Size-selective fishing drives species composition in the Celtic Sea

Fishing alters community size structure by selectively removing larger individual fish and by changing the relative abundance of different-sized species. To assess the relative importance of individual-and species-level effects, two indices of fish community structure were compared, the relative abundance of large fish individuals (large fish indicator, LFI) and the relative abundance of large fish species (large species indicator, LSI). The two indices were strongly correlated for empirical data from the Celtic Sea and for data from simulated model communities, suggesting that much of the variability in the LFI is caused by shifts in the relative abundance of species (LSI). This correlation is explained by the observation that most of the biomass of a given species is spread over few length classes, a range spanning the factor 2 of individual length, such that most species contributed predominantly to either the small or the large component of the LFI. The results suggest that the effects of size-selective fishing in the Celtic Sea are mediated mainly through changes in community composition.

Assessing the sensitivity and specificity of fish community indicators to management action

We assessed ten trophodynamic indicators of ecosystem status for their sensitivity and specificity to fishing management using a size-resolved multispecies fish community model. The responses of indicators to fishing depended on effort and the size selectivity (sigmoid or Gaussian) of fishing mortality. The highest specificity against sigmoid (trawl-like) size selection was seen from inverse fishing pressure and the large fish indicator, but for Gaussian size selection, the large species indicator was most specific. Biomass, mean trophic level of the community and of the catch, and fishing in balance had the lowest specificity against both size selectivities. Length-based indicators weighted by biomass, rather than abundance, were more sensitive and specific to fishing pressure. Most indicators showed a greater response to sigmoid than Gaussian size selection. Indicators were generally more sensitive at low levels of effort because of nonlinear sensitivity in trophic cascades to fishing mortality. No single indicator emerged as superior in all respects, so given available data, multiple complementary indicators are recommended for community monitoring in the ecosystem approach to fisheries management.

Marine spatial planning from the perspective of a small seaside community in Ireland

Marine spatial planning (MSP) is advocated as a means of managing human uses of the sea in a manner that is consistent with the maintenance of the ecological goods and services of the marine environment. Support for the process is evident at international and national levels but the degree to which it is acceptable to local level stakeholders is not clear. An Daingean (formely Dingle) is a small sea-oriented town situated on the southwest coast of Ireland in which marine-based tourism and other relatively new uses of the sea are pursued along side traditional fishing activities. Stakeholders in An Daingean are found to be positively disposed to a local process of MSP that incorporates meaningful local involvement.

Modelling recovery of Celtic Sea demersal fish community size-structure

The Large Fish Indicator (LFI) is a size-based indicator of fish community state. The indicator describes the proportion by biomass of a fish community represented by fish larger than some size threshold. From an observed peak value of 0.49 in 1990, the Celtic Sea LFI declined until about 2000 and then fluctuated around 0.10 throughout the 2000s. This decline in the LFI reflected a period of diminishing ‘large’ fish biomass, probably related to high levels of size selective fishing. During the study period, fishing mortality was maintained at consistently high values. Average biomass of ‘small’ fish fluctuated across the whole time series, showing a weak positive trend in recent years. Inter-annual variation in the LFI was increasingly driven by fluctuation in small fish biomass as large fish biomass declined. Simulations using a size-based ecosystem model suggested that recovery in Celtic Sea fish community size-structure (LFI) could demand at least 20% reductions in fishing pressure and occur on decadal timescales.

Why the size structure of marine communities can require decades to recover from fishing

A dynamic food-web model of more than 1000 species was used to quantify the recovery trajectory of marine community size-structure under different hypothetical fishing regimes, using the Northeast Atlantic as an example. Size-structure was summarised by four indicators: the Large Fish Indicator (LFI), the Large Species Indicator (LSI), the biomass-weighted mean maximum length of fish species (EMBED Equation.3) and the biomass-weighted mean maturation length of fish species (EMBED Equation.3). Time-series of these indicators recorded recovery following release from fishing with various size-selectivities, intensities and durations. In model simulations, fishing-induced trophic cascades were observed to distort fish community size-structure, but these did not have a large influence on recovery level or duration as measured by the four indicators. However, simulations showed that local extinctions of large fish species increased in number with both fishing intensity and duration, and could strongly limit the recovery level. Recovery of fish community size-structure to near equilibrium frequently took multiple decades in simulations; these long transient periods suggest that management interventions for size-structure recovery may require much longer than previously thought. Our results demonstrate the need for community-level modelling to set realistic targets for management of community size-structure.

Footprints in the sand: a persistent spatial impression of fishing in a mobile groundfish assemblage

Fishing is well known to curtail the size distribution of fish populations. This paper reports the discovery of small-scale spatial patterns in length appearing in several exploited species of Celtic Sea demersal 'groundfish'. These patterns match well with spatial distributions of fishing activity, estimated from vessel monitoring records taken over a period of 6 years, suggesting that this 'mobile' fish community retains a persistent impression of local-scale fishing pressure. An individual random-walk model of fish movement best matched these exploitation 'footprints' with individual movement rates set to <35 km per year. We propose that Celtic Sea groundfish may have surprisingly low movement rates for much of the year, such that fishing impact is spatially heterogeneous and related to local fishing intensity.

Estimating biomass, fishing mortality, and "total allowable discards" for surveyed non-target fish

Demersal fisheries targeting a few high-value species often catch and discard other "non-target" species. It is difficult to quantify the impact of this incidental mortality when population biomass of a non-target species is unknown. We calculate biomass for 14 demersal fish species in ICES Area VIIg (Celtic Sea) by applying species-and length-based catchability corrections to catch records from the Irish Groundfish Survey (IGFS). We then combine these biomass estimates with records of commercial discards (and landings for marketable non-target species) to calculate annual harvesting rates (HR) for each study species. Uncertainty is incorporated into estimates of both biomass andHR. Our survey-based HR estimates for cod and whiting compared well with HR-converted fishing mortality (F) estimates from analytical assessments for these two stocks. Of the non-target species tested, red gurnard (Chelidonichthys cuculus) recorded some annual HRs greater than those for cod or whiting; challenging "Pope's postulate" that F on non-target stocks in an assemblage will not exceed that on target stocks. We relate HR for each species to two corresponding maximum sustainable yield (MSY) reference levels; six non-target species (including three ray species) show annual HRs >= HRMSY. This result suggests that it may not be possible to conserve vulnerable non-target species when F is coupled to that of target species. Based on biomass, HR, and HRMSY, we estimate "total allowable catch" for each non-target species.