Simulating environmental effects on brown shrimp production in the northern Gulf of Mexico

Brown shrimp (Farfantepenaeus aztecus) support a commercially important fishery in the northern Gulf of Mexico, and juveniles use coastal estuaries as nurseries. Production of young shrimp from any given bay system, and hence commercial harvest of sub-adults and adults from the Gulf, is highly variable from year to year. We describe development of a spatially-explicit, individual-based model representing the cumulative effects of temperature, salinity, and access to emergent marsh vegetation on the growth and survival of young brown shrimp, and we use the model to simulate shrimp production from Galveston Bay, Texas, U.S.A. under environmental conditions representative of those observed from 1983 to 2012. Simulated mean annual (January through August) production ranged from 27.5 kg ha(-1) to 43.5 kg ha(-1) with an overall mean of 34.3 kg ha(-1) (0.70 kg ha(-1) SE). Sensitivity analyses included changing values of key model parameters by +/- 10% relative to baseline. Increasing growth rates 10% caused a 16% increase in production, whereas a 10% decrease resulted in an 18% decrease in production. A 10% increase in mortality probabilities resulted in a production decrease of 15% while a 10% decrease resulted in an 18% increase in production. We also changed values of environmental input data by 10%. Mean production estimates increased 11% in response to increasing tide heights (and thus, marsh habitat access) and decreased 19% with a decrease in tide height (and marsh access). The thirty year mean production was affected negatively by both the 10% increase and decrease in air temperature (-2% and -14%, respectively). Simulations in which bay water salinities were entirely low (0-10 PSU), intermediate (10-20 PSU), or high (>20 PSU) resulted in mean baseline production rates being reduced by 55, 7, and 0%, respectively. Uncertainty in model estimates of shrimp production were related to the magnitude and the timing of postlarval shrimp recruitment to the bay system. Simulations indicated that mean production decreased when recruitment occurred earlier in the year under all environmental conditions. Mean production varied with environmental conditions, however, when recruitment was delayed. The model reproduced biomass and size distribution patterns observed in field data. Although annual variability of modeled shrimp production did not correlate well (R-2 = 0.005) with fisheries independent trawl data from Galveston Bay, there was a significant correlation with similar trawl data collected in the northern Gulf of Mexico (R-2 = 0.40; p = 0.0005). Identifying and representing spatially variable factors such as predator distribution and abundance among bays, therefore, may be the key to understanding bay-specific contributions to the adult stock. Published by Elsevier B.V.