting and dead discarding have stayed on about the same level for the last few years. However, the share of other sources of extra mortality (misreporting and recreational catches) has been more varying and has slightly increased from 2015 to This results in a smaller share originating from the reported commercial catches than in the last few years. Looking back, the last year with similar share for

184 ICES WGBAST REPORT the reported commercial catches as in 2016 was estimated to occur in 2011 (Figure ). Survival parameters In the Mps projections an autoregressive model with one year lag is fitted at the logitscale with the historical estimates of the survival parameters. Mean values of the mean of the post-smolt survival over years (16%), variance over the timeseries and the autocorrelation coefficient are taken from the analysis into future projections. Survival from M74 is assumed to stay in its long-term mean (89%) starting in 2016, and the uncertainty around this mean corresponds to the uncertainty assessed for the historic annual estimates. No autoregressive model was applied for the M74 projections. Time-series for Mps and M74 survival are illustrated in Figure Adult natural mortality (M) is assumed to stay constant in the future, equalling the level estimated from the history. Different fisheries occur at different points in time and space, and many catch only maturing salmon which has been subject to several months natural mortality within a year. Thus, in order to increase comparability of abundances and catches, the abundances at sea have been calculated by first letting M to decrease the PFA (stock size in the beginning of year) of multi-sea-winter salmon for six months. Moreover, the stock size of grilse has been presented as the abundance after the period of post-smolt mortality and four months of adult natural mortality. This period is considered because the post-smolt mortality period ends in April, after which eight months of that calendar year remain when grilse are large enough to be fished. Half of that time period, i.e. four months, is considered to best represent the natural mortality that takes place before the fishing. Calculations for F=0.1 scenarios (scenarios 4a and 4b) are also based on stock sizes which are first affected by M, as described above. Maturation The annual sea age group-specific maturation rates are given as the average level computed over the historical period, separately for wild and reared salmon. This projection starts from 2018, as the maturation rates of 2017 can be predicted based on the SST information from the early 2017 (indicating slightly higher than average maturation rates for 2015). The time-series of maturation rates are presented in Figure Releases of reared salmon The number of released reared salmon per assessment unit is assumed to remain at the same level in the future as in 2016 (Table 3.3.1) Results According to the projections, stock size on the feeding grounds in 2018 will be about 1.46 ( ) million salmon; wild and reared, 1SW and MSW fish in total (Figure ). Of this amount, MSW salmon (i.e. fish which stay on the feeding area at least one and a half year after smolting) will account for 0.69 ( ) million salmon. These MSW fish will be fully recruited to both offshore and coastal fisheries in From the predicted amount of 1SW salmon (0.76 million, million) at sea in spring 2018, a relatively large fraction (most likely 20 30%) is expected to mature and become recruited to coastal and river fisheries. It must be noted, however, that the predicted maturation rate is estimated (only) based on winter sea surface temperatures in January February 2017.

185 180 ICES WGBAST REPORT 2017 The abundance of wild salmon at sea has fluctuated in the past without any apparent trend, but the so far highest median abundances, close to 1.1 million (both 1SW and MSW wild salmon) are estimated to have occurred around In 2018 the abundance of wild salmon is predicted to increase again to approximately the same level (1.16 million), and after 2018 it is predicted to rise even higher (Figure ). Because one of the simplifying assumptions of the modelled life cycle is that all salmon die after spawning, a lower maturation rate will increase the survival of the cohort to the next year compared to years with the same abundance but with average maturation. Similarly, a high maturation rate will decrease the abundance of MSW salmon in following years. Because of this feature, it is important to note that the predicted abundance may become over- or underestimated because of the (predicted) development of maturation rates. In contrast to wild salmon, the abundance at sea of reared salmon has decreased considerably since the mid-1990s, mainly due to the decline in post-smolt survival. Currently, the abundance of reared salmon is decreasing further, mainly due to recent reductions in the amount of stocked smolts (Table 3.3.1). The combined wild and reared abundance declined substantially from mid- 1990s until late 2000s, but the abundance has increased during the 2010s (Figure ). Table illustrates the predicted total commercial sea catches (also broken down to its parts: wanted catch, consisting of reported, unreported and misreported, and unwanted catch, consisting of discarded undersized and seal damaged salmon) and recreational reported catches (see below for explanation). The same table also contains catches and number of spawners in rivers 2018, for the given fishing scenarios. The assumed amount of unreporting, misreporting and discarding in 2018 is based on expert-evaluated shares of those catch components in the 2016 fisheries. Concerning recreational catches, only the catches reported by countries to the catch database (and shown in Table 2.2.6, Section 2) are used as the basis of estimation of recreational catches in This approach needs to be chosen instead of using the new updated total recreational catch estimates (see Section 2.1.1), because no assessment runs with the updated historic time-series of recreational catches are yet available. In 2016, the reported catch (wanted, commercial) accounted for about 57% of the corresponding estimated total sea catch, being 7% lower than estimated for the total sea catch in Unreporting, misreporting, discarding and reported recreational fishing are considered to take about 6%, 14% and 7% and 17% of the total sea catch, respectively. The share of the total catch by these components (including also river fishery) for the period is illustrated in Figure It is important to keep in mind that changes in either fishing pattern or in fisheries control may easily lead to changes in the share of catch caught under the quota regulation. With the given set of scenarios (excluding zero fishing scenario), the predictions indicate that the wanted catch reported (commercial) in year 2018 would be 65 98% ( salmon) compared to the TAC of 2017 (Table ). The corresponding total sea removal (including estimated recreational fishing) would range from salmon. The amount of spawners would be about 12% higher in scenario 3 than in scenario 2, and the zero fishing scenario indicates about 69% increase in the number of spawners compared to the scenario 2. The harvest rule of F0.1 for commercial fisheries only (scenario 4a) equals scenario 2, indicating a wanted catch reported of salmon. The harvest rule of F0.1 for the joint commercial and recreational sea fisheries (scenario 4b) equals scenario 1, indicating a wanted catch reported of salmon. Figure illustrates the longer term development of (reported) future catches given each scenario.

186 ICES WGBAST REPORT Figure a d presents the river-specific annual probabilities to meet 75% of the PSPC under each scenario. Under scenarios 1 4, different amounts of fishing has only a small influence on the level and no influence on the trend of the probability of meeting 75% over time. Only the zero fishing scenario diverges clearly from the other scenarios; several of the weakest rivers show a stronger positive effect in trends than the other scenarios. As expected, changes in fishing pressure have the smallest effect to those stocks which are close to their PSPC. As the level of fishing effort is rather low in these scenarios compared to history, the levels of post smolt and adult natural mortalities will have a high relative impact on the resulting chances of reaching the management objective with a high certainty. Table compares the probabilities to reach 75% target around the year 2022/2023, which is approximately one full generation ahead from now. The probabilities to reach the target are relatively high in most of the stocks, and differences between scenarios are small except in the zero fishing scenario. Figure a b illustrates (per scenario) the rate and the direction of change in smolt abundance in 2022/2023 compared to the smolt abundance in Predictions about smolt abundance are naturally more uncertain than the estimated abundance in However, in those stocks which are close to their PSPC also the predictions are rather certain, indicating that smolt abundance will stay close to PSPC in these rivers under different fishing scenarios. Figure a h show longer term predictions in river-specific smolt and spawner abundances for three of the scenarios (1=removal which corresponds to ICES advice for 2017; 2=20% increase in removal compared to scenario 1; 5=zero fishing). The two first scenarios result in very similar smolt and spawner abundances, while the extreme scenario of zero fishing results in notable increases in smolt and spawner abundance for most of the rivers. 4.4 Additional information affecting perception of stock status Independent empirical information is important for the evaluation of model predictions and their key parameters. In the present report, several comparisons between model predictions and independent empirical information have been conducted to evaluate model results (Sections and 4.2.5). Over the years, repeated comparisons with different kinds of such independent information have been performed, and in several cases these comparisons have prompted modifications or extensions to the full life-history model. For example, a few years ago sea temperature data were successfully introduced as a covariate of age-specific maturations rates, based on the analyses and development work carried out in the last inter-benchmark protocol (IBP, ICES, 2012) and thereafter. Also, comparisons between model predictions and empirical results from genetic mixed-stock analyses (MSA) have been used over the years to verify model performance (e.g. ICES, 2014). This section focuses on other auxiliary information important to complete evaluation of the current stock status. In particular we highlight information about diseases and other factors that may affect development in stock status, but which are not taken into consideration in the current modelling. Likewise, weaknesses in input data used in the assessment model might affect the precision of status evaluations, and in the worst case they may also introduce biases. Such shortcomings in the current assessment model, when it comes to input data and ways of handling those, are also discussed under this section. An example is the ongoing work of updating prior information on production areas and potential smolt production levels in certain salmon rivers, which may affect status evaluations of individual stocks.

187 182 ICES WGBAST REPORT M74 and disease outbreaks and potential effects on stock development If the increase in M74 observed from 2016, predicted to continue further in 2017 (Section 3.4.1), should prevail, this may gradually result in decreased stock status and reduced fishing possibilities. The future M74 development is difficult to foresee, but many of the M74-fluctuations seen since the early 1990s have tended to last some years before changing direction (Figure 3.4.3). Also, the disease outbreaks reported in recent years for adult salmon in several rivers (Section 3.4) is a concern for the future. In contrast to M74, the cause of the disease is still unknown, and to accurately quantify the amount of affected or dead salmon in a river appears difficult, if at all possible. Although future effects of increased M74 and other diseases are hard to predict, the development of individual stocks is expected to mostly depend on stock status. In cases where smolt production has increased and is approaching the maximum capacity, density-dependent mortality is expected to be much higher than in populations with weaker status. Hence, in a recovered stock (i.e. with high status) elevated fry mortality may partly be compensated by the reduced density-dependent mortality among the offspring not affected by M74. For the same reason, stronger stocks may be less sensitive to a reduced number of deposited eggs due to adult female mortality. In contrast, the effects of M74 and other diseases on the future smolt production of weaker populations are expected to be more pronounced because of the lack of the above described buffer effect of density dependence in the reproduction dynamics. Despite the recent increase in M74 and disease outbreaks, the average salmon 0+ parr densities in 2016 remained at high levels in the most AU 1 4 rivers. The only clear exception was Vindelälven, where the average 0+ density dropped drastically from ca. 40 parr/100 m 2 in 2015 to only ca. 1 parr/100 m 2 in 2016; the lowest density observed since the 1990s (Table ). Exact reasons for the decline are unclear, but it likely reflects a combination of factors. In 2015, the proportion of females counted in the Norrfors fish ladder, which all ascending salmon must pass, was only 11% (28% among MSW salmon; Table ). There are no indications of such a skewed sex ratio in the sea or at the river mouth. Hence, the recent disease problems in Ume/Vindelälven (and other rivers) may for some reason have prevented particularly females from reaching the fish ladder. Additional disease related mortality before spawning is also likely to have occurred further upstream in the river. Moreover, the low level of thiamine of the 2015 spawners (Section 3.4) may have contributed further to the drastic reduction in parr density by reducing female performance, in addition to an expected increase in M74-related fry mortality (19% of Ume/Vindelälven females used as brood stock at the local hatchery displayed M74 in 2015; Table 3.4.1). The very low parr density in Vindelälven 2016 is expected to result in reduced smolt production within some years (mainly in 2019). Since the latest successfully run assessment model only uses data up to 2015 (Section 4.1), the recent drop in 0+ parr density has yet not been accounted for in the present stock assessment. As a consequence, estimates of future smolt production and stock status in Vindelälven are likely overestimated by the presently used model. However, it should be noted that the estimated pre-fishery abundance of Vindelälven salmon exploited in the fishery during the advice year (2018) is not affected by the reduced parr densities in Revision of basic input data The potential smolt production capacity (PSPC) of three Swedish rivers, Piteälven, Lögdeälven and Öreälven, is expected to be underestimated by the assessment mod-

188 ICES WGBAST REPORT el, and this may introduce biases in the status assessment. Colonization of salmon to new areas further upstream in the rivers, recent restoration of river habitats and too low past perceptions about the potential smolt production per unit area are the main reasons behind the need to revise PSPCs for these three rivers. Factors affecting the PSPC include river production area, smolt production potential per unit area and mortalities during downward migration of smolts. In the analytical assessment of Baltic salmon, all these quantities are used to formulate a prior probability density function (hereafter called prior ) for PSPC which is updated by the model to a posterior PSPC when stock recruit data are included. Status of individual stocks is evaluated by comparing observed smolt production levels with posteriors of PSPC. To arrive at realistic estimates of PSPC, we have updated figures on production areas and information on maximal smolt production per unit area for the three above Swedish rivers. The updated information on production area and smolt production potential per unit area was used in combination with information on other important factors, such as mortality during smolt migration, to formulate new priors for PSPC to be included in the assessment model. The PSPC priors were formulated using expert opinions about relevant variables. Where available, the experts considered relevant data to help inform their opinion. A simple model was developed to derive a probability distribution for PSPC as a function of the expert-elicited variables. These variables were smolt density (i.e. number of smolts produced per hectare); the area of habitat suitable for salmonid production; survival during the downstream smolt migration; passage efficiencies for any obstacles to downstream migration and, where applicable, mortality rates associated with passage of migration obstacles. Experts were asked for summaries (mode, and quantiles) of distributions describing their knowledge and uncertainty about quantities of interest. After this, parametric distributions were fitted, plotted and shown to experts for feedback and possible revision. Updated median values for the PSPC priors were for Piteälven, for Öreälven and for Lögdeälven (Figure ). For a more detailed explanation of the procedure used to revise PSPC priors, see Annex 4 in ICES (2015). Unfortunately, because of model related problems (Section 4.1), the updated PSPC priors are not included in this year s assessment, and status assessment results for the three rivers under consideration should therefore be interpreted with caution (as indicated in tables and figures presenting assessment results). Especially Piteälven stock status is most likely overestimated in the current assessment. The plan is to include the updated PSPC priors in the assessment If seen necessary, also other rivers might be evaluated in the near future with respect to basic parameters such as current production area and potential smolt production per unit area. Also, stock status and future projections for Mörrumsån and Emån are presently uncertain, since updated smolt priors for these AU4 rivers (from the recent benchmark) were not included in the present assessment. 4.5 Conclusions The recent increase in the smolt production in most rivers, the improved status of wild stocks and the projected increase in the prefishery abundance at sea indicate higher fishing possibilities for The most recent data from 2016, which are not utilized in the assessment, indicate even a slightly more positive current situation than the assessment model results (Section 4.2.4). Moreover, the minor projected response of wild stocks to +/-20% change in the sea removals indicate that fishing mor-

