Catch estimation with restricted randomization in the effort survey

One common method for estimating: total catch is to multiply an estimate fo r CPUE, the catch per unit effort, by an estimate of total effort obtained from an independent second survey. In general, estimating total effort requ ires that sample times are chosen at random over the full fishing period; h owever, in practice, this may nut always be possible and the usual estimato r may be severely biased. Such a restriction in randomization is likely whe n aircraft are used to make instantaneous counts of fishing activity. This article proposes alternate estimators for use with both access and roving d esigns in conjunction with effort surveys for which sample times are not ra ndom. Ratio type estimators based on activity counts are developed under va rious scenarios and their performance examined under simulation. In additio n, optimizing strategies for use with multiple activity counts are explored . Finally. data from an in-river gill net fishery on the Fraser River is us ed to illustrate these results.