Spatially explicit estimates of stock sizes, structure and biomass of herring and blue whiting, and catch data of bluefin tuna

. The North Atlantic is a productive marine region which has supported important commercial ﬁsheries for centuries. Many of these ﬁsheries have exploited the pelagic species, including herring, blue whiting and tuna. Here we present data on the distribution of herring and blue whiting based on the international ecosystem survey in the Nordic Seas (IESNS), the bottom trawl survey in the Bay of Biscay and Celtic Sea (EVHOE) and the pelagic survey in the Bay of Biscay (PELGAS). We also present catch data on blueﬁn tuna, which has been depleted for decades but historically used to be a key predator on the other pelagic stocks during summer. The results show that there were substantial changes in the herring and blue whiting distribution during the 1990s and early 2000s. The earliest blueﬁn tuna catches noted were in 1907. The catches in the Norwegian Sea area peaked in the 1950s and there


Referenced data sets
Norwegian spring spawning herring:   1 Introduction

Key pelagic fish stocks in the northeastern Atlantic
The North Atlantic is a productive marine region which has supported important commercial fisheries for centuries. Many of these fisheries have exploited the pelagic (mainly open-water, surface-living) species, such as herring (Clupea harengus Linnaeus 1758) and blue whiting (Micromesistius poutassou Risso 1826). These species are primarily, though not exclusively, zooplanktivores (Fridriksson, 1944;Prokopchuk and Sentyabov, 2006;Pinnegar et al., 2015) and thus are important links in the food web between zooplankton and piscivores (e.g. cod, seals, whales, seabirds). The biomasses of herring and blue whiting, and a few other pelagic fish species (e.g. mackerel, Scomber scombrus Linnaeus 1758), are so high that they attract many seasonal migrants to the region. Hence, upper-trophic-level predators such as bluefin tuna (Thunnus thynnus Linnaeus 1758) and some species of marine mammals and seabirds have evolved migratory behaviours to inhabit this region for feeding in the summer months (Mather et al., 1995). Fisheries therefore must be conducted in ways that are sustainable not only for the fishing industry but also for the targeted fish species, other species in the food web that depend on them for food, and more generally for maintaining healthy, resilient ecosystems.
Maintaining sustainable populations and ecosystems generally requires direct monitoring of population status and their fisheries (Hilborn and Branch, 2013;Pauly, 2013). Such data can provide a basis for fisheries management decisions (e.g. quotas). Two of the most important types of monitoring information are (1) research vessel surveys of fish abundance and distribution and (2) records of commercial catches of targeted and bycatch species. These data can reveal when and where the fish were located as well as when and where they were being exploited. The data can then be compared in time and space and used in models with other data (e.g. fishing effort, fish sizes and ages) to estimate whether populations are stable, declining or increasing. When provided with this knowledge, fishery managers can make decisions on future quotas and identify possible conservation actions (e.g. minimum size limits, implementation of closed areas or seasons for fishing) that can reduce the risk of stock collapses and local extinctions.

Norwegian spring-spawning (NSS) herring
The NSS stock migrates widely over large areas in the northeastern Atlantic, including the Barents Sea, Norwegian Sea and Icelandic Sea (Devold, 1963;Dragesund, 1970;Røttingen, 1990;Dragesund et al., 1997;Fernö et al., 1998;Jakobsson and Østvedt, 1999;Holst et al., 2002). The life cycle of the NSS herring presently includes spawning along the Norwegian coast in late winter, nursery areas along the coast and in the Barents Sea, feeding areas in the Norwegian Sea and overwintering outside northern Norway (Fig. 1). During the last decades the stock has changed its distribution and migration pattern substantially with regards to spawning, feeding and overwintering areas (Fig. 1). Between 1968 and 1977 the NSS herring spawning stock biomass (SSB) declined to less than 0.2 Mt, with a minimum occurring in 1971 (Toresen and Østvedt, 2000). Fishing mortality, F , was as high as 1.5 during the pre-collapse period (Dragesund et al., 1980), but since 1994 it has ranged from 0.18 to 0.24 in accordance with the long-term management plan (ICES, 2013 (Toresen and Østvedt, 2000). Usually, the NSS herring fishery takes place in the Norwegian Sea during the summer and autumn and in the Norwegian coastal areas during the autumn and winter, including the spawning period (ICES, 2013).

