ASSESSMENT OF PATTERNS IN THE NAMIBIAN HAKE FISHERY BASED ON COMMERCIAL FISHERIES DATA, IN RELATION TO ENVIRONMENTAL FACTORS

ABSTRACT This thesis explores the relationships between commercial data and those obtained from research surveys in an attempt to broaden the database available for management, and to cover seasonal and inter-annual changes in density estimates of Cape hake (Merluccius capensis). It also attempts to explain hake variability of in terms of environmental indices based on satellite remote sensing, carefully chosen to reflect underlying oceanographic processes. The influence of various factors on catch rates (CPUE) is investigated using a general linear model. Results indicate that the full stepwise regression model is highly significant and accounts for about 51 % of variation in CPUE, with vessel size, year and area contributing 27%,9.5% and 6.9%, respectively. The spatial distribution of effort of part of the Namibian hake fleet is investigated by looking at: competitive abilities among vessels, spatial allocation of effort, competition among vessels and equalization of CPUE based on the Ideal Free Distribution (IFD). Results show that interference competition does not occur among vessels. Catch rates are equalized among areas. Since there is no interference competition among vessels, the catch ability coefficient (q) is not affected by this factor. Density estimates are compared from research surveys (which do not cover all seasons) and from commercial data with temporal and spatial overlap. Results of the stepwise linear fit to the data indicate that season and year are significant to the model fit. Density estimates from the whole year reveal the least inter-annual variability and those from the first quarter, the greatest. Thus commercial data do augment research data, indicating the extent of inter- and intra-annual variability in hake availability; they could be used for tuning models and identifying the risk and uncertainty in the production models used for management. iii The extent of the differences in hake density is related to the strength of the seasonality of the ecosystem through the temporal patterns in sea surface temperature (SST). The findings show that the maximum inter-annual variability in SST takes place during summer. The catchability of Namibian hake shows a strongly seasonal pattern correlated with seasonality in SST and its anomalies. Multinomial logistic regression analysis is used to calculate the probability of strong, average or weak recruitment of two-year old hake. The model includes environmental indices describing the extent of warm water intrusion from Angola during January to March in the year of spawning, as well as the up welling intensity from 17°S to 29°S during May to September in the year of spawning and the following year. The model accounts for 79% of the variance in hake recruitment and correctly predicts the category of hake recruitment in 4 years out of 5.