General enquiries on this form should be made to: - Defra, UKrandd.defra.gov.uk/...M1101_finalreportv3(2).docx · Web viewThis form is in Word format and the boxes may ... we used

General Enquiries on the form should be made to:Defra, Procurements and Commercial Function (Evidence Procurement Team)E-mail: [email protected]

Evidence Project Final Report

NoteIn line with the Freedom of Information Act 2000, Defra aims to place the results of its completed research projects in the public domain wherever possible. The Evidence Project Final Report is designed to capture the information on the results and outputs of Defra-funded research in a format that is easily publishable through the Defra websiteAn Evidence Project Final Report must be completed for all projects.

This form is in Word format and the boxes may be expanded, as appropriate.

ACCESS TO INFORMATIONThe information collected on this form will be stored electronically and may be sent to any part of Defra, or to individual researchers or organisations outside Defra for the purposes of reviewing the project. Defra may also disclose the information to any outside organisation acting as an agent authorised by Defra to process final research reports on its behalf. Defra intends to publish this form on its website, unless there are strong reasons not to, which fully comply with exemptions under the Environmental Information Regulations or the Freedom of Information Act 2000.Defra may be required to release information, including personal data and commercial information, on request under the Environmental Information Regulations or the Freedom of Information Act 2000. However, Defra will not permit any unwarranted breach of confidentiality or act in contravention of its obligations under the Data Protection Act 1998. Defra or its appointed agents may use the name, address or other details on your form to contact you in connection with occasional customer research aimed at improving the processes through which Defra works with its contractors.

Project identification

1. Defra Project code MF1101

2. Project title

Evaluating shelf-wide spatial and temporal changes in fish larval distribution over the last half century in relation to environmental factors and adult distributions

3. Contractororganisation(s)

Dr Sophie PitoisCefasPakefield roadNR33 0HTLowestoftSuffolk

54. Total Defra project costs £ 885,159(agreed fixed price)

5. Project: start date................. 01/07/2007

EVID4 Evidence Project Final Report (Rev. 06/11) Page 1 of 40

mailto:[email protected]

end date.................. 31/03/2012


6. It is Defra’s intention to publish this form. Please confirm your agreement to do so...................................................................................YES NO (a) When preparing Evidence Project Final Reports contractors should bear in mind that Defra intends that

they be made public. They should be written in a clear and concise manner and represent a full account of the research project which someone not closely associated with the project can follow.Defra recognises that in a small minority of cases there may be information, such as intellectual property or commercially confidential data, used in or generated by the research project, which should not be disclosed. In these cases, such information should be detailed in a separate annex (not to be published) so that the Evidence Project Final Report can be placed in the public domain. Where it is impossible to complete the Final Report without including references to any sensitive or confidential data, the information should be included and section (b) completed. NB: only in exceptional circumstances will Defra expect contractors to give a "No" answer.In all cases, reasons for withholding information must be fully in line with exemptions under the Environmental Information Regulations or the Freedom of Information Act 2000.

(b) If you have answered NO, please explain why the Final report should not be released into public domain

Executive Summary7. The executive summary must not exceed 2 sides in total of A4 and should be understandable to the intelligent

non-scientist. It should cover the main objectives, methods and findings of the research, together with any other significant events and options for new work.

The development of an ecosystem-based approach to fisheries management presents a great challenge to scientist and policy makers given the substantial uncertainties relating to habitat distribution, natural variation in the marine environment, impact of anthropogenic effects and climate change. There is increasing concern that climatic changes are having significant effects on marine ecosystems and fisheries, such as declines in traditional commercial species such as cod and increases in warmer water species. However, there is considerable uncertainty about the resilience of fish to environmental change and many of the observed shifts in distribution of fish could be the result of exploitation patterns as well as environmental changes. Fish larvae suffer high rates of mortality and are particularly sensitive to temperature, food availability (plankton) and predation. The survival of larvae has a strong effect on the subsequent strength of cohorts and the long-term sustainability of populations. Larval survival is highly dependent on the degree to which the distribution and timing of fish larval production and their planktonic food co-occur – the match/mismatch hypothesis. The main purpose of this project was to gain new insights into long-term changes in fish larval ecology in relation to factors such as sea temperature, ocean currents, fishing pressure and prey availability (phytoplankton, zooplankton). To achieve this we created the world’s largest dataset (in terms of geographical and temporal extent) on the distribution and abundance of larval fish based on samples collected in the north-east Atlantic since 1948 by the Continuous Plankton Recorder (CPR). Specific objectives were set out below:

Objective 1: To transfer hard-copy records of larval fish species from 1948-1978 into a modern relational database (records from approximately 10,000 CPR samples).

Objective 2 To identify larval fish in CPR samples for the north-east Atlantic from 1979-2006 (approximately 10,000 samples) and enter the records into the electronic database. The fish larvae project was an ambitious study to identify the fish in CPR samples covering the period 1979-2005. This required re-analysing over 10,000 archived CPR samples and recording over 32,000 identified specimens. The project has resulted in the creation of a unique fish larvae database of the NE Atlantic, North Sea and adjacent areas from 1948-2005, containing a total of 134,260 sampling points. In terms of diversity, the CPR fish larvae database consists of 75 taxonomic groups or species. By far the most common taxonomic groups over the European shelf are Clupeids and sandeels, followed by blue whiting and Atlantic mackerel. Early data exploration revealed some limitations resulting from the sampling methodology: (i) only the North Sea and the Celtic Sea area were sampled consistently and intensively enough to allow for further analysis of the data; (ii) under-sampling of the CPR for fish larvae resulting in a


high number of zero-values within the dataset. In order to obtain robust time-series, we had to further limit statistical analysis to the most abundant taxonomic groups within a given area.

Objective 3 To gather and process available environmental and trawl survey data into suitable format for matching and statistical analysis alongside the new fish larval data.

In order to examine links between spatio-temporal changes in fish larval abundance and distribution with external factors, we identified and gathered the following data: (i) climate indices over the northeast Atlantic; (ii) hydrographic indices (sea surface temperature, salinity, thermocline depth, oceanic currents, sub-surface temperature and turbulence); (iii) information on plankton as potential prey for fish larvae, including a general index of primary production and abundances of various key groups of plankton (diatoms, dinoflagellates, copepods, decapod larvae); (iv) fisheries data, including abundances of adult fish from international surveys and stock assessment indices.

Objective 4 To use the new larval database to describe the long-term changes in abundance, distribution and phenology of fish larvae (1948-2006) across the UK Shelf Seas. An atlas of fish larvae (Fish larvae atlas) was produced for all species caught in at least 150 CPR samples (1948-2005) over the European shelf: clupeids (herring, sardines, sprats), sandeels, Atlantic mackerel, blue whiting, dab, whiting, lantern fishes, pipefish, cod, dragonets, Norway pout, gobies, saithe, plaice, Atlantic horse mackerel and haddock. Long-term trends in abundance and distribution, as well as potential phenological and latitudinal change were described for each of the above types of fish larvae. The fish larvae atlas provides a useful insight in the spatio-temporal changes undergone by fish larvae in the last 60 years. Unfortunately the monthly resolution of the dataset makes it difficult to detect potential match-mismatch between fish larvae and their prey resulting from phenological and latitudinal shifts as these tend to be subtle. Furthermore, basic exploration of the long term trends of fish larvae abundances in relation to adult data from IBTS survey and stock assessments showed that gaps in the coverage and limitations of the CPR as a tool to catch fish larvae result in a limited number of taxa that can be more extensively studied. However, as the CPR samples continuously, it can still pick up larvae often enough to give a representation of areas with high and low abundance; for the most abundant taxa at least, the CPR is a reliable tool for capturing the spatial patterns and the relative abundances between areas.

Objective 5 To investigate linkages between fish larval abundance, geographical ranges (spawning site stability) and phenology (timing of production); mature adult distributions and environmental factors (sea temperature, ocean currents, phytoplankton, zooplankton and fisheries).

We initially adopted a general approach by investigating the impact of environmental changes in the North Sea on five selected taxa of fish larvae during the period 1960–2004. The analysis revealed four periods of time (1960-1976; 1977-1982; 1983-1996; 1997-2004) each reflecting a specific ecosystem status. The larvae of clupeids, sandeels, dab and gadoids changed concomitantly with the plankton, while the larvae of migratory species such as Atlantic mackerel responded more to hydrographic changes. Climate variability is more likely to influence fish populations through bottom-up control cascading through the ecosystem; i.e. changes in the NAO impact the hydrodynamic features of the North Sea, which impact on the plankton (i.e. food) available for fish larvae. We also identified regime shifts in 1986 and 1994 in the Celtic Seas ecoregion. Preliminary studies indicate that these regime shifts appear coupled with phenological changes in copepod species. The earlier regime shift was characterised by a change in the climate, followed by a rise in phytoplankton standing stock in the Irish Sea and north eastern Celtic Sea and a fall in the recruitment of plaice and cod stocks. The later shift was triggered by a rise in temperature, and coincided with a general rise in phytoplankton standing stock, a fall in total fish larvae sampled by the CPR in the Celtic Sea and an increase in the western Channel, and, two years later, by a widespread fall in total copepod abundance.We then examined further the relationships between the larvae of sandeel, dab and mackerel with existing data on egg surveys and fisheries-based assessments. The CPR has the clear advantage of covering a longer spatio-temporal span of interest compared to existing datasets. In each case we showed that the CPR can produce reliable time-series in areas where larvae are consistently abundant and the data can inform recruitment levels for the North Sea stock assessments. In the case of mackerel, we further designed a larval index data. The resulting timeseries documents the significant decrease of spawning from before 1970 to recent depleted levels in the North Sea. Spatial distributions of the larvae, and thus the spawning area, showed a shift from early to recent decades, showing that the central North Sea is no longer as important as the areas further west and south. These results provide a consistent and unique perspective on the dynamics of mackerel in this region and can potentially resolve many of the unresolved questions about this stock.

Overall, we have shown that, the CPR fish larvae dataset, for the most abundant taxa at least, is a reliable tool for capturing the spatio-temporal patterns and the relative abundances of fish larvae between


areas. This new dataset clearly offers a unique opportunity to investigate the responses of fish populations to past changes in climate, including abrupt changes such as regime shifts; and can also inform recruitment levels for stock assessment purposes. Nevertheless, the dataset also contains some limitations, mainly the impossibility to separate the several species of clupeids, and the lack of information post 2005. However, both these limitations could theoretically be overcome; as CPR samples post 2005 have been archived and recent advances molecular analysis of CPR archived material now allow molecular identification of fish larvae, and genetic probes have been developed for herring and anchovy.

Project Report to Defra8. As a guide this report should be no longer than 20 sides of A4. This report is to provide Defra with details of

the outputs of the research project for internal purposes; to meet the terms of the contract; and to allow Defra to publish details of the outputs to meet Environmental Information Regulation or Freedom of Information obligations. This short report to Defra does not preclude contractors from also seeking to publish a full, formal scientific report/paper in an appropriate scientific or other journal/publication. Indeed, Defra actively encourages such publications as part of the contract terms. The report to Defra should include: the objectives as set out in the contract; the extent to which the objectives set out in the contract have been met; details of methods used and the results obtained, including statistical analysis (if appropriate); a discussion of the results and their reliability; the main implications of the findings; possible future work; and any action resulting from the research (e.g. IP, Knowledge Exchange).


Introduction

Impacts on climate change on marine ecosystems have become increasingly evident from polar to tropical environments and affect many marine organisms and ecosystem functions. Biological responses reported, include: changes in phenology (timing of appearance) leading to mismatch between successive trophic levels1; changes in the spatial distribution of primary and secondary pelagic production placing additional stress on already depleted fish populations2; regime shifts and changes in community structures of plankton species3; latitudinal shifts in marine fish with potential impacts on fisheries and species richness4; local species extinctions and invasion of non native species5; and changes in spawning and migration patterns of some fish species with consequences for fisheries management 6. Climate change clearly affects many marine organisms and ecosystem functions. In order to promote biodiversity and sustainable fisheries, it is important to further our knowledge and understanding of how climate variability influences the early life history of fish via impacts on marine foodwebs and the interactions between species and their changing environment. The abundances of fish stocks are highly variable over time 7, and recruitment processes are particularly difficult to monitor. The health of a fish stock depends on the number of larvae recruited to the adult population each year, and the survival of larvae is key to recruitment success and variability8. As part of the plankton, fish larvae are sensitive to physical and biological changes in their environment9.

The Continuous Plankton Recorder (CPR) survey, which started in the North Sea in 1931, is one of only a few long-term biological monitoring programmes worldwide, and the only one that gives systematic spatial coverage of the North Sea10,11. The utility of this dataset has been established with the many, and increasingly complex, studies published, aiming to relate spatio-temporal changes in plankton distribution to ecosystem changes. It has also provided valuable information concerning the scale and nature of the

processes affecting fish stocks12. As well as zooplankton and phytoplankton, fish larvae were analysed until the 1970s13,14, but then economic constraints caused the analysis of fish larvae in CPR samples to be stopped.The CPR (Fig. 1) is towed by ships of opportunity at speeds of 15–20 knots, at an approximate depth of 10 m. Water enters the recorder through an aperture of 1.27 cm2, and is filtered through a continuously moving band of silk with an average mesh size of 270 μm. Each sample represents ~3 m3 of filtered seawater. Methods of counting and data processing have previously been described15,16.

