Dynamics and exploitation of unstable percid populations

The gill net fishery on perch (Perca fluviatilis) and especially the highly valued pikeperch (Stizostedion lucioperca) in Lake IJssel is characterised by large variations in the yield. These variations are caused by variations in yearclass strength in combination with the high exploitation rate. In the present study a start has been made to explain the causes for variations in year-class strength. In addition, possible management measures are evaluated for their effectiveness in optimising yield. The possibilities for yield optimisation depend on the dynamics of the stocks. Dynamics are determined by recruitment, growth and mortality. The development in the stocks and the fishery has been documented in the past to some extent. Surveys have been conducted since 1966, mainly using a bottom trawl. However, the information obtained through these surveys on the fish stocks had hardly been analysed until now. The present study comprises analyses of these historical trawl surveys on the one hand and of detailed field observations conducted from 1987-1989 on the other. Results were used as input for a size- and age-structured simulation model. This model is developed to evaluate the effectiveness of various management measures to enlarge and stabilise the yield.Bottom trawl surveys are carried out every autumn to estimate the relative abundance of six major fish species, with particular reference to pikeperch and perch. It is generally supposed among fishermen that the catchability of these species is influenced by light intensity at the bottom, which in this study was characterised by water depth and water clarity as measured by a Secchi-disk. In autumn, water clarity may change greatly from day to day because of windmediated re-suspension of sediments. Catchability of ruffe (Gymnocephalus cernua) and age 0 pikeperch showed a significant inverse relationship with light intensity at the bottom and, therefore, a correction should be made when catch data for these species are used to estimate population size or year-class strength. Results were not consistent for perch, while for smelt (Osmerus eperlanus), roach (Rutilus rutilus) and bream (Abramis brama) the influence of light intensity upon catchability was not significant. Corrected and uncorrected estimates of the abundance of age 0 pikeperch, based on trawl samples, were compared to demonstrate the effect of' water clarity. The correction was marginal for the historical trawl surveys, because sampling was mainly done when conditions were good. Increased water clarity by net avoidance behaviour may lead to zero catches and consequently diminish the effectiveness of sampling programmes. Based on the present analysis, surveys can be optimised by interrupting them when zero catches are likely to occur. For the assessment of age 0 pikeperch abundance, surveys should be interrupted when the ratio between water depth and water clarity is smaller than eight.Recruitment, indexed as the age 0 abundance in bottom trawls, of perch varied 400-fold and of pikeperch 70-fold over the period 1966-89. Year-class strength variation of perch in Lake IJssel was higher than observed in Perca spp. in other waters according to the literature, while for pikeperch it fell within the observed range of Stizostedion spp.. Hypotheses were formulated on the possible relationship between recruitment and stock biomass, prey fish availability, water temperature and wind velocity. For perch 36 % to 71 % of the variation in recruitment could be explained by stock biomass of piscivorous fish, water temperature, and year of sampling; for pikeperch 38% to 86% could be explained by water temperature, wind velocity and year of sampling. For both species strong year-classes were favoured by low temperatures in April and by high summer temperatures; a late increase in water temperature to above 8°C and a high warming rate thereafter, probably resulting in postponed spawning and rapid development, favoured recruitment of perch. Stockrecruitment relationships could not be demonstrated. The impact of the commercial fishery, which possibly accounted for the significant correlation between year of sampling and year-class strength, and of fish-eating birds might have clouded interactions among fish species. Based on this analysis, it is concluded that recruitment will be very difficult if not impossible to manage.The positive effect of warm summers on recruitment of age 0 pikeperch is most probably caused by its faster growth at higher temperatures than that of its potential prey, age 0 smelt. Mean length of age 0 pikeperch in November correlated strongly with mean summer temperature. Depending on the environmental conditions (especially water temperature and availability of food) the unimodal length frequency distribution of age 0 pikeperch developed into a positively skewed, bimodal or negatively skewed distribution towards the end of the summer. Strong year-classes were characterised by larger mean lengths, and a negatively skewed frequency distribution, while weak year-classes were smaller and positively skewed. Body energy content increased with fish length and differences in body constituents were more pronounced later in the season. The condition of non- piscivorous age 0 pikeperch was low and decreased over time, while that of piscivorous pikeperch increased over time. The onset of piscivory, favoured by high temperatures during summer, has a direct positive effect on the growth and survival of age 0 pikeperch.Also for perch the onset of piscivory has a positive effect on growth. Growth and size distributions of perch were studied in relation to prey choice and water temperature. Growth of age 0 perch is density-dependent and positively influenced by high summer temperatures. Between 10 and 20 cm perch switched from a diet of zooplankton and macrofauna (Neomysis integer and Gammarus spp.) towards fish. Age 0 European smelt was the main fish prey. The onset of piscivory occurred at a larger length when mean length of daphnids, was large or when age 0 smelt densities were low. Growth was faster for piscivores than for non-piscivores. Consequently, size distributions Of cohorts of which only the larger ones were piscivorous became positively skewed, and initial growth differences between males and females were enhanced. Besides age 0 smelt, perch also fed on age 1 smelt, age 0 perch and age 0 ruffe, but Ivlev indices revealed a preference for age 0 smelt. Cannibalism by older age-groups was rare when recruitment was low, but became common when age 0 densities of perch were high. It is concluded that smelt plays a key role in the population dynamics of both pikeperch and perch, because it facilitates the onset of piscivory thereby enhancing growth and in case of pikeperch also survival. Moreover it buffers cannibalism.A size- and age-structured simulation model is presented for evaluation of management measures for multispecies gill net fisheries. Data on the dynamics and exploitation of pikeperch and perch were used as parameters for the model. The model takes size distributions within age-groups into account. In the model, growth is determined by length and temperature and dispersion in size distributions is controlled by the 'fractional boxcar train' method. The model is applied to evaluate the integrated short-term and long-term effects of management measures such as mesh size regulations, fishing effort limitations and combinations of both on the biomass, the size and age structure of the stock and of the yield. Changes in yield during the transitional period after a management measure has been implemented, can be quantified.Finally a simple method is used, which relates the variation in the yield to the variation in recruitment and the exploitation pattern per species. This model neglects stock-recruitment relationships and other density-dependent processes, i.e. the model takes variations in year-class strength for granted and evaluates its consequences for management. It is shown that in case of the Lake IJssel gill net fishery a larger number of age-groups in the catch reduces the inter-annual variation in the yield caused by variable recruitment only to a small extent. Based on their age 0 abundance during the autumn trawl surveys the yield of that year- class can be predicted. For perch the upper and lower 95 % confidence limits around this prediction differ by a factor 1.6 and for pikeperch by a factor 2.7. As a consequence of the large variations in year-class strength and the uncertainties in yield prediction, it will be almost impossible to demonstrate the effect of altered management, if the increase in yield merely comprises a higher yield per recruit. Perhaps the largest gain in yield for the gill net fishery is achieved by enlarging the spawning stock biomass of pikeperch especially. More extreme measurements like e.g. temporarily closing the fishery at the southern part of Lake IJssel might provide the insights for this.