The state of the fish community in ospar region 2, The Greater North Sea

Supplementary Material Annex 1: Analyses of groundfish survey data in sub-regions of the UK’s Exclusive Economic Zone to examine trends in the “health” of the demersal fish community.

Simon Greenstreet, Helen Fraser, John Cotter and John Pinnegar

This annex provides descriptions of the data available in each sub-region and presents the results of the analyses applied to these data. The results of these analyses, the trends in the 15 univariate community metrics, provide the basis for the sub-regional and regional overview presented in the main body of the report.

Sub-Regional Assessments

Northern North Sea

Two main international groundfish surveys co-ordinated through ICES, the International Bottom Trawl Surveys (IBTS), have been carried out across most of ICES area IV, the North Sea, in the first (Q1) and third (Q3) quarters of each year. The Q1 time series covers the period 1983 to 2008, while the Q3 survey only commenced in 1991 (as a co-ordinated survey) and 2008’s survey data were not yet available for analysis. These surveys utilise a Grande Ouverture Verticale (GOV) trawl and sample tows are generally of 30min duration. Trawl samples are quantified to provide estimates of the numbers of each 1cm length class of each species caught at each sample location. Catch numbers were converted to densities by dividing by the area swept by the trawl (mean distance between the wings multiplied by distance the gear was towed). Where wing-spread distance was not known, this was estimated using a depth dependent relationship derived for the standard GOV.

Correspondence between metric time series derived from the Q1 and Q3 survey data was generally high and, in 12 of the15 cases, was improved by the introduction of a plus 1-year time lag to the Q3 data (Table S1). No clear trend in LFI over the last decade was provided by either data set, while mean fish size has varied markedly, but shown no systematic trend over the ten years. Both metrics, however, suggest that community size composition has shifted towards smaller fish now compared with the start of the time series. However, abundance, biomass and productivity have all declined over the last ten years and are currently at similar levels to those present during the mid to late 1980s. Little consistent trend was apparent in any of the five metrics of species richness and species evenness over the last decade, but all five metrics indicate higher species diversity now than during the 1980s. Life-history trait metrics have oscillated in value throughout the time series, with no systematic trend obvious over the last decade, and little to suggest that the community life-history composition is any different now than at the start of the time series. There is perhaps some evidence of a recent increase in fish age at maturity (Figure S1).

Southern North Sea

Data from the same two North Sea surveys that provided information for the Northern North Sea were analysed to determine metric trends in the Southern North Sea.

Correspondence between the metric time series was generally poor, and in this instance not markedly improved by the introduction of a time lag to the Q3 data (Table S2). Despite this apparent dichotomy in the detailed temporal variation in the two time series, trends in both data sets were generally consistent (Figure S2). Over the last decade the LFI has declined sharply, but then recovered, while little trend was apparent in the mean size of fish. Biomass, abundance and productivity all showed opposite changes to the LFI, such that over the 10 year period, no trend was obvious. While abundance, biomass and productivity are currently at similar levels to those prevalent in the mid to late 1980s, the proportion of large fish and mean weight of fish are both markedly lower. The recent recovery therefore needs to be maintained. Both surveys suggest recent increases in demersal fish species richness and evenness. Species evenness is currently about on a par with levels present at the start of the time series, while species richness is currently higher. Trends in the four life-history metrics were variable, but tend to suggest that the community now contains a higher percentage of individuals of opportunistic species, and fewer climax species fish. Over the longer term, however, there was no obvious difference between now and the start of the time series. This recent trend therefore needs to be reversed before a problem emerges.

Eastern Channel

French Groundfish Survey of the Channel (FGSC) has been undertaken in the fourth quarter of the year over the period 1988 to 2008. This survey utilises a Grande Ouverture Verticale (GOV) trawl and sample tows are generally of 30min duration. Trawl samples are quantified to provide estimates of the numbers of each 1cm length class of each species caught at each sample location. Catch numbers were converted to densities by dividing by the area swept by the trawl (mean distance between the wings multiplied by distance the gear was towed). Where wing-spread distance was not known, this was estimated using a depth dependent relationship derived for the standard GOV.

