A computer-based statistical model of the food and feeding patterns of the Southern brown shrimp Farfantepenaeus subtilis under culture conditions

Abstract

In less intensive shrimp culture systems (b35 shrimp/m2 ), an intrinsic relationship exists among natural food availability, shrimp feeding and overall system management. The objective of the present study was to develop and simulate onedimensional statistical models of the food and feeding patterns of Farfantepenaeus subtilis by using the dynamic modelling language StellaR II. Parameters employed in simulations were estimated based on relationships and data derived from the literature. Results indicated that shrimp hourly food intake reacted directly to feeding frequency and shrimp body weight (BW), but it was mostly affected by changes in feed ration. As feeding frequency reduced, a lower number of feeding sequences occurred over the simulation period, while at higher rations the total amount of feed consumed increased. Food intake levels progressively increased with shrimp BW, but declined after reaching a peak at 11-g shrimp. Maximisation of shrimp feed consumption was achieved when total ration administered was equivalent to ingestion rate per hour, and feeding frequency was set at four times/day. At higher shrimp stocking densities, total shrimp population feed consumption increased, but it also changed according to time of the day and feeding method used. More feed was consumed in the afternoon, particularly when the broadcast method was set during simulations. Polychaete population abundance was highly affected by both initial shrimp and polychaete density. At an increased shrimp density or under a low initial polychaete abundance, polychaete population was either depleted or reached critical levels over the simulation period. Feed supply played a role in sustaining polychaete growth, mainly at higher shrimp stocking densities, but it was not required at an increased initial polychaete abundance. Shrimp stocking density produced greater and inverse effects in the model outputs when compared to the initial polychaete density. Overall, dynamic modelling was able to integrate previously collected data and produce expected forecasts of F. subtilis food intake and its predatory effects on the abundance of pond polychaete populations.