Stock model and multivariate analysis for prediction of semi-intensive production of shrimp Litopenaeus vannamei as a function of water quality and management variables: A stochastic approach

• We used a stock model, multivariate analysis and a stochastic approach to predict shrimp biomass.
• Our analysis allowed studying the relationships between production parameters, water quality, and management variables.
• We conclude that the approach is useful for studying the variability of semi-intensive shrimp production.

We use a stock model, multivariate analysis, and a stochastic approach to predict shrimp production under commercial semi-intensive conditions as a function of water quality and alternative management schemes. Larger final weight of shrimp was obtained when temperature and duration of cultivation increased. Increases in the mortality of shrimp were associated with lower dissolved oxygen levels, shorter durations of cultivation, and higher stocking densities. There was a direct relationship between temperature and stoking density, while dissolved oxygen was inversely related with stocking density and duration of cultivation. Stocking density was inversely correlated with pond size and directly correlated with duration of cultivation. The lowest yields were predicted, using the lowest stocking densities and shortest duration of cultivation; the highest yields were predicted using the highest stocking densities and longest duration of cultivation. Yields increased from 938 to 2326 kg ha−1 (spring production cycle), and from 982 to 1907 kg ha−1 (summer production cycle). Improved management resulted in increased shrimp production and diminished variability. Sensitivity analysis indicates that final weight of shrimp and stocking density were the major factors affecting variability of shrimp yields. We conclude that stock models, multivariate analysis, and a stochastic approach constitute an effective method for studying the relationships between production parameters, water quality, and management variables, and, for analyzing variability of semi-intensive shrimp production.