Indicators for ecological status of angiosperms and charophytes in Danish coastal waters

Similar documents
Appendix A. Depth limits of selected species

National monitoring of beach litter in Denmark 2018 Amounts and composition of marine litter on reference beaches

Justification for the Swedish method for BQE Macroalgae and angiosperms

Aquatic Plant Point-Intercept Survey for Cates Lake, Scott County, Minnesota

MaxDepth Aquatics, Inc.

SECTION 2 HYDROLOGY AND FLOW REGIMES

Legendre et al Appendices and Supplements, p. 1

Monitoring Cruise Report

Chateaugay Lakes Milfoil Control Program

Annex 7: Request on management areas for sandeel in the North Sea

Jason Blackburn, Paul Hvenegaard, Dave Jackson, Tyler Johns, Chad Judd, Scott Seward and Juanna Thompson

Municipal waste management in Cyprus

Florida Seagrass Integrated Mapping and Monitoring Program

Aquatic Plant Point-Intercept Survey for Pike Lake, Scott County, Minnesota

Two types of physical and biological standards are used to judge the performance of the Wheeler North Reef 1) Absolute standards are measured against

Advice June 2014

Summary of Research within Lamlash Bay No-Take Zone - Science report for COAST July

Improving Accuracy of Frequency Estimation of Major Vapor Cloud Explosions for Evaluating Control Room Location through Quantitative Risk Assessment

Nearshore Habitat Mapping in Puget Sound Using Side Scan Sonar and Underwater Video

Exemplar for Internal Assessment Resource Geography Level 3. Resource title: The Coastal Environment Kaikoura

Monitoring Cruise Report

Bottom habitat mapping in Plateliai lake applying remote underwater video techniques

Analysis of the Article Entitled: Improved Cube Handling in Races: Insights with Isight

Refined Designated Uses for the Chesapeake Bay and Tidal Tributaries

Analysis of performance at the 2007 Cricket World Cup

EU request on management areas for sandeel in the North Sea

Characterising the status of the Western Port recreational fishery in relation to biodiversity values: Phase 1 Greg Jenkins and Simon Conron

2016 ANNUAL FISH TRAWL SURVEY REPORT

THE EFFECT OF VARIOUS DIKE REINFORCEMENT ALTERNATIVES ON THE FLOOD SAFETY ASPECTS OF THE DELTADIKE CONCEPT

CHAPTER 8 ASSESSMENT OF COASTAL VULNERABILITY INDEX

PUBLISHED PROJECT REPORT PPR850. Optimisation of water flow depth for SCRIM. S Brittain, P Sanders and H Viner

3.4.3 Advice June Barents Sea and Norwegian Sea Cod in Subareas I and II (Norwegian coastal waters cod)

Wake effects at Horns Rev and their influence on energy production. Kraftværksvej 53 Frederiksborgvej 399. Ph.: Ph.

RESOLUTION MEPC.136(53) Adopted on 22 July 2005 DESIGNATION OF THE BALTIC SEA AREA AS A PARTICULARLY SENSITIVE SEA AREA

INFLUENCE OF ENVIRONMENTAL PARAMETERS ON FISHERY

Eelgrass and Macroalgae Presence/Absence Preliminary Surveys BHP Proposed Grays Harbor Potash Export Facility A Task 400

STOCK STATUS OF SOUTHERN BLUEFIN TUNA

IMPROVING POPULATION MANAGEMENT AND HARVEST QUOTAS OF MOOSE IN RUSSIA

THE WAVE CLIMATE IN THE BELGIAN COASTAL ZONE

Upwelling and Phytoplankton Productivity

ASSESSMENT OF THE WEST COAST OF NEWFOUNDLAND (DIVISION 4R) HERRING STOCKS IN 2011

Assessment Summary Report Gulf of Mexico Red Snapper SEDAR 7

TEACHER VERSION: Suggested Student Responses Included. Upwelling and Phytoplankton Productivity

Competitive Performance of Elite Olympic-Distance Triathletes: Reliability and Smallest Worthwhile Enhancement

A Combined Recruitment Index for Demersal Juvenile Cod in NAFO Divisions 3K and 3L

PHYSICAL AND NUMERICAL MODELLING OF WAVE FIELD IN FRONT OF THE CONTAINER TERMINAL PEAR - PORT OF RIJEKA (ADRIATIC SEA)

Country report. Austrian bathing water quality in Austria. May Photo: Peter Kristensen

Analyses and statistics on the frequency and the incidence of traffic accidents within Dolj County

Country report. Czech bathing water quality in Czech Republic. May Photo: Peter Kristensen

New Zealand Fisheries Assessment Research Document 98/21. Not to be cited without permission of the authork) Malcolm Clark

Assessing the spatial dynamics of Small Scale Coastal Fisheries of the Baltic Sea. A Geographic Information System based approach

Sandeel (Ammodytes spp.) in Divisions IIIa, IVa, and IVb, SA 3 (Skagerrak and Kattegat, North and Central North Sea)

Conservation Limits and Management Targets

Reduction of Speed Limit at Approaches to Railway Level Crossings in WA. Main Roads WA. Presenter - Brian Kidd

EVALUATING THE EFFECTS OF BIVALVE SHELLFISH AQUACULTURE AND ITS ECOLOGICAL ROLE IN THE ESTUARINE ENVIRONMENT IN THE UNITED STATES

GRASS CARP SPAWNING IN LAKE ERIE TRIBUTARIES: WHEN AND WHERE?

Zooplankton community structure in the northern Gulf of Mexico: Implications for ecosystem management

6.4 Stock summaries Advice June 2012

Monitoring cruise with r/v Gunnar Thorson in the Sound, the Arkona Sea, the Belt Sea and the Kattegat

STRIP EDGE SHAPE CONTROL

North East Atlantic Fisheries Baltic Sprat Whitepaper March 2011

Sprat (Sprattus sprattus) in subdivisions (Baltic Sea)

Certification of AMS acc. EN 15267, Part 3 - Overview and First Experience -

Protect Our Reefs Grant Interim Report (October 1, 2008 March 31, 2009) Principal investigators: Donald C. Behringer and Mark J.