189 184 ICES WGBAST REPORT 2017 tality is currently at a fairly low level compared to other (natural) sources of mortality. The observed positive trends among most of the AU 4 6 stocks (which are generally considered to be weaker than the AU 1 3 stocks) indicate that the current fishing mortality level, given the current natural mortality, allow gradual recovery of also these stocks. Therefore, the WG could not see any need for decreasing current fishing mortality. Obviously, probabilities to reach the objectives are higher for scenarios with low exploitation, but differences between scenarios are small except for the one with zero fishing. There are, however, also concerns for the short-term stock development. Because the WG was not able to run assessment using the very latest data (2016), the evaluation of the current status of stocks contains some extra uncertainty which is also carried to the stock projections. Moreover, there are indications of currently increasing M74- mortality, as well as reported deaths of spawners due to disease problems (Section 4.4.1). Finally, recreational trolling fishing of salmon appears to be far more extensive, especially on the salmon feeding grounds, than reported to the WG in the past years (see Section 2.1.1). This growing mixed-stock fishing is not yet explicitly included in the assessment and the stock projections. It is important to note, however, that wild reproduction among most of the stocks have improved in spite of the simultaneous increase of the trolling fishery, as summarized in the previous paragraph. Given the still large number of stocks with weak current status, any positive effects of a higher-than-expected post-smolt survival would need to be directed to the increase of spawners in these stocks, rather than to increasing fishing possibilities. This holds especially for fisheries which take place on the migration routes of the weakest stocks (e.g. close to the river mouth) where their share in catches becomes higher. If M74- mortality will increase it may easily counteract any positive effects of higher-thanexpected post-smolt survival, especially in weaker stocks (Section 4.4.1). It is also important to remember that special actions (not only fishery-related ones) directed to the weakest stocks are required, especially in AUs 4 6 at the adviced TAC levels. Despite signs of improvement in status of most southern stocks in the last few years, a large majority of these rivers are still far below a good state. In contrast, a majority of northern stocks are close to or above the MSY-level (2016 smolt production; Table ) and the surplus produced by these stronger stocks could in theory be directed towards stock-specific fisheries. However, the current management system, with a single TAC for SD that is set at a relatively low level to safeguard weaker salmon stocks, prevents this surplus to be fully utilised by the commercial sea fishery. In a similar way, the surplus of reared salmon cannot be fully utilised today because reared salmon is also included in the TAC. Stock-specific management could be developed further by implementation of more flexible systems for regulation of commercial fisheries, with the aim of steering exploitation towards harvesting of reared salmon and stronger wild stocks, through e.g. area-specific quotas and/or exclusion of certain single-stock fisheries from the quota system (such as fisheries in estuaries of rivers with reared stocks). Also noncommercial coastal fishing, not regulated by international quotas, could be steered towards stock-specific harvesting. In contrast, the increasing recreational trolling in Main Basin is a true mixed-stock fishery where stock-specific harvesting is not possible, although regulations that only allow landing of finclipped (reared) salmon, as has been implemented in Sweden since 2013, can somewhat reduce fishing mortality for wild stocks (given that post-release mortality is relatively low). A higher degree of

190 ICES WGBAST REPORT stock-specific exploitation will also be necessary in the future if different management objectives should be decided upon for individual stocks (e.g. if to allow for a higher number of spawners than needed to fulfil the MSY-level in certain wild rivers). 4.6 Benchmark and a road map for development of the stock assessment Benchmark The benchmark of Baltic salmon (WKBaltSalmon 2017) convened into two workshops at ICES (Copenhagen, Denmark). The data evaluation workshop was held November It focused on mapping and evaluating available monitoring and fisheries data, and searched for possibilities to improve the coverage and quality of data for Baltic salmon assessment. WKBaltSalmon also made an effort to correct shortcomings and inconsistencies in the dataseries of salmon and sea trout catches, discards and efforts from , and to transfer previously collected data into InterCatch (the ICES web-based database). The data evaluation workshop resulted in an updated description of available river monitoring data. In addition, shortcomings in fisheries data were identified and some outlining for an improvement work plan was done. Transfer of commercial fisheries data into InterCatch largely failed, however, because all countries did not manage to submit all required data and the whole intended time-series. As a result, InterCatch contains at present less fisheries data than the catch database that is used as a basis for the WGBAST assessment. The aim is to solve the catch data issues in connection to next data call in late The benchmark process also resulted in a time-series of preliminary trolling catch estimates trough an expert elicitation (Section 2.1.1). The intention is to submit also recreational catch data into InterCatch in the future. The Method evaluation workshop was held in 30 January 3 February The workshop focused on three major aspects: stock recruitment model selection, an evaluation of the 75% objective as a proxy for MSY, and development of a smolt production model for southern Baltic Sea rivers. The benchmark workshop evaluated the present modelling work and outlined a work plan ( road map ) for complementation of the model development (see Section 4.6.2). The aim is to complete the work in autumn 2018 when the next benchmark is planned to take place. The report of WKBaltSalmon will be become available in May Road map for development of the assessment The tasks listed below refer to ongoing, planned and potential updates of the assessment method. Issues included in the benchmark assessment (WKBaltSalmon) are indicated. Ongoing work Evaluation of alternative stock recruitment models for Baltic salmon (part of WKBaltSalmon): The current assessment model uses a Beverton Holt stock recruitment function, and the stock recruit parameterization in use could give rise to a population that is not at demographic equilibrium (constant population size and age structure) in the absence of fishing. New parameterization of stock recruitment dynamics and new priors from hierarchical meta-analysis of Atlantic salmon stock recruit data has therefore been developed and explored during WKBaltSalmon. Also, migration of the model

191 186 ICES WGBAST REPORT 2017 from WINBUGS to JAGS environment has been initiated. Migration of the model to a new software environment, as well as new parameterizations of stock recruit dynamics, will be evaluated further during 2017, and implemented in the assessment model in Development of smolt production models for southern stocks and for Piteälven (part of WKBaltSalmon): Important updates of production areas and priors for potential smolt production capacities for Mörrumsån and Emån were carried out in So far, the yearly smolt production in these rivers has been estimated from parr densities and previous expert opinions about production areas and within river mortalities, and the quality of these estimates is questionable. In AU 5 6 rivers, smolt production is also estimated from similar, simple models. The work to develop a more rigorous smolt production model, similar to the one used for AU 1 3 stocks, including available smolt counts and updated information on habitats and within river mortalities, was initiated in 2015 and a first version has been finalized during WKBaltSalmon. Mörrumsån and Emån have been in focus in this work, and updated smolt production estimates of these two rivers will be implemented in the 2018 assessment. The idea is also to expand the model to include Testeboån, which will facilitate inclusion of this AU 3 river in the full life-history model in the coming assessment. In the future, when data become available, the model can also be applied to AU 5 6 stocks. Estimation procedures for smolt production and the prior for potential smoltproduction capacity for Piteälven are of poor quality and probably give biased results. An updated prior for potential smolt production capacity has been developed during the benchmark (Section 4.4.2), but the procedure to estimate current smolt output will be developed in beginning of 2018 and hopefully included in the assessment in Evaluation of the 75% objective as a proxy for MSY (part of WKBaltSalmon): The use of the current management objective (smolt production that corresponds to 75% of PSPC) as a proxy for MSY was evaluated during WKBaltSalmon. A modified version of the scenarios code was used for forward projections of population dynamics, to find effort that maximises long-term catch based on an optimisation routine. Results indicate that the MSY level differs between rivers and that the 75% objective may in many cases be too low and not consistent with precautionary approach. However, this needs to be evaluated further once an assessment model (in JAGS environment) including all rivers and data is available, and when a new parameterization of stock recruit dynamics has been finalized, i.e. during Inclusion of the recreational sea fishery (mainly trolling) (part of WKBaltSalmon): Because of the increase in the recreational trolling fishery at sea, it is important to include it as an independent fishery to the model framework. This will require good quality data and catch estimates from countries with a significant recreational trolling fishery. During the benchmark, a plan for how to collect data and include recreational catches in the assessment model as a separate fleet was discussed and decided upon. An updated time-series of trolling catch estimates (benchmark expert elicitation; Section 2.1.1) will also be included in the 2018 assessment and onwards.

192 ICES WGBAST REPORT Important issues, planned to be dealt with in the next 2 3 years Inclusion of AU5 and 6 stocks in the full life-history model: At present, these stocks are treated separately from the AU 1 4 stocks. Inclusion in the full life-history model will require updated information regarding e.g. smolt age distributions, maturation rates, exploitation rates, post-smolt survival and information about exploitation of stocks from Gulf of Bothnia and Main Basin in the Gulf of Finland (and vice versa). In addition, increased amounts of basic biological data (e.g. smolt and spawner counts, additional electrofishing sites) may be needed for some rivers. The smolt-production model for southern stocks that has been developed could be used also for AU 5 6 stocks in the future, to produce smolt production priors for the life-history model. Continuing the work of including data from established index rivers: This e.g. includes fitting the life-history model to spawner counts from River Mörrumsån. In parallel with data collection in index rivers additional data collection of smolt and spawner abundances in other wild rivers will reduce biases, and improve precision in assessment results. Therefore, it would be desirable to initiate a rolling sampling programme that regularly collects abundance data from rivers where no such data are presently available. Improved estimates (for input into assessment model) of the exploitation of stocks in the coastal fishery: A spatially- and temporally-structured Bayesian population dynamics model that tracks the migration of Baltic salmon stocks from their feeding grounds in the Baltic Sea to their natal rivers has been developed (Whitlock et al., in prep.). The model uses information about the proportions of different stocks in catch samples from Swedish and Finnish coastal fisheries at different points in space and time, as well as finclipping information about the proportion of wild and reared fish in catches (also for catches where no genetic data are available). The potential for incorporation of tag recapture data to further inform migration patterns will also be investigated. The model will be used for estimation of stock-specific exploitation rates in the coastal fisheries and as input data in the current assessment model. Furthermore, the migration catch model is a suitable tool to evaluate (by simulations) effects of changes in fishing patterns/management on the exploitation and development of wild salmon stocks. Such a tool is anticipated to become very important in the work to develop a stock-specific management, and necessary to fulfil requirements in a new multiannual management plan (cf. COM/2011/0470 final). Improving precision in short-term projections: Improving estimates of postsmolt survival by fitting the model to explanatory variables like information on herring recruitment and development in sea surface temperatures will increase precision in short-term projections. Less urgent issues, good to keep in mind Inclusion of data on composition of stocks at sea: The life-history model has already been fitted to information on proportions of wild and reared salmon in Main Basin as determined from scale readings. The next step would be to include genetic information on proportions of fish from different AUs, separating also wild and reared salmon from those areas. Subsequently, in-

193 188 ICES WGBAST REPORT 2017 formation on the representation of single stocks may be included. See more on future MSA in ICES (2015), Section 4.7. Further use of scale-reading data: In addition to wild/reared proportions, age data from catch samples could be used to get improved knowledge on year-class strength, maturation and natural mortality rates. 4.7 Needs for improving the use and collection of data for assessment Because requirements for data will always exceed available resources, preferences must be given. The identification and prioritisation of new data collection, or modifications to ongoing monitoring work, should be based on end-user needs, particularly ICES assessment needs, and is of importance with respect to planning of the next multiannual period within the European data collection framework (DCF), scheduled to begin in 2018 (new acronym: EU-MAP). Over the years the WG has repeatedly highlighted and discussed various needs for data collection (e.g. ICES, 2014; 2015; 2016). For example, two years ago (ICES, 2015) the need for genetic analysis to study stock composition in catch samples (MSA) was reviewed, with suggestions provided regarding future studies. In last year s report (ICES, 2016) comments were given to a comprehensive list of proposals for Baltic salmon data collection produced at an earlier ICES workshop in Further, the need for at least one index river per assessment unit was highlighted, with suggestions given on potential candidates in AUs 5 6 (where full index rivers are still missing). Finally, as part of a recent benchmark for Baltic salmon (WKBaltSalmon; Section 4.6) all different types of information needed as input for the Baltic salmon stock assessment (fisheries statistics, biological data, etc.) were reviewed with respect to needs, availability and quality, and the outcome of that evaluation is to be reported in This report could be used as a basis for prioritisations and decisions about data collection included in EU-MAP.

