NEAFC request to evaluate the use and quality of VMS data

9.3.2.4 NEAFC request to evaluate the use and quality of VMS data Request Regarding deepsea fisheries, NEAFC requests ICES to evaluate the use and quality of VMS data and records of catch and effort to be received from NEAFC in order to provide information on the spatial and temporal extent of current deep-water fisheries in the NE Atlantic. If data quality allows such analyses, these should be provided with particular emphasis on activity in the NEAFC Regulatory Area. ICES response Vessel monitoring system data (VMS) can be used to gain valuable insights into many questions related to fishing pressure in specific locations. In most cases, the minimum information transmitted through VMS is the time of the signal, vessel identity, and vessel location at the time of transmission. Vessel speed maybe transmitted, but it can also be computed as the distance covered between successive locations. It is usually assumed that fishing is occurring when the computed vessel speed is less than about 3 knots. However, this could vary by country and gear type being used. Gear information may or may not be included in the VMS record but could possibly be implied by vessel speed. This information is not always accurate, however, as the vessel may be drifting because of malfunction, processing the catch, or experiencing other problems. It may be assumed that catch is proportional to time fishing and thus could be assigned based on time and distance between polling points. This assumption may not always be accurate. 1. Available data: VMS data and catch and effort data VMS data that clearly depict vessel location may be used to perform spatial and temporal analysis of deep-water fisheries in the NEAFC Regulatory Area, for example to: a) estimate the number of fishing vessels entering the NEAFC areas with a breakdown of their nationalities; b) evaluate the seasonality of fishing spatial patterns; c) infer spawning aggregation areas, such as for blue ling; d) track the historical development and changes in behaviour of the fishing fleets. When linked to catch and effort data (including vessel speed, gear type being used, amount of gear, catch record, and discards), VMS data can be used to better understand fleet activity, fishery-specific behaviour, spatial distribution of catches, and fishing effort. For example: a) catch and effort data can be presented as density maps; b) fisheries can be characterized in terms of single-target species or in terms of bycatch species in mixed fisheries. 2. Quality of VMS data and catch and effort data ICES received approximately 1 115 000 individual VMS data points covering the period 2002-2006. The basic format of this data was a flat file with a number of descriptive fields. Information on gear type deployed or speed of vessel is not transmitted. A total of approximately 1400 vessels provided VMS reports from the NEAFC area over this period (although some vessels may have changed codes between years). Reports of catch on board are meant to be transmitted weekly (or on entering and leaving the NEAFC area), but only 27% of the vessels that transmitted VMS data had ever reported catch. ICES is aware of analyses that compare the apparent position indicated by the VMS signal with positions as recorded by satellite overflight which show that VMS signals may be altered to give false positions. Equally, it is possible to misreport catch. ICES was unable to assess the quality of the NEAFC VMS data from this perspective of accuracy and truth and has assumed that the data are fully accurate in subsequent analyses. Based on the above, ICES advises that NEAFC could further improve the usefulness of the VMS data by: a) requiring transmission of vessel speed and gear in use (if applicable); b) increasing the transmission frequency; c) increasing frequency of transmission (maybe to once per day) and completeness (substantially more than 27% of the fleet) of catch data. ICES Advice 2007, Book 9 25

3. Analyses with particular emphasis on activity in the NEAFC Regulatory Area Comprehensive analysis of these data is likely to require significantly greater amounts of time and resources than are available to ICES at present. ICES has therefore undertaken a few example analyses of the NEAFC VMS data. These examples demonstrate a possible methodology for making detailed inferences about the likely impact of fishing activities on sensitive habitats (see Technical annex). The two case studies are on areas known to be important fishing grounds and areas for which there is separate data suggesting the presence of sensitive habitats such as hard and soft corals. ICES would be happy to discuss with NEAFC possible ways of carrying out further analyses of this data. Sources of information Report of the Working Group on the Biology and Assessment of Deep-Sea Fisheries Resources. 3-10 March 2008, ICES Headquarters, Copenhagen (ICES CM 2008/ACOM:14). Report of the ICES-NAFO Joint Working Group on Deep-Water Ecology. ICES headquarters, 10-14 March 2008 (ICES CM 2008/ACOM:45). 26 ICES Advice 2007, Book 9