Report on BPC Special Project regarding the Future Study Recommendations and Peer Review of the Salmonid Fish Tagging Study (Hearn, et al.

Green crabs: invaders in the Great Marsh Featured scientist: Alyssa Novak from the Center for Coastal Studies/Boston University

Advice June 2012

CHAPTER 6 DISCUSSION ON WAVE PREDICTION METHODS

Monitoring Cruise Report

At each type of conflict location, the risk is affected by certain parameters:

CHAPTER 10 TOTAL RECREATIONAL FISHING DAMAGES AND CONCLUSIONS

United States Commercial Vertical Line Vessel Standardized Catch Rates of Red Grouper in the US South Atlantic,

An update of the 2015 Indian Ocean Yellowfin Tuna stock assessment for 2016

Chapter 12 Practice Test

Appendix F: Ecology F-6 Methodology for Estimating Potential Hydroacoustic Impacts to Abundant Hudson River Fish Species and Shortnose Sturgeon from

AMERICAN ACADEMY OF ACTUARIES MEDICARE SUPPLEMENT WORK GROUP SUBTEAM REPORT ON LOSS RATIO CURVES FOR REDETERMINATION OF REFUND BENCHMARKS

BLACK SEA WHITING, MERLANGIUS MERLANGUS EUXINUS NORDMANN

Ocean and Plume Science Management Uncertainties, Questions and Potential Actions (Work Group draft 11/27/13)

Insurance. The practical application. of credibility models in the rating of Health

WFD Intertidal Seagrass Assessment

Advice May Herring in Subdivisions and 32 (excluding Gulf of Riga herring)

Status and trend of four commercially important coastal cephalopods in China Seas: an overview with implications for climate change

Rostock. Meeresbiolog. Beitr. Heft Rostock 2009

OPERATIONAL AMV PRODUCTS DERIVED WITH METEOSAT-6 RAPID SCAN DATA. Arthur de Smet. EUMETSAT, Am Kavalleriesand 31, D Darmstadt, Germany ABSTRACT

Beach Wizard: Development of an Operational Nowcast, Short-Term Forecast System for Nearshore Hydrodynamics and Bathymetric Evolution

Tutorial for the. Total Vertical Uncertainty Analysis Tool in NaviModel3

Results of a Suspended Solids Survey at the Whites Point Quarry, Little River, Digby County, Nova Scotia

9.4.5 Advice September Widely distributed and migratory stocks Herring in the Northeast Atlantic (Norwegian spring-spawning herring)

Imperfect information: Status classification in a world of uncertainty

Standardized catch rates of U.S. blueline tilefish (Caulolatilus microps) from commercial logbook longline data

Ad Hoc Review Group IP(06)12 FINAL. Implementation Plan. European Union (Denmark)

and found that there exist a significant overlap between the billfish resources and the exploitation activities targeting tunas and mahi mahi.

COASTAL PROTECTION AGAINST WIND-WAVE INDUCED EROSION USING SOFT AND POROUS STRUCTURES: A CASE STUDY AT LAKE BIEL, SWITZERLAND

Ecological Interactions in Coastal Marine Ecosystems: Rock Lobster

5.1 Introduction. Learning Objectives

Fishing mortality in relation to highest yield. Fishing mortality in relation to agreed target

Section I: Multiple Choice Select the best answer for each problem.

Review of A Detailed Investigation of Crash Risk Reduction Resulting from Red Light Cameras in Small Urban Areas by M. Burkey and K.

Citation for published version (APA): Canudas Romo, V. (2003). Decomposition Methods in Demography Groningen: s.n.

Title: Modeling Crossing Behavior of Drivers and Pedestrians at Uncontrolled Intersections and Mid-block Crossings

Transcription:

Indicators for ecological status of angiosperms and charophytes in Danish coastal waters Research note from DCE - Danish Centre for Environment and Energy Date: 7 March 2018 Jacob Carstensen & Dorte Krause-Jensen Department of Bioscience Claimant: The Danish Nature Agency No. of pages: 24 Referee: Karsten Dahl Quality assurance, DCE: Susanne Boutrup Tel.: +45 +45 8715 0000 E-mail: dce@au.dk http://dce.au.dk

Contents Summary 3 1 Introduction 4 2 Data for angiosperm indicators 6 2.1 General description of data 6 2.2 Data set for vegetation depth limits 6 2.3 Data set for vegetation cover 7 3 Statistical analysis 10 3.1 Indicators for depth limits 10 3.2 Indicators for cumulative cover 10 3.3 Linking depth limits and cumulative cover to environmental variables 11 4 Results and discussion 13 4.1 Trends of depth limit indicators 13 4.2 Trends of cumulative cover indicators 14 4.3 Relation to eutrophication pressure 17 5 Indicator discussion 21 References 23 2