194 ICES WGBAST REPORT Table Posterior probability distributions for steepness, alpha and beta parameters of the Beverton Holt stock recruit relationship and eggs per recruit (EPR, millions) for Baltic salmon stocks. Posterior distributions are summarised in terms of their mean and CV (%). Steepness Alpha parameter Beta parameter EPR Mean cv Mean cv Mean cv Mean cv Assessment unit 1 1 Tornionjoki Simojoki Kalixälven Råneälven Assessment unit 2 5 Piteälven Åbyälven Byskeälven Kågeälven Rickleån Sävarån Ume/Vindelälven Öreälven Lögdeälven Assessment unit 3 14 Ljungan Assessment unit 4 15 Emån Mörrumsån

195 190 ICES WGBAST REPORT 2017 Table Posterior probability distributions for the smolt production capacity (* 1000) in AU1 4 rivers and the corresponding point estimates in AU5 6 rivers. The posterior distributions are described in terms of their mode or most likely value, the 90% probability interval (PI) and the method by which the posterior probability distribution was obtained. These estimates serve as reference points to evaluate the status of the stock. For the updated estimates of AU1 4 rivers except Testeboån, medians as estimated by last years stock assessment are also shown. This enables comparison of how much the estimated medians have changed compared to last year. Note that recently updated priors of potential smolt production capacities (Piteälven, Lögdeälven, Öreälven) and current smolt production (Emån, Mörrumsmån) were not used, and posterior estimates for these rivers are uncertain or likely biased (Section 4.4.2). Smolt production capacity (thousand) Method of Last year s median % change Mode Median Mean 90% PI estimation Assessment unit 1 1 Tornionjoki % 2 Simojoki % 3 Kalixälven % 4 Råneälven % Total assessment unit 1 Assessment unit % 5 Piteälven * % 6 Åbyälven % 7 Byskeälven % 8 Kågeälven % 9 Rickleån % 10 Sävarån % 11 Ume/Vindelälven % 12 Öreälven * % 13 Lögdeälven * % Total assessment unit % Assessment unit 3 14 Ljungan % 15 Testeboån Total assessment unit 3 Assessment unit % 16 Emån * % 17 Mörrumsån * % Total assessment unit % Total assessment units 1-4 Assessment unit % 18 Pärnu % 19 Salaca % 20 Vitrupe % 21 Peterupe % 22 Gauja % 23 Daugava % 24 Irbe % 25 Venta % 26 Saka % 27 Uzava % 28 Barta % 29 Nemunas river basin % Total assessment unit % Assessment unit 6 30 Kymijoki % 31 Luga % 32 Purtse % 33 Kunda % 34 Selja % 35 Loobu % 36 Pirita % 37 Vasalemma % 38 Keila % 39 Valgejögi % 40 Jägala % 41 Vääna % Total assessment unit % Total assessment units % * River w ith recently updated prior for potential or current smolt production (not included in present assessment).

196 ICES WGBAST REPORT Table Salmon smolt production in Baltic rivers with natural reproduction of salmon grouped by assessment units. Median number (x 1000) of smolts from natural reproduction with the associated uncertainty (90% Probability interval). Note that in WGBAST report 2011 and earlier, distributions were described in terms of their modes (single most likely value) instead of their medians. From 2015 also reproductive areas are shown as medians with 90% PI (previously modes with 95% PI). Note that old priors for PSPC and/or smolt production are used for Piteälven, Öreälven, Lögdeälven, Emån and Mörrumsån, and the presented estimates of PSPC and/or smolt production for these rivers are therefore likelly biased. Testeboån has not yet been included in the full life-history model and therefore no posterior estimates of smolt production are given (smolt priors are presented in Table ). Assessment unit, sub-division, Reprod. area (ha, median) Potential (*1000) Method of Pred Pred Pred estimation Pot. Pres. prod. prod. country Category Gulf of Bothnia, Sub-div : Finland Simojoki wild % PI Finland/Sweden Tornionjoki;Torneälven wild % PI Sweden Kalixälven wild % PI Råneälven wild % PI Assessment unit 1, total % PI Piteälven wild % PI Åbyälven wild % PI Byskeälven wild % PI Rickleån wild % PI Sävarån wild % PI Ume/Vindelälven wild % PI Öreälven wild % PI Lögdeälven wild % PI Kågeälven wild % PI Assessment unit 2, total % PI Ljungan wild % PI Testeboån wild na na na na na na na na na na na % PI na na na na na na na na na na na na na

197 192 ICES WGBAST REPORT 2017 Table Continued. Assessment unit 3, total % PI na Total Gulf of B., Sub-divs % PI Method of Assessment unit, Pred Pred Pred estimation sub-division, Reprod. area Potential Pot. Pres. country Category (ha, median) (*1000) prod. prod. Sweden Emån wild % PI Mörrumsån wild % PI Assessment unit 4, total % PI Estonia Pärnu wild , 4 Latvia Salaca wild Vitrupe wild Peterupe wild , 5 Gauja mixed , 5 Daugava*** mixed , 6 Irbe wild Venta mixed , 5 Saka wild Uzava wild Barta wild Lithuania Nemunas river basin wild , 4 Assessment unit 5, total Total Main B., Sub-divs Method of Assessment unit, Pred Pred Pred estimation sub-division, Reprod. area Potential Pot. Pres. country Category (ha, median) (*1000) prod. prod. Finland: Kymijoki mixed 151)+602) 20 1) +80 2) Russia: Neva mixed Luga mixed SE Estonia: Purtse mixed Kunda wild 1.9 2,1(2,8) Selja mixed Loobu mixed ,5 (17) Pirita mixed , 3 SE Vasalemma wild Keila wild 3.5 5,4 (12) Valgejögi mixed Jägala mixed Vääna mixed Assessment unit 6, total Gulf of B.+Main B.+ Gulf of F., Sub-divs % PI

198 ICES WGBAST REPORT Table Overview of the status of the Gulf of Bothnia and Main Basin stocks in terms of their probability to reach 50 and 75% of the smolt production capacity in Stocks are considered very likely to have reached this objective in case the probability is higher than 90%. They are likely to have reached the objective if the probability is between 70 and 90%, uncertain when the probability is between 30 and 70 % and unlikely if the probability is less than 30%. For the AU1 4 stocks except Testeboån, the results are based on the assessment model, whilst the categorization of Testeboån and AU5 6 stocks is based on expert judgments; for those rivers there are no precise probabilities (column 'Prob'). Note five stocks (cells in grey) where current status is uncertain (see Section 4.4.2). Prob to reach 50% Prob to reach 75% Prob V.likely Likely Uncert. Unlikely Prob V.likely Likely Uncert. Unlikely Unit 1 Tornionjoki 1.00 X 0.79 X Simojoki 0.92 X 0.50 X Kalixälven 1.00 X 0.93 X Råneälven 0.96 X 0.76 X Unit 2 Piteälven* 1.00 X 0.98 X Åbyälven 0.98 X 0.87 X Byskeälven 1.00 X 0.91 X Kågeälven 0.45 X 0.27 X Rickleån 0.10 X 0.01 X Sävarån 0.94 X 0.79 X Ume/Vindelälven 0.99 X 0.77 X Öreälven* 0.74 X 0.48 X Lögdeälven* 0.95 X 0.79 X Unit 3 Ljungan 0.92 X 0.78 X Testeboån n.a. X n.a. X Unit 4 Emån* 0.02 X 0.00 X Mörrumsån* 1.00 X 1.00 X Unit 5 Pärnu n.a. X n.a. X Salaca n.a. X n.a. X Vitrupe n.a. X n.a. X Peterupe n.a. X n.a. X Gauja n.a. X n.a. X Daugava n.a. X n.a. X Irbe n.a. X n.a. X Venta n.a. X n.a. X Saka n.a. X n.a. X Uzava n.a. X n.a. X Barta n.a. X n.a. X Nemunas n.a. X n.a. X Unit 6 Kymijoki n.a. X n.a. X Luga n.a. X n.a. X Purtse n.a. X n.a. X Kunda n.a. X n.a. X Selja n.a. X n.a. X Loobu n.a. X n.a. X Pirita n.a. X n.a. X Vasalemma n.a. X n.a. X Keila n.a. X n.a. X Valgejögi n.a. X n.a. X Jägala n.a. X n.a. X Vääna n.a. X n.a. X * Status uncertain. Updated priors for PSPC and/or smolt production not in present assessment.

199 194 ICES WGBAST REPORT 2017 Table Estimates of wild smolt production (*1000) in Baltic salmon rivers updated with 2016 data. The estimates are based either on the river model (most of the AU1 3 rivers) or other methods considered appropriate by experts of these rivers (all AU 4 6 rivers). For rivers in AU1 4, median values and the associated 90% probability intervals (PI) are presented. For most of the other rivers no measure of uncertainty is available, but for some rivers + SE are presented. In earlier years, when full assessment was carried out, these timeseries from AU1 4 rivers were used as input in the Full Life-History Model (see Section and Stock Annex). Wild smolt production (thousand) Method of estimation Assessment unit 1 1 Tornionjoki *) 1,2 90% PI Simojoki ,2 90% PI Kalixälven % PI Råneälven % PI Total assessment unit % PI Assessment unit 2 5 Piteälven % PI Åbyälven % PI Byskeälven % PI Kågeälven % PI Rickleån ,2 90% PI Sävarån ,2 90% PI Ume/Vindelälven ,2 90% PI Öreälven % PI Lögdeälven ,2 90% PI Total assessment unit % PI Assessment unit 3 13 Ljungan % PI Testeboån % PI Total assessment unit % PI

200 ICES WGBAST REPORT Table Continued. Assessment unit estimation 15 Emån ,4 90% PI Mörrumsån ,4 90% PI Total assessment unit % PI Assessment unit 5 17 Pärnu ,4 18 Salaca Vitrupe Peterupe ,5 21 Gauja ,5 22 Daugava ,6 23 Irbe Venta ,5 25 Saka Uzava Barta Nemunas river basin ,4 Total assessment unit 5 Assessment unit Kymijoki Neva Luga ±SE Purtse Kunda Selja Loobu Pirita ,3 ±SE Vasalemma Keila Valgejögi Jägala Vääna Total assessment unit 6 Wild smolt production (thousand) *) Electrofishing data from 2016 is not used at all in the R. Tornionjoki, which effectively prevents prediction of 2019 smolt production in this river (see Section for details) Method of estimation of current smolt production 1. Bayesian linear regression model, i.e. river model (see Stock Annex) 2. Sampling of smolts and estimate of total smolt run size. 3. Estimate of smolt run from parr production by relation developed in the same river. 4. Estimate of smolt run from parr production by relation developed in another river. 5. Inference of smolt production from data derived from similar rivers in the region. Method of

201 196 ICES WGBAST REPORT 2017 Table Key assumptions underlying the stock projections. The same post-smolt survival scenario and M74 scenario are assumed for all effort scenarios. Survival values represent the medians to which Mps and M74 are expected to return. Scenario Total commercial removal (dead catch) for year Removal that corresponds to ICES advice for fishing year % increase to scenario % decrease to scenario 1 4(a) F0.1 approach (commercial removal) 4(b) F0.1 approach (total removal) 5 zero fishing In all scenarios we assume that the commercial removal (wanted catch reported) covers 56% of the total sea fishing mortality, whereas 44% of this mortality consists of discards, misreported, unreported and recreational sea fisheries. (See text for details) Post-smolt survival of wild salmon Average survival between (16%) Post-smolt survival of reared salmon Same relative difference to wild salmon as on average in history M74 survival Historical median (89%) Releases Same number of annual releases in the future as in 2016 Maturation Age group specific maturation rates in 2016 are predicted using january-april 2016 SST data, and in 2017 using january-february 2017 SST-data. For other years, average maturation rates over the time series are used, separately for wild and reared salmon.

202 ICES WGBAST REPORT Table Scenario Total commercial catch at sea Commercial catches (thousands of fish) at sea in SD in 2018 Wanted Catch Reported (% of 2017 EU TAC) Unwanted Catch (Dead+Alive) Undersized Seal damaged Wanted Catch Unreported Wanted Catch Misreported % % % (a) % (b) % % Recreational Scenario catch at sea (a) 28 4(b) Total sea catch (comm. + recr.) 2018 River catch 2018 Spawners Table River-specific probabilities in different scenarios to meet 75% of PSPC in 2022/2023 (depending on the assessment unit) Probabilities higher than 70% are presented in green. Note five stocks (cells in grey) where status is uncertain (see Section 4.4.2). Probability to meet 75% of PSPC River Year of Scenario comparison (a) 5 Tornionjoki Simojoki Kalixälven Råneälven Piteälven* Åbyälven Byskeälven Rickleån Sävarån Ume/Vindelälven Öreälven* Lögdeälven* Ljungan Mörrumsån* Emån* Kågeälven * Status uncertain. Updated priors for PSPC and/or smolt production not in present assessment.

203 198 ICES WGBAST REPORT 2017 Figure Post-smolt survival for wild and hatchery-reared salmon. Estimates from 2015 stock assessment are illustrated in grey for comparison.

204 ICES WGBAST REPORT Figure Proportion maturing per age group and per year for wild and reared salmon.

205 200 ICES WGBAST REPORT 2017 Figure a. These graphs show the distributions for egg abundance (million), plotted against the smolt abundance (thousand) for stocks of assessment units 1, 2, 3 and 4. Blue dots present the posterior distributions of annual smolt and egg abundances, red curves indicate the distributions of stock recruit relationship.

206 ICES WGBAST REPORT Figure b. These graphs show the distributions for egg abundance (million), plotted against the smolt abundance (thousand) for stocks of assessment units 1, 2, 3 and 4. Blue dots present the posterior distributions of annual smolt and egg abundances, red curves indicate the distributions of stock recruit relationship.

207 202 ICES WGBAST REPORT 2017 Figure c. These graphs show the distributions for egg abundance (million), plotted against the smolt abundance (thousand) for stocks of assessment units 1, 2, 3 and 4. Blue dots present the posterior distributions of annual smolt and egg abundances, red curves indicate the distributions of stock recruit relationship.

208 ICES WGBAST REPORT Figure d. These graphs show the distributions for egg abundance (million), plotted against the smolt abundance (thousand) for stocks of assessment units 1, 2, 3 and 4. Blue dots present the posterior distributions of annual smolt and egg abundances, red curves indicate the distributions of stock recruit relationship.

209 204 ICES WGBAST REPORT 2017

210 ICES WGBAST REPORT Figure a. Prior and posterior probability distributions of the potential smolt production capacity obtained in the assessment in 2015 (thin line) and 2017 (bold line). Prior distributions are illustrated with dotted lines.