A. Technical annex to the ICES response Vessel monitoring system data (VMS) can be used to gain valuable insight into many questions pertaining to fishing pressure in specific locations. VMS is used in commercial fishing to allow environmental and fisheries regulatory organizations to monitor fishing vessel activities. Each VMS unit is assigned to a particular vessel and is required to be secured and operating effectively. Typically the information transmitted through VMS includes identity information, location of the vessel, and time of signal. Vessel speed may be computed as the distance covered between locations. It is usually assumed that fishing is occurring when the computed vessel speed is less than 3 knots. However, this could vary by country and gear type being used. Gear information can be included in the VMS record (it is not in NEAFC data) but could possibly be implied by vessel speed. This though relies on some assumptions that are not always met. Vessel speed may be affected through malfunction or other problems, or processing of the catch. It is typically assumed that catch is proportional to the time fished and catch could thus be estimated based on the time and distance between transmission points. This assumption may not always be reasonable. A.1. Evaluation of data use ICES received VMS data and catch and effort data from NEAFC to perform spatial and temporal analysis of deep-water fisheries in the NEAFC Regulatory Area. By using vessels VMS signal, the position of fishing vessels can be displayed in a GIS environment and mapped to visualize the spatial distribution of vessel activity (Fig. 9.3.2.4.1). In addition, the break down of the information stored in the VMS dataset allows the characterization of the different fleets in a spatial and temporal content to give: Number of fishing vessels entering the NEAFC areas with a breakdown of their nationalities; Seasonal activity of spatial patterns; Historical development and behaviour of the fishing fleets. VMS may also be used to monitor vessel activities in relation to regulatory actions involving fishing quotas, harvesting limits, position in relation to areas closed to fishing, special management of fishing zones, and license limitations. The speed of the vessels is calculated by dividing the distance from one VMS position to the next with the time difference between the two points. This analysis can be used to differentiate between steaming and fishing activities and therefore show the spatial extent of the actual fishing activity. Assumptions on gear type in use can be made on the basis of the derived speed and direction. These analyses can provide spatial patterns of usage of different gear, which in turn is needed to examine gear-habitat interactions and gear-specific environmental impact analyses of fishing activity. The addition of logbook data to the VMS data increases the potential of the analyses. The vessel and its positional data can be linked to declared catches. The catch data provided by NEAFC embedded in the VMS record contains a weekly breakdown of the weight and the species composition caught by vessels operating within the NEAFC area, with some vessels reporting more frequently. If this catch data is linked to the VMS positional data, it is possible to spatially locate specific fisheries. This can substantially enhance analyses of fleet activity, fishery-specific behaviour, and effort. Catch and effort data can be spatially presented and density maps of fishing effort and fishing areas for different species can be shown. For example, Fig. 9.3.2.4.2 shows maps of density of effort from VMS data of the Hatton Bank area by year from 2002 to 2006. One can clearly see that fishing effort in the southeast of the area has dramatically increased over the years. The annual spatial variability of fishing effort is also evident. For example, the area in the centre of the chart was heavily fished in 2005, but not in 2003 or 2006. In addition, cluster analysis can be performed on the catch composition of the individual vessels to allow the characterization of fisheries in terms of single-target species and bycatch species in mixed fisheries. Vessels can be grouped with similar catch composition and when combined to gear characterization, a meta-analysis can be performed. Information from VMS records, logbooks, fishers, and from scientific literature has been used to infer areas of spawning for southern blue ling. Using additional VMS data available one may be able to define other unknown spawning areas in greater spatial detail. Such analyses are of great use in defining the most appropriate areas to close (or otherwise manage) for any specific management purpose. Onboard observer reports further complement both the VMS and logbook records and can add another layer of reliability of vessel activities. The observer reports are likely the most accurate source of information on bycatch and discards. It would be possible to overlay information on the seafloor and/or oceanographic conditions with VMS records, logbook reports, and onboard observer assessments to examine patterns of fishing, intensity of fishing effort, and the association of fishing efforts to bathymetry, oceanographic conditions (e.g., temperature, salinity, currents) and habitat (e.g., areas of cold-water corals). This would enable any habitat-fisheries links to be determined, which may in turn lead to better fisheries management decisions. The comparison of vessel fishing activities with known or likely vulnerable marine ecosystems may lead to regulatory measures to protect these ecosystems. ICES Advice 2007, Book 9 27

