Predicting herring recruitment from young-of-the-year densities, spawning stock biomass, and climate

Because fish are key organisms in most aquatic ecosystems, we seek to understand what determines their highly variable reproductive success. From our work, it appears that yearclass strength of Baltic Sea herring (Clupea harengus L.) can be predicted from young-of-the-year densities in a small coastal area (hydroacoustic data), a climate index (the North Atlantic Oscillation), and the spawning stock biomass. These three factors explained 93% of the variation in the number of age 2 herring during 1985–2000. By predicting year-class strength 3 yr before the fish enter the fishery, we provide managers with the opportunity to adjust fishing pressure per upcoming year classes and manage the fishery by multiannual catch quotas. Overfishing is recognized as one of the most serious human impacts on marine ecosystems (e.g., Pauly et al. 1998; Jackson et al. 2001). This failure of management derives, at least in part, from uncertain stock size assessments and a poor understanding of variability in fish recruitment. Furthermore, yearclass strength typically is only assessed on entry into a fishery, clearly too late to manage effort appropriately. The difficulty in understanding and predicting fish recruitment reflects a classical problem in fish biology and fisheries research (e.g., Ricker 1954; Beverton and Holt 1957; Cushing 1996). Fish populations typically display large natural variations in abundance. Periods with unusually large populations can persist for decades (Alheit and Hagen 1997; Toresen and Östvedt 2001). For naturally short-lived marine species or intensively fished populations the time between high and low population size can be shorter and even based on a single or a few consecutive strong or weak year classes (e.g., Cushing 1996). These fluctuations can be self-perpetuating with large stocks of adult fish producing numerous offspring and small stocks producing few. However, for most marine pelagic fish species, the stock–recruitment relationship is generally weak, as high fecundity is combined with high and variable egg and larval mortality. Early life stages suffer highest mortality, subject to variability in water temperature, feeding conditions, and wind stress (Houde 1994; Cushing 1996), suggesting that year-class strength is set generally by the time young fish have passed larval and early metamorphosed stages (Bradford 1992; Cushing 1996). Accordingly, in most marine pelagic species, year-class strength can be estimated after young fish have passed these stages. Herring (Clupea harengus L.) and sprat (Sprattus sprattus L.) are the dominant pelagic species in the Baltic Sea and are subjected to intensive fishing. Although the Baltic Sea is an open system and adult herring migrate extensively (e.g., Parmanne et al. 1994), a growth gradient from south to north reveals their sedentary nature within regions (Arrhenius and Hansson 1993). Herring is fished mainly offshore but spawn in shallow coastal areas, predominantly in spring and early summer, all along the Swedish east coast. In late summer, the archipelago areas support dense populations of youngof-year herring (Rudstam et al. 1992; Hansson 1993). To predict recruitment of Baltic Sea herring shortly after spawning, we compared data on year-class strength (YCS) of age 2 herring with young-of-the-year herring densities (YOY) and herring spawning stock biomass (SSB, fish age $ 3 yr). Because mild winters can positively affect yearclass strength (e.g., Brielmann 1989; Rajasilta et al. 1996; Laine et al. 1998), we also included climate data (the North Atlantic Oscillation index, NAO) in our analysis. Material and methods—Our data were derived from subdivision 27 (SD27, 26,700 km2), a standard assessment area