Comparison of horse mackerel length frequencies obtained from research vessels and commercial midwater trawlers: implications for biomass estimation

The validity of abundance estimates from hydroacoustic surveys relies, inter alia, on the ability of the fishing gear on the research vessel to sample non-selectively. This study compares the length frequencies of Cape horse mackerel Trachurus trachurus capensis taken in Namibian waters by the R.V. Welwitchia and commercial midwater trawlers. Results indicate that the length distributions of catches taken by the Welwitchia were significantly different from those from commercial trawlers, with a greater proportion of fish >25 cm being sampled by the latter. Also, the biomass estimated per length-class from hydroacoustic surveys over the period 1994–2000 was compared with those from annual landings of the commercial fishery. The length distribution of horse mackerel in purseseine catches compares favourably with those from the research vessel. However, comparisons with the midwater trawl catches indicated that the length frequencies obtained from research vessels during the years 1994–1997 underestimated the number of large fish in the population and biased the biomass in those years. From 1998 to 2000 the bias was negligible. The magnitude of the error varied between years, but it remained fairly low. To compensate for this bias, length distribution data from midwater trawlers should be integrated into the acoustic biomass calculation procedure. Keywords: acoustic estimates, avoidance, horse mackerel, length frequencies, trawl catchesAfrican Journal of Marine Science 2001, 23: 265–27