Diel variation in bottom trawl catch rates of hake (Merluccius capensis and M. paradoxus) was studied by applying a stochastic model to logbook records of the Namibian trawler fleet and data from the standard hake assessment surveys. The commercial CPUE standardized by the vessel-day average were about 3.6 times higher around noon compared to CPUE during the night, and the diel variations was higher in areas dominated by M. capensis than in those mainly inhabited by M. paradoxus. The day-night difference in commercial CPUE decreased with depth and increased with density of hake and varied with latitude, whilst no significant seasonal or annual patterns were observed. Further, the time of transition from night to day level was correlated with time of sunrise which changes with season. Trawling in shallow waters at night is avoided during the Namibian hake surveys, and the systematic survey design complicated the diel variation modelling of the survey data. However, by utilizing the parameters which describe the diel variation transition estimated from the commercial CPUE data and compensating for the survey design effect, the survey hake catch rates at noon were estimated to be 27% and 86% higher than at night for M. capensis and M. paradoxus, respectively. More important, we also demonstrate that the transition period from night-level catches to maximum day-level is long, which implies that the catch efficiency in the morning and the afternoon are in general lower than at noon, and consequently may bias the survey estimates.

Keywords: Hake, Diel variation, Namibia, Trawl surveys, Logbook data.