211 206 ICES WGBAST REPORT 2017 Figure b. Prior and posterior probability distributions of the potential smolt production capacity obtained in the assessment in 2015 (thin line) and 2017 (bold line). Prior distributions are illustrated with dotted lines.

212 ICES WGBAST REPORT Figure Posterior probability distribution (median and 90% PI) of the total smolt production within assessment units 1 4 and in total. Vertical lines show the median (solid line) and 90% PI (dashed lines) for potential smolt production capacity (PSPC).

213 208 ICES WGBAST REPORT 2017 Figure Probability of reaching 50% of the smolt production capacity for different stocks of assessment units 1 4. Figure Probability of reaching 75% of the smolt production capacity for different stocks of assessment units 1 4.

214 ICES WGBAST REPORT Figure Estimated posterior distributions of catches in comparison to corresponding observed catches. Observed catches refer to reported commercial catches recalculated to take into account unreported catches in the longlining fishery. Note that the latest observation in each time-series (indicated with a square) was not included in the simulation model run.

215 210 ICES WGBAST REPORT 2017 Figure Estimated model proportions of wild salmon in offshore catches in comparison to wild proportions observed in the catch samples among 2SW and 3SW salmon (from scale reading). Note that the latest years in time-series (indicated with squares) were not included in the simulation model run. The x-axis shows year of smoltification.

216 ICES WGBAST REPORT Figure Estimated posterior distributions of the amount of spawners (in thousands) in each river versus observed numbers of spawners in fish counters. River observed numbers indicated with triangles are used as input in the full life-history model.

217 212 ICES WGBAST REPORT 2017 Figure a. Estimated posterior distributions of the harvest rates (harvested proportion of the available population) in offshore driftnet and offshore longline fisheries separately for one-seawinter and multi-sea-winter salmon. Note that the driftnet harvest rate in 2008 is not zero, since due to computational reasons it contains fishing effort from the second half of year 2007.

218 ICES WGBAST REPORT Figure b. Estimated posterior distributions of the harvest rates (harvested proportion of the available population) in other coastal fisheries than driftnetting in AU1 and in coastal driftnetting (all AUs together) separately for one-sea-winter and multi-sea-winter salmon.

219 214 ICES WGBAST REPORT 2017 Figure c. Estimated posterior distributions of the harvest rates (harvested proportion of the available population) in the river fishery separately for one-sea-winter and multi-sea-winter salmon. Figure Combined harvest rates (harvested proportion of the available population) for offshore and coastal fisheries for MSW wild salmon in

220 ICES WGBAST REPORT Smolts (thousands) Life history model, data up to 2015 River model/expert estimates 5 % 95 % 5 % 95 % Smolts (thousands) Life history model,data up to 2015 River model/expert estimates 5 % 95 % 5 % 95 % Smolting year Smolting year Smolts (thousands) Life history model, data up to 2015 River model/expert estimates 5 % 95 % 5 % 95 % Smolts (thousands) Life history model, data up to 2015 Expert estimates 5 % 95 % 5 % 95 % Smolting year Smolting year Figure Development of total production of wild smolts in the AU 1 4 rivers, based on the last full assessment (data up to 2015) and the river model and expert estimates updated with 2016 data (red lines, from Table ).

221 216 ICES WGBAST REPORT Fish counting vs FLH model index Fish counting index FLH index Fish counter index Figure Annual average strength of the spawning run calculated as the relative proportion of ascending spawners in relation to 2009 in seven rivers (Tornio, Simo, Kalix, Pite, Åby, Byske and Ume/Vindel) with fish counting, compared with same index based on corresponding spawner abundance estimates from the full life-history model (with data up to 2015) Fish counting vs FLH model index ( ) 16 Fish counting index y = x R² = FLH index 14 Figure Regression between the average strength of the spawning run index based on fish counting and the corresponding spawner abundance estimates (same data as in Figure ).

222 ICES WGBAST REPORT Smolt production relative to PSPC (%) % level 50% level Barta Irbe Peterupe Saka Salaca Uzava Vitrupe Nemunas Pärnu Figure Wild smolt production level in relation to the potential in AU 5 wild salmon populations. Smolt production relative to PSPC (%) % level 50% level Venta Daugava Gauja Figure Wild smolt production level in relation to the potential in AU 5 mixed salmon populations.

223 218 ICES WGBAST REPORT Smolt production in thousands Figure Total smolt production level in AU 5 wild salmon populations Smolt production in thousands Figure Smolt production level in AU 6 wild salmon populations.

224 ICES WGBAST REPORT % Smolt production relative to PSPC 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Keila Vasalemma Kunda 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Figure Smolt production level in relation to the potential in AU 6 wild salmon populations. Note that the potential is calculated only up to the lowermost migration obstacle and that rivers have substantial rearing habitat areas above migration obstacles.

225 220 ICES WGBAST REPORT 2017 Smolt production in relation to potential (%) % level 50% level Purtse Selja Loobu Pirita Valgeõgi Jägala Vääna Figure Smolt production level in relation to the potential in Estonian AU 6 mixed salmon populations. Note that the potential is calculated only up to the lowermost impassable migration obstacle and that many rivers could have considerably higher total potential. Smolt production in relation to potential (%) % level 50% level Kymijoki Luga Figure Wild smolt production level compared to potential in river Kymijoki (Finland) and in river Luga (Russia).

226 ICES WGBAST REPORT The sahare of adipose fin-clipped salmon Figure Share of adipose finclipped salmon caught on the southern coast of the Gulf of Finland.

227 222 ICES WGBAST REPORT 2017 Figure a. Harvest rates (median values and 90% probability intervals) for wild multi-sea winter salmon in offshore longline fishery within scenarios 1 3 and 4a.

228 ICES WGBAST REPORT Figure b. Harvest rates (median values and 90% probability intervals) for wild multi-sea winter salmon in coastal trapnet fishery within scenarios 1 3 and 4a.

229 224 ICES WGBAST REPORT 2017 Figure Median values and 90% probability intervals for post-smolt survival of wild and reared salmon and M74 survival assumed in all scenarios.

230 ICES WGBAST REPORT Figure Median values and 90% probability intervals for annual proportions maturing per age group for wild and reared salmon in all scenarios.

231 226 ICES WGBAST REPORT 2017 Figure a. Pre-fishery abundances of MSW and 1SW wild salmon and wild and reared salmon combined, based on scenario 1. PFA s reflect the abundance that is available to the fisheries. In case of MSW salmon, natural mortality is taken into account until end of June of the fishing year, and in case of post-smolts until end of August (four months after post-smolt mortality phase). See text for details.

232 ICES WGBAST REPORT Figure b. Pre-fishery abundances of MSW and 1SW wild salmon and wild and reared salmon combined based on scenario 5 (zero fishing). PFA s reflect the abundance that is available to the fisheries. In case of MSW salmon natural mortality is taken into account until end of June of the fishing year, and in case of post-smolts until end of August (four months after post-smolt mortality phase). See text for details.

233 228 ICES WGBAST REPORT 2017 Figure Estimates of reported commercial sea catches (all gears) based on scenarios 1 3 and 4a.

234 ICES WGBAST REPORT Figure a. Probabilities for different stocks to meet an objective of 75% of potential smolt production capacity under scenarios 1 3, 4a and 5. Fishing in 2018 affects mostly years

235 230 ICES WGBAST REPORT 2017 Figure b. Probabilities for different stocks to meet an objective of 75% of potential smolt production capacity under scenarios 1 3, 4a and 5. Fishing in 2018 affects mostly years

236 ICES WGBAST REPORT Figure c. Probabilities for different stocks to meet an objective of 75% of potential smolt production capacity under scenarios 1 3, 4a and 5. Fishing in 2018 affects mostly years

237 232 ICES WGBAST REPORT 2017 Figure d. Probabilities for different stocks to meet an objective of 75% of potential smolt production capacity under scenarios 1 3, 4a and 5. Fishing in 2018 affects mostly years

238 ICES WGBAST REPORT Figure a. Predicted smolt production in 2023 under fishing scenarios 1 3 and zero fishing scenario 5 (thin lines) compared to estimated production in 2016 (bold line). Medians of distributions are illustrated with vertical lines.

239 234 ICES WGBAST REPORT 2017 Figure b. Predicted smolt production in 2023 under fishing scenarios 1 3 and zero fishing scenario 5 (thin lines) compared to estimated production in 2016 (bold line). Medians of distributions are illustrated with vertical lines.

240 ICES WGBAST REPORT Figure a. Median values and 90% probability intervals for smolt and spawner abundances for rivers Tornionjoki, Simojoki, Kalixälven and Råneälven in scenarios 1 (black), 2 (red) and 5 (blue).

241 236 ICES WGBAST REPORT 2017 Figure b. Median values and 90% probability intervals for smolt and spawner abundances for rivers Piteälven, Åbyälven, Byskeälven and Rickleån in scenarios 1 (black), 2 (red) and 5 (blue).

242 ICES WGBAST REPORT Figure c. Median values and 90% probability intervals for smolt and spawner abundances for rivers Sävarån, Vindelälven, Öreälven and Lögdeälven in scenarios 1 (black), 2 (red) and 5 (blue).

243 238 ICES WGBAST REPORT 2017 Figure d. Median values and 90% probability intervals for smolt and spawner abundances for rivers Ljungan, Mörrumsån, Emån and Kågeälven in scenarios 1 (black), 2 (red) and 5 (blue).

244 ICES WGBAST REPORT number of salmon river recreational sea dead discard sea unreported sea misreported sea commercial sea Figure Share of commercial and recreational catches at sea, river catches (including misreporting and also some commercial fishing), and discard/unreporting/misreporting of total sea catches in Subdivisions in years

245 240 ICES WGBAST REPORT 2017 Figure Old and updated PSPC priors for Piteälven (top), Lögdeälven (centre) and Öreälven (bottom). Dotted red line, old PSPC prior (from Stock Annex, ICES, 2014); thick solid black line, new PSPC prior. These priors are rather uninformative (reflecting the uncertainty) and are updated in the full life-history model when stock recruit data are included. Because of methodological problems, the old priors are still used in this year assessment. The updated priors will be used in next year assessment and onwards.

246 ICES WGBAST REPORT Sea trout The assessment of sea trout populations in the Baltic is based on a model developed by the Study Group on Data Requirements and Assessment Needs for Baltic Sea Trout, SGBALANST (ICES, 2011), first implemented at the assessment in 2012 (ICES, 2012). For the evaluation of model results, which are based on electrofishing densities, other basic observations such as i.a. tagging data, spawner counts and catch statistics are also taken into account. 5.1 Nominal catches The highest total (commercial and recreational) catches, above 1300 tonnes, were taken in the early and late 1990s (Table 5.1.1). In the new century they have been decreasing to the level of tonnes in recent years, and even 300 tonnes in 2015 and 341 tonnes in 2016 (but without Danish recreational catches) (Tables and 5.1.3). The total commercial catch of sea trout in the Baltic Sea increased from 192 tonnes in 2015 to 232 tonnes in A majority (79%) of this catch was caught in the Main Basin. The Main Basin catch dropped from 954 tonnes in 2002 to 236 tonnes in After two years of somewhat higher catches, around 450 tonnes, the total commercial catch again fell, reaching a minimum of 145 tonnes in In 2016, the catch again increased a little. The commercial catch in the Gulf of Bothnia was 29 tonnes in 2016, close to the level in 2015, but below the ten year average catch (53 tonnes). In the Gulf of Finland, sea trout catches have dropped to below 20 tonnes in the last three years (Table 5.1.2). About 82% of the commercial catch of sea trout in the Main Basin was taken by Polish vessels. In 2016 they reported 151 tonnes, which is 45% more than in 2015 (104 tonnes) but clearly below the ten year average (239 tonnes). Out of the total Polish commercial catch in 2016, 57% was taken by the coastal and 41% by the offshore fishery. In the Gulf of Bothnia, the 2016 commercial catch was the lowest in thirty years (29 tonnes); it was almost exclusively a coastal catch. The catch of sea trout in the Gulf of Finland increased a little from 2015 to 2016 (from 17 to 19 tonnes). Recreational river catches in 2016 were 25 tonnes, taken mainly in Swedish Gulf of Bothnia rivers. This is a much smaller catch than in 2014, and less than the ten years average (50 tonnes; Table 5.1.3). Most (93%) of the recreational catch in the coastal zones of the Gulf of Bothnia and the Gulf of Finland was taken by Finnish fishermen (71 tonnes). Data on recreational coastal catches from the Main Basin in 2016 only were available from Estonia and Latvia, amounting to 7.3 tonnes (Table 5.1.3). From the last few years there are also results from questionnaires on Danish coastal recreational catches; those catches increased from 224 tonnes in 2011 to 521 tonnes in In 2015, they decreased to 396 tonnes, which still constitutes about 78% of the total Baltic Sea recreational catch of sea trout Biological catch sampling Strategies for biological sampling of sea trout and the procedures are very similar to those for salmon (Section 2.5). In total, 1675 sea trout were sampled in 2016 (Table ). Most samples were collected in the Main Basin (ICES SD 22 28) from German (n=373), Latvian (n=400), Polish (n=102) and Swedish (n=83) catches. In addition, 202 samples were collected from Estonian catches in the Gulf of Finland (SD 32), and 194 from Finnish catches in SD (Table ).

