Hybrid experimental designs for joint estimation of absolute and relative retention curves for trawl gears

Abstract

Abstract Methods for jointly estimating both the size-specific relative retention and absolute retention (retention probability) of two experimental trawl gears are considered. One approach would be to undertake comparative fishing of both gears, from which their comparative catchability can be estimated directly. If comparative catchability can be corrected for differing sampling effort and fishing efficiency of the two gears, then relative retention can be estimated. Furthermore, if a suitable functional form of the absolute retention curve is assumed, then estimates of absolute retention can potentially be obtained. However, these estimated retention curves are seen to have extremely large statistical variability and will not be useful in practice. On the other hand, if the gears are each fished with a non-selective control, then absolute retention can be directly estimated, from which estimates of relative retention can immediately be obtained. However, these estimates of relative retention have greater variability than would be obtained by comparative fishing with the same level of sampling effort. It is shown that a hybrid selectivity design that includes both comparative fishing of the experimental trawls and fishing with a non-selective control provides a more balanced approach. An extensive simulation study is conducted using scenarios based on recently published work on selectivity experiments targeting cod (Gadus Morhua) and hoki/blue grenadier (Macruronus novaezelandiae).