Blue whiting
Blue whiting is widely distributed over the northeastern Atlantic with the dominant spawning area situated to the west of the British Isles. It has a major feeding migration and summer distribution north all the way up to Svalbard and into the Barents Sea and southward down to the Bay of Biscay and the Iberian coast ( Fig. 2) (Monstad, 2004  The fishery for blue whiting has displayed a dramatic "boom and bust" dynamic over the past two decades (ICES, 2013). The modern commercial fishery of this stock began in the late 1970s and early 1980s. Landings during the 1980s and early 1990s were typically between 0.5 and 1 Mt. However, the late 1990s and early 2000s saw a succession of extremely strong year classes, starting with the 1995 cohort, typically a factor of 4 to 10 times greater than that observed during the previous 15 years. When the first of these cohorts reached maturity in 1998, the spawning stock biomass expanded rapidly and the landings from the fishery nearly doubled from one year to the next (0.64 Mt in 1997, 1.13 Mt in 1998). The fishery continued to grow into the early 2000s on the back of the strong year classes, and in 2004 landings reached 2.4 Mt, an increase of 400 % in just 7 years. At this point, blue whiting was the largest fishery in the North Atlantic, ahead of herring, and the third largest marine capture fishery in the world (FAO, 2010). The subsequent decline of the fishery has, however, proved to be equally dramatic. Year class strengths from the 2005-2009 cohorts were comparable to, or even lower than, those prior to 1995 (ICES, 2013). The 2010-2012 cohorts are then apparently above average size. The extremely large fishery could not be sustained under these conditions of reduced productivity and rapid reductions in allowable catch followed. Landings were reduced by 75 % in the space of 5 years.

Bluefin tuna
Bluefin tuna have historically migrated into the Norwegian and North seas, preying on the high concentrations of pelagic fish in this area during summer (MacKenzie and Myers, 2007;Mather et al., 1995). While they were in this region they were targeted by fishing vessels from several nations. The fishery developed in the 1920s-1940s, peaked in the 1950s and declined in the mid-late 1960s before ending in the mid-1970s. The species has since then been rarely seen and has not supported commercial or recreational fisheries. Reasons for the disappearance are unclear, but probably due to a combination of fishing and climatic and ecosystem factors (Fromentin, 2009;Tiews, 1978).

Objectives
Here we provide data on the spatial distribution of herring and blue whiting in the northeastern Atlantic that have been compiled by research vessels operated by several nations in the region. For NSS herring, data from the IESNS (international ecosystem survey in the Nordic Seas) survey in May in the Nordic Seas are presented (period: 2004-2012). Furthermore, data from three surveys are presented here for blue whiting: the IESNS survey in May in the Nordic Seas (period: 2004-2012), the French pelagic survey (PELGAS; period: 2000-2012) and the demersal bottom trawl surveys (EVHOE;period: 1997. We have also compiled and presented a long time series of the commercial catches of bluefin tuna in this region resolved by different spatial regions and countries (period: 1906-2010). These data will help with studies into understanding the causes of fluctuations in abundance and distribution of these species and contribute to the wider objectives of a large European Union project (EURO-BASIN) investigating how climate variability and fisheries affect food webs and biogeochemical fluxes in the off-shelf regions of the northeastern Atlantic Ocean. Additional brief backgrounds of the species biology and fisheries are given in the Methods section. The survey and catch data reported here have been submitted to the PANGAEA website.

Data sources and geographic region of coverage
The survey and catch data described here cover a wide area of the northeastern Atlantic. The region is divided into smaller units for fishery and ecosystem management purposes by ICES (International Council for the Exploration of the Sea) ( Figs. 3 and 4). The historical data on NSS herring and blue whiting in the northeastern Atlantic originate from various national and international surveys aimed at mapping the major distribution and estimating abundances and demographic structure of these large pelagic planktivorous fish species (e.g. ICES, 2009ICES, , 2011. The vast majority of data on NSS