Summary Until now, indicators of ecological state of Danish coastal waters on the basis of marine vegetation have encompassed eelgrass (Zostera marina) and red, brown and green macroalgae growing on hard substratum. By contrast, charophytes (a division of macroalgae) and angiosperms other than eelgrass have not previously been explored in this context. In this research note, we fill this gap by investigating, based on Danish monitoring data, the potential for developing ecological indicators of angiosperm species other than eelgrass as well as charophyte species. For these groups we explored the potential to use maximum depth limits (i.e. the deepest occurrence of shoots), main depth limits (i.e. the deepest occurrence of patches of 10 % coverage), and the cumulative cover (i.e. sum of all species' cover) at a given water depth. Transect data on species-specific cover and depth limits obtained from the national monitoring programme have been analysed using statistiscal models resolving the spatial and temporal heterogeneity of these data. Trends of cumulative cover and depth limits for angiosperms (excluding eelgrass) and charophytes have been calculated and investigated in relation to Secchi depth and nutrient levels. Charophyte data are generally sparse and restricted to areas of relatively low salinity, with just one area (Guldborgsund) having sufficient data for analyses of trends in depth limits and just four areas having sufficient data for exploring trends in cover. Therefore, indicators were associated with large uncertainty and were not linked to eutrophication pressures. Consequently, the monitoring data do not support the development of charophyte indicators for assessing ecological status of Danish coastal waters. Angiosperms were more common than charophytes in the most brackish Danish coastal ecosystems, and indicators developed on angiosperm data were relatively precise, particularly depth limits that also exhibited clear linkage to Secchi depth. The precision of angiosperm depth limits can be improved by increasing the number of monitored transects or increasing the focus on examining their depth limits along existing monitoring transects. Angiosperm cover data were more variable and did not exhibit the same strong linkage to pressures such as light conditions. On this basis, we conclude that angiosperm depth limits are useful for assessing ecological status of brackish and sheltered coastal ecosystems. 3

1 Introduction According to the European Water Framework Directive, the ecological status of a water body should be assessed from the status of a number of Biological Quality Elements (BQE) through the one-out-all-out combination rule. For coastal water bodies, the BQEs are phytoplankton, benthic vegetation and fauna. The BQE benthic vegetation includes the status of both angiosperms and macroalgae. For Danish coastal waters, a range of macroalgae indicators for hard bottom substratum have been developed previously (Carstensen et al. 2008, 2014) and are currently under implementation (Carstensen 2017), but for angiosperms there is only one indicator available at present, the depth limit of eelgrass. Since Danish coastal waters inhabit more angiosperm species than eelgrass (Zostera spp.), the potential for deriving additional ecological indicators, based on other angiosperm species, should be investigated. Furthermore, since charophytes (a division of macroalgae) are also present on soft seafloors of the more brackish coastal water bodies, potential ecological indicators based on charophyte species should also be analysed. Therefore, the objective of the current note is to investigate the potential for such indicators. Eight criteria have been identified for the quality of an indicator (Quieros et al. 2016): 1. It should be underpinned by a strong scientific basis of causal linkages to given pressures. 2. It should be relevant to the ecosystem in question. 3. It should be responsive to one or multiple pressures, implying that significant changes in the indicator are expected within realistic ranges of pressures. 4. It should be possible to set targets for the indicator, i.e. reference conditions and class boundaries. 5. It should have a response time to pressures that enables mitigation. 6. It should be quantitatively measureable and precise. 7. It should be cost-effective. 8. It should be assessable over time, implying that longer time series should be available. These eight criteria will be used for assessing the quality of the proposed indicators in the concluding section discussing the potential for indicators. In relation to criterion #2, charophytes and angiosperms are known to have ecological relevance in shallow freshwater and brackish ecosystems (Schneider et al. 2015). With respect to criterion #1 and #3, charophytes, in particular, are known to be very sensitive to reduced water quality and, hence, have the 4

potential to be responsive indicators of ecological quality (Blindow 2000). Indeed, the presence and depth penetration of charophytes and angiosperms are highlighted as important indicators of ecological quality in the brackish and protected German boddens where they thrive in clear waters with low nutrient concentrations, while they do not occur along the coastline outside the boddens (Selig et al. 2007). Charophytes are also known as sensitive indicators of water quality in Danish freshwater environments (Baastrup Spohr et al. 2013; Sand-Jensen et al. 2017), and there are reports on reduced occurrence of charophytes in response to eutrophication in e.g. Randers Fjord (Nielsen et al. 2003.). Here, we further explore the sensitivity of depth limits and cover of charophytes and angiosperms in Danish coastal waters based on monitoring data. 5

2 Data for angiosperm indicators 2.1 General description of data Angiosperms and charophytes have been monitored in parallel with eelgrass along transects in Danish estuaries and coastal waters. Angiosperm species in Danish coastal waters mainly belong to the genera Zostera, Ruppia, Zannichellia and Potamogeton, whereas charophytes are almost entirely species from the genus Chara (Manscher & Krause-Jensen 2010). Since eelgrass indicators have already been developed and the focus of this research note is to explore and potentially develop indicators for other soft-bottom vegetation, we will refer to angiosperms as those species mentioned in Table 2.1 and charophytes as those species mentioned in Table 2.2. Data extracted from the national database (ODA) were collected as part of the National Monitoring and Assessment Programme for the Aquatic and Terrestrial Environment (NOVANA) and regional monitoring activities. The monitoring of soft-bottom vegetation in Danish coastal waters involves recording of coverage and depth limit of individual species along transect lines based on diver or video surveys. Until 2001, divers visually assessed the cover of the various species as averages for a depth interval but since 2001, the recordings are done continuously for consecutive points along the transects. The transects are T-shaped: The main transect extends perpendicularly from the coast until the maximum depth limit (i.e. at the deepest shoot) is reached, after which 5-7 additional point observations of the maximum depth limit are obtained along a line parallel to the shore, resulting in the T-shape. The main depth limit is assessed along the main transect as the deepest occurrence of 10 % vegetation cover. 2.2 Data set for vegetation depth limits Depth limits for angiosperms and charophytes were not monitored as consistently as for eelgrass, but nevertheless there were >10000 recorded max depth limits for angiosperms and >2000 for charophytes. There were only 1485 main depth limits for angiosperms and 237 for charophytes. For angiosperms, depth limits were recorded for 15 taxonomical classifications within the genera Potamogeton, Ruppia and Zannichellia in addition to the species Batrachium baudotii, Myriophyllum spicatum, Nuphar lutea and Sparganium emersum. For charophytes, depth limits were recorded for four taxonomical classifications within the genus Chara as well as for the species Tolypella nidifica. Multiple depth limits may be recorded on a single transect (when the diver swims perpendicularly to the transect at the outer depth limit) and these were averaged to provide one observation per transect per day of investigation. This resulted in 1090 and 321 max depth limits for angiosperms and charophytes, respectively, and similarly in 1440 and 351 main depth limits for angiosperms and charophytes. With the exception of a single max depth limit, all depth limits were recorded after the method change in 2001 so that they represent the period 2001-2016. The depth limits were observed across a broad range of coastal sites, though with few observations in each of the sites. In order to have sufficiently large data sets for analysis, we restricted the analysis of depth limits to coastal sites with at least 32 observations, corresponding to ~2 per year for the study period (2001-2016). Consequently, depth limits of angiosperms were analysed for 10 coastal sites, whereas there was only one 6