Figure 1: CPR instrument

The newly available fish larvae dataset from the CPR survey offers a unique opportunity to investigate long-term changes over decadal scales in the abundance and distribution of fish larvae in relation to physical and biological factors. We designed our study with the aim to evaluate how different species of fish have responded to past environmental changes at the critically important larval stage. We first describe the long-term changes in abundance, distribution and phenology of the selected fish larvae (1948-2005) across the UK Shelf Seas; we then assess the reliability of the CPR to catch fish larvae and evaluate the use of the dataset as a tool to aid in the assessment of commercial fish stocks, with case studies on selected key species. Finally, we attempt to link variations in climate, oceanography and plankton production to these changes, and evaluate whether fish larvae respond directly to hydro-climatic variability or indirectly through interaction with other trophic levels.

I. Objectives 1 & 2: Transfer of paper records of fish larvae (1948-1978) to database and identification of fish larvae in samples (1979-2006).(achieved)

I.1. Content of the fish larvae database

The CPR survey records over 500 taxonomic entities and although the number of fish larvae on each sample is counted, they are not routinely identified to species or taxonomic group. The fish larvae project was an ambitious study to identify the fish larvae and transfer the results to a database incorporating the data from an earlier CPR Fish Atlas17. This required re-analysing over 10,000 archived CPR samples and recording over 32,000 identified specimens. The project has resulted in the creation of a unique fish larvae database of the NE Atlantic, North Sea and adjacent areas from 1948-2005, containing a total of 134,260 sampling points, including 128,314 points covering the European shelf delimited by latitudes 45°N to 65°N and longitudes 20°W to 10°E (Fig. 2). Unfortunately fish larvae identification had to be carried out on a consultancy basis, following staff departures from SAHFOS and as a result of budget constraints the measurement of fish larvae could not be done as part of this project.


Figure 2: location of 134,260 sampling points from CPR fish larvae dataset over the period 1948-2005

In terms of diversity, the CPR fish larvae database consists of 75 taxonomic groups or species. By far the most common taxonomic groups over the European shelf are Clupeids and Sandeels. Blue whiting and Atlantic mackerel also appeared in over 1000 samples (Fig. 3).

Young fish are some of the largest organisms sampled by the CPR survey, typically ranging from 2 mm to 50 mm (mean size ~12 mm), with the CPR (although selective) sampling a comparable part of each population from year to year. Due to the size of the fish larvae and the sampling method, they can often be damaged and identification to species level is not always possible using traditional microscopic methods as attempted in this project. The difficulty in visually analyzing fish larvae in CPR samples is highlighted by 3620 occurrences where identification was not possible, 1893 occurrences where clupeid and sandeel larvae could not be distinguished from each other and 775 occurrences where gadoids could not be identified to species level.

I.2. Potential biases and limitations of the CPR fish larvae dataset

I.2.1 Spatio-temporal coverage Different areas were sampled with different intensities (SAHFOS atlas). From early data exploration, it became clear that only the North Sea, English Channel and Celtic sea were sampled with enough intensity and consistency to undergo further data analysis:

The Bay of Biscay and NE Atlantic region is poorly and sparsely sampled (~20,000 records), making large spatio-temporal gaps in the dataset unavoidable, and impossible to use consistent statistical analysis. This area could therefore not be studied as an entity.

The English Channel and Celtic Sea comprises ~17,000 records, but the area is small with fairly consistent sampling.

The west of Scotland and Irish Sea region (~37,000 records) is sparsely sampled with major spatio-temporal gaps in the data, making it also difficult to study as an entity.

The North Sea is the most intensively and consistently sampled area (~55,000 records) and therefore represent the area of highest value from an analytical perspective.


ClupeidsSandeels

UnidentifiedClupeids/SandeelsAtlantic mackerel

Blue whitingDab

RocklingsDragonets

Unidentified gadoidsWhiting

FragmentsScopelids (myctophyd)

Norway poutLantern fishes

GobiesSaithe

European plaiceAtlantic horse mackerel

CodPleuronectids

SoleidsHaddock

Pearlsides (lantern fishes)Triglids

0 1000 2000 3000 4000 5000 6000

58174009

292118931800

1143839825791775747699

546461429427

311295266200184180164161156

Number of positive samplesFigure 3: Relative contribution of taxonomic groups to CPR fish larvae dataset over the European shelf seas. Only the taxa appearing in at least 150 out of the 128,314 samples with fish larvae present.

I.2.2. Undersampling of the CPR instrument for fish larvae The CPR instrument with its very small aperture and fixed sampling depth is bound to substantially undersample fish larvae. This is demonstrated by the very high number of zero-values (non occurrence) within the dataset. All identified taxa could be mapped, but for the less common taxa, as the number of positive occurrences decreases within the samples, the likelihood of the signal being a random noise rather than a true representation of that taxon’s presence increases. In order to obtain robust time-series to ensure that further data analysis as set by the objectives of this project could be performed, we considered the following when selecting fish larvae taxa for study: (i) the number of positive samples and density of these samples’ occurrences, and (ii) the spatial and temporal distribution of fish larvae from the CPR dataset in relation to known spawning seasons and locations. Consequently only the most abundant taxa within a given area could be studied. The efficiency at catching fish larvae and the reliability of the data is discussed within Objectives 4 with a case studies on sandeel larvae.

I.2.3. Pycnocline depth Due to the subsurface sampling of the CPR, the survey cannot take into account long-term changes in the depth of the pycnocline (boundary separating surface water from deeper water defined by large density difference which effectively prevents vertical currents, and is influenced by temperature and salinity). If there were significant changes to the depth of the pycnocline over a decadal time scale, this could potentially bias CPR results. Due to the lack of sufficient physical data at the same scales of this study, this potential bias could not be tested. Therefore, generalizations based on this study concerning long-term changes in ichthyoplankton taxa must be treated with some caution. It is worth noting, however, that the water immediately behind a large, fast-moving vessel is likely to be mixed and homogenized well below the CPR towing depth16.


II. Objective 3: Environmental, biological and trawl survey data.(achieved)

The following data were gathered and used to complete Objectives 4 and 5. Whenever possible, these datasets had to be transformed to a consistent format covering spatio-temporal scales matching the CPR dataset. When the data was already on a gridded format, it was thought better to maintain these formats rather than re-grid the data to match the grid for CPR data. Adult fish data (abundances per haul) were gridded on the same grid as the CPR data (see Objective 4 for procedure).

II.1 Climate indices

The following climate indices with centres of action over the Northeast Atlantic were downloaded from http://www.cgd.ucar.edu/cas/jhurrell/indices.html:

North Atlantic Oscillation Index (NAO): normalized sea-level pressure difference between a station on the Azores and one on Iceland18.

East Atlantic pattern (EA): second prominent mode of low-frequency variability over the eastern north Atlantic. Although the pressure dipole of the EA is shifted southward relative to that of the NAO it is structurally similar, with the notable exception of a strong subtropical link in the EA pattern that is not found in the NAO.

Scandinavian pattern (SCA): primary circulation centre over Scandinavia with weaker centres of opposite sign over Western Europe and Eastern Russia/Western Mongolia.

Atlantic Multidecadal Oscillation (AMO): defined by long term unsmoothed variability in sea surface temperature, and determined from the Kaplan SST V2 dataset19 calculated at NOAA/ESRL/PSD1.

II.2 Hydrographic indices

II.2.1 Sea Surface Temperature (SST) Gridded monthly SST data, covering the North Atlantic for the period 1960-2005, were obtained from the ICOADS 1-degree enhanced dataset provided by the NOAA–CIRES Climate Diagnostics Center (http://www.cdc.noaa.gov). We also used temperature data obtained from CTD profiles at sampling locations whenever possible.

II.2.2 Surface Salinity Gridded monthly surface salinity data for 0.111° × 0.167° (latitude × longitude), covering the North Atlantic for the period 1960-2004, were obtained from the POLCOM model through collaboration with Dr Jason Holt, from the National Oceanography Centre (UK). We also used salinity data from CTD profiles at sampling locations whenever possible.

II.2.3 Thermocline data The thermocline is the boundary beneath the relatively warm, well-mixed surface layer of water in which interval temperatures diminish steadily. The deep waters below the thermocline layer decrease in temperature much more gradually toward the seafloor. Thermocline depth data for the period 1948-2005 were processed from a long-term ECOSMO model run20. The thermocline data are provided on a grid 0.1° x 1/6° grid (latitude x longitude) as monthly averages.

II.2 4 Oceanic currents, sub-surface temperature and turbulence As a basic of calculations of drift and advection of fish larvae, we apply an established hydrographic backtracking technique21. The backtracking calculation was performed using a model forced with hourly physical fields (currents, temperature and turbulence) derived from the NORWECOM model22. These fields were available from 1970 to 2005.

II.3 Biological data

II.3.1 Plankton chlorophyll index (CHLO) from CPR CHLO was extracted, covering the North Atlantic and the period 1948-2007. This index of greenness provides an indication of the quantity of primary production and hence phytoplankton, and has also been referred to as the Plankton Colour Index.

II.3.2 Dinoflagellate and diatom abundance from CPR Information on the abundance of the larger groups of phytoplankton, namely diatoms and dinoflagellates was extracted, covering the North Atlantic and the period 1948-2007. Abundances of individual species were collated within each of these two groups to create 2 separate indices of abundance.


http://www.cdc.noaa.gov/

http://www.cgd.ucar.edu/cas/jhurrell/indices.html

Limitations and potential bias: Due to the mesh size of CPR silks, many phytoplankton species are only semi-quantitatively sampled due to the small size of the organisms. There is thus a bias toward recording larger armoured flagellates and chain-forming diatoms, and smaller species abundance estimates from cell counts will probably be underestimated in relation to other sampling methods. However, the proportion of the population that is retained by the CPR silk reflects the major changes in abundance, distribution, and composition (i.e., the percentage retention is roughly constant within each species even for very small-celled species23).

II.3.3 Zooplankton abundance (from CPR)

We extracted the abundance data for individual species of copepods and cladocerans, covering the North Atlantic and the period 1948-2007. These two groups were selected because they are important prey of many fish larvae in their early stages of life. The CPR underestimates zooplankton abundance compared with other datasets24. Because we thought it important to take into account the relative contribution of each particular potential prey species to the total zooplankton abundance, undersampling was corrected using species-specific WP2/CPR ratios25. We gathered together these species-specific corrected abundances to create an index of zooplankton abundance. Data for decapods larvae were extracted as a separate index.

Limitations and potential bias: Fish larvae caught by the CPR are very small; for example, the larvae of Atlantic mackerel caught by the CPR were reported to be <7 mm long26, on average 5 mm. Such larvae are very young and likely to feed on the smallest prey available, such as phytoplankton and copepod eggs and nauplii. The zooplankton index from the CPR has been corrected for undersampling of the CPR, but it is still only based on the larger copepodite and adults and does not include naupliar stages. For lack of better information, it seems reasonable to assume that abundances of copepod nauplii are linked to similarly high abundances of copepods during summer, as shown here.

II.4 Fisheries data

II.4.1 International Bottom Trawl Survey in North Sea (NS-IBTS) and West of Scotland (WS-IBTS) Information on the relative abundance of principally demersal and benthic adult fish species fish was extracted for all four quarters, but temporal coverage varies from quarter to quarter, data from the first quarter (January to March) offer the most extensive spatio-temporal coverage (Table 1).

Survey Coverage Number of datapointsQuarter 1 (January to March) NS-IBTS 1967-2010 45,901

WS-IBTS 1985-2010 11,455Quarter 2 (April to June) NS-IBTS 1991-1997 6,863

WS-IBTS 1995-2010 137Quarter 3 (July to September) NS-IBTS 1991-2010 19,201

WS-IBTS 0Quarter 4 (October to December) NS-IBTS 1991-2010 3,583

WS-IBTS 1990-2010 7,265Table 1: IBTS survey dataset


Information on maturity and growth pattern could not be obtained in a complete manner; exploration of the DATRAS database revealed major gaps in the data and inconsistencies between surveys. Other sources of information exist, but only cover major commercial species. Similarly, we attempted to obtain data prior to the 1970s, but these also covered major commercial species only, which are not best sampled by the CPR. Both data also contain major gaps and the sampling technique used was not consistent with that of the IBTS. The poor quality of the data combined with the limitation of the CPR data lead us to decide not to use these for further analysis.

II.4.2 Stock assessment data from ICES Stock and recruitment indices along with international landings and mean mortality data were obtained for commercial species and downloaded from ICES website at http://www.ices.dk/datacentre/StdGraphDB.asp

III. Objective 4: long-term changes in abundance, distribution and phenology of fish larvae (1948-2005) across the UK Shelf Seas(achieved)

III.1 Data preparation

The CPR survey collects samples at different times of day and at locations that do not follow a regular grid. All CPR data therefore need to be regularized in time and space before being subjected to numerical analyses10. This was undertaken on a 50 x 50 nautical mile grid using Inverse Distance Weighing Interpolation27. CPR data also show major seasonal and diel patterns28 and these were taken into account when calculating annual values. The samples covering our area and period of study were grouped into day and night periods for each month and year; day/night periods were defined from calculation of sun zenith angle at each sampling location. Spatial interpolation was performed for each month for both day and night, resulting in 24 matrices per year that were then combined to produce monthly and annual matrices covering the European shelf29. To calculate monthly and yearly means within a specific area, we averaged the data from the above matrices produced within that area.