Figure 9.3.2.4.1 Activity maps using raw VMS data. 2002 2003 2004 2005 2006 Figure 9.3.2.4.2 Maps showing the density of effort from VMS data of the Hatton Bank area by year from 2002 to 2006. Black and red areas = highest density of fishing vessels. 28 ICES Advice 2007, Book 9

A.2. Quality of NEAFC VMS data ICES received approximately 1 115 000 individual data points of VMS covering the period 2002-2006. The basic format of this data was a flat file with a number of descriptive fields (Table 9.3.2.4.1). Information on gear type deployed or speed of vessel is not transmitted. A total of approximately 1400 vessels provided VMS reports from the NEAFC area over this period (although some vessels may have changed codes between years). Reports of catch on board are meant to be transmitted weekly (or on entering and leaving the NEAFC area), but only 27% of the vessels that transmitted VMS data had ever reported catch. Table 9.3.2.4.1 FIELD Date Time Vessel ICES Area Latitude Longitude Type of report FAO code for fish Kg of fish Fields in VMS data supplied by NEAFC. COMMENT Reporting code Position, catch, transfer, entrance, exit ICES is aware of analyses that compare the apparent position indicated by the VMS signal with positions as recorded by satellite overflight which show that VMS signals may be altered to give false positions. Equally, it is possible to misreport catch. ICES was unable to assess the quality of the NEAFC VMS data from the perspective of accuracy and truth and has assumed that the data are fully accurate in subsequent analyses. ICES identified a number of issues with the NEAFC VMS data. These can be summarized as follows: The dataset does not contain any gear information. Broad gear types can be estimated by their position and speed as explained above. This, however, can not be carried out with absolute certainty as it depends heavily on the frequency and regularity of the transmission. The frequency of VMS records ranged from 1 to 2 hours. Preliminary studies indicate that this frequency of recording and reporting might not be unreasonable for trawlers, when the haul duration exceeds that interval, which is generally the case in deep-sea fisheries. However, there is some evidence that a 1-2 hour frequency does not allow the identification of fishing operations and the calculation of fishing effort at a satisfactory precision for vessels using passive gear. A higher reporting frequency is recommended. A higher frequency of recording would be desirable for monitoring fishing activities in relation to protected areas. In addition there are limitations when linking the catch report data to the positional VMS data. The catch is only reported weekly, while the positional data is recorded on a two-hourly basis. It is therefore not possible to ascertain the position of effort and catches at a higher resolution than a weekly window (or the specific period when the vessel enters and exits the NEAFC area). According to NEAFC data only about 22% of the boats operating in the area in 2006 have provided their associated catch information (catch or transfer reports). For the period from 2002 to 2006 only about 27% of the boats had associated catches. A more in-depth analysis of catch composition and fishing effort can therefore only be carried out on a subsection of the vessels and it is not known whether these vessels are representative of all activities or whether such an analysis can bias certain gear types or nationalities. A.3. Examples of analysis ICES aimed to summarize the potential and the limitations of the possible data analysis to map fishing activities with VMS data and logbook data in the NEAFC regulatory area. Comprehensive analysis of these data, including cluster analysis to identify different fisheries, is likely to require significantly more resources than are available to ICES at present. ICES presents here two examples of the use of VMS and catch data demonstrating ways of inferring the likely impact of fishing activities on sensitive habitats. The two case studies cover areas known to be important fishing grounds and areas for which there is data suggesting the presence of sensitive habitats such as hard and soft corals. Introduction It is important to realize the limitations of VMS data and to analyse on appropriate geographic scales. On broad scales, VMS data can identify the regions that are being fished heavily (Fig. 9.3.2.4.2). Within these regions, areas that contain ICES Advice 2007, Book 9 29