coastal site (Guldborgsund) with sufficient depth limits for charophytes. Hence, the limited data for charophytes did not allow a general development of indicators. 2.3 Data set for vegetation cover Over the years, many observations of vegetation cover have been recorded for both angiosperms (Table 2.1) and charophytes (Table 2.2). Most of the cover data before 2001 were assessed for depth intervals, whereas data after 2001 are depth-specific (Krause-Jensen et al. 2001). Compared to Manscher & Krause- Jensen (2010), who analysed data from 1982 to 2010, the number of species has increased slightly, and the number of cover observations has almost doubled. Importantly, our analysis also included zero observations, i.e. depth-specific observations where the species could be expected but did not have any cover at that specific sampling point. We calculated the cumulative cover of angiosperms and charophytes by summation over all species listed in Table 2.1 and Table 2.2, respectively. Table 2.1. Number of cover observations for different species of angiosperms (ex. Zostera spp.) extracted from ODA (1982-2016). Cover was observed for depth intervals (older data) and as point observations (specific depths; newer data). Species N (depth interval) N (depth) Batrachium baudotii 410 40 Batrachium circinatum 28 44 Blomsterplanter, uspec. 92 Callitriche sp. 3 36 Ceratophyllum demersum 4 210 Elodea canadensis 6 179 Myriophyllum spicatum 292 960 Nuphar lutea 11 115 Nymphaea alba 1 4 Potamogeton crispus 36 359 Potamogeton filiformis 1 Potamogeton obtusifolius 19 93 Potamogeton pectinatus 3721 32663 Potamogeton perfoliatus 81 498 Potamogeton polygonifolius 1 Potamogeton praelongus 1 Potamogeton pusillus 7 18 Potamogeton sp. 1024 27284 Ruppia cirrhosa 2812 20352 Ruppia maritima 2359 23978 Ruppia sp. 1564 68964 Sparganium emersum 52 Zannichellia major 855 572 Zannichellia palustris 150 19278 Zannichellia pedunculata 17 229 Zannichellia repens 46 22 Zannichellia sp. 936 34213 Total 14477 230163 Aggregating the species-specific data over these two groups poses a challenge. For some depth-specific observations, angiosperms or charophytes may not have been recorded and hence, summation will lead to a missing value. If the depth-specific observation is part of the main transect, then the missing value should be replaced with a zero value, because no angiosperm 7

or charophyte species were observed, but they would have been recorded if they had been present. However, on the T-part of the transect, where usually only eelgrass is recorded, it may not be correct to assume a zero value for the cover of the angiosperm or charophyte species. Therefore, missing values for the cover of angiosperm or charophyte were replaced with zero observations along the main transect, but not along the T-part (identified by the database flag for recording max depth limit). Table 2.2. Number of cover observations for different species of charophytes extracted from ODA. Cover was observed for depth intervals (older data) and as point observations (specific depths; newer data). Species N (depth interval) N (depth) Chara aspera 349 304 Chara baltica 119 1887 Chara canescens 111 Chara sp. 556 32689 Charophyceae 4 Kransnålalger, uspec. 21 Lamprothamnium papulosum 40 1 Tolypella nidifica 522 1181 Total 1701 36083 The spatial distribution of occurrences of Ruppia spp., Zannichellia spp., Potamogeton spp. and Chara spp. was investigated by Manscher & Krause-Jensen (2010) and showed that they were observed in estuaries and coastal waters of the Belt Sea area, The Sound and the Baltic Sea. In this research note, we only focused on coastal sites that had at least 200 cumulative cover observations of angiosperms or charophytes, and where at least 10 % of these observations displayed cover (i.e. >0 % cover). This resulted in 17 and 4 coastal sites with sufficient data for analysing potential indicators for angiosperms and charophytes, respectively. Furthermore, only data after 2001were considered due to the shift in monitoring from depth intervals to point observations. The data analysis focused exclusively on angiosperms and charophytes from the depth range where physical disturbance was no longer a major controlling factor for cover by adopting the definition used for attached macroalgae (Carstensen et al. 2008) (see Figure 2.1). Carstensen et al. (2008) employed depth cut-off values of 1 m for weakly exposed areas, 3 m for moderately exposed areas and 5 m for highly exposed areas. Examination of cover versus depth for the different coastal sites revealed that light limitation started at shallower depths for angiosperms and charophytes than for attached macroalgae. The more exposed coastal sites typically had a cut-off value around 2 m and coastal sites with low exposure had a cut-off value around 1 m. The cut-off was even as low as 0.5 m for Hjarbæk Fjord. 8

