Forecasting Shrimp and Fish Catch in Chilika Lake over Time Series Analysis DOI: http://dx.doi.org/10.5772/intechopen.85458

forecasts of the output variables of a dynamic system [6]. Very well-known Box Jenkin's [7] Auto Regressive Integrated Moving Average (ARIMA) model is used to analyse and forecast the univariate time series data. SARIMA model is composed of ARIMA model including seasonal component of the time series data and very frequently used for time series modelling. Applications of ARIMA and SARIMA models were used to analyse and forecast fisheries data [12–14]; illustration on catch and effort data for the skipjack tuna fleet at Hawaii [8], Saila [9] modelled on monthly average catch per day of rock lobster, Sathianandan and Srinath [10] modelled of all India marine fish landings using ARIMA model, while NunoPrista et al. [11] described the application of SARIMA models for data-poor fisheries with a case study on sciaenid fishery of Portugal. Modelling and forecasting of marine fish production in Odisha, India, using the SARIMA Model [12].

When exogenous variable is included in the ARIMA model, it is known as ARIMAX model. Similarly, SARIMAX model is developed using well-known seasonal ARIMA model with external regressors associated with the time series data. SARIMAX model performs with better accuracy in fish catch prediction than the SARIMA model with minimum error variance [14]. However, the environmental factors' influence as a regressor in the time series ARIMA (SARIMA) model could be quantified by identifying the underlying patterns in periodic time series data using ARIMAX (SARIMAX) model. The influence of water quality parameters of Chilika lake on monthly catch of commercial fisheries (Beloniformes: order) and total fisheries of the lake was modelled using SARIMAX modelling approach [13, 14]. In this case study, SARIMA model for modelling and forecasting the 15 years quarterly time series shrimp catch data of Chilika lagoon of post restoration period since 2001 was developed as case study and further using this model to forecast shrimp catch landings up to 2018. Very few studies have been conducted based on time series model based forecasting of fish species in general and shrimp in particular of Chilika lagoon. The aim of this study is to develop the best fitted time series forecasting model for shrimp landing in Chilika lagoon and further model based catch forecast up to the period 2018. Since catch prediction in advance would be necessary for appropriate planning and designing of the national fishery development plan for sustainable exploitation of fisheries of the said water bodies, this shrimp forecasting study in the Chilika lagoon will be immensely useful in decision making for the policy makers and lagoon managers for sustainable fisheries production and management. In this chapter, a case study of shrimp prediction model using quarterly time series data of the Chilika lagoon was developed, and further prediction model of fisheries and commercial fish (Beloniformes: order) in the Chilika lagoon was also discussed.
