Environmental and biological factors associated with the stage-specific abundance of brown shrimp (Penaeus aztecus) in Louisiana: applying a new combination of statistical techniques to long-term monitoring data

Abstract

Several short-term and small-scale correlative studies have associated brown shrimp (Penaeus aztecus) recruitment with high densities of sub-adults, high salinity, warm temperature, low river flow, and low precipitation. In this paper, we address criticisms of traditional correlative studies by using a spatially and temporally extensive dataset, by comparing stepwise multiple regression (SMR) to Bayesian model averaging (BMA), and by investigating nonlinear relationships with generalized additive models (GAMs). We use this combination of statistical methods to examine relationships between annual, stage-specific abundance estimates and environmental factors. BMA and SMR resulted in models with similar explanatory power, but BMA suggested fewer linear predictors. GAMs did not suggest nonlinear relationships among stage-specific abundance estimates. Postlarval abundance was not well described by any model. Juvenile abundance was partially described by environmental variables such as temperature, water clarity, and water level. Adult abundance was well described by early-juvenile abundance, salinity, and temperature. These results suggest that juvenile abundance may be the critical component in determining year-class strength of brown shrimp. Identifying mechanisms that regulate juvenile production within the estuary will be a critical step in effectively managing Louisiana's brown shrimp resource.