Figure 2.1. Illustration of the hypothesis that vegetation cover in shallow water is reduced due to physical exposure while from intermediate water depth towards deeper water, vegetation cover is reduced in parallel to reductions in light availability. Consequently, maximum vegetation cover is found at intermediate water depths and is located deeper in more exposed areas. Arrows indicate cut-off depths for different levels of exposure. In addition to the cover of angiosperms and charophytes, the cover of suitable soft-bottom substrate was also recorded. Since the cover of rooted vegetation can also be limited by lack of suitable substrate, we restricted the analysis of cumulative cover of angiosperms and charophytes to observations with at least 75 % suitable soft-bottom substrate, similar to the threshold used by Wikström et al. (2016). Finally, the analysis was restricted to months from May to October, when most of the transects were surveyed; a few transects were carried out in November but these observations had substantially lower cover compared to the other months. Although the technical guidance recommends that sampling should be carried out between June and September, in practice the sampling season extends a bit longer. 9

3 Statistical analysis Observations of cumulative cover of angiosperms and charophytes were used to develop indicators to describe changes over time, while accounting for the heterogeneity of the monitoring data with respect to their sampling in time (month), depth and transect as well as the diver-specific variation. Changes over time and among sites were subsequently linked to relevant environmental variables to assess if the indicator responded to eutrophication pressure. 3.1 Indicators for depth limits Max (D max ) and main (D main ) depth limits of angiosperms were analysed with a statistical model describing variations over time and space: D = site i + month j + year k + site i year k + STATION l (site i ) + DIVER m where site i described differences between the 10 coastal sites, month j described a common seasonal variation, year k described the time trend common to all sites, site i year k described the site-specific time trend, STATION l (site i ) described the random variation among monitoring transects within a site, and DIVER m described the random variation caused by using different divers in the monitoring programme. With this model, variations in the depth limits were partitioned across the different factors, and means for the site-specific depth limit, common trend and site-specific trends were calculated from the parameter estimate of the model. This means that site-specific means represent an average across all months (May-Oct) and years (2001-2016), and the common trend represents an average across all months (May-Oct) and sites (10 sites). 3.2 Indicators for cumulative cover Variations in the cumulative cover of angiosperms (n = 20763) and charophytes (n = 4722) generally increased with cover levels, suggesting that these data should be described with a lognormal distribution. However, the distributions of cumulative cover were strongly zero-inflated with 43 % and 79 % of observations equal to zero for angiosperms and charophytes, respectively. Thus, we need to describe cumulative cover of angiosperms and charophytes with a distribution that accounts for the many zero observations. In such cases, a compound distribution is useful that combines a binomial distribution with a lognormal distribution. Therefore, we describe the observed cumulative cover as where cumcover = P(absence) 0 + P(presence) (cumcover present) P(presence) = g(site i + depth + month j + year k + site i year k ) log(cumcover present) = site i + depth + month j + year k + site i year k + STATION l (site i ) + DIVER m 10

Thus, the presence/absence of angiosperms or charophytes is a binary variable where the probability of presence is a linear function of site, depth, month and year of the observation translated through the logistic link function g(). This model assumes that all sites display the same seasonal pattern and decreasing probability of presence with depth, although these relationships are displaced by the factor site i describing that some sites have generally higher probability of presence and some sites have generally lower probability of presence. The factors year k and site i year k describe the common and sitespecific interannual variations that are of interest for describing changes of time in indicators of soft-bottom vegetation. The model for cumulative cover when present (cumcover present) contains the same factors and two random factors in addition that explain variation among transects within sites (STATION l (site i )) and variation among divers surveying the transect (DIVER m ). In other words, cumulative cover is modelled using two models: 1) one presence/absence binary model using all data and 2) a lognormal model for observations with cumulative cover >0. Distributions for the annual means of the common and the site-specific trends were calculated for probability of presence and cumulative cover when present. The means were calculated to represent a depth of 2 m, an average of all six months (May-Oct), an average of all transects within the given site and an average of all divers involved in the monitoring. Subsequently, these distributions were combined to describe the overall distribution of cumulative cover by Monte Carlo simulation, i.e. simulating 1000 outcomes of probability of presence and cumulative cover when present. The mean and 95 % confidence interval for the resulting distribution of the overall cumulative cover were calculated from the product of these two simulated distributions. Thus, the resulting trends account for variations in monitoring data regarding both presence/absence and cumulative cover when present. The mean cumulative cover for all sites was calculated similarly from the parameters of the model. 3.3 Linking depth limits and cumulative cover to environmental variables Data on salinity, Secchi depth and nutrient concentrations (total nitrogen and total phosphorus) were extracted from the national database ODA using representative stations for the 15 sites with sufficient angiosperm observations and 6 sites with sufficient charophyte observations. Observations were averaged for the surface layer (0-10 m) and only observations after 2001 (corresponding to the criteria for vegetation) were used. Nutrient concentrations were log-transformed prior to statistical analysis. Variations in these environmental variables were described using a model similar to the above Y or log(y) = site i + station j (site i ) + month k + year l + site i year l where site i described variations among sites, station j (site i ) described variation among monitoring stations within sites, month k described the seasonal variation for the same months as for vegetation, year l described the common trend for all sites, and site i year l described the site-specific trends. After partitioning the variations of the environmental monitoring data into spatial and temporal sources of variation, site-specific means and trends as well as the common trend were calculated as marginal means from the model. For investigating the relationship between indicators of depth limits for angiosperms and cumulative cover for angiosperms and charophytes versus potential pressures, we combined 1) site-specific means, 2) common trends for 11

12 all sites, and 3) site-specific trends. Potential relationships were investigated for cumulative cover versus Secchi depth, TN or TP. Salinity was also tested as an additional explanatory variable, but was not found significant for any model. No significant relationships were found for the cumulative cover of charophytes and therefore, for comparison, we present the relationship to the most significant explanatory variable for cumulative cover of angiosperms.