III.2 Atlas of fish larvae

An atlas was produced by SAHFOS. This atlas contains the distributions and time-series for the 5 most common taxa from the central and eastern North Atlantic) with an additional 4 included of particular current interest. Taxa included in the atlas are: clupeids (herring, sardines, sprats, anchovies), sandeels (Ammodytidae), Atlantic mackerel (Scomber scombrus), blue whiting (Micromesistius poutassou), dab (Limanda limanda), whiting (Merlangus merlangius), lantern fishes (Myctophydae), pipefish (Syngnathidae), and cod (Gadus morhua). We have further extended this atlas to the following taxa: rocklings (Onos sp.), dragonets (Callionimus sp.), Norway pout (Trisopterus esmarkii), gobies (Gobidae), saithe (Pollachius virens), plaice (Pleuronectes platessa), Atlantic horse mackerel (Trachurus trachurus), and haddock (Melanogrammus aeglefinus). The remaining taxa were identified in less than 150 samples only and were not thought to be abundant enough for mapping and further study. With each taxon we have included a general description of long-term spatio-temporal changes and potential shifts in phenology. The atlas can be access via this link: Fish larvae atlas.pdf. Overall a general decrease in the abundance of fish larvae caught by the CPR can be seen from the period 1948-1985 to 1986-2005; out of the 17 taxa presented in the atlas, only three increased in abundance: (i.e. pipefish, plaice and horse mackerel). The increase in pipefish is very recent, and plaice has further undergone a phenological shift with the larval peak (1986-2005) appearing 2 months earlier than during the period 1948-1985.

III.3 Environmental changes, match/mismatch with potential prey

The distributions and time-series for SST, surface salinity, total zooplankton and the three indices of phytoplankton (i.e. CHLO, dinoflagellates and diatoms) are presented in a similar format as that for fish larvae and were included in the material produced. Overall, SST increased during the winter and summer months while salinity decreased. A noticeable increase in phytoplankton can be seen from the first (1948-1985) to the second period (1986-2005), but at the same time abundances of diatoms and dinoflagellates decreased, suggesting that smaller phytoplankton species contributed to that increase in chlorophyll. Zooplankton generally decreased in particular during the period May to October. In order to examine match-mismatch between fish larvae and their potential prey, we visually analysed maps of monthly distribution for each fish larvae taxa in comparison with equivalent maps for zooplankton, primary production, dinoflagellates and diatoms. The results were inconclusive because in many cases the


http://www.ices.dk/datacentre/StdGraphDB.asp

resolution of the CPR coverage was too limited and/or the abundance of fish larvae was too low, or too widespread. Hence, we look at the catchability of the CPR instrument and the reliability of the dataset produced and each of the main taxa caught by the CPR.

III.4 Catchability of the CPR

III.4.1 Effect of increasing ship speed Overall, the speed of ships towing the CPR has steadily been increasing from approximately 10.5 knots in 1952 to 14.8 knots in 1999. This increase in tow speed was shown to have no effect on the depth of sampling and the mechanical efficiency of the internal mechanism, but at the highest tow speeds there is some evidence that flow may be reduced, but this cannot yet be quantified16.

III.4.2 Case study using sandeel larvae The reliability of a plankton sampler with such a small aperture as the CPR to study fish larvae is debatable. Therefore it was necessary to verify the reliability in the CPR data against other available data sources. As a case study we selected sandeels as an abundant taxon, second only to clupeids and with less of an issue regarding species identification, and the North Sea as the area with the most consistent sampling coverage.

In order to verify the spatial patterns the CPR data were compared statistically and visually to ICES egg and larval surveys in the North Sea. These surveys took place between June and April 2004 and 2009 30. GULF samplers and BONGO nets were deployed in double oblique hauls for a minimum of 15 minutes (multiple hauls in shallow water), down to two metres above seafloor or a maximum of 100 m. The spatial structure in the CPR data (mean abundance/m3 by ICES rectangle, using February-June data for all available years 1950-2005) was compared to that observed during February-April in the international plankton surveys (Fig. 4). A Mantel correlogram31 was calculated to test the significance of spatial correlation between the two datasets given the spatial structure in the data (Table 2).

Figure 4: Sampling areas and sandeel abundance (A) CPR sample locations (light grey points) and presence of sandeel larvae (dark grey circles) from all transects 1950-2005; solid black lines show North Sea sandeel assessment regions, the sub-region ‘South of Shetland’ and additional areas in the Irish Sea and Celtic Sea (dashed black lines). (B) Mean abundance of larvae in CPR samples; with cells in the: North Sea and eastern Channel (east of 2°W) based on February-June data only; Celtic Sea (south of 52.5°N) and western Channel (west of 2°W), March-July data; Irish Sea, north of 52.5°N, April-August data.

Positive spatial correlation between CPR (1950-2005) and ICES (2004 and 2009) surveys was significant (p<0.01) for distances up to 125 km (i.e. approximately an ICES rectangle) suggesting that these independent surveys show similar spatial structure in the Ammodytidae larval distribution in the North Sea (Fig. 4). Visual comparison of the maps for 2004, based on CPR and ICES survey data exhibited similar centres of abundance (Fig. 5). Overall larvae were more abundant in the ICES surveys, peaking at 17.5/m 3

in 2004, and were mostly abundant along the margins of the North Sea.


Class (km) Mantel (Pearson) α (Pearson) α (Holm corrected)Correlation

75 +0.11 0.001 0.001125 +0.09 0.001 0.002175 +0.03 0.042 0.042250 +0.03 0.095 0.095350 -0.02 0.217 0.217450 -0.04 0.011 0.043600 -0.03 0.049 0.155Table 2: Mantel correlogram statistics for comparisons between the CPR data (Feb-Jun 1950-2005) and ICES surveys (Feb-April, 2004 and 2009) for sandeel larval abundance by ICES rectangle. Values in bold are significant at α = 0.01 after Holm correction for multiple comparison, based on 9999 permutations.

The similarity in larval distributions between ichthyoplankton surveys of the entire water column and the CPR does suggest that the limited depth of this sampling device is not a serious constraint to the mapping of the spatial distributions of sandeel larvae. The concentrations of sandeel larvae around the Northern Isles and off the eastern UK are also consistent with past plankton surveys32. The comparison of abundances per m3, as calculated from CPR and ICES survey data shows that the two main concerns when sampling fish larvae with a CPR, namely the small aperture and the constant and relatively shallow depth of the sampling device, seemingly have an effect on the estimation of the abundance in absolute numbers. Yet, the CPR is useful in capturing the spatial patterns and the relative abundances between areas, as with both data sets the centres of abundance are generally the same. Several studies24,33 have compared the CPR to samplers with larger apertures and similarly found large differences in sampling efficiency. These large differences were attributed to active avoidance and emphasized that this is species specific, which is a problem for describing a plankton community based on CPR, but might be immaterial for a single species. The shallow sampling depth is likely problematic when sampling for those larvae which exhibit strong diurnal vertical migrations, leading to less capability of accurately estimating the abundance due to smaller sample size. Both concerns when sampling with a CPR may be partially alleviated by the much longer deployment time, compared to the 15 minutes in the plankton surveys. As the CPR samples continuously it can still pick up larvae often enough to give a representation of areas with high and low abundance.

Figure 5: Distribution of sandeel larvae from ICES surveys (left panel) and the CPR in 2004 (right panel)

III.5 Relationships between CPR fish larvae with stock assessments – case studies on key species

III.5.1 Clupeidae (herring, sardine, sprat) By far the most numerous fish larval group recorded by the CPR survey are the Clupeidae. In European waters Clupeidae caught by the CPR mainly comprise the herring (Clupea harengus), sardine (Sardina pilchardus) and sprat (Sprattus sprattus). The broad seasonal range and wide distribution reflects the multi-species composition. Most of the larvae in the north (northern North Sea northwards) are herring. Intervening regions of the southern North Sea, Irish Sea and English Channel are a mixture of sprat and sardine (with the possibility of some anchovy in the German Bight).


Herring is numerically one of the most important pelagic species in several North Atlantic ecosystems and intensive exploitation goes back several centuries. Stocks have fluctuated enormously in the past in response to both natural environmental variations and human exploitation. It is a key species considered to have major impact as prey and predator to most other fish stocks, as plankton feeders they form an important part of the food chain up to the higher trophic levels34.Herring stocks comprise many populations that may be isolated through specific spawning grounds and specific spawning seasons (Fig. 6, Table 3).Figure 6: Herring spawning grounds

Sprat is also an important source of food for many predators, including predatory fish, marine mammals and seabirds. Sprat is mainly exploited for industrial processing, but a small market exists

for human consumption (smoked sprat and whitebait). Sprat is found widely in the region. Spring egg surveys in the North Sea show sprat eggs are widely distributed and their distribution, as well as spawning season (May-August) overlaps with that of herring. This makes it impossible to distinguish between the two species from data analysis. Although landings data are collected for the North Sea and English Channel but there are no analytical assessments of sprat stock sizes. Sardine has increased in the English Channel in the last few years and known spawning grounds are located in the southwest of England.

Population Spawning time Spawning ground Nursery groundDowns Nov-Feb South North Sea Dutch coast and

English Channel German BightBank Aug-Oct English coast and English coast and

Central North Sea off DenmarkBuchan Aug-Sept east coast of Scotland east coast of Scotland

and Skagerrak, KattegatOrkney/Shetland July-Sept Orkney, Shetland east coast of Scotland

and Skagerrak, KattegatCeltic Sea Autumn-Winter inshore waters Celtic Sea, Irish SeaIrish Sea Sept-Nov northern Irish Sea northern Irish SeaTable 3: Herring stocks around the UK

Overall, there has been a decrease in number of clupeids caught by the CPR over the last few decades and the spawning grounds have become more concentrated in the south-western North Sea and eastern Channel perhaps reflecting an increase in sprat and sardine abundance compared to herring. Long term trends in clupeid abundances, in relation with adult abundance from IBTS survey (herring, sprat in North Sea and West of Scotland) and stock assessment data (herring) were examined for the South West; West of Scotland and for the North Sea datasets only. Bay of Biscay show too few records.

South West area (ICES divisions VIIe-h & VIIj): The spawning grounds for herring in the Celtic Sea are well known and are located inshore close to the coast (Fig. 6). Herring to the south of Ireland in the Celtic Sea and in Division VIIj comprise both autumn and winter spawning components. Larvae from the spawning grounds in the western part of the Celtic Sea may be transported into VIIj. In the western Channel, herring spawns in spring. The strong seasonal signal shown by clupeid larvae caught by the CPR mostly in May-July (Fig. 7), is at odds with this and raises some doubts as to whether herring contribute the most to clupeid larvae relatively to sprat and sardines. Unfortunately no assessment of either sprat or sardine exists for this area.


Figure 7: Clupeids in the south west area

West of Scotland (ICES division VIa): There are two main herring spawning grounds: north-west of Ireland and further north to the west of Scotland (Fig. 8). Herring to the northwest of Ireland comprise both autumn and winter/spring spawning components, while herring in the west of Scotland is considered to be an autumn spawning stock. Historically herring were also significant in the Clyde as spring spawners but this stock is now depleted and is not formally assessed. Data from the CPR show two seasonal peaks for clupeid larvae abundance: April and November (Fig. 8), in agreement with herring spawning season.

Figure 8: Clupeids in the West of Scotland area (ICES division VIa). Note: time series from IBTS is too short to be used in further analysis.

North Sea and eastern Channel area (ICES division IVa-c & VIId)The North Sea herring comprises several stocks most of which spawn in the autumn/early winter. There is a spring spawning stock also (Blackwater herring) but this is relatively small. The general pattern of clupeid larval seasonality (Fig. 9) is generally in agreement with herring spawning season, although the peak season of clupeid larvae still appears to be July-September, thus raising some doubts and suggesting that a substantial part of the clupeid larvae in the CPR North Sea samples may be sprat and not herring.

Figure 9: Clupeids in the North Sea and eastern Channel (ICES division IVa-c,


VIId)

In terms of Spawning Stock Biomass (SSB), all herring stocks appear to dip around 1970 onwards, corresponding with increased fishing mortality at the same time (Figs. 7-9). In all three areas, the relationship between fish larvae abundances and SSB and recruitment is significant although not strong (Table 4), and in the North Sea, a positive significant relationship exists with both adults of sprat and herring. The present results show that sprat, and sardine in the southwest area, and sprat in the North Sea may contribute substantially to the clupeid larvae caught by the CPR. It is unfortunately not possible to distinguish these three species from each other visually; and in the absence of any formal assessments of sprat and sardine recruitment in those areas, genetic analysis of the CPR samples seems to be the only solution to differentiating between clupeids within the samples. Without that knowledge, the use of CPR data for clupeid larvae as a tool to assist the assessment of relevant commercial fish stock is limited.

Clupeids larvae (vs) North Sea & East channel West of Scotland South West(1960-2005) (1956-2005) (1957-2005)

Herring SSB 0.39*** 0.49*** 0.33**Herring Recruits 0.43*** 0.33** 0.31**Herring Fishing Mortatlity F -0.34** -0.23 -0.23Herring (IBTS) 0.43*** -- -Sprat (IBTS) 0.36** -- -Table 4: Pearson correlation coefficients calculated between yearly mean abundances values for clupeid larvae from the CPR with stock assessment indices (SST, Recruits, mean F) and adult abundances, where: *** p < 0.01; ** p < 0.05; and * p < 0.10.