Few obvious trends were apparent in any of the fifteen univariate metrics applied to the data, but some evidence of a major methodological change in 1998 was apparent making interpretation difficult. The analysis provided no evidence to suggest that any major changes in the composition, structure, and functioning of the demersal fish community had occurred in the eastern English Channel during the last decade, although fish size may have increased, particularly in the last few years (Figure S3). Comparison of the current state of the community with the situation at the start of the time series was hampered by the apparent methodological problem, however, taken at face value, the data suggest the current community is more dominated by a few species, and consists of smaller fish. Species evenness may also be lower now than at the start of the time series, but the possible methodological issue may have influenced these apparent long-term trends.

Western Channel and Celtic Sea

Data from three surveys using a GOV trawl were available for the Western Channel and Celtic Sea sub-region. Two commenced comparatively recently, both carried out in the fourth quarter (Q4) of the year; the French Celtic Sea Groundfish Survey (FCSGS) started in 1997, while the English Southwest Bottom Trawl Survey (ESWBTS) only started in 2003. Both surveys are still extant and the FCSGS at least covers the decade of main interest for this report. The third survey, the English Celtic Sea Groundfish Survey (ECSGS), carried out in Q1, provides a longer time series, extending back to 1984, but coverage of this sub-region by this survey ceased in 2004. Trawl samples were quantified to provide estimates of the numbers of each 1cm length class of each species caught at each sample location. Catch numbers were converted to densities by dividing by the area swept by the trawl (mean distance between the wings multiplied by distance the gear was towed). Where wing-spread distance was not known, this was estimated using a depth dependent relationship derived for the standard GOV. Correspondence between in the metric trends derived from these three surveys was generally weak. When comparing Q1 and Q4 trends, introduction of a plus one-year lag to the Q4 data had marginal effect: eight correlations were improved, while seven decreased. However, for two metrics, the LFI and the P/B ratio, introduction of the lag resulted in significant correlations between the Q1 ECSGS and Q4 FCSGS data (Table S3).

Despite the lack of detailed (year on year) agreement between the different metric time series, examination of the general trends provided greater similarity between the metric signals (Figure S4), rendering an assessment of the state of the demersal fish community in the sub-region possible. Biomass, abundance and productivity all appear to have increased, with a clear increase in the last ten years evident. A long-term increase in the P/B ratio was clearly indicated. These changes have been associated with a marked reduction in the proportion of large fish and mean weight of fish in the demersal fish community. This trend may have been halted in the last decade, but not reversed. The French Q4 survey suggests an increase in both species richness and species evenness over the last decade, a change that is to some extent corroborated by the shorter five-year span of the English Q4 time series. Because of this, demersal fish species diversity (evenness and richness) may well now be higher than when data were first available. However, this involves a qualitative comparison of these two surveys with the longer lived, but now ended, English Q1 data. The four life-history trait metrics suggest a recent return to a more “climax” type community with increases in average ultimate body-length, age and length at maturity and decreased average growth rate all indicated. These four metric time series all indicate long-term improvement in this aspect of the composition, structure and function of the demersal fish community in the sub-region.

Irish Sea

Data from three surveys were available for the Irish Sea sub-region, all of which commenced comparatively recently. Two surveys conducted in Q1 and Q3 by Northern Ireland, the Northern Irish Groundfish Survey (NIGS) commenced in 1992. The third survey, carried out in Q1 by Scotland (the SWCGS) provides data since 1996, but this survey ceased coverage in this sub-region in 2006. These surveys use a GOV trawl. Trawl samples were quantified to provide estimates of the numbers of each 1cm length class of each species caught at each sample location. Catch numbers were converted to densities by dividing by the area swept by the trawl (mean distance between the wings multiplied by distance the gear was towed). Where wing-spread distance was not known, this was estimated using a depth dependent relationship derived for the standard GOV. Correspondence between the two Irish surveys was relatively weak, although the four abundance/biomass/productivity metric time-series were all significantly correlated. Adding a plus one-year time lag to the Q3 data weakened these four correlations, but strengthened eight of the remaining eleven relationships, rendering the LFI, average ultimate body length, growth rate and length at maturity correlations statistically significant. Twelve of the fifteen correlations between metric time series derived from the NIGS and SWCGS Q1 data sets were positive, and seven of these significantly so (Table S4).

