Statistical relationships between palagic fish catches and long-term series of environmental conditions in the southern Benguela region

Bibliography: leaves 182-198.Three different techniques are used to established statistical relationships between annual pelagic fish species catches viz. pilchard, horse mackerel and chub mackerel, and monthly environmental indices. The three techniques are Spearman's Rank correlation, multiple regression and cross-correlations of the Box-Jenkins time series approach. The first method yielded interesting results but only in terms of the nature of the relationship not in terms of the response effect. In the same way the multiple regression analysis showed inherent problems in the interpretation of the results due to serial correlation in all fish catches and environmental data series. The Time-Series analysis yielded more coherent results, presumably due to the fact that the data series were pre- whitened to remove serial correlation. Monthly means of sea surface temperature, north-south wind component, west- east wind component and sea level were used as environmental indices. Over the period 1950 to 1985, annual pilchard catches are found to be negatively correlated to sea surface temperature in the Namaqualand and Agulhas Bank areas, but are positively related to sea surface temperature in the south western Cape area. With regard to wind components, northerly and westerly winds in the Namaqualand and south western Cape areas are found to be significantly correlated to annual pilchard catch, whereas in the Agulhas Bank southeasterly - and southwesterly wind are found to be predominant. Warmer waters improved horse mackerel catches in all areas, immediately in the Agulhas Bank area but affecting catch the following year in the other two areas. In the Namaqualand area predominantly northwesterly winds are found to favour horse mackerel catch immediately and this response is probably due to availability. The same wind orientation is found in the south western Cape area but with one-year-lag. In the Agulhas Bank area the favourable wind orientation is southerly with a zero lag and westerly with one-year-lag. Cooler waters influence chub mackerel catches favourably in all areas and in all seasons. In the Namaqualand area the effect of the sea surface temperature is immediate in all but the winter season. Northwesterly winds are most important in this area having an immediate effect from winter to summer. Cooler temperatures in the south western Cape area the previous year improve the annual chub mackerel catch. Northerly and easterly winds favour catch immediately during summer while northwesterly winds are most important during spring. In the south coast area, southwesterly winds improve catches immediately during spring and summer but south-westerly winds during autumn are related to improved catches the following year. These results are not easy to interpret because of problems in the quality of the catch data. These problems may be identified as constraints due to the effect of the different age-classes to the catch and the variability in the effort exerted by the fishery onto the resource. Furthermore, age-length distributions have changed recently for all three species, so that it is difficult to establish relationships which link up to specific effects of the environment on specific age- classes. It is assumed, however, that a zero-lag in the relationships between fish catches and environmental variables represents availability rather than environmental effects on spawning, recruitment or stock abundance