Pelagic Fish Distribution and Dynamics in Coastal Areas in the Baltic Sea Proper

Pelagic fish distribution and diel behaviour patterns were studied in coastal areas in the north-western Baltic Sea Proper to understand more about how fish distribution and behaviour might affect planning and analyses of results of hydroacoustic surveys (Papers I and II). The vertical distribution of fish at night from spring to autumn showed seasonal and annual trends that could be explained by predictable and consistent seasonal changes, e.g., in temperature and stratification. Horizontal fish distributions did not show any trends probably owing to a lack of such seasonal characteristics. The observed vertical fish distribution over the diel cycle showed that hydroacoustic surveys at night were to be preferred over daytime surveys. At night, fish did not school and were generally less aggregated resulting in less variable hydroacoustic backscattering values and a higher percentage of single echo detections. By starting the surveys one hour after sunset and stopping one hour before sunrise, confusion between day- and nighttime behaviour in fish could be avoided. At night, fish occupied mid-water layers to a higher extent than surface and bottom layers, which was beneficial for the quality of the hydroacoustic data, particularly with respect to the hydroacoustic blind and dead zones (i.e. surface and bottom, respectively). To quantify seasonal changes in pelagic fish abundance, densities and size distributions, nighttime hydroacoustic surveys were done every second week from spring through autumn in 2000 and 2001 (Paper III). There was a drastic increase in fish abundance and densities that started in early July and peaked in mid-August in both years. Analyses of the hydroacoustic data in relation to gillnet and trawl catches showed that the increase was caused mainly by young-of-the-year (YOY) herring. This age class is commonly not well represented in catches using traditional sampling methods like gillnets and trawling. Consequently, hydroacoustic data that have high precision and accuracy may improve quantitative estimates and our understanding of the biology in coastal nursery areas. Baltic herring spawn in coastal areas and the density of metamorphosed YOY individuals may provide an early estimate of year-class strength. By analysing the relationship between parameters known to affect recruitment success and year-class strength in age 2 herring (YCS) a model that predicted herring recruitment was developed (Paper IV). The model explained 93 % of the variation in the number of age 2 herring over the period 1985-2000 and included the parameters YOY densities, climate (North Atlantic Oscillation index) and spawning stock biomass (SSB). Thus YCS could be predicted two years earlier than today and three years before entering the fishery. Up to the present, three new years (2001-2003) have become available for testing the model. For one of these years the predicted YCS was notably different from the assessed YCS. The reason for this is not fully understood, but for all three years SSB was outside the range used in the original model. Including the three new years into the data series resulted in a poorer explanation of the observed recruitment variation (55 %). A comparison of the standardized regression coefficients of both models showed increased significance for the parameter YOY (from 0.47 to 0.61)