Effects of walleye predation on variation in the stock-recruitment relationship of Lake Erie yellow perch

Abstract

Abstract Stock-recruitment relationships (SRRs) may vary over time due to temporal variation in ecological conditions, reducing confidence in projections from stock-recruitment models. We examined whether the time-varying SRRs detected for yellow perch (Perca flavescens) in the western basin of Lake Erie between 1977 and 2013 could be attributed to variation in yearling walleye (Sander vitreus) predation, indexed by variation in density. Annual variation in yearling walleye density was strongly associated with yellow perch recruitment dynamics, and positively correlated with temporal variation in density-dependence of yellow perch SSRs. However, non-stationary SRRs persisted after accounting for effects of yearling walleye density, and the extent of temporal variation in SRRs actually increased. In simulations, we showed that time-varying SRRs may result more from variation in low-frequency ecological factors on the order of decades, than from variation in high-frequency ecological factors on the order of years (e.g., yearling walleye density) and thus may not be distinguishable from noise. Hence, of these two types of factors, the systematic identification, characterization and incorporation of those of low-frequency factors into stock-recruitment models (e.g., exotic mussel invasions and eutrophication, in the case of Lake Erie) may offer greater promise to improve the reliability of long-term forecasts for sustainable harvests in this and other fisheries in dynamic ecosystems.