An analysis of sampling variability in replicated midwater trawls off Southern California

Abstract

Four series of oblique tows were taken with a midwater trawl in the San Pedro and Santa Catalina Basins off Southern California. Comparisons of biomass, numbers, and mean sizes of fish and decapod crustaceans between day and night tows show few significant differences, indicating that daytime net avoidance is not a problem with this sampling method. Day to night differences in numbers of individuals captured were greater in Catalina than in San Pedro Basin, as was variability from trawl to trawl in overall species numbers. Rank order of abundance appears to be a relatively predictable community characteristic for the common species offish and decapods captured. Correlation matrices showed little agreement in type or degree of correlation between basins, suggesting that positive correlations may not be based on real species affinities, but on other factors in the environment. Relative variability from trawl to trawl in numbers of individuals of the various species was generally higher than the relative variability in numbers of the total multispecies populations of fish and decapods, which may indicate some overall density regulating mechanism operating in the community. Estimates made of the number of samples required adequately to predict population means indicate that a mean number of ten trawls are required.