247 242 ICES WGBAST REPORT Data collection and methods Monitoring methods Monitoring of sea trout populations is carried out in all Baltic Sea countries. The intensity and period during which monitoring has been going on varies (ICES, 2008c). Some countries started their monitoring in recent years, while very long data series exist for a few streams in others (ICES, 2008c). From 2016, a new European Union (EU) regulation (2016/1251) adopting a multiannual program for the collection, management and use of data in the fisheries and aquaculture, obligated EU countries to collect sea trout catch data. Most monitoring of sea trout is carried out by surveying densities of trout parr in nursery streams by electrofishing. In Denmark, a rolling scheme is used for electrofishing-monitoring of sea trout on the national level; however, only a few sites in Baltic streams are monitored annually. In a couple of countries sampling of parr densities are used to calculate smolt production by a relation of parr to smolt survival, either developed in the same stream or in some other (ICES, 2008). In most countries (but not in Denmark and Poland) electrofishing is supplemented with annual monitoring of smolt escapement by trapping and counting in one or more streams. In total, smolt production estimates exist for rivers in the entire Baltic area, but the length of the timeseries varies very much. In only three streams/rivers (Åvaån and Vindelälven in Sweden and Pirita in Estonia) both number of spawners and the smolt run is monitored. Adult counts are determined by trapping or inspection of the ascending sea trout using an automatic counter (VAKI). In Lithuania, the spawning run is estimated by test fishing in a couple of rivers and by redd counting. In 17 rivers (ten in Sweden, two in Poland, four in Germany and one in Estonia) the number of spawners is monitored by automatic fish counters or video systems. In three rivers the total run of salmonids is determined using an echosounder. However, this technique does not allow discrimination between sea trout and salmon (or other fish species of similar size). An indication of the spawning intensity can be obtained by counting of redds. Such information is collected from a number of sea trout streams in Poland, Lithuania, Germany and Denmark (ICES, 2008). In a couple of streams in Denmark the catch in sport fisheries has also been used to estimate the development in the spawning run. Catch numbers are also available from some Swedish rivers. Tagging and marking are also used as methods to obtain quantitative and qualitative information on trout populations (see below). Evaluations of status of rivers are done based on national expert opinions, as well as on factors influencing status. Such evaluations are updated irregularly Marking and tagging In 2016, the total number of finclipped sea trout released in the Baltic Sea area was smolts, equal to approximately 31% of all released smolts, and parr (Table ). Finclipping of hatchery reared smolts is mandatory in Sweden and Estonia. The highest number of finclipped smolts was released in Sweden ( ) and Finland ( ). All released sea trout smolts have been finclipped in the Gulf of Finland since 2014, and in the Gulf of Bothnia since Finclipping in 2016 was also performed in Poland (all released smolts in the Łeba River, ind.). A small number of sea trout was finclipped in Germany (Schleswig-Holstein rivers, 980 ind.). There was no stocking of finclipped sea trout smolts in Denmark, Russia, Latvia or Lithuania.

248 ICES WGBAST REPORT A total of sea trout larvae marked using alizarin red S solution were released at the alevin and fry life stages into the Łeba River (Poland). In Finland, similarly marked trout larvae were released, mainly in SD (Table ). In 2016, the total number of Carlin tagged sea trout was Most of the tagged trout were released in SD 26 and (Table ). In addition, sea trout were tagged with T-bar (T-Anch) tags. Out of those, smolts were released to Polish Pomeranian rivers, 330 adults in Germany (SD 22) and 48 kelts in Poland (SD 25). In Finland, 8387 juveniles were tagged with T-bars, mostly in SD 32 (Table ). Additionally 6051 reared and wild juveniles (mainly smolts) were tagged internally with passive integrated transponders (PIT); 2000 in Denmark (SD 22), 980 in Germany (SD 22) and 3071 in Sweden SD 31 (Table ). 5.3 Assessment of recruitment status Methods Recruitment status The SGBALANST (ICES, 2008c; 2009b) screened available data on sea trout populations around the Baltic Sea, and proposed an assessment method for trout (ICES, 2011). The basic method, theory and development is fully described in ICES (2011; 2012), and the slightly adjusted method applied in the assessment in 2012 is briefly summarized below, together with modifications applied in the present assessment. Through screening of data availability (ICES, 2008; 2009; 2011) it was found that only data on abundance of trout from electrofishing were available from all countries. Together with habitat data, trout densities are collected annually from specific sites every year in most countries. However, at the time of the screening, the number of sites was highly variable and mostly sparse in many parts of the Baltic. From a few countries directly useable data were not available, either because there was no electrofishing programme at all, or because the information collected was not sufficiently detailed. It was also found that only little and scattered information existed on other life stages (sea migration, abundance of spawners, smolt production and survival). Also information on human influence, such as sea and river catches (especially recreational ones) of sea trout was sparse. An assessment model using electrofishing data together with habitat information collected at the same sites was proposed focusing on recruitment status as the basic assessment tool (reference point). Recruitment status was defined as the observed recruitment (observed densities) relative to the potential maximal recruitment (maximal densities that could be expected under the given habitat conditions, i.e. the predicted densities, see below) of the individual sea trout populations. Due to the significant climatic (e.g. temperature and precipitation) and geological differences found across the Baltic area, as well as the huge variation in stream sizes, the model proposed is constructed to take variables quantifying such differences into account. Differences in habitat qualities (suitability for trout) influence trout parr abundance; given that stock status is below carrying capacity, and that spawning success is not limited by environmental factors such as migration obstacles downstream to monitored sites. To allow comparison of trout abundances between sites with different habitat quality, a submodel was used, i.e. the Trout Habitat Score (THS). THS is calculated by first assigning values (scores) for the following relevant (and available) habitat parameters

249 244 ICES WGBAST REPORT 2017 for 0+ trout: average/dominating depth, water velocity, dominating substrate, stream wetted width, slope (where available) and shade. Scores assigned ranged between 0 for sites with poor conditions and 2 for best conditions (assessed from suitability curves and in part by expert estimates; see details in ICES, 2011). THS is then calculated by addition of score values resulting in a THS that can vary between 0 (very poor conditions) and 12 (10 if slope is omitted) for sites with very good habitat conditions. Finally, the THS values obtained were further grouped in four Habitat Classes ranging between 0 (poorest) and 3 (best) (Table ) (ICES, 2011). In calculations, observed parr abundance was transformed using Log10 (x+1) to minimize variation and improve fit to a normal distribution. The potential maximum recruitment for sites with a given habitat quality used in this year s assessment was the same as in 2015 (ICES, 2015). Predicted maximum densities were determined by a multiple linear regression analysis based on select sites displaying expected optimal densities (see Section in ICES, 2015). The analysis found the variables log (width), average annual air temperature, latitude, longitude and THS to be significant in determining optimal densities of 0+ trout (r 2 =0.5, Anova; F2,254=51.8, p<0.001) according to the following relation: 1 ) Log10 (0+optimal density) = (0.906*logwidth) + (0.045*airtemp) - (0.037*longitude) + (0.027*latitude) + (THS*0.033). This multiple regression relation (1) was used for calculating the potential maximal densities at the individual fishing occasions, with current Recruitment Status (2) calculated as: 2 ) Recruitment status = (Observed density / Predicted maximal density) * 100 A mean recruitment status was calculated for each assessment unit in the analysis (see below) with a 95%-confidence interval. Recruitment trends Recruitment over time was assessed for the last five year period ( ) in order to illustrate the most recent development in change of status. An indicator of Recruitment Trend was calculated as the bivariate correlation between annual recruitment status and sampling year, illustrated using Pearson s r (resulting in possible values ranging from 1 to +1). Both recruitment status and recruitment trend was calculated as average values for each unit of analysis: Assessment Areas (to raise the number of sites for some areas with poor coverage), ICES Subdivisions (SDs) and, where more countries have streams in one SD, for individual countries.

250 ICES WGBAST REPORT The Assessment Areas were defined according to the table below: Assessment Area SD Gulf of Bothnia (GoB) Gulf of Finland (GoF) 32 Western Baltic Sea (West) 27 & 29 Eastern Baltic Sea (East) 26 & 28 Southern Baltic Sea (South) For the final assessment, the outcome of the above analyses was combined with additional information gathered, most markedly from fisheries and count of spawners (where available) Data availability Information on densities of 0+ trout from 386 fishing occasions in 2015 and 2016, at sites with good or intermediate water quality and without stocking, was available for calculation of recruitment status and trends. The geographical distribution of the fishing occasions is shown in Table (number of sites per SD) and in Figure (map). Note that new, previously not available electrofishing data have been included in the assessment over time, and that some time series for average parr densities (depicted in graphs) show small differences compared to in previous WG reports. In order to also include sites from the Gulf of Finland (SD 32), data from the Finnish river Ingarskilanjoki, where sea trout smolts are stocked, were included (according to local expert opinion spawning contribution from the stocked trout is insignificant). Data were not available from Germany in 2015 or In Russia, data for 2016 were not available, and the same was the case for SD 32 in Finland in In Denmark, information from different sites was comparable to previous assessments. From Finland, data from a number of additional sites in a tributary in the middle reach of river Tornionjoki were made available for this year s assessment. In SD 29, electrofishing in Estonia is biannual; data in the present analysis were collected in 2012, 2014 and For the trend analysis, data from 1038 fishing occasions during the period were available (Table ); however, for some sites not for the complete time period. 5.4 Data presentation Trout in Gulf of Bothnia (SD 30 and 31) Sea trout populations are found in a total of 56 Gulf of Bothnia rivers, of which 28 have wild and 28 have mixed populations (Table ). The status of sea trout populations in Swedish rivers is in general considered to be uncertain (Table ). Populations are affected by human activities influencing freshwater habitats, mostly through overexploitation, damming, dredging, pollution and siltation of rivers (Table ). Average densities for rivers in the area are presented in Figure For Swedish rivers, the densities presented in this figure are from sites in rivers were also salmon are found. These rivers are therefore less suited for sea trout, and they all differ from rivers and sites used for the main trout assessment. In the Swedish sites, densities dropped after 2005, from 8 16 parr per 100 m 2 to 1 3 of 0+ parr per 100 m 2, and they have remained stable at this low level since then, however with a slight increase in 2015 (Figure ). The SD electrofishing results from Finland include three rivers

251 246 ICES WGBAST REPORT 2017 (Torne River with two tributaries, Isojoki and Lestijoki). Densities have remained low in Isojoki and Lestijoki, while they have been variable in the tributaries to Torne River, resulting in an overall average of between 2 6 sea trout 0+ parr per 100 m 2 in recent years, with an increase to more than ten in again followed by a decrease in 2016 (Figure ). The number of sea trout spawners recorded by fish counters in some of the larger rivers in Sweden is in general very low (Figure ). The average number in River Kalixälven increased somewhat after 2006, from some 100 sea trout to , with a further increase in 2014 to over 300 fish. In River Byske, the number decreased after 2005, from approximately 100 sea trout to very low levels around 25 sea trout per year, again increasing in 2015 and 2016 to almost 300 fish. From 2001, river Vindelälven has shown a cyclic population size with between approximately ascending sea trout. However, the number of ascending Vindelälven increased considerably in 2015 to more than 500, but decreased again in 2016 to 250 fish. In river Pite, a positive trend has been obvious with an increasing run from ca. 500 in 2011 to more than 1200 fish in 2016 (Figure ). Catches of wild sea trout in SD have declined considerably over a long time period, indicating very large overall reductions in population size. Although catches from 2013 and forward do not reflect actual runs, because of implemented restrictions (size limits, only sea trout with lengths between 30 and 45 cm can be retained; in R. Torne a complete ban on harvest of sea trout) the catches have declined considerably after the late 1970s and remained low until present time. As an example, Swedish catches in the rivers Torneälven and Kalixälven are presented in Figure However, in 2016 the total catch of wild sea trout from sport fishing in Swedish SD 31 rivers increased, whereas they decreased in SD 30 (Figure ). Returns from Carlin tagged sea trout show a rapid decrease since the 1990s, and after 2003 the average return rate has been below 1% (Figure ). Tagging results from rivers in the Gulf of Bothnia show a large and increasing proportion of the sea trout, often a majority, to be caught already as post-smolts during their first year in sea. Trout are mainly caught as bycatch in the whitefish fishery, by gillnets and fykenets. Based on tagging data, the proportion of fish caught as undersized fish during the first sea year has been fluctuating around 50%, even though the total effort of gillnet fishery by professional fishermen has not changed during the last decade (Figures and ). In Table estimated smolt numbers for the period are presented. In addition, the exact number of counted individuals and details on methods and catchability coefficients of the different traps are given. In river Tornionjoki, smolt trapping during the whole migration period for sea trout has only been possible in some years (the trout smolt run is earlier than for salmon, and in most years the trout smolt run is already ongoing when river conditions allow start smolt-trapping). According to tagging results, the survival of the released smolt is at present lower than the long-term average. Furthermore, tagging data show that Finnish sea trout migrate partly to the Swedish side of the Gulf of Bothnia (ICES, 2009), whereas Swedish sea trout have been caught at the Finnish coast. A Bayesian mark recapture analysis (Whitlock et al., 2017) has recently been conducted for tagging data from reared sea trout in two Finnish rivers in SD 30 and 31 (Isojoki and Lestijoki, ). The results of this study indicate substantial fishing mortality for sea trout aged three years and older from both stocks, but particularly in the case of Isojoki (Figure ). Annual total fishing mortality rate estimates ranged

252 ICES WGBAST REPORT from between 1 to 3 in most years for sea trout aged 3 and older in both rivers, corresponding to harvest rates between 0.63 and Total fishing mortality for the Isojoki stock showed a decreasing pattern over time, while the temporal pattern was fairly stable for Lestijoki sea trout. Fishing mortality was considerably higher for sea trout of ages 3+ compared with fish of age 2 in both stocks (Figure ). A decreasing pattern of survival in the first year at sea was also estimated (results not shown). Sustained high rates of fishing mortality have likely contributed to the poor status and limited reproduction of wild sea-trout stocks in the Isojoki and Lestijoki rivers (Whitlock et al., 2017) Trout in Gulf of Finland (SD 32) The number of streams with sea trout in Gulf of Finland was updated in It is now estimated that there are 101 rivers and brooks with sea trout in this region; out of these 85 have wild stocks (Tables and ). The rest have been supported by releases. Status of populations is uncertain in 30 rivers and very poor in 29 (with current smolt production below 5% of the potential; Table ). In Estonia, sea trout populations are found in 45 rivers and brooks in the Gulf of Finland region, of which 38 have wild populations (Table ). Electrofishing data from Estonian rivers show densities of up to parr per 100 m 2 in the 1980s. In more recent years, densities have in general been below parr per 100 m 2. Average densities from 1992 and onwards are presented in Figure Rivers with higher smolt production are situated in the central part of the North Estonian coast. Smolt runs for trout in River Pirita during the period varied between 100 and 4000, and after a drastic drop in 2014 it attained its record value in 2016 (Table ). Number of spawners recorded in this river by a fish counter was 26 in 2014, 125 in 2015 and 76 in Parr densities for sea trout in the Finnish River Ingarskilanjoki in the Gulf of Finland have been highly variable, with densities varying between 0 and parr per 100 m 2 in the period as shown in Figure (Ingarskilanjoki is the only Finnish river presented in this figure). Also in Gulf of Finland the recapture rate of Carlin tagged sea trout shows a continued decreasing trend for more than 20 years; in recent years it has been zero (Figure ). Finnish tagging results have shown that in general about 5 10% of the tag recoveries are from Estonia and some also from Russia. Likewise, Estonian tagged sea trout have been partly recaptured at the Finnish coast. These migration patterns have recently been confirmed in a genetic mixed stock analysis (Koljonen et al., 2014). In Russia, wild sea trout populations are found in at least 40 rivers and brooks, including the main tributaries (Table ). A majority are situated along the north coast of Gulf of Finland, but rivers with the highest smolt production are found in the southern area of GoF. In more recent years, average 0+ parr densities have in general been below 10 individuals per 100 m 2 (Figure ). The highest Russian 0+ parr densities have been observed in a few small streams and two of the River Luga tributaries. In Lemovzha (tributary of Luga River) densities increased up to 100 individuals of 0+ parr per 100 m 2 in 2014 and 56 individuals in In Solka (the other tributary of Luga River) the numbers of 0+ parr per 100 m 2 have been increased since 2011 up to 56 in 2014, but in 2015 the densities were 0. This lack of 0+ parr was due to reconstruction of the dam situated upstream near the main spawning grounds, which led to the disruption of the spawning migration in autumn There were no density estimates from Russian rivers in 2016.