Depth limits (m) Depth limits (m) Depth limits (m) Depth limits (m) 4 Results and discussion 4.1 Trends of depth limit indicators Depth limits were relatively constant over the study period (2001-2016) for all investigated coastal sites with the exception of Nissum Fjord and Ringkøbing Fjord that both displayed significant trends of angiosperms growing out to deeper depths (Figure 4.1). In Nissum Fjord, this was most pronounced for the max depth limit (D max ) expanding from 1.2 m to 1.8 m, whereas the main depth limit (D main ) only increased from 0.7 m to 0.9 m over the 16 years. In Ringkøbing Fjord, the max depth limit almost doubled from 1.4 m to 2.5 m, whereas the main depth limit increased from m to 1.5 m. Stege Bugt/Nor and Nakskov Fjord exhibited increases in both D max and D main, but the length of the time series was too short to conclude general deeper growth of angiosperms at these sites. On the other hand, Dybsø Fjord and Præstø Fjord both displayed decreasing depth limits over time. However, trends at Stege Bugt/ Nor, Nakskov Fjord, Dybsø Fjord and Præstø Fjord were not significant. The average trend across all 10 coastal sites revealed a non-significant trend for D main and a significant increasing trend for D max from 2.6 m to m over the 16-year study period. This trend was mainly driven by the results from Nissum Fjord and Ringkøbing Fjord. 6.0 Dybsø Fjord 6.0 Guldborgsund 6.0 Karrebæk Fjord 6.0 Korsør Nor Figure 4.1. Annual means of angiosperm depth limits (D max and D main ) for the 10 coastal sites with sufficient data as well as the common trend across all these sites. Error bars show the 95 % confidence interval for the annual means. Note that the scaling varies among sites. Figure 4.1. continues on next page 13

Depth limits (m) Depth limits (m) Depth limits (m) Depth limits (m) Depth limits (m) Depth limits (m) Depth limits (m) 6.0 Nakskov Fjord 6.0 Nissum Fjord 6.0 Præstø Fjord 6.0 Ringkøbing Fjord 6.0 Skælskør Fjord/Nor 6.0 All 10 sites 6.0 Stege Fjord/Nor Figure 4.1. (continued) Annual means of angiosperm depth limits (D max and D main ) for the 10 coastal sites with sufficient data as well as the common trend across all these sites. Error bars show the 95 % confidence interval for the annual means. Note that the scaling varies among sites. 14 4.2 Trends of cumulative cover indicators The 17 sites with sufficient data on cumulative cover of angiosperms displayed a broad range of trends, varying in the overall mean level as well as in the direction of change over time (Figure 4.2). Only Præstø Fjord and Ringkøbing Fjord had data for all the years 2001-2016. Comparing across sites, Præstø Fjord had the highest angiosperm cumulative cover, followed by Guldborgsund and Stege Bugt/Nor. At the other end of the scale were Hjarbæk Fjord, Korsør

Nor, Nissum Fjord and Roskilde Fjord exhibiting relatively low angiosperm cover. The overall trend across all 17 sites showed an increasing tendency from less than 3 % angiosperm cumulative cover to around 6-8 % in the last five years. However, the first years (2001-2003) were dominated by many zero observations (e.g. in Korsør Nor), which may have been incorrectly substituted in the data processing, i.e. we interpreted lack of observation as zero while the possibility exists that lack of observation meant lack of recording even though the angiosperm was present. 4 Dybsø Fjord 4 Guldborgsund 3 3 2 2 1 1 4 Hjarbæk Fjord 3 4 Karrebæk Fjord 3 2 2 1 1 4 Kertinge Fjord/Nor 3 4 Korsør Nor 3 2 2 1 1 10 Lindelse Nor 8 6 4 2 16 Nakskov Fjord 12 8 4 Figure 4.2. Annual means of angiosperm cumulative cover for the 17 investigated sites as well as the common trend across all 17 sites. Error bars show the 95 % confidence interval for the annual means. Note that the scaling varies among sites. Figure 4.2. continues on next page 15

4 Nissum Fjord 4 Odense Fjord (inner) 3 3 2 2 1 1 10 Præstø Fjord 8 6 4 2 4 Randers Fjord 3 2 1 4 Ringkøbing Fjord 4 Roskilde Fjord 3 3 2 2 1 1 4 Skælskør Fjord/Nor 4 Smålandsfarvandet 3 3 2 2 1 1 40 Stege Bugt/Nor 30 2 All 17 sites 15% 20 1 10 5% Figure 4.2. (continued) Annual means of angiosperm cumulative cover for the 17 investigated sites as well as the common trend across all 17 sites. Error bars show the 95 % confidence interval for the annual means. Note that the scaling varies among sites. 16

Cover charophyte Cover charophyte Cover charophyte Cover charophyte Similar to angiosperm cumulative cover, the four coastal sites with cumulative cover of charophytes also displayed a broad range of trends (Figure 4.3). There were many years with no charophyte cover recorded, and only Guldborgsund exhibited a trend without dominance of zero values. The few sitespecific trends and the strong influence of zero observations suggest that the monitoring data set does not support the cover of charophytes as a suitable indicator of ecological status. This is partly because their distribution is limited to the more brackish coastal sites. 10 Dybsø Fjord 4 Guldborgsund 8 3 6 2 4 2 1 4 Korsør Nor 3 2 1 10 Lindelse Nor 8 6 4 2 Figure 4.3. Annual means of charophyte cumulative cover for four investigated sites. Error bars show the 95 % confidence interval for the annual means. Note that the scaling varies among sites. 4.3 Relation to eutrophication pressure Ecological indicators are expected to respond to one or several pressures. Therefore, we investigate if angiosperm depth limits and cumulative cover were linked to Secchi depth and nutrient levels at both spatial and temporal scales. 4.3.1 Responses at the spatial scale Variations among coastal sites in angiosperm depth limits and cover were more strongly linked to the mean Secchi depth (Figure 4.4) than to mean nutrient levels. The relationships were particularly strong for depth limits. The maximum depth limit of angiosperms extended to approximately 90 % of the Secchi depth and the main depth limit was approximately 0.7 m shallower than the max depth limit. The cumulative cover of angiosperms was also related to Secchi depth, although there was more scatter in the relationship. This larger scatter could be due to relatively larger uncertainty on the site-specific means of cumulative cover, as indicated by the larger error bars for angiosperm cover, the many zero observations adding to this. Overall, the relationships document that angiosperms respond to changing light climate as expected. 17