The international ecosystem survey in the Nordic Seas (IESNS)
The objective of the IESNS survey is to study the abundance and distribution of pelagic fish in the Nordic Seas, where the main focus has been on the NSS herring in relation to environmental and ecological conditions. A detailed description of the surveys can be found in survey reports (ICES, 2009) and a survey manual (ICES, 2014). The surveys were nor-mally carried out in late April to early June by five vessels, one from a European Union country, and the others from the Faroe Islands, Iceland, Norway and Russia. The surveys covered the Norwegian Sea and adjacent waters with parallel east-west transects at 40-60 nmi (nautical miles) intervals, starting from south and heading towards north, and were carried out during 24 h each day. Trawl hauls were performed on acoustic registrations as needed in order to get information on species and length compositions with varying types and size of pelagic trawls. The catches were then sorted and weighed; fish were identified to species level, when possible, and other taxa to higher taxonomic levels. Normally a subsample of 50-100 herring and blue whiting were sexed, aged, and measured for length and weight, and their maturity status was estimated using established methods. An additional sample of 50-250 fish was measured for length. Acoustic estimates of herring and blue whiting abundance were obtained during the surveys. This was carried out by visual scrutiny of the echogram and using post-processing software (Knudsen, 1990;Korneliussen, 2004). This process involved collecting continuous acoustic recordings of fish using calibrated echo integration systems with 38 kHz as the primary frequency. The echograms were analysed at a threshold of −70 dB. The scrutinised echo traces were integrated over the water column and the allocation of NASC (nautical area-scattering coefficient, m 2 nmi −2 ) (MacLennan et al., 2002) values to herring, blue whiting and other acoustic targets were based on the composition of the trawl catches and the appearance of echo recordings in a standardised way. To estimate the abundance, the allocated NASC values were averaged for ICES rectangles (0.5 • latitude by 1 • longitude). For each statistical rectangle, the unit area density of fish in number per square nautical mile (N nmi −2 ) was calculated using standard equations (Toresen et al., 1998) Note that, in 2012, the TS function was changed for blue whiting in all the main acoustic surveys within ICES (TS = 20 log(L) -65.2 dB) (ICES, 2011;Pedersen et al., 2011), but the abundance estimates in the PANGAEA database are based on the previously used function. To estimate the total abundance of fish, the unit area abundance for each statistical square was multiplied by the number of square nautical miles in each statistical rectangle and then summed for all the statistical rectangles within defined subareas and over the total area. Biomass was calculated by multiplying abundance in numbers by the average weight of the fish in each statistical rectangle then summing all rectangles Earth Syst. Sci. Data, 7, 35-46, 2015 www.earth-syst-sci-data.net/7/35/2015/ within defined subareas and over the total area. The Norwegian BEAM software (Totland and Godø, 2001) was used to make estimates of total biomass and numbers of individuals by age and length in the whole survey area and within different subareas.

The pelagic survey in the Bay of Biscay (PELGAS)
Acoustic surveys were carried out every year in the Bay of Biscay in May from 2000 to 2012 (except 2001) onboard the research vessel Thalassa, which is equipped with a Simrad EK60 echo sounder operating at five frequencies (18, 38, 70, 120 and 200 kHz; 7 • beam angle at −3 dB and 1.024 ms pulse length for all frequencies) at 6 m depth on the fixed vessel keel. Only the data collected at 38 kHz were used here. The survey protocol for acoustic data collection has been stable since 2000. Systematic parallel transects (12 nmi distance) perpendicular to the French coast were carried out. The survey covered the continental shelf from 20 m depth to the shelf break about 200 m (in certain years more offshore). Acoustic data were only collected during daytime.
During night the pelagic target species are usually dispersed and found close to the sea surface, and therefore they "disappear" in the blind layer of the echo sounder, which extends between the surface and 8 m depth. The calibration method has been stable over time.
Acoustic data were acquired with the Movies+ and Hermes software and archived in the international hydroacoustic data format (HAC) with a −100 dB threshold applied. The identification of species and size classes comprising fish echo traces heavily depends on identification via trawl hauls performed by R/V Thalassa using a pelagic trawl (two doors; headline: 76 m; foot rope: 70 m). Echograms were scrutinised in real time and trawl hauls were performed as often as possible. The criteria for performing an identification haul included observation of numerous fish echotraces over several elementary distance sampling units (EDSUs) or of very dense fish echotraces in one EDSU, changes in the echotrace characteristics (morphology, density or position in the water column), and observation of an echotrace type fished on previous transects but never fished on the current transect.
The scrutinised echo traces were integrated over the water column by EDSU, providing NASC values. For deriving biomass and abundance estimates, acoustic energies were converted by applying catch ratios, length distributions and weighted by abundance of fish in the area surrounding haul. Further information on the survey and the data analysis methods can be found in Doray et al. (2010).
The objective of the PELGAS survey is to study the abundance and distribution of pelagic fish in the Bay of Biscay. The main target species of the survey are anchovy and sardine, but sprat, horse mackerel, mackerel and blue whiting are also covered. The identification of species and size classes comprising fish echotraces heavily depends on iden- tification via trawl hauls performed by R/V Thalassa using the pelagic trawl (Doray et al., 2010).