III.5.2 Ammodytidae (sandeels) Sandeels play an important role in the North Sea ecosystem and are a highly commercially important group. Sandeels are a key prey species in the ecosystem and they have been exploited in the North Sea by small meshed industrial fisheries, for oil and fishmeal, since 1953. Sandeel larvae are also numerous in CPR samples during spring and they are mostly found in shallow waters around the UK. Highest concentrations of sandeel are found off the east coast of Scotland and south of the Shetland islands. There is also a sizable population at the Dogger Banks in the southern North Sea. Generally there has been an increase in sandeel abundance recorded by the CPR from the first period (1948-1985) to the second (1986-2005) in the North Sea. No latitudinal or phenological shift could be detected.

A limitation of the current analysis of CPR fish larvae is that the five sandeel species in the North and Irish Seas cannot be distinguished. In the North Sea, the most abundant sandeel species is the commercially important lesser sandeel (Ammodytes marinus). Spawning of lesser sandeel occurs in the North Sea between December and January and in the Irish Sea between January and March. Three additional species of Ammodytidae spawn in the North and Irish Seas: greater sandeel Hyperoplus lanceolatus spawns in April and May; smooth sandeel Gymnammodytes semisquamatus spawns from April to July; and small sandeel Ammodytes tobianus spawns in the autumn.

The scientific and fisheries information available to inform the assessment differs by region and so analytical assessments are only available for the most commercially important of these35,36. In sandeel assessment areas 1 (Dogger Bank) and 2 (Wadden Sea, South Eastern North Sea) a dredge survey index from 2004 (collected in area 1 only) is used as a tuning index for an analytical assessment. In sandeel region 3 (Central Eastern North Sea) the assessment is based on commercial indices only. A dredge survey from 2008 onward informs the trends-only assessment in area 4 (Wee Bankie, Central Western North Sea). No assessment is made in areas 5 (north-eastern North Sea, Viking and Bergen Bank) or 6 (Kattegat) where the landings are too low or sporadic for an assessment. A recruitment index is available from 1985 to 2007 for area 7 (Shetland) although fishing in these coastal waters are now highly regulated and no fishery has operated in recent years. No assessments are made in the Irish (ICES division VIIa) or Celtic Sea (ICES divisions VIIfg) due to the lack of large scale sandeel fisheries in those areas. Given that the sandeel stocks are highly dependent on the incoming year classes, short term forecasts of year-class strength may be improved using estimates of larval abundance from the Continuous Plankton Recorder


(CPR) survey.

To assess the interannual variability in CPR larval data, larval abundances were compared with recruitment indices from sandeel assessments and with other surveys where available.

Long-term trends, sensitivity and relationship to recruitment data The North Sea was split into regions (Fig. 4) based on the current stock assessment areas, which were proposed by WKSAN35. Ammodytidae data were summarised by region on a monthly basis and used to explore the seasonality of the planktonic larval period in CPR samples (Fig. 10). The statistical distribution underlying the variability in the larval data was investigated and found to be well characterised by a negative binomial distribution due to the many zeros in the data with occasional high counts. The overall mean abundance per m3 was mapped by ICES rectangle for all years combined for those months in each region determined to lie within the planktonic larval period (Fig. 4B).

Annual mean abundances during the larval period with estimates of variability were calculated for each region (Fig 11) and annual maps of the mean abundance by ICES rectangle were plotted 37. Pearson correlation coefficients were calculated between Ammodytidae larval data and recruitment indices (1983 onward) for North Sea sandeel stock assessment units to determine whether or not the data show similar temporal patterns.

The long-term trends in CPR data indicate increases in Ammodytidae larval abundance in the North Sea and Irish Sea and decreases in the small numbers seen in Celtic Sea (Fig. 11). In the North Sea the long term trend is dominated by an increase in the Wee Bankie region since 1978. The second most abundant area for Ammodytidae larvae is the commercially important Dogger Bank region. Other North Sea areas show sporadic high abundances of Ammodytidae larvae without trend.

Figure 10: Mean abundance of sandeel larvae in CPR samples by region: left column average seasonal cycle for all years combined (1950-2005, except Irish Sea 1971-2005); right column and legend monthly


means. Regions (see Fig. 4) from top row to bottom are: R1 Dogger & Southern Banks; R2 Wadden Sea; R3 Fisher and Klondyke; R4 Wee Bankie; R5 South of Shetland; Celtic Sea box; Irish Sea box.

The time-series of larval Ammodytidae abundance in the North Sea from the CPR data is significantly correlated with the total recruitment index for the North Sea (recruitment summed from all assessment areas where data available (Fig. 12, R = 0.56, p = 0.005) from the ICES assessment for sandeels 35. This correlation is due solely to the Dogger Bank area where the vast majority of the commercial catch is taken. No correlation was found between recruitment indices and larval abundance in the Wadden Sea or in the Fisher and Klondyke region. A notable difference between the larval data and the recruitment data concerns the high recruitment event in 1996, which the CPR data suggest continued into the subsequent year in both the Dogger Bank and Wadden Sea regions. The CPR also picked up this recruitment event in 1996 in the South of Shetland area, second only to the larval abundance during the earlier high recruitment year, 1985 (Fig. 11). When set in the long-term context by the CPR data, the high recruitments of 1983, 1985 and 1996-1997 in the Dogger Bank region were likely the greatest events since the high larval abundances recorded in 1962 and the early 1950s.

Figure 11: Mean larval abundance in North Sea during Feb-Jun, Celtic Sea during Mar-Jul and Irish Sea during Apr-Aug. Shaded areas indicate the 95% CI based on the standard error in the mean.

Figure 12: Recruitment of sandeels in the North Sea population from stock assessment, where: R1, the Dogger & Southern Banks (black); R2 is the Wadden Sea i.e. south-eastern North Sea (light grey); and R3 is the Fisher and Klondyke Banks in the central eastern North Sea (dark grey).


No correlation was evident between the larval indices and spawning stock biomass (1983-2005) with a lag of 0, 1 or 2 years for the Dogger Bank, Wadden Sea or Fisher and Klondyke region. Similarly, no correlation was evident between the larval indices and estimated catch (1955-2005) in Dogger Bank, Wadden Sea or Wee Bankie areas with and without a one year lag in each region. However, significant correlations were evident between the highly variable larval indices in both the Fisher and Klondyke region and the Shetland regions and the catch in the previous year (R = 0.44, p = 0.028 and R = 0.47, p < 0.001 respectively, Fig. 13).

The correlations between Ammodytidae larvae abundance and North Sea sandeel recruitment indices show that the CPR sampler can produce reliable time-series of abundance in areas where Ammodytidae larvae are consistently abundant (e.g. Dogger and Southern Banks), and the data can inform recruitment levels for the North Sea sandeel stock assessment. However, for the Wadden Sea and Fisher and Klondyke regions either the CPR data or the indices of recruitment are not robust or the processes determining larval survival are so great that they obscure any relationship.

Figure 13: Annual sandeel catch (black line and circles), SSB blue dashed line and Ammodytidae larvae abundance (red line). Y axes scaled by minus the mean and divided by the standard deviation.

III.5.3 Atlantic mackerel (Scomber scombrus)Mackerel is one of the most abundant and widely distributed fish species in the North East Atlantic. Mackerel plays an important ecological role by feeding on zooplankton and on the pelagic larval and juvenile stages of a number of commercially important fish stocks. Large changes in mackerel abundance and distribution have therefore substantial effects on ecosystems as well as economies.


Results show two separate spawning components of mackerel in the central North Sea and south-west of the UK. During the first period (1948-1985), in the North Sea the larvae were caught predominantly between June and August with the highest number occurring in July, while they were caught predominantly in June during the second period (1986-2005). In the Celtic sea mackerel spawn earlier from March to July with highest number occurring in April and May (Fig. 14). Overall mackerel larvae have declined substantially over the last few decades, mostly reflecting the decline of first the North Sea component in the 1960s and 70s. There is an apparent shift in phenology to later larval occurrence in the south-western component.Figure 14: Abundance of mackerel larvae for the period1948-1985 and 1986-2005.

Radical changes in abundance and distribution of mackerel have been observed throughout the northeast Atlantic and North Sea during the last century. Mackerel that spawn in the North Sea are considered to form a stock distinct from the currently larger western mackerel stock. Fish larvae maps from CPR data show that the change in spawning of mackerel in agreement with current scientific knowledge. The North Sea spawning stock was large and lightly fished up to the late 1960s, where the development of modern sonar, power blocks and single-vessel purse seining led to a ten-fold increase in mackerel landings 38. This fishery was unsustainable and resulted in a collapse of the stock in the 1970s. Despite subsequent regulations of the fishery designed specifically to protect this stock, it never rebuilt to its former level. In the last decade the spawning stock biomass has been 150-230 kt, compared to over 2 500 kt in the beginning of the 1960s39,40. It is currently unknown why the North Sea stock has not rebuilt to former levels. Unfortunately, documentation of the historic development is based on fragmented information sources that do not consistently cover the whole period from before to after the collapse. This is a hindrance for addressing key questions about the lack of stock rebuilding and the consequences of these changes in distribution and abundance. The CPR fish larvae time-series, with its consistent and broad temporal span offer a unique opportunity to investigate long term changes in abundance and distribution of mackerel larvae, and aid in the understanding of the development of this stock.

Here, we use the mackerel larvae data from the CPR, covering the North Sea, to verify the spatial origin of the larvae through use of a hydrographic backtracking model for all sampled larvae. Using a technique not previously applied to CPR data, we then construct a larvae index considering catchability as well as spatial and temporal autocorrelation. Considering the larvae abundance as a proxy for number of spawned eggs and spawning biomass, we compare it with existing egg survey data and fisheries-based assessments with a focus on the decline around the 1970’s. Finally, we review the possible applications of this time series, including supplementing or improving the mackerel stock assessment and the international mackerel egg survey with data from the CPR survey.

Effect of larval drift (hydrographic backtracking model) The positions of mackerel larvae captured by the CPR survey do not necessarily correspond to the actual location where spawning took place. In some regions of the North Sea, ichthyoplankton can be advected away rapidly from their spawning location: the magnitude and direction of this drift can vary appreciably between years. As a first step in the analysis of the larval dataset, we attempt to estimate the magnitude of this advection, and thereby check for a potential bias introduced by drift processes forced by physical fields derived from the NORWECOM model (see Objective 3).

For each location (in time and space) where mackerel larvae were observed in the CPR survey, 100 particles representing mackerel “larvae” were released in the model, uniformly distributed throughout the water column. Time in the model was then run backwards to determine a range of possible trajectories along which the larvae could have originated. No active-behaviour was applied to the particles – the “larvae” were mixed throughout the water column following the modelled turbulence as passive tracers. No explicit attempt was made to account for ontogenetic changes during this time (e.g. changes in egg buoyancy, hatching of eggs, changes from endogenous to exogenous feeding of larvae). The duration of the backwards-advection scheme was based upon an estimate of time-since-spawning. Mackerel larvae in the CPR survey have a mean length of 4.8 mm (s.d. 2.0mm)41. Under good temperature and food conditions, mackerel larvae grow from a typical hatch size of 3mm to 4.8 mm in approximately 2.4 days. Mackerel eggs are pelagic and therefore drift of the eggs also needs to be accounted for: typically 50% of mackerel eggs hatch after 6.7 days at 11°C42. We therefore estimate that, on average, approximately 10 days have passed since the larvae captured by the CPR were spawned. The simulated mackerel particles


were therefore advected backwards in time for 10 days. At the completion of this period the geographical distance between the site of capture and the end point was calculated for each particle and the median of the distance distribution calculated. The process was then repeated for all larval observations in the CPR and the distribution of advection-distances across all observations generated. This distribution was then used to assess the magnitude and importance of advection processes in shaping the distribution of larvae.Hydrographic drift simulations showed that advection of the larvae between the estimated spawning time and capture by the CPR was generally minor (Fig. 15). 90% of the larvae caught by the CPR had drifted less than 60 km from the spawning site and 75% have drifted less than 35 km (Fig 15b). Advection of mackerel eggs and larvae between spawning and capture in the CPR, and therefore any interannual variability associated with it, can reasonably be assumed not to induce a significant bias in the spawning distribution when looking for changes at the scale of the North Sea basin. The CPR larval observations can therefore be used as proxies for the spawning distribution of North Sea mackerel.

Figure 15: Backtracking simulations. a) Examples of backtracked trajectories for six observations of larval in the CPR distributed across the North Sea. Red circles mark capture points in the CPR, blue circles the end points of particles after 10 days of backtracking. Black lines connect the two points for visual reference. Text denotes the CPR label code. b) Distribution of particle displacements after 10 days drift. Left axis (grey bars) depict the frequency (number of CPR observations containing larvae) for each 10km class bin. Black-line with black dot (right axis) shows the empirical cumulative distribution function.

Mackerel larvae model (larval index) and comparison with egg survey and stock assessment data The distribution of larvae captured in the CPR survey were analysed using the log-gaussian cox process (LGCP) model43. The spatio-temporal distribution of larvae is not completely random: aggregation in both

space (“patches”) and time can be expected. Also, some degree of continuity from day to day and from year to year would be expected because the abundance of larvae are expected to be related to the stock size of the mackerel and mackerel live and spawn for multiple years. It is assumed that spawning and hence larvae abundance follows a fixed seasonal pattern within the year, modelled here as gaussian. However, the yearly level is considered as a random effect.

