MODELING TRENDS AND PREDICTING MARINE FISH CATCH IN NIGERIA USING ARIMA TECHNIQUES (1960–2015)

A 56-year (1960 to 2015) catch data of Nigerian marine fisheries was analysed, and an ARIMA (3,1,3) model was employed to generate a short-term forecast (2016–2045). The Akaike information criterion (AIC), Schwarz information criterion (SIC), and Hannan-Quinn criterion (HQC) in conjunction with the log likelihood (Log L) and R2 statistics were employed to select the optimal model for the time series. Similarly, three benchmarks were considered: Root Mean Square Error (RMSE), Mean Absolute Error (MAE) and Theil Inequality Coefficient (TIC) to compare the in-sample and out–of –sample forecasts' performances of the estimated ARIMA (3,1,3) model, evaluating its forecasting ability and determining which forecast is superior for the model. The forecast results indicated that, everything remaining equal, approximately 27,021,461 metric tons of marine fish will be caught in Nigeria between 2016 and 2045 with an annual average of 900,715 metric tons. Temporal pattern of the forecasted data exhibited inherent fluctuations and an upward trend, similar to the original data. These fluctuations in catch trends are attributed to various factors, such as government policy, piracy, illegal activities of foreign trawlers, war conflicts, and the remote location of fishing villages. To obtain robust and reliable data from Nigerian marine fisheries, the negative impacts of these factors must be mitigated.