The Bay of Biscay and Celtic Sea bottom trawl survey (EVHOE)
The EVHOE survey has been conducted onboard R/V Thalassa annually in autumn since 1987 in the Bay of Biscay and since 1997 in the Celtic Sea. The vessel was changed in 1997, but this did not impact the catchability of blue whiting (Pelletier, 1998). Between 50 and 100 stations are trawled with a grande ouverture verticale (GOV) bottom trawl according to a stratified random design based on bottom depth and latitude. The catch is identified to species level and individually measured. Further details on the survey protocol can be found online at https://datras.ices.dk/Documents/ Manuals/EVHOEManual.doc.

Bluefin tuna catch data in northern European waters (ICES areas II-VII)
Bluefin tuna catch data were extracted from the ICES catch databases for the time periods 1906-1949 and 1950-2010. The catch database versions used in this report were those available on the ICES website (http://www.ices.dk/marine-data/dataset-collections/Pages/ Fish-catch-and-stock-assessment.aspx) and downloaded on 21 January 2014. The data in the ICES catch databases are derived from ICES Bulletin Statistiques; these data are annual statistical reports containing fishery data for northeastern Atlantic countries (Lassen et al., 2012). The data contained in those reports have been entered by ICES into the existing catch database. The data contained in the ICES catch database are resolved by species, year, country and sea region (known as ICES areas). For this report, we were most interested in the sea areas north of the Bay of Biscay, because this was the former summer foraging habitat until its disappearance in the 1970s and because catches in this region have undergone large fluctuations (see above) whose reasons remain unclear (Tiews, 1978;Mather et al., 1995;MacKenzie and Myers, 2007;Fromentin, 2009). We extracted data by year, country and region and made time series plots of the data to illustrate and compare temporal- spatial trends and to identify which countries and/or regions had the largest catches.
We are also aware of some historical catch data which are not presently included in the ICES catch databases; these data have been identified in various fishery reports, museum records and scientific literature (MacKenzie and Myers, 2007), and are included in the database described here and available at the PANGAEA website. The data sources not included in the ICES catch database are identified with notes in the online data files.
The ICES areas considered are only part of the entire spatial domain over which the bluefin tuna fishery is managed, which covers the North Atlantic east of 45 • longitude and includes the Mediterranean and Black seas (ICCAT, 2012). Catches for these areas (i.e. northeastern Atlantic and Earth Syst. Sci. Data, 7, 35-46, 2015 www.earth-syst-sci-data.net/7/35/2015/ Mediterranean, including the Black Sea) are available from the International Commission for the Conservation of Atlantic Tunas (ICCAT), which is the agency responsible for management of bluefin tuna in the Atlantic and Mediterranean. We extracted and plotted the ICCAT catch data (total international landings) to enable comparison of the total international landings in the entire stock management area with the ICES data for the northern region. In principle, the catch data reported to ICES should be the same as those reported to ICCAT, but in some instances there may be some minor differences (Lassen et al., 2012), which are not considered further here.