The seasonal peak of the larvae abundance was found to be in mid-July (Fig. 16). Since we estimated mean larval age to be approximately 10 days, this corresponds to a peak in spawning at the start of July. This is comparable to egg survey based estimates from 1982-2008, where the peak spawning were found to be 8-20 days earlier. A difference in this direction were expected because our study period includes cooler decades than the period from 1982 to 2008


and spawning is known to be earlier in warm years44.

Figure 16: Seasonal effect on CPR larvae index

Because the CPR operates horizontally in a fixed depth of approximately 10 m, the catchability of the recorder can be expected to be sensitive to changes in vertical distribution of the larvae. Small mackerel larvae, such as those caught by CPR, have been observed to stay above the thermocline where they migrate towards the surface at night45,46. We therefore tested the systematic effects; diurnal migration and thermocline depth in the model. Of the two catchability effects; thermocline depth was found to be significant (p<0.001) whilst the diurnal catchability pattern (hour effect) was not. Consequently only thermocline depth was retained in the final model. Catchability peaked in areas where the CPR was sampling just above the thermocline. Larvae were rarely caught when the thermocline was below 45 m. Having

corrected for catchability effects, we assume that the CPR catches represents the true larvae concentration plus random sampling error. Furthermore, active avoidance of the sampling gear can also potentially affect catchability. This is more pronounced for larger larvae46, but since the larvae caught by the CPR are small, we assumed that this effect was negligible. The fitted model was used to predict the larvae concentration at any point in the North Sea, through each day in the period 1948-2005. From this dataset we produced yearly distribution maps and a time series of yearly indices of larvae abundance, by calculating the posterior mean of the spatially integrated intensity for each year. The hypothesis of a change in abundance from before 1970 to after 1990 was tested by a likelihood-ratio hypothesis test.

Annual larvae abundance index is illustrated for the whole study period in Fig. 17. We found a significant (p<0.001) shift in the mean larvae index of 6.1 from before 1970 to 1.6 after 1990. There is unfortunately too much variability in the CPR larval index to precisely pinpoint the onset and completion of this decline. Nevertheless, the broad pattern of a systematic decline in abundance between 1970 and the mid-1980s shown here agrees with data from other independent sources e.g. standardized catch rates in the Dutch commercial spring fishery and catch/tagging based assessments indicate a decline beginning in the late 1960s (Fig. 18). The decline continues through the 1970s, as also indicated by the catch/tagging based ICES assessment and early mackerel egg surveys suggest that the decline ends in the mid 1980’s. The CPR larval index is therefore in broad agreement with the piecewise picture available from other data sources: however, it also has the clear advantage of covering the entire time-span of interest.

Figure 17: Larvae abundance index with 95 %confidence interval.


Figure 18: Long term mackerel trends in the North Sea,43,47,48. Loess smoothed trend lines with span=0.5.

Spatial distributions obtained from the model showed a shift in

spawning area from early to recent decades (Fig. 19), suggesting that the central North Sea is no longer as important as the areas further west and south. This change is in line with the results from the international mackerel egg surveys; although these surveys do not cover the extreme south and southeast (Objective 4, mackerel maps). Spawning in the North-western North Sea was, as also observed in the egg surveys, at a very low level in all periods.

Figure 19. Modelled spatial distribution of mackerel larvae caught by CPR. Color scale from white (low abundance) represents quantile 0.1-0.2, 0.2-0.3, etc. up to red 0.9-1.0 (high abundance).


Applications of CPR data for mackerel larvae For the first time, a single unbroken time series describing the dynamics of the North Sea mackerel population has been presented. The time series covers the full time span of interest, from 1948, through the 1970s and 1980s stock collapse, all the way up to 2005. This index is based on a novel analysis of CPR observations, using powerful modern statistical techniques. The resulting perspective is both unique and gives a broad view of the dynamics of this population where previously only brief glimpses were available. Our results confirmed the long-term development of the North Sea stock, previously based on assessments of spawning stock size and egg abundance covering part of the time span. Furthermore we found a spatial shift corresponding to a similar observation in egg distribution. This provides some validation for all approaches and suggests that the larvae index, at least on longer time scales, is a usable proxy for egg abundance and spawning stock size in the North Sea.

It is noteworthy that the interannual variability in the CPR index was extremely high. Several sources of variability seem possible: i) high statistical uncertainty such as random sampling error that increase due to the few larvae being captures in the later years, ii) variation in fecundity, iii) variation in mortality during the approximately 7 days of egg phase and 2 days of larval phase, iv) poor spatial sampling coverage in the central North Sea in later years, v) lack of sampling in Skagerrak/Kattegat. However, our conclusion on the decline from before 1970 to after 1990 seems robust to these uncertainties. Even though sampling intensity in the central North Sea has been reduced in the later decades, the sampling that did take place in this area did not result in catch rates comparable to those in the earlier decades. Furthermore, analysis of the spatial patterns also suggests that the central North Sea is no longer as important as the areas further west and south. However, a spatial shift back towards the central North Sea in the future might not readily be detected with the present survey design. Improved spatial coverage in this region would therefore improve the precision of the CPR larval index and further increase the value of this time series for scientific community as well as stock advice and management. Spawning is also known to take place in Skagerrak/Kattegat. The importance of this area is possibly limited to approximately 5% of the North Sea mackerel spawning47. However, this estimate is highly uncertain as the area has never been properly covered by the CPR or egg survey.

The CPR survey covered parts of the North Sea outside the egg survey area, providing an opportunity to evaluate the spatial coverage of the North Sea egg survey (Fig. 19). Modelled distribution of larvae in the whole North Sea showed that the Southern North Sea has been a relatively important spawning area in the North Sea through the last decades. This result suggests that the area covered by the mackerel egg survey does not cover the entire spawning distribution, and may need to be expanded. The new time series developed herein has the potential to address several outstanding problems regarding the mackerel stock in the North Sea. The most significant of these is: “Why has the North Sea spawning stock not rebuilt despite decades of protection from commercial fisheries?” We propose four hypothesis that may explain this observation:

i) Changes in environment or predation pressure have reduced the productivity of the stock. This was tested (see part IV.1), and our results suggest that changes in climate-induced hydrographic features indeed played a dominant role in the drop of mackerel larvae caught by the CPR during the 1970s-80s.

ii) The fishing pressure is still too high due to by-catches in herring fisheries and/or in the large fishery for western mackerel in the northern North Sea.

iii) The North Sea mackerel is not a separate natal homing stock and the observed crash was merely a change in distribution of one large north eastern Atlantic panmictic mackerel population.

iv) The North Sea mackerel was a separate natal homing stock up to the collapse but subsequent modification of the genotype and behaviour resulted of intermixture between the small North Sea stock and the larger western stock44.

Whilst it is not possible to address these questions directly here, there is little doubt that the CPR larval index can make a valuable contribution to understanding this, and other, pressing questions regarding this stock. Applying state-of-the-art statistical models such as the present log-gaussian cox process model provides numerous advantages over the more simple deterministic raising algorithms, such as increased usage of the precise spatio-temporal sample information, uncertainty estimates, statistical hypotheses tests of parameter significance, and increased resolution in maps and time series.

III.5.4 Blue whiting (Micromesistius poutassou)


Blue whiting is a small pelagic fish that is widely distributed in the North-East Atlantic. It is an important link in the marine food chain, feeding on plankton crustaceans and small fish and itself prey to a wide range of predatory fish, squid and marine mammals. Spawning takes place in spring (March-May) along the shelf edge west of the British Isles and west of the Rockall Bank. Overall there has been a steep decline in fish larvae caught by the CPR after the 70s.

Figure 20: Blue whiting over UK shelf

Blue whiting is currently one of the largest fisheries in the North-east Atlantic, and all blue whiting in this area is treated as one stock. As a fishery resource, blue whiting is still young – the stock was only “discovered” in the late 1960s, and the fishery developed in the 1970s. During most of the 1980s and 1990s, the catches were rather stable. However, the catches increased rapidly in the late 1990s, and a new catch record was set almost every year - with total catch over 2 million tonnes in 2003-2005. However, recruitment of the stock has fallen to a low level over the last decade. This has caused a decline in spawning stock and some strong reduction in catch quotas. Results from the CPR survey parallel this decline in spawning stock, but do not agree with recruitment data from the assessment (Fig. 20). This discrepancy is likely due to movement offshore (an area poorly sampled by the CPR) by the stock during the 1990s resulting from changes in the oceanic gyre49.

III.5.5 Dab (Limanda limanda) Dab is one of the most abundant demersal species in the North Sea with the centre of distribution in the German Bight in the southern North Sea. This is expected given fishery catch information. It has a sedentary nature and has proven to be a valuable indicator in eco-toxicological studies. The abundance of dab larvae was high during the 1950s but during the 1970s and 1980s recruitment would have been very poor generally (Fig. 21). Following the low recruitments of the 1970 and 1980s, larval abundances rose in the central North Sea (IVb) during the 1990s with a high recruitment event occurring in 1993 in both the southern and central areas. Since 2000, high larval abundances have been confined to the southern area IVc. In the final year of the current dataset (2005) larval abundances in IVc were average relative to the mean abundance since 1995 while the abundance in IVb was low.

There is no information on the stock identity of this species. Landing data are not complete and are probably not indicative for catches since discard rates are variable. The mixed TAC with flounder reduces the accuracy of catch statistics per species. Different surveys show a stable to increasing total biomass for the main area (IV, North Sea) in which the fisheries are conducted. Dab is a bycatch in the fishery for flatfish, shrimp and demersal species, mainly in the beam trawl fisheries. Dab catches are generally discarded based on the availability of target species and market price.

Figure 21: Dab larvae, (left) annual spring/summer means for the whole North Sea and (right) North Sea


subareas, where red: central North Sea (IVb), green: southern North Sea (IVc), grey: northern North Sea (IVa)

III.3.6 Whiting (Merlangius merlangus) Whiting is one of the most numerous and widespread gadoid species found in the North Sea. In the North Sea, numbers of larvae declined from the mid-1960s then increased from the 1980s but again decreased from the late 1990s; while in the South west area there seems to be an increase in whiting larvae from the late 1970s in the western Channel and to the south of Ireland. The spawning season extends from March to July, with a prominent peak in April in the North Sea in the early 1990s. Whiting has recently become more of a management issue with the general decline of other fish stocks. Whiting is one of the main predators of other commercial fish species such as Norway pout, sprat, sandeel, herring, cod and haddock.

Figure 22: Whiting in North Sea (ICES areas IV and VIId) and south west area (ICES areas VIIe-k)

We compared yearly trends for whiting larvae with those for adult fish and recruitment data. Whiting larvae are mostly caught between March and August in the North Sea and in April-May in the South west area (Fig. 22). In the North Sea, significant correlations were found between CPR fish larvae yearly abundances and Adults from IBTS data (R=0.57, p<0.01), SSB (R=0.68, p<0.01) and Fishing mortality (R=0.50, p=0.05). Although the CPR high larval abundances in the southwest area in some high recruit year, the sparsity of whiting larvae in the samples rendered the correlations inconsistent and therefore unreliable.

IV. Objective 5: linkages between fish larval abundance and distribution with environmental effects and fisheries.(achieved)

In order to achieve this objective, we used a regional approach with a study on the impacts on climate change on fish larvae in the North Sea, and a study on regime shifts in the Celtic Sea.

IV.1 Towards a better understanding of the impacts of bottom-up effects derived by climate variability on fish populations.

We investigated the combined effects of several selected biotic and abiotic environmental variables on fish larvae in the North Sea over the period 1960-2004, to ensure consistency across all datasets. We first tried to identify regimes with certain environmental conditions that are more or less favourable to the selected fish larvae, and second, to identify potential linkages between the environment and fish larvae. We focussed on the five most abundant groups of fish larvae: clupeids, sandeels, Atlantic mackerel (Scomber scombrus), dab (Limanda limanda), and gadoids. Gadoids consisted of the followings taxa collated: Merlangius merlangus, unidentified gadoids, Trisopterus esmarkii, Gadus morhua, Pollachius virens,


Melanogrammus aeglefinus, Trisopterus minutus, Micromesistius poutassou, Pollachius pollachius, Raniceps raninus and Trisopterus luscus.

As fluctuations in the abundance of fish larvae are likely to be influenced by several factors, correlated or independent, we used a multivariate technique to investigate to combined effects of the seven environmental and biological variables: SST, surface salinity, NAO, CHLO, abundances of diatoms, dinoflagellates and zooplankton) (Objective 3). A Principal Component Analysis (PCA) was performed on the matrix of annual values (1960-2004) of the selected parameters for the survival and growth of the five selected taxa of fish larvae in the North Sea. The aim of this PCA was to identify major patterns of year-to-year changes in the environmental composition of the North Sea. Second, a PCA was performed on the matrix of monthly values of the same parameters (PCA, 540 years–months × 7 variables). The aim of this PCA was to also consider seasonal variability of the selected variables and identify long-term monthly changes in environmental composition of the North Sea ecosystem, and their impacts on the selected taxa of fish larvae.

