Main Article Content

Modelling and forecasting of catfish species yield from Mangochi artisan fisheries of Lake malawi in Malawi


M. Mulumpwa
W. Jere
A. Mtethiwa
T. Kakota
J. Kang’ombe

Abstract

Most of the wild fish stocks in Malawi either are fully or over exploited. This challenge underpins importance of forecasting using available data to support sustainable fisheries management. The study aimed at modelling and forecasting Catfish (Mlamba) species yield from artisan fishery on Lake Malawi in Mangochi District as they are becoming important food fish due to decline of more important fish species such as Oreochromis
(Chambo). The study was based on secondary data on fish catches between1976 and 2012, collected from Fisheries Research Unit of the Department of Fisheries in Malawi. The study considered Autoregressive Integrated Moving Average (ARIMA) processes to select an appropriate stochastic model for forecasting the species yield. Appropriate models were chosen based on ARIMA (p, d, q). Autocorrelation function (ACF), Partial
autocorrelation (PACF), Akaike Information Criteria (AIC), Box-Ljung statistics, correlogram of residual errors, distribution of residual errors, ME, RMSE, MAPE and MAE. Selected model was ARIMA (0, 0, 1) for forecasting artisan landings of Catfish from Lake Malawi in Mangochi District from 2013 to 2022. Based on the chosen model, forecast for artisan Catfish landings showed mean of 352 tonnes and mean of actual catches was 362 tonnes. However, catches in year 2022 are projected to be 360 tonnes, slightly below the actual catches mean but above 236 tonnes in 2010, assuming other factors remain constant. Confidence intervals of the forecasts included a zero and as such over exploitation of the species cannot be ruled out. Landings of the fishery will increase to 360 tonnes and remain stable through year 2022 necessitating fisheries management
consideration to improve the trend. Policy makers should secure sustainable exploitation of Catfish species, among artisan fishery in the study area by controlling all fishing effort that lands the species such as gillnets, beach seines, open water seines among others.


Journal Identifiers


eISSN: 1684-5374
print ISSN: 1684-5358