Results and discussion
3.1 Spatially explicit estimates of stock sizes, structure and biomass of NSS herring The centre of gravity for adult NSS herring shifted substantially from 1996 to 2011 (Fig. 5). From 1996 to 2003 the centre of gravity moved north/northeast by almost 4 • latitude. This represents a distance of 445 km. From 2003 to 2011 a similar shift of 4 • latitude (445 km) in centre of gravity in the opposite direction towards south/southwest was evident, even though the survey coverage has remained rather stationary during this time (Utne et al., 2012). Maps of herring distribution and aggregation patterns, primarily from the international May survey from 1995 to 2011 are provided below based on acoustic estimates (NASC values) in forms of mean 5 nmi values (Fig. 6). NSS herring has changed its distribution and migration pattern substantially during the last decades in terms of spawning, feeding and overwintering areas (Fig. 1). Data on herring distribution and aggregation patterns from the IESNS survey from 2004 to 2012, are provided in PANGAEA. Examples of the data on herring distribution from this survey are shown in Fig. 6. For a more comprehensive review of the distribution of NSS herring, see Utne et al. (2012). There are various sources of uncertainty related to acoustic estimates for pelagic fish stocks. For NSS herring, inadequate coverage of the distribution area and active vessel avoidance by the herring are the two most important factors contributing to uncertainty in the acoustic abundance estimate (Løland et al., 2007).

Spatially explicit estimates of stock sizes, structure and biomass of blue whiting
The blue whiting SSB has fluctuated widely over the last three decades. Examples of the spatial distribution of the stock during the IESNS in May are shown in Fig. 7. In 2004 there are some high-density areas outside northern Norway and further south between the Shetland and Faroe islands and Iceland. In 2009, on the other hand, the stock size is much lower and consequently no comparable high-density areas are seen, and the stock is less widespread than in 2004.

Spatial blue whiting estimates in the Bay of Biscay
Spatial blue whiting biomass estimates in the Bay of Biscay for two contrasting years are shown in Fig. 8. The year 2002 was a typical year with little blue whiting biomass in the water column on the shelf, while in 2010 blue whiting was found in high abundance on the continental shelf. It should be noted that, as the survey is targeting anchovy and sardine, the transect lines end in general where schools of blue whiting are encountered. As adult blue whiting are distributed along the shelf edge further offshore, the PELGAS surveys do not cover the full distribution of this species in the Bay of Biscay. Hence it is difficult to distinguish a situation where blue whiting are more spread-out over the continental shelf because of favourable environmental conditions from a situation where  spreading is caused by higher densities, or a combination of both.
The gridded average distribution of blue whiting across the years 1997-2011 is shown in Fig. 9. This map of the demersal part of the blue whiting stock, primarily made up of young of the year, shows that they are distributed along the  outer parts of shelf edge of the Bay of Biscay and Celtic Sea, similar to the pelagic adult part of the population in Fig. 8 There is a notable omission of data from the version of the ICES catch database downloaded for this work. The data omitted are those from Norway. These data are, however, present in the original ICES Bulletins Statistiques and have been reported earlier in the literature (e.g. MacKenzie and Myers, 2007), and they are also included in the ICCAT database. The omitted landings are substantial because Norway had the largest catches of all countries in the region considered in this report until the fishery in this area declined in the mid-1960s to early 1970s. Moreover, Norway was the first country in this region to report its landings to ICES (starting in 1927;ICES, 1903ICES, -1972. The Norwegian landings which are in the ICES , but which are omitted from the ICES catch database, are those for 1927-1949 for areas II, III and IV. Aside from the omission of early officially reported Norwegian landings, there are additional landings by Norway (Tangen, 1999) and a few other countries from the early decades of the 20th century which are not included in the ICES  or the present version of the ICES catch database. The earliest bluefin tuna catches noted (1907) were those by a French herring boat fishing at Dogger Bank in the southern North Sea. Several more years of French tuna catches in this area are available in French fishery reports (Statistiques de Pêches Maritimes; MacKenzie and Myers, 2007). Norwegian, Swedish and Danish fishermen also caught bluefin tuna in the Norwegian Sea and Skagerrak-Kattegat in the 1910s-1920s, and thus before these governments began reporting the landings to ICES (Fig. 12). In comparison with available catch data from ICCAT, the ICES data compilation presented here covers a longer time period and has higher spatial resolution (e.g. to regional sea level); however its geographic range is limited to waters of northern Europe, whereas the ICCAT database contains data from countries of southern Europe, including the Mediterranean Sea. A summary of historical bluefin tuna catches in the northwestern European waters with maps of catch locations is presented in LeGall (1927).
The vast majority of the catches (96 % by weight in the whole time period 1906-2010) were taken in areas II-IV, and the rest were in areas VI-VII. Norway, Denmark, Germany, Sweden and France were responsible for 99.8 % of the total reported landings during 1906-2010; Norway's share was largest (73 %), followed by Denmark (11 %, Fig. 13).