IV.1.1 PCA results and ecosystem status PCA results, for the first three principal components on matrix of yearly values are as follows (Table 5):

PCA loadings and variability PC1 PC2 PC3By componentPCA, years × variablesDiatom abundance 0.47 -0.05 -0.35Dinoflagellate abundance 0.42 0.32 0.31CHLO 0.37 -0.31 -0.34SST 0.29 -0.53 -0.06Surface salinity 0.25 0.54 0.18Winter NAO 0.18 -0.44 0.79Zooplankton abundance 0.53 0.20 -0.03Percentage total variance explained 37.59 27.41 11.21Cumulative variance 37.59% 65.00% 76.20%

Table 5: Principal Component Analysis (PCA) scores on matrix of yearly values

For PCA on the matrix of annual environmental values, we produced a score plot (Fig. 23). Four distinct periods of time are revealed: 1960–1976 (Period I), 1977–1982 (Period II), 1983–1996 (Period III) and 1997–2004 (Period IV). Periods I and III are similar in the sense that the scores are scattered in the middle of the plot, but these two periods are different in their respective environmental compositions. Period I is generally characterised by the highest values of the second principal component (PC2), the most important drivers being high salinity and low NAO and SST, and to a lesser extent higher zooplankton and dinoflagellate abundance and lower CHLO, whereas Period III is more scattered around the origin. Period II is characterised by low values of both PC1 and PC2, the most important drivers being low diatom and zooplankton abundance. Note that the lowest values of six of the seven variables selected were during Period II (Fig. 24). Period III seems to show a recovery of the system from Period II, with high abundance of dinoflagellates and zooplankton, and high values of chlorophyll and the NAO, combined with low diatom abundance. Period IV is characterised by the highest values of PC1 and low values of both PC2 and PC3, and a resulting combination of high SST, chlorophyll and diatom abundance with low surface salinity and zooplankton abundance.


Figure 23: Scores plot of the first three principal components for the analysis performed on the matrix of annual values of seven selected environmental variables and biological parameters (PCA, 45 years × variables). Four distinct periods of time can be seen: Period I, 1960–1976 (open circles); Period II, 1977–1982 (black triangles); Period III, 1983–1996 (grey circles); Period IV, 1997–2004 (black squares). (a) PC2 vs. PC1; (b) PC3 vs. PC1.

period 1 2 3 4

Figure 24: Standardized anomalies in annual fluctuations in environmental variables and abundances of fish larvae, zooplankton and phytoplankton, 1960–2004. Dashed vertical lines show the boundaries between the four periods of time (I–IV) identified by the PCA. Visual examination of long-term monthly changes in PC1 and PC2 (Figs. 25a,b) also revealed distinction between the four periods. Similarly to results obtained from PCA analysis performed on annual matrices, Periods II and IV are particularly distinguishable even when the seasonal variability is taken into account, indicating that interannual variability can be stronger than seasonal variability, and that long-term environmental changes are critical in defining the status of the ecosystem.


Figure 25a: Long-term monthly variability (1960–2004) in the seven environmental variables used simultaneously as first principal component (PC1, 54.03% of total variability). The blue line represents the annual abundance of each of the 5 fish larvae taxa superimposed on monthly PC1 anomalies.

Figure 25b: Long-term monthly variability (1960–2004) in the seven environmental variables used simultaneously as first principal component (PC2, 15.15% of total variability). The blue line represents the annual abundance of each of the 5 fish larvae taxa superimposed on monthly PC2 anomalies. Interestingly, Period II coincided with the first Great Salinity Anomaly (GSA) or cold, low salinity event of the early 80s50, and Period III with the second GSA or warm, high salinity event in the late 1980s – early 1990s51. The biological response to this later event was so dramatic as to be termed “regime shift” 52-54. That period encompassed a shift in the ecosystem towards a warmer dynamic equilibrium55. As for period IV, the system shifted to an even warmer but with low salinity status (Fig. 24). The NAO was also important in PC2 and explained most of PC3. The positive correlation between the


NAO and SST has been recognised in many studies. The NAO also affects water masses, currents and circulation patterns in turn affecting the flow of Atlantic water into the North Sea 56; this will in turn affect the mixed layer depth57, salinity levels and the flux of nutrients needed for phytoplankton growth58.

Our results from PCA analysis with the variables selected suggest a cascading effect from climate variability (as encapsulated by the NAO) to plankton via hydrographical changes in temperature and salinity within the North Sea: The NAO (as the largest contributor to PC3) appears to be a fundamental driving force on the physical status of the North Sea ecosystem as described by PC2. The status of the physical environment would in turn impact on the ecological status of the North Sea ecosystem as displayed by PC1. This is in agreement with other studies57 but the complexity of the mechanisms by which the NAO may influence ecological processes means that these haven’t yet been fully understood and require further study.

IV.1.2 Environmental variability and the survival of fish larvaeWe looked at correlations between larval fish abundance and both PC1 and PC2 from yearly analysis (Table 6):

Clupeids Atlantic mackerel Sandeels Dab Gadoids

PC1 0.37 (0.01,0.17)* –0.12 (0.43) 0.31 (0.04,0.18)* 0.31 (0.04, 0.25)* 0.46 (0.00, 0.02)*PC2 0.13 (0.39) 0.54 (0.00, 0.02)* -0.32 (0.06) 0.24 (0.11) 0.34 (0.07)

Table 6: Pearson’s correlations coefficients (with associated p-values) between annual abundance of fish larvae and PC1 PC2 from PCA on the annual average to match the seasonality of fish larvae: clupeids, sandeels and gadoids (January–December), Atlantic mackerel (June–August), dab (March–September). When a statistically significant relationship was found (*95% confidence interval), the p-value was adjusted taking into account autocorrelation; an emboldened value indicates that the relationship is still significant.

There were positive correlations between PC1 and the annual abundances of four of the larvae. There was no correlation between annual values of PC2 and the abundance of these larvae. These results indicate that the relationship between PC1 and fish larvae was mainly attributable to the abundance of zooplankton and phytoplankton (as indicated by their contributions to PC1).

The results illustrate the complexity of the mechanisms involved in the survival of fish larvae; many biotic and abiotic processes interact with each other, supplementing or cancelling out each other. Despite the direct effects of both temperature and salinity on fish larvae, the results show a low frequency link between clupeid, sandeel, dab and gadoid larval abundance through PC1 only, although the statistical significance of the relationship holds for gadoids only when adjusted for autocorrelation, indicating the year to year variability is important to this taxa only. PC1 is driven mainly by plankton indices, suggesting that for these four fish taxa, changes in climate, hydrography and temperature in the North Sea, are more likely to affect larval survival and hence recruitment indirectly through the plankton rather than directly, supporting previous analyses of cod, plaice and sole larvae59-61.

In the case of Atlantic mackerel, the results show no link with PC1 but a link with PC2. PC2 is driven by physical parameters, rather than plankton indices. There are two main differences between mackerel and the other four taxa: the narrower time-window of larval presence and the migratory characteristic of mackerel. The larvae of mackerel are caught during a very short period of time from June to August (Objective 4, atlas); this may make mackerel more sensitive, than the other fish taxa, to changes in phenology in their potential preys. All potential prey for mackerel dropped in abundance during period II (Fig. 26), but their subsequent increase during Period III was not followed by a recovery of larval abundance, and we do not see any phenological change that would lead to a temporal mismatch between the larvae and their potential prey during the period of larval presence (Figs 24,26). Visual analysis of changes in monthly spatial distribution of mackerel larvae and the selected plankton indices does not reveal a spatial mismatch either (Objective 4). Our results lead us to think that the drop and non-recovery of mackerel larvae in the North Sea in the late 1960s-70s was not primarily caused by a decrease in prey availability. Mackerel is a migratory species, and it might be that other factors linked to hydrographic variability, or fishery were playing a prominent part in influencing their spawning in the North Sea.


Figure 26: Long-term monthly variability (1960–2004) in the main contributors to PC1 and PC2: (a) zooplankton abundance, (b) CHLO, (c) diatom abundance, (d) dinoflagellate abundance, (e) SST and (f) surface salinity.

IV.1.3 Species reaction and adaptation to environmental variability It has been reported that the regime shift linked to the GSA event of the late 1980s radically changed the food environment for larval cod in a way that reduced the survival of young fish59. The taxa studied here seemed to fare differently during each period, suggesting that different species react differently to environmental changes and in particular to changes in the availability and quality of prey. The diet of the larvae of each taxon analysed here have been studied in detail before. After hatching, larvae start feeding on phytoplankton and copepod nauplii and gradually switch to copepodites as they grow. In the case of sandeels, the literature indicates that early larval stages <8 mm long feed exclusively on phytoplankton. Although the prey items are similar for all five taxa, the relative proportions of prey from gut analysis vary from fish to fish and study to study62-68. Our results agree with the results of other studies of larval diet, except in the case of Atlantic mackerel.

The increase in phytoplankton biomass in the North Sea from the late 1980s (Fig. 26b) has been attributed to increases in SST58. This increase seems to have benefitted sandeel, clupeid, dab and gadoid larvae, all of which increased markedly in abundance at the start of Period III (Fig. 24). The abundance of zooplankton also increased dramatically at the same time, but the trend was reversed in the early 1990s and zooplankton abundance has been decreasing ever since. Clupeid abundance remained high as that of zooplankton decreased, suggesting that clupeid larvae do adapt to changes in their prey environment and can switch between different prey, depending on availability. The ability of fish larvae to switch their diet to lower trophic levels when necessary has been reported previously69. In a comparative study on sandeel and dab larvae, they found that for dab, smaller size classes can switch to lower trophic levels and handle the greater variability of prey quality better than sandeels, and they suggested that although vulnerable to changing prey fields, fish larvae may be able to react flexibly as they evolve in a fast-changing environment. This vulnerability and flexibility seems to be species specific, and the more adaptable larvae should be able to survive when the quality of the feeding environment decreases, by switching to lower trophic level prey such as phytoplankton, microzooplankton, protozoans and small metazoans. Indeed, there is growing evidence that microzooplankton, including the protozoans, plays an important role and is regularly used as prey by fish larvae or at least to substitute their commonly accepted crustacean based diet70,71.

Larval mortality Pre-recruit natural mortality was investigated through multiple regression analyses of log transformed recruitment indices regressed against sandeel larvae, climate (AMO and NAO), zooplankton biomass and phytoplankton indices during February-June as explanatory variables37. Recruitment with a one year lag was also included in the model to account for potential cannibalism by one year old sandeels.In each of the three stock assessment areas, low values of the AMO (i.e. low temperatures) proved a significant predictor of high recruitments (p< 0.05). In both the Wadden Sea and the Dogger Bank regions


the previous year’s recruitment was significant indicating potential high levels of cannibalism. A high abundance of larval Ammodytidae was a significant predictor of high recruitments in the Dogger Bank region only, while an abundance of dinoflagellates was indicative of high recruitments in the Fisher and Klondyke Banks region only.

Long-term relationships between sandeel fish larvae and copepod prey abundance, phytoplankton (diatoms and dinoflagellates) and climate indices (sea surface temperature and NAO) indicate bottom up forcing of larval abundances by climate, in agreement with our previous results using PCA analysis.

IV.2 Regime shifts in the Celtic sea ecoregion

Recent research has suggested that climate change may have resulted in long-term changes in the diversity of zooplankton and evidence is growing to suggest that the phenology of phytoplankton and

zooplankton is altering in response to global warming. Regime shifts have been documented in the North Sea phytoplankton and zooplankton from studies using CPR data, notably following the GSA event of the late 1980s as mentioned above.

The spatial smoothing technique used to incorporate spatial and temporal variability as explained in Objective 4 is highly successful for data rich areas, such as the North Sea, or over larger areas, such as the North Atlantic, where data are sparse but where interpolation can be considered robust. However, as described in Objectives 1&2, data for the shelf seas to the west of the British Isles are relatively sparse and thus the region has thus attracted much less research. For work in the Celtic sea region, we applied a statistical method that is applicable under such circumstance is that of smoothing via Generalized Additive Models (GAMs)72.For plankton data, annual means by ICES divisions (Fig. 27) were determined for years with samples in ≥8 months, which is adequate to determine the average seasonal cycle.

Figure 27: Map showing areas studied and ICES division labels.

IV.2.1 Trends in abundance (non-parametric analysis of monthly means by ICES division) Fish larvae species with peak larval abundance in March-May were categorised as spring spawners: Clupeidae, dragonets (Callionymus sp.), pollock (Pollachius pollachius), cod (Gadus morhua), rockfish (Sebastes sp.) lemon sole (Microstomus kitt). Species with peak larval abundance in Jun-August were categorised as summer spawners: gobies (Gobiidae), wrasse (Labridae), saithe (Pollachius virens), scaldfish (Arnoglossus laterna), sole (Soleidae), sea robins (Triglidae) and pouting (Trisopterus luscus).To investigate change within plankton communities, monthly mean abundance data for CHLO, decapod larvae and species of copepod and fish larvae were further investigated for long-term trends (1960-2007 for crustacean zooplankton and 1950-2005 for fish larvae) and changes in seasonal patterns using robust non-parametric statistical approaches (seasonal Kendall’s tau and Sen’s slope). Table 7 shows an increase in phytoplankton standing stock across all areas. The zooplankton data indicate a pattern whereby increases were observed in warm shelf-edge species (Centropages typicus and Calanus helgolandicus) in coastal areas and decreases in deep waters. The larger sized copepods Para/pseudocalanus sp. and Calanus finmarchicus generally decreased in all areas, with the following exceptions: Calanus finmarchicus increased in the western Channel (VIIe) and the southern Celtic Sea (VIIh). The particularly small copepods Oithona sp. decreased to the west of Scotland (VIa) and southern and western Celtic Sea (VIIh,j) but increased in the central areas: including the western Channel (VIIe), Bristol Channel (VIIf), northern Celtic Sea (VIIg) and Irish Sea (VIIa). Metridia lucens, an indicator of oceanic water, decreased over much of its core area to the west of Scotland and in both the western and northern Celtic Sea (VIIg,j) but increased weakly in the southern Celtic Sea (VIIh), Bristol Channel (VIIf), western Channel (VIIe).