Over the last decade, the abundance, biomass and productivity of the demersal fish community in the Irish Sea have all declined in line with an increase in the proportion of large fish present. At least one of the three surveys indicates an increase in species richness and all three time series suggest that species evenness has increased in recent years. Of the four life-history trait metrics, only age at maturity appears to have increased, while no trend is apparent over the last ten years in the other three metrics. Over the longer term, species diversity (richness and evenness) appears higher now than when data were first available. The four life-history metrics all suggest a shift towards a more mature community. The abundance, biomass and productivity metrics suggest a current situation similar to that prevalent when surveys began, while the proportion of large fish may be higher now than in the 1990s (Figure S5).

Minches and Western Scotland

Scotland operates two surveys in the southwestern part of the Scottish Continental Shelf, the Scottish West Coast Groundfish Surveys (SWCGS) carried out in the fourth (Q4) and first quarters (Q1) of the Year. Data from the Q1 survey were available from 1986 to 2008 and in Q4 from 1996 to 2007. These surveys utilise a Grande Ouverture Verticale (GOV) trawl and sample tows are generally of 30min duration. Trawl samples are quantified to provide estimates of the numbers of each 1cm length class of each species caught at each sample location. Catch numbers were converted to densities by dividing by the area swept by the trawl (mean distance between the wings multiplied by distance the gear was towed). Where wing-spread distance was not known, this was estimated using a depth dependent relationship derived for the standard GOV.

Trends in each of the community metrics were again generally positively correlated, and the level of correlation tended to be enhanced by the introduction of a 1 year time lag to the Q4 data so that Q4_year=x+1 was related to Q1_year=x (Table S5). Trends in the 15 metrics generally indicate an improving situation for the demersal fish community in the Hebridean Continental Shelf sub-region. Since 2000, the proportion of large fish in the area has increased. Species diversity appears to have increased; both surveys indicate increased evenness of abundance and reduced dominance and the Q4 survey indicates an increase in species richness. Trends in average life-history trait composition in the community are more equivocal; variation was high and little trend was discernable (Figure S6). Over the whole period for which data area available, species evenness and life-history trait composition of the demersal fish community appear to be in better state now than at the beginning of the data time series. There is however, some suggestion that species richness in the sub-region is currently lower than when surveys first began. The abundance/biomass/productivity and size composition aspects of community composition, structure and function appear to be little different now than when data were first available (Figure S6).

Scottish Continental Shelf

This sub-region straddled the boundary between ICES areas IV and VI, and because of this none of the groundfish surveys available for analysis adequately represented the situation across the entire region. Differences in survey design and timing precluded aggregation of the different data sets. It was necessary therefore to further sub-divide the sub-region into two separate areas and assess the state of the demersal fish community in each area using the different survey data sets available.