253 248 ICES WGBAST REPORT 2017 The smolt run in River Luga during the period varied between 2000 and 8000 wild trout smolts (Table ). After an increase to a record level of smolts in 2015, almost three times higher than the average for the total monitoring period (ca smolts), it again decreased to 2600 in Total smolt production in the Russian part of Gulf of Finland has been estimated to about Genetic studies showed that 6 9% of the sea trout caught along the southern Finnish coast was of Russian origin (Koljonen et al., 2014) Trout in Main Basin (SD 22 29) In the Main Basin, when including populations within larger river systems (Odra, Vistula and Nemunas), there are 491 rivers and streams with sea trout populations, and out of these 401 are wild (Tables and ). This does not include Germany, where the actual number of sea trout streams/rivers has not yet been evaluated, although it is estimated that it could be approximately 70. In Sweden, 207 sea trout rivers are found in the entire Main Basin. Out of these 200 have wild sea trout populations whereas seven are supported by releases. In Denmark, 139 out of 173 trout rivers are wild, and majority of them as classified as in good condition. In Poland, the number of populations was revised in 2013; sea trout are found in 25 rivers (twelve in SD 26), mainly in Pomerania (ten) but also in Vistula (six) and Odra (six) river systems (including the main rivers). All Polish sea trout populations are mixed due to stocking for many years. There are three sea trout rivers flowing to the Main Basin in Russia (the Kaliningrad Oblast). All are wild and their status is uncertain. In Lithuania, sea trout are found in 19 rivers, eight belonging to the Nemunas drainage basin. In four Lithuanian rivers there are wild populations while the rest are supported by releases. In Latvia, sea trout populations are found in 28 rivers, about half of them wild. In Estonia, sea trout occurs in 36 rivers and brooks discharging into the Main Basin. All of them are small and have wild populations. The status of sea trout populations in the Main Basin was partially revised in 2014, and it was found to be uncertain in 218 rivers with wild populations. Status of 24 populations (wild and mixed, including tributaries in large systems) are poor with an estimated production <5% (Table ), mainly due to habitat degradation, dam buildings and overexploitation (Tables and ). Trout in eastern Main Basin Average densities in eastern Estonian rivers (SD 29) have in recent years been up to approximately 40 sea trout parr per 100 m 2 and with an up-going trend (Figure ). However, densities tend to vary much between years, partly because of varying water flow. In Latvia, average densities of 0+ parr have varied from four to twelve per 100 m 2 (Figure ). The rivers Salaca, Gauja and Venta have the highest estimated wild smolt production in Latvia. Estimated sea trout smolt production in 2016 for all Latvian rivers combined was about smolts, above the five year average (57 700). In river Salaca estimated smolt numbers from smolt-trapping have varied between in the period (Table ). In 2016 the estimated number was , which is higher than the ten year average (11 500). In Lithuania, average parr densities for 0+ trout have varied from six to ten parr per 100 m 2 during the last few years (Figure ). The estimated total natural smolt production in 2016 increased to individuals, compared to in the year before. Monitoring of salmonids in small streams and rivers is carried out every three years. The estimated overall number of spawners has for a number of years been relatively

254 ICES WGBAST REPORT stable (Kesminas and Kontautas, in Pedersen et al., 2012) varying between 5500 and 8000 individuals. In Poland, the density of 0+ parr in Subdivision 26 rivers (Reda, Zagórska, Struga) has been highly variable, with densities up to more than 90 parr per 100 m 2, and in the last four years the average density has stayed above 70 (Figure ). Note that the very low density observed in 2007 was based on data from one site only. In a newly rebuilt fish-pass at the Wloclawek dam in Vistula River, there has been documented upstream migration of 811 sea trout spawners in 2016, which was half of the run in 2015 (Figure ). Trout in western Main Basin The average densities of 0+ trout parr in Swedish river Emån are very low; since the mid-1990s they have varied between 0.2 and eleven parr per 100 m 2 (Figure ). Catches in Emån (SD 27) are presented in Figure The sport fishing harvest of sea trout in Emån has been declining, and in recent years it has been only been between 20 and 40 landed fish annually. However, since catch and release is not included in these numbers, they do not give a correct picture of the total catch (or run) of sea trout. Densities in the Swedish river Mörrumsån have been below parr per 100 m 2 since the mid-1990s (Figure ). Results from smolt trapping shows that the production in the upper part of Mörrumsån (approximately 15 km from outlet) has varied between 3500 and smolts during the last six years (Table ). Nominal catches (landed) from sport fishing in Mörrumsån (SD 25) are presented in Figure The harvest has varied around 500 sea trout annually for several years. However, the landed catch has declined to 84 sea trout in the last five years (not including catch and release). It is estimated that the number of wild smolt produced in Danish rivers in Subdivision is presently approximately smolt annually. Electrofishing data from Danish streams used to show average parr densities of between 50 and just under 200 parr per 100 m 2 since around year 2000 (Figure ). Annual smolt migration in one stream on the Island of Bornholm (Læså, length 17 km, productive area 2.46 ha) was on average 6300 smolts in ; however, with very high variation among years ( ) due to varying water levels (Table ). This smolt-trapping has not continued after Since 2002 a monitoring programme has been established in Mecklenburg Western Pomerania (Germany) based on electrofishing, to evaluate recruitment and stocking success. From initially about 25 rivers and tributaries in 2002, 17 rivers remain to date in the Mecklenburg Western Pomerania stocking programme. No specific monitoring or assessment of stocking survival was performed in Schleswig-Holstein (the second German state at the Baltic Sea) before in Recently, an electrofishing survey for 0+ and 1+ parr, based on the Trout Habitat Parr Index method (THS) as recommended by the ICES WGBAST, was initiated in Schleswig-Holstein. This approach has now been followed also in the Mecklenburg Western Pomerania parr monitoring since In 2016, nine rivers that flow into the Baltic Sea have been analysed by electrofishing at 37 stations. The parr abundance in 2016 was higher than in 2015 (Figure ). Spawner numbers have been collected by video counting in three German streams with wild populations in SD 22 and 24 (Figure ). In 2011, the counts were incomplete

255 250 ICES WGBAST REPORT 2017 due to flooding events. In the Peezer Bach (SD 24) the number of spawners has remained almost identical over the last three years (650). Hellbach (SD 22) showed the highest count in 2013 with 2300 adult trout, whereas in 2013 the count was In Tarnewitz (SD 22) counts varied between 140 and 380 adults during the period Data from 2015 and 2016 are not available. Average densities of 0+ parr on spawning sites in five Polish rivers in SD 25 in recent years varied between 22 and 114 parr per 100 m 2 (Figure ). The spawning run is monitored by fish counters in the Slupia River; the average number of migrants during the past ten years has been about In 2016, the amount of fish counted was extremely low, below 1500, less than in the year before (Figure ). A dermatological disease of sea trout spawners in most of Polish Pomeranian rivers has continued (outbreak in 2007). In summary, densities of parr in southern rivers in western Baltic were in average similar in all countries and had an increasing trend in 2016, unlike for most of rivers in the rest of the Baltic Sea. The observed numbers of spawners in southern Baltic rivers are much higher than in the large northern rivers, even if some of are very small and have much smaller productive areas. The number of years with observations in some of the rivers is too short to evaluate trends. However, in the rivers with the highest numbers of spawners there was a decrease in the spawning run. 5.5 Recruitment status and trends Results from the updated analyses of recruitment status and trends for sea trout in rivers and streams around the Baltic Sea are shown in Figures to Recruitment status In the Gulf of Bothnia the recruitment status is in general relatively good, although it is higher in SD 31 than in SD 30 (Figures and 5.5.2). This difference mainly reflects a moderate status in Swedish SD 30, from where the number of included sites is much higher than from Finland. In SD 31, where sites from Finland dominate, the average status is relatively good (Figure 5.5.3). In the Gulf of Finland (SD 32) the overall status is good (Figure 5.5.1), but with clear differences between the three countries in the area; the situation is best in Estonia, where status in general is close to optimal, being much better than in both Finland and Russia (Figure 5.5.3). In the eastern part of the Baltic Sea (SD 26 and 28) the overall status is on average low, but at the same time it varies much between SD s and countries (Figures to 5.5.3). The overall low status is due to low trout densities on Lithuanian sites in SD 26, while it is reasonable good in E Poland and Latvia, and reasonably good in Sweden (i.e. at the Island of Gotland, SD 28). In the western part of the Baltic (SD 27 and 29) the overall status is reasonable, being very good in Estonia and reasonable in Sweden. In the southern part of the Baltic Sea (SD 22 25), overall status is reasonable. Status is good or reasonably good in both Sweden and Poland, below average in the westernmost SD 22, but low in SD 24. It should be noted, that status for sea trout populations in Germany (SD 24 25) are still unknown.

256 ICES WGBAST REPORT Recruitment trends The trend in the development of population status is overall good or very good in all Assessment Areas (Figure 5.5.4). However, the overall positive trends in the larger Assessment Areas covers large variations between countries and subdivisions (Figures and 5.5.6). A positive trend is especially evident in the northernmost areas (Gulf of Bothnia) where progress is observed in both Finland and Sweden. In SD 32 (Gulf of Finland) the overall positive trend is a result of progress in Estonia, in contrast to negative trends seen in both Finland and Russia (Figure 5.5.6). In the eastern part of the Baltic Sea (SD 26 and 28) the overall progress covers large differences in the northern part (SD 28), where a strong positive trend can be seen in both Estonia and Latvia, as opposed to a negative trend in Swedish SD 28 and partly also in Lithuania and Poland (SD 26). In the western part of the Baltic (SD 27 and 29) the progress also covers differences between SD 27, where the trend was strongly negative, and SD 29, where progress was observed in both Sweden and Estonia. In the southern part of the Baltic (SD 22 25) the slightly positive trend is due to progress especially in SD 22, but also in SD 23, while there was no development in SD 24 and 25 (Figure 5.5.5). 5.6 Reared smolt production Total number of reared sea trout smolts released 2016 in the Baltic Sea (SD 22 32) was , which is more than last year ( ) and above the ten years average ( ). Out of this total, smolts were released into the Main Basin, into the Gulf of Bothnia and into the Gulf of Finland (Table 5.6.1). In Finland, trout smolt production is mainly based on reared brood stocks supplemented by spawners caught in rivers. In the past ten years the average number of smolts released has been In 2016, the number of smolts was , whereof 57% were stocked into the Gulf of Bothnia and 27% into the Gulf of Finland. In Sweden, the number of trout smolts stocked in 2016 was , close to the average level of last few years. A majority of the Swedish smolts were released into Gulf of Bothnia (69%). Estonia released 3000 smolts into the Gulf of Finland in In Poland, juvenile fish are reared from spawners caught in each river separately; only a part of the Vistula stocking is of reared brood stock origin. A total of smolts were released into Polish rivers in 2016, little less than the ten years average of Denmark released in 2016, substantially more than in 2015, and then a few years average. Latvia released smolts in 2016, much more than the ten years average ( ). Russia released smolts in 2016 into the Gulf of Finland, which was double previous year s release. The German level of stocking has been smolts per year since In addition to direct smolt releases, trout are also released as eggs, alevins, fry and parr (Table 5.6.2). The estimated number of smolts originating from releases of younger life