Fish larvae data indicates significant increases in the north-eastern Celtic Sea (VIIg), Irish Sea (VIIa) and western Channel (VIIe). In the western Channel the increases in fish larval abundance are largely attributable to the summer spawning fish (dominated by gobies and wrasse), while spring spawning fish (largely clupeids) dominate the trends in the north-eastern Celtic Sea and Irish Sea.

Celtic Seas EcoregionHabitat Taxa VIa VIIj

DeepVIIhMid-

VIIgMid

VIIaMid-

VIIfCoast

VIIeCoast

CHLO 0.15*** 0.15*** 0.21*** 0.17*** 0.27*** 0.25*** 0.27***Calanus nauplii 0.03 0.06* 0.20*** 0.07* 0.12*** 0.04 0.19***

WarmShelf-edge

Calanus helgolandicus

-0.01 -0.03 0.10*** 0.06* 0.03 0.23*** 0.16***

WarmShelf-edge

Centropages typicus

-0.05 -0.14*** -0.06* 0.12*** 0.30*** 0.15*** 0.17***

Temora Longicornis

-0.03 -0.02 -0.02 0.10*** -0.03 -0.10*** -0.01

Oithona sp. -0.28*** -0.14*** -0.06* 0.12*** 0.30*** 0.15*** 0.17***Cold, deep

Metridia lucens -0.21*** -0.20*** 0.06* -0.14*** 0.00 0.12*** 0.12***

Cold, deep

Calanus finmarchicus

-0.23*** -0.07* 0.07* 0.01 -0.15*** -0.02 0.09**

Para/pseudo calanus sp.

-0.15*** -0.27*** -0.20*** -0.15*** -0.16*** 0.01 -0.08**

Cold, deep

Acartia sp. -0.19*** -0.26*** -0.11*** -0.06* -0.13*** -0.17*** -0.12***

Decapod larvae -0.16*** -0.09** -0.02 -0.05` -0.02 -0.02 0.02Total fish larvae -0.05 0.01 0.03 0.11*** 0.14** -0.01 0.08***Summer fish larvae -0.11 -0.01 0.06 0.02 -0.01 -0.05 0.12***Spring Fish larvae -0.06 0.02 -0.02 0.08* 0.09 0.02 -0.02

Table 7: Seasonal Mann-Kendall trend test (tau statistic) for monthly mean indices of phytoplankton colour, crustacean zooplankton and fish larval abundance (total and split by summer / spring spawners) where: *** p < 0.01; ** p < 0.05; and * p < 0.10.

IV.2.2 Identification of regime shifts

Regime shifts were identified using two complementary methods: exploratory ‘hierarchical clustering’ and statistical ‘breakpoint analysis’ based on Principle Components Analysis (PCA)73.

Figure 28: Shifts in the Celtic Seas ecoregion identified in several principal components from multiple ICES divisions for each level of the ecosystem. Starting from the top: CPR fish larvae (PC1 66% of variability in VIIeh), Phytoplankton Colour Index or CHLO (PC1 62% of variability in VIIaegh), SST (PC1, 83% of variability in SST for VIIaegh), and climate indices (PC2, 29% of variability in NAO, EA, SCA, AMO).


Regime shifts in 1986 and 1994 (Fig. 28) were identified statistically in the Celtic Seas ecoregion (including ICES divisions VIa, VIIa and VIIe-j). The earlier regime shift was characterised by a change in the climate (increase in the NAO and decrease in the SCA), followed by a rise in phytoplankton standing stock (colour index or CHLO) in the Irish Sea (VIIa) and northern Celtic Sea (VIIg) and a fall in the recruitment of plaice Pleuronectes platessa and cod Gadus morhua stocks. The later shift was triggered by a rise in temperature, in both the AMO and local SST, and coincided with a general rise in phytoplankton standing stock, a fall in total fish larvae (dominated by clupeids) sampled by the CPR in the southern Celtic Sea (VIIh) and an increase in the western Channel (VIIe).

IV.3 Predicting species-specific response from climate variability

The work above in the North and Celtic Seas suggest that the availability of prey seems to be the main driver for the survival of fish larvae under various environmental conditions. The cascading effects of warming temperature, through the plankton ecosystem, in recent decades has been demonstrated in many previous studies, and the results here similarly suggest that this indirect effect of temperature on the survival of larval fish is the main pathway of climate change effects on fish recruitment. In the case of Atlantic mackerel, and possibly other migratory species, however, hydrographic features seem to play a more dominant role by controlling migration of the adults to their spawning ground 53,74. Nevertheless, these two effects are not mutually exclusive. Recent research has shown that that the main spawning locations of fish such as cod, haddock, plaice, long rough dab and sandeel in the North Sea are linked to recurrent hydrographic features75, and that Atlantic mackerel recruitment is heavily dependent on the production of the copepod nauplii that contribute extensively to the diet of its larvae76. In addition, there may be other factors not considered here also playing an important role. These results illustrate the complexity of the mechanisms involved in the appearance and survival of fish larvae.

V. Conclusion

The results here suggest a cascading effect from climate variability to plankton (including fish larvae) via hydrographical changes. Climate acts as a fundamental driving force on the environmental status of the marine ecosystem; the status of the environment will, in turn, impact on the ecosystem’s ecological status. Therefore, changes in climate, hydrography and temperature are likely to indirectly affect larval survival and hence recruitment through the plankton. However for migratory species such as mackerel, factors linked to hydrographic variability likely play a more prominent part in influencing spawning of the adults. Ideally, this hypothesis should be verified with other similar species.

Our results indicate that responses to environmental variability as well as the ability of fish larvae to adapt to these changes, are species-specific, and with further, in-depth analysis, it might be possible to develop some species-specific indices of “larval survival potential” under specific environmental conditions and state of the ecosystem resulting from climate change. However, the complexity of the mechanisms involved, highlighted within the scope of this research, is further increased by the added complication of the impact of fisheries. Overfishing not only reduces stock abundance, but also renders stocks more sensitive to additional stress from either fishing pressure (top-down control) or environmental variability (bottom-up control)77-81. The effects of climate and fishing interact with each other in a manner that climate may cause failure of a fishery management plan, while exploitation may disrupt the ability of a fish population to withstand or adapt to climate change. The interacting effects of fishing and climate make the prediction of the response of marine fish to climate variability or climate change scenarios an immense challenge. So far, attempts to separate these effects empirically have not been successful; it might well be that the complexity of such interactions and the mechanisms involved in these processes makes it an unrealistic possibility. For future fisheries research, It might be more feasible to adopt a more integrated approach and support studies orientated towards understanding the interactions between climate and fishing, rather than trying to separate them, as recently suggested67,80.

We have shown that, the CPR fish larvae dataset, with its extensive spatio-temporal coverage provides a valuable insight in the spatio-temporal changes undergone by fish larvae in the last 60 years. For the most abundant taxa at least, the CPR is a reliable tool for capturing the spatial patterns and the relative abundances of fish larvae between areas. This new dataset clearly offers a unique opportunity to investigate the responses of fish populations to past changes in climate, including abrupt changes such as regime shifts; and can also inform recruitment levels for stock assessment purposes. Furthermore, our study on mackerel larvae shows that the dataset can valuable contribution to understanding the dynamic of fish populations. However, the dataset also contains some limitations, mainly the impossibility to separate the several species of clupeids, and its end date of 2005. However, both these limitations could theoretically be overcome; as CPR samples post 2005 have been archived and recent advances molecular analysis of


CPR archived material now allow molecular identification of fish larvae, and genetic probes have been developed for herring.

...For future work, we suggest catching up on the analysis of fish larvae from the year 2006 with the view to make the analysis a routine part of CPR sample processing. This would permit the use of the data to aid in the assessment of commercial species such as mackerel, sandeels and dab; and aid policy makers to design medium to long-term management plans for these species. In the case of clupeids, further genetic analysis using molecular techniques would be required. The probes have already been developed and the technique is becoming cheaper all the time. The value of the data obtained would most likely offset the cost involved....In particular, we have demonstrated that the larval index built for mackerel could be used as a proxy for eggs, with the added value of its extended coverage compared with the existing egg surveys. The availability of the post-2005 data for mackerel larvae, would allow us to extend the index and this would make a valuable contribution to understanding the several potential problems regarding the mackerel stock, in combination with other existing datasets such as the acoustic data.

The Team:

Clive Fox (SAMS). Martin Edwards (SAHFOS): Instigating the proposal.Nick Halliday (MBA): painstaking visual analysis of CPR samples.SAHFOS team: production of fish larvae atlas.Clive Fox (SAMS): research on clupeids.Christopher Lynam (Cefas): research on sandeels and ecosystem regime shifts.Teunis Jansen (DTU Aqua): research on mackerel.Sophie Pitois (Cefas): Project manager, fish larvae atlas, research on climate change impact on fish

populationsJason Holt (POL): providing formatted temperature and salinity data issued from the POLCOM model.Steve Mackinson (CEFAS), Nick Dulvy (SFU): Previous project managers.Kieran Hyder (CEFAS): Project Sponsor.


References to published material9. This section should be used to record links (hypertext links where possible) or references to other

published material generated by, or relating to this project.Publications generated by this project

Lynam, C., Pitois S. And Edwards M. (2010) Spatial patterns and trends in abundance of larval Ammodytidae from Continuous Plankton Recorder surveys in the North Sea: 1950- 2005. Report for use by the ICES Workshop on Sandeel.

Edwards, M., Helaouet, P., Halliday. N., Beaugrand, G., Fox, C., Johns, D.G., Licandro, P., Lynam, C., Pitois, S., Stevens, D, & Coombs, S. 2011. Fish Larvae Atlas of the NE Atlantic. Results from the Continuous Plankton Recorder survey 1948-2005. Sir Alister Hardy Foundation for Ocean Science. 22p. Plymouth, U.K. ISBN No: 978-0-9566301-2-7

Sophie G. Pitois, Christopher P. Lynam, Teunis Jansen, Nick Halliday, Martin Edwards (in press). Towards a better understanding of the impact of bottom-up effects derived by climate variability on fish populations, using fish larvae data from Continuous Plankton Recorded. MEPS.

Teunis Jansen; Kasper Kristensen; Mark Payne; Martin Edwards; Corinna Schrum; Sophie Pitois Submitted). Long-term Retrospective Analysis of Mackerel Spawning in the North Sea: A New Time Series and Modeling Approach to CPR Data. Plos One

Christopher P. Lynam, Nicholas C. Halliday, Cindy J. G. van Damme, Martin Edwards, Hannes Höffle, Peter J. Wright, Sophie Pitois (in prep). Spatial patterns and trends in abundance of larval sandeels in the North Sea: 1950-2005.

Christopher P. Lynam, Kieran Hyder, David Johns, Martin Edwards, Sophie Pitois (in prep.). Ecosystem regime shifts in the Celtic and North Seas Ecoregions.

Oral presentations

Pitois, S. G., Lynam, C. P., Halliday, N. C. and Edwards, M. Long-term changes in the distribution and abundance of selected fish larvae from the CPR (1950-2005) over the UK shelf, in relation to biological and environmental factors. Presented at 5th International Symposium on Zooplankton Production, Pucon, Chile (March 2011).

Lynam C. P., Pitois S, Halliday N. C., van Damme C. and Edwards M. Spatial patterns and trends in abundance of larval Ammodytidae from Continuous Plankton Recorder surveys of the North Sea: 1950-2005. Presented at ICES Annual Science Conference 2011, Gdańsk, Poland. 19-23 September 2011.

Poster presentations

Lynam, C. P., Pitois, S. G., Halliday N. C, Edwards M. and van Damme C. Spatio-temporal patterns in the abundance of larval sandeels from CPR surveys in the North and Irish Seas. Presented at the International conference on Zooplankton – Pucon, Chile – March 2011.