sensitive habitats may often occur in close proximity to areas that do not. Different fishing practices have greater or lesser impact than others (see Section 9.3.2.3). The impact of bottom trawl is far more detrimental to the seabed than static gear such as gillnets and non-contact operations such as pelagic trawling. It is important to a) have finely resolved spatial information in particular areas of interest in relation to known occurrences of sensitive habitats, and b) have the data carefully filtered with respect to the speed and behaviour of the vessel. The likelihood that the vessel was travelling between fishing grounds, dodging in poor weather, bottom trawling, or pelagic trawling can all be estimated based on the average speed profile of the boat calculated from the distance between two consecutive points. Further validation checks can be made by analysing the vessel s path in relation to the seabed contours (usually isobathic for bottom trawling, but not for pelagic) and the patterns of movement for example the vessel will return to the exact same spot to retrieve pots or anchored gillnets. In this term of reference ICES takes advantage of recent advances in the processing of VMS data. VMS data sources and analysis All vessels greater than 24 m are required to carry VMS transmitters if they are to be permitted to fish in NEAFC waters. NEAFC provided ICES with VMS data for the period 2002 2006. No reference to a vessel s true identity was provided in the data. Analysing the frequency of time spent at varying speeds can assess the type of activity being undertaken by the vessel. The following speed criteria were used to help assign a type of fishing activity: a) if the vessel was moving at speeds greater than 5 knots it was assumed to be travelling between fishing grounds; b) if the vessel was travelling between 4 and 5 knots it was assumed to be pelagic trawling; c) if the vessel was travelling at speeds between 2 and 3 knots it was assumed to be bottom trawling; d) if the vessel was travelling between 1 and 2 knots it could either be dodging in poor weather, laying gillnets, or laying longlines. If none of the above criteria were clearly reached, investigation of the vessel s path in relation to the seabed contours (usually isobathic for bottom trawling, but not for pelagic) and the movement pattern were undertaken. For example, if the vessel returned to the exact same positions it was likely to be retrieving pots or anchored gillnets. Sensitive habitat data sources ICES holds data on the occurrence of species considered sensitive (e.g. the cold-water coral, Lophelia pertusa) and of sensitive habitats. Two case study sites were considered in which there was both evidence of fishing activity and the presence of sensitive deep-water habitat to illustrate how the approach can be developed in the future and extended to larger areas. Case study 1: Lousy Bank Lousy Bank, or Outer Bailey Bank, is a seamount that lies approximately at 60º25.02'N and 12º34.98'W, to the northeast of Hatton Bank. It lies at the limits of the European and Faroese fishing zones and the NEAFC areas and has been fished for blue ling that aggregate to spawn on its lower flanks in early spring. There are a number of historical records of Lophelia pertusa on the seamount (Fig. 9.3.2.4.3). An early study was made of the sessile fauna of the bank in the 1930s. It is likely to still contain sensitive deep-water habitats and organisms. NEAFC VMS data for three vessels was examined (Fig. 9.3.2.4.3). The first vessel a (Fig. 9.3.2.4.4a) is likely to be operating a bottom trawl since its speed profile shows peaks of movement at around 3.2 knots (typical bottom trawling speed) or around 1.5 knots (typical dodging speed in poor weather). The second vessel (Fig. 9.3.2.4.4b) is likely to be operating gillnets, as its speed was mostly lower than 2 knots. The activity of third vessel (Fig. 9.3.2.4.4c) is more difficult to interpret, as there is a broad peak in speeds between 1 and 4 knots, suggesting it was unlikely to be bottom trawling. This vessel could be longlining. All the vessels were clearly targeting a small area within a narrow depth band, probably blue ling spawning aggregations. Such intense fishing activity in such a small area would undoubtedly have an impact on a sensitive habitat such as coral reefs. However, there appears to be little overlap between the fishing activities and Lophelia records. 30 ICES Advice 2007, Book 9

