Modeling of Lake Malombe Annual Fish Landings and Catch per Unit Effort (CPUE)

Forecasting, using time series data, has become the most relevant and effective tool for fisheries stock assessment. Autoregressive integrated moving average (ARIMA) modeling been commonly used to predict general trend fish landings with increased reliability precision. In this paper, ARIMA models were applied Lake Malombe annual catch per unit effort (CPUE). The CPUE trends first observed both non-stationary. first-order differencing was transform non-stationary data into stationary. Autocorrelation functions (AC), partial autocorrelation function (PAC), Akaike information criterion (AIC), Bayesian (BIC), square root of mean error (RMSE), absolute (MAE), percentage standard prediction (SEP), relative variance (ARV), Gaussian maximum likelihood estimation (GMLE) algorithm, efficiency coefficient (E2), determination (R2), persistent index (PI) estimated, which led identification construction models, suitable in explaining forecasting. According measures forecasting accuracy, best (0,1,1) (0,1,0). These had lowest values AIC, BIC, RMSE, MAE, SEP, ARV. further displayed highest GMLE, PI, R2, E2. “auto. arima ()” command R version 3.6.3 (0,1,0) as best. selected satisfactorily forecasted 2725.243 metric tons 0.097 kg/h by 2024.