Survey augmentation using commercial vessels in the Mid-Atlantic Bight: Sampling density and relative catchability

Abstract

Summary A series of side-by-side tows was conducted between a survey vessel and a commercial vessel in two seasons, spring and fall (autumn), to examine the use of commercial vessels to increase sampling density in trawl-based stock surveys. Both vessels caught more fish offshore in the spring. The commercial vessel caught more fish than the survey vessel in both seasons. Catches of nearly all species were contagiously distributed in the spring. Most were contagiously distributed in the fall; however, somewhat more species were characterized by random or even distributions. The variance-to-mean ratio was consistently higher for most species for commercial vessel catches, regardless of season. As both vessels sampled in the same region at the same time, the increased predilection for the survey vessel to assess the distribution pattern as less patchy than the commercial vessel must accrue from some difference in sampling dynamics rather than variation in species distribution. A simulated decrease in sampling effort from 59 to 30 or 15 hauls increased the variance-to-mean ratio. Reduced sampling effort increased the tendency for occasional large catches to vary the estimate of domain biomass. The sampling program included an onshore–offshore gradient in station density. Domain biomass was considerably underestimated with reduced station density for six species characterized by large catches offshore in that portion of the survey domain characterized by low station density. In this study, a factor of two variation in domain biomass became more likely in 40% of species when sampling effort was reduced to 15 hauls from 59. A factor of two in biomass may distinguish a sustainable fishery from one in which a species is overfished. As survey sampling effort in this area was 18 hauls, increasing sample number by inclusion of commercial vessel tows would be advantageous. A regression between paired tows failed to adequately predict catches of one vessel from catches of the other. Standardization of vessel catches by the ratio-of-mean catches provided a more realistic comparison because large catches accounted for a significant fraction of domain biomass; however, a single conversion coefficient between boats could not be used for both sampling periods. The underlying impediment in developing a general conversion factor between the two vessels seems to be rooted in the differential in variance-to-mean ratios of the catches; this differential exists despite sampling of the same distribution of fish.