Figure 9.3.2.4.3 Fishing activity and the historical locations of coral reefs on Lousy Bank. Positions of three fishing vessels (different coloured dots) are shown in relation to historical records of Lophelia pertusa (black symbols) on the bank. Little overlap between the fishing grounds and Lophelia reefs can be observed. ICES Advice 2007, Book 9 31

a) b) c) Figure 9.3.2.4.4 Lousy Bank fishing vessels (a) Vessel 1 bottom trawler, (b) Vessel 2 gillnetter, (c) Vessel 3 possible longliner. 32 ICES Advice 2007, Book 9

Case Study 2: Lyoness Bank (East Hatton) The Lyoness Bank is a volcanic rock outcrop that rises out of the Rockall-Hatton basin to the southwest of Hatton Bank (approximate position 56 N 18 W). The igneous centre forms a dome covered with a veneer of sediment through which igneous rocks outcrop, forming pinnacles that stand 100 140 m above the surrounding sea floor. Two UK surveys took place in 2006. They involved multibeam bathymetry mapping and biological groundtruthing using video and still photography. The seabed was typically coarse sand, with extensive areas of bedrock and boulder reef. The rock is covered with encrusting fauna such as sponges, sessile sea cucumbers (Psolus squamatus), Serpulid worms, saddle oysters, and occasional anemones. A diverse range of corals including scleractinians (Lophelia pertusa and Madrepora oculata), antipatharian corals (Stichopathes sp., Leiopathes sp.), stylasterid corals, soft corals (Anthomastus grandiflorus), and gorgonians were also frequently observed. Extensive areas of dead coral framework and coral rubble were also observed, with characteristic associated epifauna as described above. VMS data indicated that fishing vessels visit the area on numerous occasions (Fig. 9.3.2.4.5). The vessels often follow the contour around the base of the Bank with the vessel tracks forming a crescent shape. The VMS records of 2 vessels were chosen for detailed analysis (Fig. 9.3.2.4.6). The 2 vessels can clearly be identified as not bottom trawling. The first appears to be operating gillnets since its peak speed is usually less than 1 knot (Fig. 9.3.2.4.6a), while the other vessel is likely to be operating pots for deep-water crabs. This conclusion was reached on the basis that the vessel always returned to exactly the same spot in order to set and retrieve the pots. However, although we did not find evidence for bottom trawling in these 2 vessels it is likely that bottom trawling does occur around the base of the bank also. Scottish landings data report large catches of blue ling by bottom trawl in the statistical rectangle that contains this Bank. It can be seen there is very little overlap between the vessel operating crab pot grounds and the reefs; however, it would appear that there is overlap with the vessel thought to be operating gillnets. While not as damaging as bottom trawling, the dragging of anchors and retrieval of gillnets can cause substantial damage and the overlap of coral records with gillnet fishing activity is a cause for concern in this area. Discussion The recent advances in the analysis of VMS data provide a much more detailed and objective assessment of fishing activity and its likely impact on sensitive deep-water habitats. It is only by looking in detail and in fine scale that one can get an accurate estimate of the amount of overlap between areas known to contain sensitive habitats and potentially damaging fishing activity. Of course we cannot assume that a lack of information equates to a lack of sensitive habitat and this is a shortfall of this approach. ICES recommends that this approach is further developed over the coming year. In particular the set of the criteria for assigning types of fishing activity needs to be refined and expanded. ICES Advice 2007, Book 9 33

Figure 9.3.2.4.5 Positions of two fishing vessels (blue dots = vessel fishing with crab pot, yellow dots = fishing thought to be gillnetting) are shown in relation to recent records of Lophelia pertusa on the bank. 34 ICES Advice 2007, Book 9

a) b) Figure 9.3.2.4.6 Lyoness Bank. (a) vessel 1 likely to be a gillnetter. (b) vessel 2 likely to be operating crab pots. ICES Advice 2007, Book 9 35