Max depth limit angiosperms Main depth limit angiosperms y=0.87x+0.414 n=10 P=036 R 2 =0.6733 y=0.92x-0.319 n=10 P=005 R 2 =0.7970 0 1 2 3 4 5 0 1 2 3 4 5 Secchi depth (m) Secchi depth (m) 3 2 y=41x-31 n=15 P=044 R 2 =0.4792 1 0 1 2 3 4 5 Secchi depth (m) Figure 4.4. Relationships between site-specific means of max and main depth limits as well as cumulative cover of angiosperms at 2 m depth versus Secchi depth. Error bars show the 95 % confidence interval for the site-specific means. 4.3.2 Responses at the temporal scale The annual means across all coastal sites were more strongly related to TN or TP changes over time than to changes in Secchi depth (2001-2016; Figure 4.5). However, the relationships between annual means were not significant for angiosperm depth limits and the significance for cumulative cover of angiosperms was mainly driven by a single year. The annual nutrient means ranged ~50 %, which is less than the spatial variation in Secchi depth ranging by a factor of 3 (Figure 4.4). The relatively low variability in annual means could explain the lack of significance in relating changes in angiosperm indicators to water quality changes over time. 18

Max depth limit angiosperms Main depth limit angiosperms 3.5 y=-28x+3.73 n=16 P=673 R 2 =0.2194 y=-51x+2.85 n=15 P=0.1017 R 2 =0.1926 2.5 2.5 20 25 30 35 40 Total Nitrogen (µmol L -1 ) 15% 1 y=-035x+0.167 n=16 P=077 R 2 =0.4089 1.5 1 1.2 1.4 1.6 Total Phosphorus (µmol L -1 ) 5% 20 25 30 35 40 Total Nitrogen (µmol L -1 ) Figure 4.5. Relationships between annual means (2001-2016) of max and main depth limits as well as cumulative cover of angiosperms at 2 m depth versus nutrient levels. Error bars show the 95 % confidence interval for the annual means. Considering all the site-specific annual means of angiosperm indicators versus Secchi depth produced expected patterns of increasing depth limits and cover with increasing Secchi depth, but these relationships were primarily driven by spatial differences rather than interannual variations (Figure 4.6). It is also noteworthy that the scatter was typically ~1 m for the site-specific annual depth limit means around the expected common relationship, which is of similar magnitude as the uncertainty of the site-specific annual means. This implies that a large part of the scatter (Figure 4.6) can be explained by the high uncertainty in the annual mean estimates, and that this scatter can be reduced if more depth limits were available to provide estimates that are more precise. The relatively larger scatter for cumulative cover of angiosperms could also be explained by larger uncertainty of the site-specific annual means (Figure 4.2). Although there were considerably more cover observations than depth limits, the variability in cover observations is so large that it severely affects the confidence of the cumulative cover estimates. It is also possible that assessing the cover for individual species adds to the overall uncertainty of the aggregated cumulative cover observations, and that assessing the overall total cover of soft-bottom vegetation could reduce this uncertainty. 19

Main depth limit angiosperms Max depth limit angiosperms Figure 4.6. Relationships between site-specific annual means of max and main depth limits as well as cumulative cover of angiosperms at 2 m depth. Dybsø Fjord Guldborgsund Karrebæk Fjord Korsør Nor Nakskov Fjord Nissum Fjord Præstø Fjord Ringkøbing Fjord Skælskør Fjord/Nor Stege Fjord/Nor 0 1 2 3 4 5 6 Secchi depth (m) Dybsø Fjord Guldborgsund Karrebæk Fjord Korsør Nor Nakskov Fjord Nissum Fjord Præstø Fjord Ringkøbing Fjord Skælskør Fjord/Nor Stege Fjord/Nor 6 4 2 0 1 2 3 4 5 6 Secchi depth (m) 0 1 2 3 4 5 6 Secchi depth (m) Dybsø Fjord Guldborgsund Hjarbæk Fjord Karrebæk Fjord Kertinge Fjord/Nor Korsør Nor Lindelse Nor Nakskov Fjord Nissum Fjord Odense Fjord (inner) Præstø Fjord Randers Fjord Ringkøbing Fjord Roskilde Fjord (inner) Skælskør Fjord/Nor Stege Fjord/Nor Storstømmen 20

5 Indicator discussion Observations of species-specific cover and depth limits of angiosperms (excluding Zostera spp.) and charophytes have been analysed using statistical models that account for the heterogeneous temporal and spatial sampling properties of the monitoring data. These observations can potentially produce indicators for angiosperms and charophytes that describe changes over time at regional and local levels. Data on charophyte cover and depth distribution were too few to produce reliable indicator estimates, with just one area (Guldborgsund) having sufficient data for analysing trends in depth limits and just four areas having sufficient data for analysing trends in cover. Consequently, the monitoring data does not support the development of charophyte indicators for assessing ecological status of Danish coastal waters in general. Indicators based on angiosperm cover and depth distribution data are more applicable for assessing ecological status, fulfilling the criteria for good indicator quality: 1. Angiosperms respond to the present light climate in a similar manner as eelgrass. Although there are many scientific studies documenting the causal linkages between angiosperms and light conditions, the scientific basis is dominated by seagrass studies. 2. Angiosperms typically grow in brackish and sheltered coastal ecosystems and may constitute the most important benthic vegetation in such systems. This study shows that angiosperms are found in coastal sites with salinity less than 15, although not common along open coastal stretches with low salinity such as Fakse Bugt and Hjelm Bugt. Hence, angiosperms (other than eelgrass) are important component of brackish coastal ecosystems. 3. Spatial variations in angiosperm cover and depth limits among coastal sites were clearly linked to prevailing light regimes. Our results suggest an almost proportional response in depth limits to changing Secchi depth, whereas the response in cumulative cover was weaker and more uncertain. 4. Ecological targets for angiosperm indicators have not been determined for Danish coastal waters. It is possible that such targets can be found from correlation analysis with eelgrass depth distributions, since reference conditions and class boundaries already exist for eelgrass depth limits. Also targets defined for depth limits of angiosperms in German boddens are relevant to consider (e.g. Selig et al. 2007). 5. Similar to eelgrass, it is difficult to assess the response time to changing pressures. Experiences from Nibe-Gjøl Bredning and Køge Bugt have shown that eelgrass responds to improved light conditions within a few years. However, the response to light conditions could be specific to the ecosystem. Only two coastal sites exhibited changes in angiosperm depth limits (Ringkøbing Fjord and Nissum Fjord). Although increasing depth limits appear to coincide with improving light conditions for some years 21