(Southwest)
Scotland operates two surveys in the southwestern part of the Scottish Continental Shelf, the Scottish West Coast Groundfish Surveys (SWCGS) carried out in the fourth (Q4) and first quarters (Q1) of the Year. Data from the Q1 survey were available from 1986 to 2008 and in Q4 from 1996 to 2007. These surveys utilise a Grande Ouverture Verticale (GOV) trawl and sample tows are generally of 30min duration. Trawl samples are quantified to provide estimates of the numbers of each 1cm length class of each species caught at each sample location. Catch numbers were converted to densities by dividing by the area swept by the trawl (mean distance between the wings multiplied by distance the gear was towed). Where wing-spread distance was not known, this was estimated using a depth dependent relationship derived for the standard GOV.
Trends in each of the community metrics were generally positively correlated, and the level of correlation tended to be enhanced by the introduction of a 1 year time lag to the Q4 data so that Q4_year=x+1 was related to Q1_year=x (Table S6). Trends in the 15 metrics generally indicate an improving situation for the demersal fish community in the Hebridean Continental Shelf sub-region. Since 2000, the proportion of large fish in the area has increased. Species diversity appears to have increased; both surveys indicate increased evenness of abundance and reduced dominance and the Q4 survey indicates an increase in species richness. The proportion of larger-sized, slow-growing, late-maturing species has increased. As a consequence of these changes, abundance and growth productivity have declined (Figure S7). Over the whole period for which data area available, species evenness and life-history trait composition of the demersal fish community appear to be in better state now than at the beginning of the data time series. The other three aspects of community composition, structure and function appear to be little different now than when data were first available (Figure S7).
(Northeast)
Two main international groundfish surveys, co-ordinated through ICES, the International Bottom Trawl Surveys (IBTS), have been carried out across most of ICES area IV, the North Sea, in the first (Q1) and third (Q3) quarters of each year. The Q1 time series covers the period 1983 to 2008, while the Q3 survey only commenced in 1991 (as a co-ordinated survey) and 2008’s survey data were not yet available for analysis. These surveys utilise a Grande Ouverture Verticale (GOV) trawl and sample tows are generally of 30min duration. Trawl samples are quantified to provide estimates of the numbers of each 1cm length class of each species caught at each sample location. Catch numbers were converted to densities by dividing by the area swept by the trawl (mean distance between the wings multiplied by distance the gear was towed). Where wing-spread distance was not known, this was estimated using a depth dependent relationship derived for the standard GOV.
Correspondence between the two survey trends was generally low but, with the exception of the of the biomass, abundance, productivity and P:B ratio metrics, this was markedly improved by the introduction of a plus 1-year time lag to the Q3 time series (Table S7). Over the last decade, variability in the LFI has increased in both data sets, but generally the data suggest an increase in the proportion of large fish in the community; a result corroborated by the mean weight index. Both data sets suggest that biomass, abundance, productivity and the P/B ratio all declined at first and then increased, so that over the 10 year period no real trend was observed. Similarly, over the recent decade, no obvious trends in the two species richness, three species evenness, and four life-history trait metrics was discernable in either of the two survey time series (Figure S8). Comparison of the current situation with state of the demersal fish community at the start of the time series reveals a similar story. Values of most metrics over the last five years were similar to those observed in the first 8 years when data were available; only species richness appears higher now than during the mid to late 1980s (Figure S8).

Rockall trough and Bank

The Rockall Trough and bank sub-region covers a large area, much of which consists of deeper, non-continental shelf waters. Very little systematic monitoring of fish communities has been carried out in the region and only one set of survey data were available for analysis of trends in the health of the fish community in the region; the Scottish Rockall Groundfish Survey (SRGS). This survey was first undertaken in 1986, but biannual surveys only commenced in 2001, later becoming annual surveys in 2005. 2008 data were not available at the time of writing. The survey essentially operates on the Rockall bank, so does not really cover deep water as such. Some trawl samples were taken from water in excess of 250m depth, but too few for serious analysis. All trawl samples collected from depths >250m were therefore excluded from analysis to avoid these unusual samples influencing the values of the metrics calculated from the rest of the data. The sub-region in which this survey operates is shown in Figure S9, but in essence the trawl samples collected by the SRGS were obtained from a much more restricted area. Being essentially a shallow plateau, the area surveyed is totally unrepresentative of the fish communities that occupy the more wide-spread deeper parts of the sub-region.

The survey utilises a Grande Ouverture Verticale (GOV) trawl and sample tows are generally of 30min duration. Trawl samples are quantified to provide estimates of the numbers of each 1cm length class of each species caught at each sample location. Catch numbers were converted to densities by dividing by the area swept by the trawl (mean distance between the wings multiplied by distance the gear was towed). Where wing-spread distance was not known, this was estimated using a depth dependent relationship derived for the standard GOV.
Trends in the fifteen metrics applied to the SRGS data set suggest that the proportion of large fish on the Rockall plateau has increased in recent years, associated with declines in the abundance, biomass and productivity metrics. Mean fish size appears slightly down. The five species richness and species diversity metrics all suggest increasing trends, but the four life-history trait metrics all imply that the community has become increasingly composed of opportunist rather than climax community species. A similar story emerges with regard to longer-term change in the composition, structure and function of the Rockall Bank demersal fish community, except that recent changes in size composition have simply achieved parity with the earlier data. Given the scarcity data, particularly the gap between 1986 and the start of the systematic surveys in 2001, confidence in these assessments is low (Figure S10).