257 252 ICES WGBAST REPORT 2017 stages (i.e smolt equivalents) over time is presented in Table In 2016, the estimated smolt number expected from these releases was around , mainly in Main Basin rivers. The prediction for 2017 is approximately smolts for the whole Baltic, of which will migrate into the Main Basin (Table 5.6.3). Total number of smolt equivalents from enhancement releases in recent years has been lower than in the very beginning of the 20th century (Table 5.6.3). 5.7 Recent management changes and additional information Management changes In the Bothnian Bay (SD 31), Bothnian Sea (SD 30) and Gulf of Finland (SD 32) fishing in the sea is mainly directed towards other fish species, but using gears that catches also young age groups of sea trout (e.g. gillnets aimed for whitefish). To improve the situation for the poor sea trout stocks in SD 31, a number of changes were implemented in the Bothnian Bay from July 1st 2006 in both Sweden and Finland. The minimum landing size for sea trout was raised from 40 to 50 cm in the sea. In the Finnish part of the Gulf of Bothnia (SD 29 31) the minimum size of all sea trout has been raised further to 60 cm. From 2019, only released sea trout (finclipped) can be harvested. In the Finnish economic zone in the Gulf of Finland all reared sea trout must be adipose finclipped, and all sea trout with an intact adipose fin must be released back to sea. Minimum landing size in the Gulf of Finland (for finclipped sea trout) was in 2014 increased to 60 cm, and to 65 cm in village owned waters (but not in private waters). Minimum bar length in gillnets that target sea trout is 80 mm. In all bottom gillnets with less than 80 mm bar length only monofile nets are allowed, and the diameter of the fibre must not exceed 0.20 mm. In the river Tornionjoki harvest of sea trout is completely banned since 2013 (see below). In Sweden a ban of fishing with gillnets at a water depth of less than 3 meters during the periods 1 April 10 June and 1 October 31 December has been enforced, in order to decrease the bycatch of trout in other fisheries. In the period 1 31 October, only gillnets with a mesh size of less than 37 mm (knot to knot) is allowed. New restrictions for the rivers in Bothnian Bay (SD 31) were implemented in 2013, to further strengthen the protection of sea trout. This included shortening of the autumn period for fishing with two weeks, resulting in a fishing ban from 1 September to 31 December (in some rivers also between 1 May 18 June), and restrictions of catch sizes to a minimum 50 cm in sea areas in SD 30 and a slot limit between cm in some rivers (30 40 cm in Subdivision 31). The size restrictions differ between rivers. The new regulation also includes a bag limit of one trout per fisherman and day. As a part of the bilateral agreement between Sweden and Finland on fishing in the River Torne (Border River and area outside river mouth), a total ban on landing trout was decided and implemented in spring From 2013 the Swedish offshore fishery targeting salmon and sea trout has been phased out. In Russia, catch of sea trout is banned entirely. In Estonia, new restrictions were established in 2011: 1) the closed area for fishing around the river mouth was extended from 1000 to 1500 m for the time period 1 September 31 October for rivers Kunda, Selja, Loobu, Valgejõe, Pirita, Keila, Vääna, Vasalemma and Purtse; and 2) in rivers Selja, Valgejõgi, Pirita, Vääna and Purtse recreational fishery for salmon and sea trout is banned from 15 October 15-November.

258 ICES WGBAST REPORT In Germany the management varies between the federal states Mecklenburg Western- Pomerania and Schleswig-Holstein, and between the commercial and the recreational fishery: the minimum landing size for trout is 40 or 45 cm and a closed season starts between 1 September and 2 October and ends between 14 December and 31 March. In Denmark, the minimum legal size is 40 cm. In the sea it is prohibited to set gillnets m from the coastline. In the sea, coloured sea trout are protected from 16 November 15 January, and in freshwater all sea trout are protected during the same period. All fishing is prohibited in closed areas (500 m radius) around river mouths all year in streams wider than 2 m at the outlet, and from 16 September 1 March (in some places 16 September 15 January) in smaller streams. A map displaying closed areas is available on the Internet (also for smartphones). In addition to these general rules, special closed areas have been established. Around the island of Bornholm, rules for sea trout fishing was tightened from 1 January A bag limit of three sea trout per angler or per boat was established. Harvest of maturing sea trout is prohibited 15 September 28/29 February. The use of nets by recreational fishermen was also restricted, both in space (sea depth), type of gear and mesh-size Additional information In recent years predators on sea trout such as cormorants (Phalacrocorax carbo) have increased dramatically in the Baltic area. Studies have shown that cormorants can have severe effects on fish stocks (Bzoma, 2004; Leopold et al., 1998; Koed et al., 2006; Skov et al., 2014). Where large cormorant colonies occur in the vicinity of important salmonid rivers, there are good reasons to investigate whether cormorants have a significant negative impact on the stock. In Poland research focused on investigation of genetic polymorphism in order to support fisheries resources management and biodiversity protection has been taken in the NMFRI in the 1990s. Differentiation between and within sea trout populations inhabiting Polish rivers was analysed with various methods, first with RFLP analysis of mtdna and later using more selective methods like nuclear microsatellite and SNP analyses. Recent results have shown that the genetic composition of sea trout sampled in the Vistula river system, resembled the non-admixed hatchery population that is composed of descendants of the winter-run stock with origin from the Vistula River. Trout parr otolith core strontium/calcium (Sr/Ca) ratios have been used to determine whether parr has an anadromous or resident maternal parent. Trout parr were sampled from 21 Estonian and three Finnish short, coastal streams. In Estonia, 92% of the analysed parr were of anadromous and 8% of resident maternal origin. In Finland, the respective proportions were 79% and 21%. It could be concluded that easily accessible spawning areas, situated up to 30 km from the sea, largely contained progeny of sea trout. This study demonstrated that the Sr/Ca-method provides an effective means to distinguish between the progeny of sea trout and resident brown trout (Rohtla et al., 2017). However, if an anadromous female enters into freshwater several months prior to spawning, a dilution effect (i.e. marine Sr:Ca ratio is diluted with low freshwater Sr:Ca ratio in the developing eggs) can lower the Sr:Ca core otolith values, and the anadromic mark among the offspring may become unnoticeable. High freshwater Sr:Ca values may also make the initial core Sr:Ca value undistinguishable between marine and freshwater phases. Therefore additional studies are needed to clarify regional characteristics.

259 254 ICES WGBAST REPORT Assessment result In general, a positive development is observed in sea trout populations around most of the Baltic Sea. In spite of this general result, populations in some areas are still considered fragile, and many uncertainties remain. In the Gulf of Bothnia, counts of spawners shows a continued increase, however, with high variation among rivers. Also, the absolute spawner numbers are still very low considering the size of the rivers. Parr densities have also improved, as has the recruitment status, reflected in the positive trends in development. The restrictions in the Swedish sea fishery, that have now been in effect for a number of years, and more recent restrictions in the Finnish sea fishery, are likely to be the main reason for this positive development. Also the ban on catch of sea trout in the River Torne is believed to have contributed positively to the development in this river. The complete ban of harvest of wild sea trout in Finnish waters (that will come into effect 2019) is expected to contribute further to the positive development. However, the continued fishery for other species (e.g. whitefish) with fine meshed nets that also catch post-smolt and young sea trout is still problematic, and expected to either limit the level of wild sea trout populations in the area, or at least to delay their recovery. For this reason, trout in SD 30 and 31 are still considered to be vulnerable. It is recommended to further reduce the fishing mortality in the fishery targeting other species, and to maintain the present restrictions. In the Gulf of Finland, a positive development has been observed in Estonia, where trout populations in general seem to be in a good shape. In contrast, the trout populations in Russia and Finland remain relatively weak. Recent catch restrictions for wild sea trout in Finland are expected to improve sea survival of trout from all countries in the area. It is recommended to continue with the present management restrictions in both Finland and Estonia. In Finland catch restrictions of wild sea trout in private waters should also be established. In Russia illegal catch of sea trout may be the reason the continued poor status for the populations in this area. In the Western Main Basin (SD 27 and western part 29), the trout populations in general seems to be in a good or reasonably good state. In spite of the negative trend seen for the development in SD 27 in recent years, there is at the moment no reason for concern. In the Eastern Main Basin (SD 26, 28 and eastern part of 29) populations are in general in good shape. However, in Lithuania the recruitment status has for the last few years been low. The main reason is believed to be low water flows during the spawning period in recent years. It is still uncertain if this is the only reason for the low status. In the Southern Baltic Sea (SD 22, 23, 24 and 25) the status is not optimal in the Danish populations included in the analysis (SD 22 and especially 24). However, this is not believed to represent the general situation for Danish sea trout populations, and does not raise concern at the moment. The remaining populations seem to be in a reasonable or good state. The present situation in Germany is unknown, and needs to be evaluated Future development of model and data improvement While the assessment model employed so far has been based on habitat characteristics taking climatic and habitat variables into account, additional information has now been collected. This includes (a) information on the distance from the sites used for assess-

260 ICES WGBAST REPORT ment to the sea, possibly reflecting increased mortality during smolt downstream migration and upstream migration of adults, and, (b) the number of migration obstacles between sites and the sea. The latter information also includes the type of barrier(s) and the expect impact on both downstream and upstream sea trout migration (based in part on expert opinion). The possible inclusion of the two above variables (a, b) is planned to be evaluated in coming years. By using data already provided combined with new additional data from streams in SD31, it also needs to be evaluated how well the current assessment model, that is based primarily on information from more southern areas, describes status in the northernmost part of the Baltic Sea. 5.9 Compatibility of the EU-MAP with the data needs for WGBAST A better geographical coverage than hitherto is needed, with a sufficient number of fishing sites from typical trout streams. This is especially relevant in the northern parts of the Baltic, where most present electrofishing sites are situated in larger rivers, fished primarily for monitoring of salmon populations. This is relevant both for the actual assessment as such, and in order to evaluate how well the application of the assessment model employed work in these areas. Also in the southwest Baltic (Denmark) there is currently a lack of sites being collected annually. The concept of the current assessment model builds on a comparison of observed densities with estimated maximum densities at sites with good conditions, no migration obstacles and no or low impact from fishing. This is to a large extent depending on expert judgment. If an array of trout index rivers, with counts of both smolt and returning adults, was established it would, from recruitment at MSY (i.e. with spawning targets reached) be possible to express both the recruitment as a proportion of the maximum density and include spawning targets in this approach Recommendations The method used by Rohtla et al. (2017) for determination of parr origin on the basis of Sr:Ca ratio should be applied on a broader scale to enable the distinction between resident and migrant forms of trout, and their respective contributions to parr production. Sufficient data coverage of sea trout parr densities from typical trout streams should be collected in all countries. Continued (annual) sampling from these sites for longer time periods is required. Sea trout index rivers should be established to fulfil assessment requirements with respect to geographical coverage and data collection needs. Data on recreational sea trout catches should be consistently collected, taking into account the potentially high impact of recreational fisheries on sea trout stocks and the lack of these data in several countries.

261 256 ICES WGBAST REPORT 2017 Table Nominal catches (commercial + recreational; in tonnes round fresh weight) of sea trout in the Baltic Sea in years Commercial catches after 2000 are presented in T and recreacional catches after 2000 in T Year Main Basin Total Gulf of Bothnia Total Gulf of Finland Total Grand Denmark 1,4 Estonia Finland 2 Germany 4 Latvia Lithuania Poland Sweden 4 Main Finland 2 Sweden Gulf of Estonia Finland 2 Gulf of Total S + C C S S + C R C S + C R C R S 9 S + C R S 6 C 6 R Basin S C R S 6 C 6 R Bothnia C S C R Finland na 10 na na na na na na na na na na 6 na na 11 na na na na na na na na na na 87 na na 51 na 5 na na na na na na na na na na na na na na na na na na 50 na 14 na na na na na na na na na 66 na 9 na na na na na na na na na 62 na 9 na na na na na na na na na 53 na 8 na na na 1 24 na na 66 na 2 na na na 1 26 na 150 na na 99 na 8 na na na na na na na na , na na na na na , na na na na , na 6 na na na na , na na na na , , na na na , , na na na , na na , na na na , , , , ,339 1 Additional sea trout catches are included in the salmon statistics for Denmark until 1982 (table 3.1.2). 2 Finnish catches include about 70 % non-commercial catches in , 50 % in , 75% in Rainbow trout included. 4 Sea trout are also caught in the Western Baltic in Sub-divisions 22 and 23 by Denmark, Germany and Sweden. 5 Preliminary data. 6 Catches reported by licensed fishermen and from 1985 also catches in trapnets used by nonlicensed fishermen. 7 Finnish catches include about 85 % non-commercial catches in ICES Sub-div. 22 and Catches in included sea and coastal catches,since 1998 costal (C) and sea (S) catches are registered separately na=data not available + Catch less than 1 tonne.