References cited

[1] Edwards M, Richardson AJ (2004) Impact of climate change on marine pelagic phenology and trophic mismatch. Nature 430:881–884

[2] Richardson AJ, Schoeman DS (2004) Climate impact on plankton ecosystems in the Northeast Atlantic. Science 305:1609–1612

[3] Beaugrand G (2004) The North Sea regime shift: evidence, causes, mechanisms and consequences. Prog Oceanogr 60:245–262

[4] ter Hofstede R, Hiddink JG, Rijnsdorp AD (2010) Regional warming changes fish species richness in the eastern North Atlantic Ocean. Mar Ecol Progr Ser 414:1–9

[5] Cheung WWL, Lam VWY, Sarmiento JL, Kearney K, Watson R, Pauly D (2009) Projecting global marine biodiversity impacts under climate change scenarios. Fish Fish 10:235–251

[6] Jansen T, Gislason H (2011) Temperature affects the timings of spawning and migration of North Sea mackerel. Cont Shelf Res 31:64–72

[7] Houde ED (2008) Emerging from Hjort's shadow. J Northwest Atl Fish Sci 41:53–70[8] Horwood J, Cushing D, Wyatt T (2000) Planktonic determination of variability and sustainability of

fisheries. J Plankton Res 22:1419–1422


[9] Hays G, Richardson AJ, Robinson C (2005) Climate change and marine plankton. Trends in Ecol Evol 20:337–344

[10] Beare DJ, Batten SD, Edwards M, McKenzie E, Reid PC, Reid DG (2003) Summarizing spatial and temporal information in CPR data. Prog Oceanogr 58:217–233

[11] Reid PC, Colebrook JM, Matthews JBL, Aiken J (2003) The continuous plankton recorder: concepts and history, from plankton indicator to undulating recorders. Prog Oceanogr 58:117–173

[12] Brander KM, Dickson RR, Edwards M (2003) Use of Continuous Plankton Recorder information in support of marine management: applications in fisheries, environmental protection, and in the study of ecosystem response to environmental change. Prog Oceanogr 58:175-191

[13] Coombs SH (1975) Continuous plankton records show fluctuations in larval fish abundance during 1948–72. Nature 258:134–136

[14] Lindley JA, Halliday NC, Stevens DP (2007) Fish data from the Continuous Plankton Recorder survey. ICES CM 2007/E: 24. 16 pp

[15] Colebrook JM (1975) The continuous plankton recorder survey: automatic data processing methods. Bull Mar Ecol 8:123–142

[16] Batten SD, Clark R, Flinkman J, Hays G, John E, John AWG, Jonas T, Lindley JA, Stevens DP, Walne A (2003) CPR sampling: the technical background, materials and methods, consistency and comparability. Prog Oceanogr 58:193–215

[17] Coombs SH (1980) Continuous plankton records: a plankton atlas of the North Atlantic and North Sea: Supplement 5 - Young fish, 1948-1972. Bull Mar Ecol 8: 229-281

[18] Hurrell JW (1995) Decadal trends in the North Atlantic Oscillation: regional temperatures and precipitation. Science 269:676–679

[19] Enfield DB, Mestas-Nunez AM, Trimble PJ (2001) The Atlantic Multidecadal Oscillation and its relationship to rainfall and river flows in the continental U.S. Geophys Res Letters 28: 2077-2080

[20] Schrum C, Alekseeva I, St John M (2006) Development of a coupled physical-biological ecosystem model ECOSMO - Part I: Model description and validation for the North Sea. J Mar Syst 61: 79-99

[21] Christensen A, Daewel U, Jensen H, Mosegaard H, St. John M, et al. (2007) Hydrodynamic backtracking of fish larvae by individual-based modelling. Mar Ecol Prog Ser 347: 221-232

[22] Skogen M, Søiland H. (1998) A user´s guide to NORWECOM v.2. The NORWegian ECOlogical Model system. Fisk Havet 18.

[23] Robinson GA (1970) Continuous plankton records: Variation in the seasonal cycle of phytoplankton in the north Atlantic. Bull Mar Ecol 6: 333–345

[24] Clark DR, Aazem KV, Hays GC (2001) Zooplankton abundance and community structure over a 4000 km transect in the North-east Atlantic. J Plankton Res 53:365–37

[25] Pitois SG, Fox CJ (2006) Long-term changes in zooplankton biomass concentration and mean size over the Northwest European shelf inferred from Continuous Plankton Recorder data. ICES J Mar Sci 63:785–79

[26] Johnson PO (1977) A review of spawning in the North Atlantic mackerel, Scomber scombrus L. Fish Res Tech Rep 37. MAFF, Lowestoft, UK

[27] Legendre P, Legendre L (1998) Spatial analysis. In: Legendre P, Legendre L (eds) Numerical ecology: Developments in environmental modelling. Elsevier, Amsterdam

[28] Beaugrand G, Ibañez F, Lindley JA (2003) An overview of statistical methods applied to CPR data. Progr Oceanogr 58:235–262

[29] Beaugrand G, Ibañez F, Lindley JA (2001) Geographical distribution and seasonal and diel changes in the diversity of calanoid copepods in the North Atlantic and North Sea. Mar Ecol Progr Ser 219:189–203

[30] ICES (2007). Results of the spring 2004 North Sea ichthyoplankton surveys. ICES Coop Res Rep No. 285

[31] Borcard D, Gillet F, Legendre P (2011) Numerical Ecology with R, Springer, New York, 306 pp[32] Munk P, Wright PJ, Pihl NJ (2002) Distribution of the Early Larval Stages of Cod, Plaice and Lesser

Sandeel across Haline Fronts in the North Sea. Estuar Coast Shelf Sci 55:139–149[33] Richardson AJ, John EH, Irigoien X, Harris RP, Hays GC (2004) How well does the continuous

plankton recorder (CPR) sample zooplankton? A comparison with the Longhurst Hardy Plankton Recorder (LHPR) in the northeast Atlantic. Deep Sea Res I 51:1283-1294

[34] Dickey-Collas M, Nash RDM, Brunel T, Van Damme CJG, Marshall CT, Payne MR, Corten A, Geffen AJ, Peck MA, Hatfield EMC, Hintzen NT, Enberg K, Kell LT, Simmonds EJ (2010) Lessons learned from stock collapse and recovery of North Sea herring: a review. ICES J Mar Sci 67: 1875-1886

[35] ICES (2010) Report of the Benchmark Workshop on Sandeel (WKSAN), Copenhagen, Denmark. ICES CM 2010/ACOM:57

[36] ICES 2011. Advice Book 6. Section 6.4.21[37] Lynam CP, Pitois SG, Halliday NC, van Damme CJG, Edwards M, Höffle H, Wright PJ. Spatial

patterns and trends in abundance of larval sandeels in the North Sea: 1950-2005. In preparation[38] Hamre J. (1978) The effect of recent changes in the North Sea mackerel fishery on stock and yield.

Rapp p-v Reun (CIEM) 172: 197-210[39] ICES (2011) Report of the Working Group on Widely Distributed Stocks (WGWIDE). ICES CM


2011/ACOM:15[40] ICES (2010) Report of the Working Group on Widely Distributed Stocks (WGWIDE). ICES 376 S CM

2010/ACOM:15[41] Bartsch J (2002) Modelling the temperature mediation of growth in larval fish. Fish Oceanogr 11: 310-

314[42] Mendiola D, Alvarez P, Cotano U, Etxebeste E, De Murguia AM (2006) Effects of temperature on

development and mortality of Atlantic mackerel fish eggs. Fish Res 80: 158-168[43] Lewy P, Kristensen K (2009) Modelling the distribution of fish accounting for spatial correlation and

overdispersion. Can J Fish Aquat Sci 66: 1809-1820[44] Jansen T, Gislason H (2011) Temperature affects the timing of spawning and migration of North Sea

mackerel. Cont Shelf Res 31: 64-72.[45] Hillgruber N, Kloppmann M (2001) Small-scale patterns in distribution and feeding of Atlantic

mackerel (Scomber scombrus L.) larvae in the Celtic Sea with special regard to intra-cohort cannibalism. Helgol Mar Res: 149

[46] Röpke A (1989) Small-scale vertical distribution of ichthyoplankton in the Celtic Sea in April 1986. Meeresforsch 32: 192-203

[47] Iversen SA (1981) Spawning and trends in spawning stock size of the North Sea mackerel during the period 424 1973-1980. ICES CM 1981/H:16

[48] ICES. (1979) Report of the mackerel working group. ICES CM 1979/H:5[49] Hátún H, Payne MR, Beaugrand G, Reid PC, Sandø AB, Drange H, Hansen, B, Jacobsen JA, Bloch

D (2009) Large bio-geographical shifts in the north-eastern Atlantic Ocean: From the subpolar gyre, via plankton, to blue whiting and pilot whales. Prog Oceanogr 80:149-162

[50] Dickson RR, Meincke J, Malmberg S-A, Lee AJ (1988) The "great salinity anomaly" in the northern North Atlantic 1968–1982. Prog Oceanogr 20:103–151

[51] Belkin IM, Levitus S, Antonov J, Malmberg S-A (1998) "Great Salinity Anomalies" in the North Atlantic. Progr Oceanogr 41:1–68

[52] Edwards M, Beaugrand G, Reid PC, Rowden AA, Jones MB (2002) Ocean climate anomalies and the ecology of the North Sea. Mar Ecol Progr Ser 239:1–10

[53] Reid PC, Borges Mde F, Svendsen E (2001) A regime shift in the North Sea circa 1988 linked to changes in the North Sea horse mackerel fishery. Fish Res 50:163–171

[54] Beaugrand G (2004) The North Sea regime shift: evidence, causes, mechanisms and consequences. Prog Oceanogr 60:245–262

[55] Beaugrand G, Ibañez F (2004) Monitoring marine plankton ecosystems. 2. Long-1 term changes in North Sea calanoid copepods in relation to hydro-climatic variability. Mar Ecol Progr Ser 284:35–47

[56] Ottersen G, Planque B, Belgrano A, Post E, Reid PC, Stenseth NC (2001) Ecological effects of the North Atlantic Oscillation. Oecologia 128:1-14

[57] Stenseth NC, Ottersen G, Hurrell J, Belgrano A (eds) Marine Ecosystems and Climate variation: the North Atlantic. Oxford University Press, Oxford

[58] Edwards M, Reid P, Planque B (2001) Long-term and regional variability of phytoplankton biomass in the Northeast Atlantic (1960–1995). ICES J Mar Sci 58:39–49

[59] Beaugrand G, Brander KM, Lindley JA, Souissi S, Reid PC (2003) Plankton effect on cod recruitment in the North Sea. Nature 426:661–664

[60] Beaugrand G, Kirby R (2010) Climate, plankton and cod. Glob Change Biol 16:1268–1280[61] Kirby RR, Beaugrand G (2009) Trophic amplification of climate warming. Proc R Soc Lond B

276:4095–4103[62] Last JM (1980) The food of twenty species of fish larvae in the west–central North Sea. MAFF,

Directorate of Fisheries Research, Lowestoft, UK[63] Checkley DMJ (1982) Selective feeding by Atlantic herring (Clupea harengus) larvae on zooplankton

in natural assemblages. Mar Ecol Progr Ser 9:245–253[64] Cohen RE, Lough RG (1983) Prey field of larval herring Clupea harengus on a continental shelf

spawning area. Mar Ecol Progr Ser 10:211–222[65] Fossum P (1983) Digestion rate of food particles in the gut of larval herring (Clupea harengus L.).

FiskDir Skr Ser HawUnders 17:347–357[66] Peterson WT, Ausubel SJ (1984) Diets and selective feeding by larvae of Atlantic mackerel Scomber

scombrus on zooplankton. Mar Ecol Progr Ser 17:65–75[67] Munk P, Kiørboe T (1985) Feeding behaviour and swimming activity of larval herring (Clupea

harengus) in relation to density of copepod nauplii. Mar Ecol Progr Ser 24:15–21[68] Simonsen CS, Munk P, Folkvord A, Petersen SA (2006) Feeding ecology of Greenland halibut and

sandeel larvae off West Greenland. Mar Biol 149:937–952[69] Malzahn AM, Boersma M (2009) Trophic flexibility in larvae of two fish species (lesser sandeel,

Ammodytes marinus and dab, Limanda limanda). Sci Mar 73:131–139[70] de Figueiredo GM, Nash RDM, Montagnes DJS (2005) The role of the generally unrecognised

microprey source as food for larval fish in the Irish Sea. Mar Biol 148:395–404[71] Pedersen T, Fossheim M (2008) Diet of 0-group stages of capelin (Mallotus villosus), herring (Clupea

harengus) and cod (Gadus morhua) during spring and summer in the Barents Sea. Mar Biol


153:1037–1046[72] Wood SN (2008) Fast stable direct fitting and smoothness selection for generalized additive models. J

Roy Stat Soc B:70(3):495-518[73] Andersen T, Carstensen J, Hernandez-Garcıa E, Duarte CM (2008) Ecological thresholds and regime

shifts: approaches to identification. Trends Ecol Evol 24(1):49-57[74] Jansen T, Kristiansen K, Payne M, Halliday N, Pitois S (in review) North Sea spawning mackerel and

the stock crash in northern North East Atlantic mackerel[75] Munk P, Fox CJ, Bolle LJ, van Damme CJG, Fossum P, Kraus G (2009) Spawning of North Sea

fishes linked to hydrographic features. Fish Oceanogr 18:458–469[76] Castonguay M, Plourde S, Robert D, Runge JA, Fortier L (2008) Copepod production drives

recruitment in a marine fish. Can J Fish Aquat Sci 65:1528–1531[77] Brander KM (2005) Cod recruitment is strongly affected by climate when stock biomass is low. ICES J

Mar Sci 62:339–343[78] Ottersen G, Hjermann DØ, Stenseth NC (2006) Changes in spawning stock structure strengthen the

link between climate and recruitment in a heavily fished cod (Gadus morhua) stock. Fish Oceanogr 15:230–243

[79] Anderson CNK, Hsieh C-H, Sandin SA, Hewitt R, Hollowed A, Beddington J, May RM, Sugihara G (2008) Why fishing magnifies fluctuations in fish abundance. Nature 452:835–839

[80] Perry RI, Cury P, Brander K, Jennings S, Möllmann C, Planque B (2010) Sensitivity of marine systems to climate and fishing: concepts, issues and management responses. J Mar Syst 79:427–435

[81] Planque B, Fromentin J-M, Cury P, Drinkwater KF, Jennings S, Perry RI, Kifani S (2010) How does fishing alter marine populations and ecosystems sensitivity to climate? J Mar Syst 79:403–417


Documents

General enquiries on this form should be made to: - Defra, UKrandd.defra.gov.uk/...M1101_finalreportv3(2).docx · Web viewThis form is in Word format and the boxes may ... we used