Forecasting the Anchovy Kilka Fishery in the Caspian Sea Using a Time Series Approach

Özet

Forecasting the status of fish landings is a major tool for fisheries managers and policy makers in order to decide onsustainable management issues. In this paper, yearly landings kilka data from 1990 to 2014 were analyzed using time seriesmodel. Autoregressive (AR), Moving Average (MA) and Autoregressive Integrated Moving Average (ARIMA) wereconsidered through the analysis to select appropriate model for forecasting. Based on Autocorrelation function (ACF), PartialAutocorrelation function (PACF) and degree of differentiation, ARIMA (0, 2, 3) model with the lowest normal Bayesianinformation criterion (BIC) and Akaike information criterion (AIC) value was selected. Results showed that Kilka catch willincrease gradually in the coming years. However, the hypothesis that the commercial catches have reached their zero pointcould not be rejected. In conclusion, results of this study revealed despite government reduced fishing mortality in the recentyears, potential risk of population collapse is still remained.