262 ICES WGBAST REPORT Table Biological sea trout samples collected in NUMBER OF SAMPLED FISH BY SUBDIVISION COUNTRY MONTHS (NUMBER) FISHERIES GEAR TOTAL Estonia 1 12 Coastal Gillnet Finland 4 9 Coastal All gears Latvia 3 11 Coastal, River Gillnet, trapnet Poland 1 12 Offshore, Coastal Trawl, Longline, Gillnet/fyke net Sweden 6 9 Coastal All gears Sweden 6 9 River All gears Germany 1 12 Coastal Rod&Line, Gillnet Total 1675

263 258 ICES WGBAST REPORT 2017 Table Nominal commercial catches (in tonnes round fresh weight) of sea trout in the Baltic Sea ( ). S=Sea, C=Coast and R=River. Main Basin Total Gulf of Bothnia Total Gulf of Finland Total Grand Year Denmark Estonia Finland Germany Latvia Lithuania Poland Sweden Main Finland Sweden Gulf of Estonia Finland Russia Gulf of Total S C S C S C S S C R S C R S C R S C R Basin S C C R Bothnia C S C R Finland

264 ICES WGBAST REPORT Table Nominal recreational catches (in tonnes round fresh weight) of sea trout in the Baltic Sea ( ). S=Sea, C=Coast and R=River. Main Basin Total Gulf of Bothnia Total Gulf of Finland Total Whole of the Baltic Grand Denmark Estonia Finland Latvia Poland Sweden Main Finland Sweden Gulf of Estonia Finland Gulf of Finland Total Year C C R C R R R Basin R C R Bothnia C R Finland C 2001 n.a n.a n.a n.a n.a n.a n.a n.a n.a n.a n.a n.a n.a n.a

265 260 ICES WGBAST REPORT 2017 Table Adipose finclipped and tagged sea trout released in the Baltic Sea area in Country Sub- River Age Number Tagging Other Methods division fry parr smolt Carlin T-Anch PIT ARS (1)* n.n Denmark 22 Gudsoe Mollebak Denmark 22 Egåen Germany 22 Lipping Au 1 and Germany 22 Trave spawners 330 Poland 25 Łeba 188,200 Poland 25 Łeba 100,600 Poland 25 Łeba spawners 48 Poland 25 Łeba 1 62,230 Sweden 25 Bräkneån 1 1,100 Sweden 25 Mörrumsån 1 4,000 Sweden 25 Mörrumsån wild smolts 289 Poland 25 Wieprza 1 2,000 Poland 25 Rega 1 4,000 Poland 25 Łeba 1 4,000 Poland 26 Vistula 2 10,000 Sweden 27 Åvaån 1 1, Sweden 27 Åvaån 1 147,500 Sweden 27 Åvaån 2 7,000 Sweden 27 Åvaån 2 9,825 Sweden 27 Motala Ström 1 16,000 Latvia 28 Venta 1 1,000 Finland 29 at sea 2 40,052 Finland 29 at sea 3s 1,186 Finland 30 at sea 2 7, Finland 30 Lapväärtinjoki 1 6, Finland 30 Lapväärtinjoki 2 15,617 1,998 2,091 Finland 30 Karvianjoki 1 17,456 Finland 30 Karvianjoki 2 4,794 Finland 30 Kokemäenjoki 1 2,085 2,085 Finland 30 Kokemäenjoki 2 33,600 33,600 Sweden 30 Skellefteälven 1 7,000 Sweden 30 Ångermanälven 2 39, Sweden 30 Ångermanälven 2 19, Sweden 30 Indalsälven 1 108,963 Sweden 30 Ljungan 2 10,000 1,000 Sweden 30 Ljusnan 1 15,478 13,918 Sweden 30 Ljusnan 2 45,090 Sweden 30 Dalälven 2 1,200 Sweden 30 Dalälven 1 30,191 Sweden 30 Dalälven 2 54,773 1,500 Sweden 30 Testeboån Wild smolts 567 Finland 31 Lestijoki 1 22,544 8,500 Finland 31 Lestijoki 2 9,386 1,898 Finland 31 Perhonjoki 1 7,060 Finland 31 Perhonjoki 2 10, ,056 Finland 31 Siikajoki 1 12,400 Finland 31 Oulujoki 2 10,335 1,997 10,335 Finland 31 Kiiminkijoki 1 20,000 20,000 Finland 31 Iijoki 2 66,410 1,000 66,410 Finland 31 Kuivajoki Finland 31 Simojoki 1 1,200 Finland 31 Kemijoki 2 84, ,985 Finland 31 Kemijoki Finland 31 Tornionjoki 607 Finland 31 Tornionjoki 1 5,204 5,204 Finland 31 Tornionjoki 2 4,233 4,233 Sweden 31 Luleälven 1 15,487 Sweden 31 Luleälven 2 51,116 2,000 Sweden 31 Skellefteälven 1 27,421 Sweden 31 Umeälven 1 21,087 Sweden 31 Umeälven 2 7,923 Sweden 31 Lögdeälven Wild smolts 94 Sweden 31 Rickleån Wild smolts 121 Estonia 32 Pudisoo 2 3, Finland 32 at sea 2 113,630 1, ,130 Finland 32 Vehkajoki 1 3,252 3,252 Finland 32 Summanjoki 2 1,763 1,763 Finland 32 Kymijoki 2 70,443 1,986 70,443 Finland 32 Koskenkylänjoki 1 5,815 5,815 Finland 32 Vantaanjoki 6,750 6,750 Finland 32 Vantaanjoki 1 10,047 10,047 Finland 32 Vantaanjoki Finland 32 Fiskarsinjoki 12,538 12,538 Finland 32 Karjaanjoki Finland 32 Ingarskilajoki 2 7, Total sea trout 19, ,203 1,147,236 20,487 18,765 6, ,674 - (1) ARS = Alizarin Red Staining Table Number of Carlin-tagged sea trout released into the Baltic Sea in Country Total Latvia 1, Finland 1,998 3, Sweden 386 3, Poland 10, Total , , ,498 3, ,487

266 ICES WGBAST REPORT Table Habitat classes and corresponding Trout Habitat Scores (THS). See text for details. Habitat Class THS including slope THS omitting Slope 0 <6 < Table Number of fishing occasions/sites (in ) available for assessment of trout recruitment status, distributed in ICES subdivisions (SD). Number fishing occasions ICES SD Recruitment Trend Total

267 262 ICES WGBAST REPORT 2017 Table Status of wild and mixed sea trout populations. Partial update in Potential Smolt production (% of potential production) Area Country smolt <5 % 5-50 % > 50 % Uncertain Total production wild mixed wild mixed wild mixed wild mixed wild mixed Gulf Finland < of Bothnia * > Uncertain 0 0 Total Sweden** < > Uncertain Total Total Gulf Estonia < of Finland > Uncertain 0 0 Total Finland*** < > Uncertain 0 0 Total Russia < * > Uncertain Total Total

268 ICES WGBAST REPORT Table Continued. Main BasinDenmark < > Uncertain 0 0 Total Estonia < > Uncertain 0 0 Total Latvia < > Uncertain Total Lithuania < > 100* 0 0 Uncertain 0 0 Total Poland < > Uncertain 0 0 Total Russia < > Uncertain Total Sweden** < > Uncertain Total Total Grand total * includes data from large river systems ** data from 2006 *** in 7 wild rivers it is not known if releases are carried out

269 264 ICES WGBAST REPORT 2017 Table Factors influencing status of sea trout populations. Partial update in Country Potential Number of populations Area smolt Over Habitat Dam Pollution Other Uncertain production exploitation degradation building Gulf of Finland < Bothnia* > Uncertain Total Total Gulf of Finland < Finland > Uncertain Total Russia < > Uncertain Total Estonia < > Uncertain Total Total Main Estonia < Basin* > Uncertain Total Latvia < > Uncertain Total Lithuania < > 100** Uncertain Total Poland < ** > Uncertain Total Russia < > Uncertain Total Denmark < > Uncertain Total Total Grand total * data from Sweden were unavaila** includes large river systems, see Table

270 ICES WGBAST REPORT Table Sea trout smolt estimates for the period SD Country DK SE LV LV LT LT LT LT SE SE FIN RU RU EE EE River name Læså Mörrum Salaca Salaca R. Mera R. Mera R. Siesartis R. Siesartis Sävarån Rickleån Tornionjoki Luga Luga Pirita Pirita Method 1) 2) 3) 4) 5) 6) 5) 6) 7) 8) 9) 10) 11) 12) 13) n.d. 243 n.d n.d n.d. trap moved 348 n.d n.d * 23* n.d. trap moved n.d. n.d n.d n.d. trap moved n.d. = no data 1) based on smoltrap - directly counted number of smolts, varying efficiency over years due to water level (probability level data available) 2) Median values of Bayesian estimates are only for the upper part of the river! 3) estimated smolt output on the base of counted smolts and mean trap efficiency (2014=8.5%; 2015=5.9%; 2016=9.5) 4) directly counted number of smolts during trapping season 5) estimated output derived by electrofishing data. (assumed surval probabilities to smolts: 0+ --> 40%; >0+ --> 60%) 6) counted number of individuals smolts in trap. Assumed trap efficiency almost 100% 7) simple Peterson estimates - trap moved to river Rickleån in Year ) Trap located close to river mouth, so this is the total estimated production 9) estimated smolt output. Trap efficiency in 2016 from efficiency for salmon smolt 10) estimated number of smolt output based on results of floating trap-netting- 2.9 % in 2016, due to high water only part of migration period covered 11) directly counted number of smolts in trap 12) figures for the period recalculated 13) counted number of individuals smolts in trap-experimentally derived catch efficiency in 2014 = 20% *) due to high waterlevel counts individual numbers presumably too low

271 266 ICES WGBAST REPORT 2017 Table Status of wild and mixed sea trout populations in large river systems. Country River Potential Smolt production (% of potential production) (Area) smolt <5 % 5-50 % > 50 % Uncertain Total production wild mixed wild mixed wild mixed wild mixed wild mixed Lithuania Nemunas < (Main Basin) > Uncertain 0 0 Total Poland Odra < (Main Basin) > Uncertain 0 0 Total Poland Vistula < (Main Basin) > Uncertain 0 0 Total Russia Luga < (Gulf of Finland) > Uncertain Total Finland Tornion- < joki (Gulf of Bothnia) > Uncertain 0 0 Total

272 ICES WGBAST REPORT Table Factors influencing status of sea trout populations in large river systems. Partial update in Country River Potential Number of populations smolt production Overexploitation Habitat degradation Dam building Pollution Other No influence Lithuania Nemunas < (Main Basin) > Uncertain Total Poland Odra < 1 0 (Main Basin) > Uncertain 0 Total Poland Vistula < 1 0 (Main Basin) > Uncertain 0 Total Russia Luga < (Gulf of Finland) > Uncertain Total Finland Tornion- < joki (Gulf of Bothnia) > Uncertain Total

273 268 ICES WGBAST REPORT 2017 Table Sea trout smolt releases (x1000) into the Baltic Sea by country and subdivision in Note that project based fisheries enhancement releases included. yea r country age M ain DE 1yr Basin 22-2yr DK 1yr yr EE 1yr yr FI 1yr yr yr LT 1yr yr 3 1 LV 1yr yr PL 1yr yr SE 1yr yr Main Basin Total Gulf of FI 1yr Bothnia 30-2yr yr SE 1yr yr Gulf of Bothnia Total Gulf of Finland 32 EE 2yr FI 1yr yr yr RU 1yr yr Gulf of Finland Total Grand Total

274 ICES WGBAST REPORT Table Releases of sea trout eggs, alevins, fry and parr into Baltic rivers in The expected number of smolts is added to Table as "enhancement". Region Egg Alevin Fry Parr Expected number of smolt 1- s old 1- y old 2- s old 3-s old Total Sub-divs (1) (1) (4) (6) (9) (10) (10) Denmark - - 3,639 6, Estonia Finland 4, ,400-18, ,500-15,375 2,674-18,049 Germany , ,550-29,550 Latvia , ,828-6,828 Poland - 4,673,700 2,691, , ,479 Sweden ,300 2, ,379-2,379 Lituania , ,840-3,840 Total 4,000 4,826,100 3,882, , ,500-15, , ,643 Sub-divs (2) (3) (5) (7) (8) (8) (10) Finland - 194,800-34, ,304 24, ,356 4,980 21,336 Sweden 206, ,300 42,800 49,600 39, ,644 9,107 19,751 Total 206, , ,300 77, ,904 63, ,000 14,087 41,087 Sub-div. 32 (1) (1) (4) (6) (9) (10) (10) - Estonia Finland 71,400 4, , , ,650 Russia Total 71,400 4, , , ,650 Grand total Sub-divs ,000 5,024,900 4,157, , , ,600-17, ,023 14, ,381

275 270 ICES WGBAST REPORT 2017 Table Estimated number of sea trout smolts originating from eggs, alevins, fry and parr releases in Sub-divs Denmark 30,858 25,555 45,759 7,912 17,790 17,508 13,695 13,695 13,704 12,540 12,540 10,737 9,177 9,606 9,240 9,246 9, Estonia - - 2,100 1, , Finland ,670 33,965 19,550 18, ,445 13,815 10,350 8,100 14,375 16,260 17,787 14,349 18,313 2,674 - Germany 25,500 24,900 61,200 72,240 27,240 36,900 32,550 38,400 29,640 29,910 40,800 34,500 29,400 34,650 32,700 32,580 31,860 35,874 29,550 - Latvia 13,815 8,644 11, ,340 15,227 6,462 3,189 19,015 6,840 17,664 30,595 5,987 15,300 28,913 7,787 11,621 6,000 6,828 - Poland 167, ,500 84,240 68,400 91,000 63,236 77,690 61, ,686 84, , ,982 95, , , , , , ,479 - Sweden 13,129 39,333 42,690 5,320 29,335 2,055 27,700 4,425 1,623 2, ,385 1,737 2,940 3,258 1,368 1,380 2,379 - Lituania ,670 2,400 4,350 7,440 18,180 12,990 8,040 6,750 5,370 10,935 8,580 6,300 4,560 4,680 3,840 - Total 251, , , , , , , , , , , , , , , , , , ,268 - Sub-divs Finland 54,268 80,662 26,523 42,828 36,670 1,890 31,362 11,787 22,704 29,892 32,550 46,753 39,285 25,881 22,595 18,782 12,878 12,879 21,328 4,980 Sweden 84,237 78,440 43,614 24,092 22,921 36,170 20,207 22,756 24,561 16,690 16,497 12,811 13,026 5,456 21,906 9,073 25,850 12,996 17,203 9,107 Total 138, ,102 70,137 66,920 59,591 38,060 51,569 34,543 47,265 46,582 49,047 59,564 52,311 31,337 44,501 27,855 38,728 25,875 38,531 14,087 Sub-div Estonia ,412 2,532 4,407 2, ,536 2,098 6,552 9,486 3, , Finland 20,910 5,500 2, , ,574 8,997 4,353 5,919 5, ,747 1,632 1,050 7,716 2, Russia 3,882 3,630 7, ,630 1,281 6,690 3, , ,441 1, , Total 24,792 9,130 9,849 3,031 4,502 9,117 9,135 15,918 8,997 4,665 16,836 7,457 10,284 12,979 5,154 4,800 8,736 3, Grand total Sub-divs , , , , , , , , , , , , , , , , , , ,554 14,087

276 ICES WGBAST REPORT Figure Electrofishing sites in Subdivisions used for assessment of sea trout recruitment status (2017).