Depth limits (m) Depth limits (m) at these sites, the pattern is not consistent for the entire period (Figure 5.1). This implies that the recovery of angiosperms is more complex, involving potential time lags and other factors such as sediment suitability. Nissum Fjord Secchi depth Ringkøbing Fjord Secchi depth Figure 5.1. Trends in angiosperm depth limits and Secchi depth for the two sites exhibiting increasing depth limits. 6. Angiosperm depth limits and cumulative cover are both measureable, but depth limits provide more precise information than cover estimates. Given that annual means of depth limits were estimated from a few transects, it should be possible to obtain more precise indicators by monitoring more transects or focusing monitoring more specifically on determining the depth limit. 7. Angiosperm monitoring data are recorded together with eelgrass surveys and therefore, monitoring angiosperms is marginally inexpensive provided that eelgrass is already monitored. Thus, the additional cost of monitoring other angiosperm species than eelgrass is small. 8. Angiosperms have been monitored for several years and 16-year trends are available for many coastal sites already. Hence, there is an existing legacy from the marine monitoring programme to build upon. Based on the above, we conclude that angiosperm depth limits are useful for assessing the ecological status in selected coastal ecosystems (brackish and sheltered), whereas cumulative cover is too variable to provide a good indicator. The monitoring data set does not offer a useful basis for developing charophyte indicators for assessing the ecological status of Danish coastal waters. 22

References Baastrup Spohr L, Iversen LL, Dahl Nielsen J & Sand Jensen K (2013) Seventy years of changes in the abundance of Danish charophytes. Freshwater Biology 58(8):1682-1693. Blindow I. (2000) Distribution of Charophytes along the Swedish Coast in Relation to Salinity and Eutrophication. International Review of Hydrobiology 85(5 6):707-717. Carstensen J, Krause-Jensen D, Dahl K, Henriksen P (2008) Macroalgae and phytoplankton as indicators of ecological status of Danish coastal waters. National Environmental Research Institute, University of Aarhus. 90 pp. - NERI Technical Report No. 683. http://www.dmu.dk/pub/fr683.pdf Carstensen J, Krause-Jensen D &Josefson A (2014) Development and testing of tools for intercalibration of phytoplankton, macrovegetation and benthic fauna in Danish coastal areas. Aarhus University, DCE Danish Centre for Environment and Energy, 85 pp. Scientific Report from DCE Danish Centre for Environment and Energy No. 93. http://dce2.au.dk/pub/sr93.pdf Carstensen J (2017) Methodology for calculating macroalgae indicators of abundance and composition. Research note from DCE - Danish Centre for Environment and Energy. 17pp. Krause-Jensen D, Laursen JS, Middelboe AL, Stjernholm M (2001) NOVA Teknisk anvisning for marin overvågning. Ch. 12. National Environmental Research Institute. Ministry of Environment and Energy, Denmark (in Danish). Manscher O & Krause-Jensen D (2010) Marine blomsterplanter og kransnålalger i danske farvande. Notat fra Danmarks Miljøundersøgelser. 20 December 2010. Nielsen K, Sømod B, Ellegaard C & Krause-Jensen D (2003) Assessing reference conditions according to the European Water Framework Directive using modelling and analysis of historical data: an example from Randers Fjord, Denmark. AMBIO: A Journal of the Human Environment 32(4):287-294. Sand Jensen K, Bruun HH, & Baastrup Spohr L (2017) Decade long time delays in nutrient and plant species dynamics during eutrophication and re oligotrophication of Lake Fure 1900 2015. Journal of Ecology 105(3):690-700. Selig U, Eggert A, Schories D, Schubert M, Blümel C & Schubert H (2007) Ecological classification of macroalgae and angiosperm communities of inner coastal waters in the southern Baltic Sea. Ecological Indicators 7(3): 665-678. Schneider SC, García A, Martín-Closas C & Chivas AR (2015) The role of charophytes (Charales) in past and present environments: An overview. Aquatic Botany 120(A):2-6. 23

Quieros AM, Strong JA, Mazik K, Carstensen J, Bruun J, Somerfield PJ, Bruhn A, Ciavatta S, Flo E, Bizsel N, Özaydinli M, Chuseve R, Muxika I, Nygård H, Papdopoulou N, Pantazi M, Krause-Jensen D (2016) An objective framework to test the quality of candidate indicators of good environmental status. Frontiers in Marine Science 3:73. doi: 10.3389/fmars.2016.00073. Wikström S, Carstensen J, Blomqvist M & Krause-Jensen D (2016) Cover of coastal vegetation as an indicator of eutrophication along environmental gradients. Marine Biology 163:257. doi: 10.1007/s00227-016-3032-6. 24