A generalized model for correcting bias due to permeability using emergent fry traps in mesocosm experiments

Abstract

Relative abundance, or CPUE, is a commonly used metric for fisheries conservation and management. Interpreting CPUE data requires understanding the probability that a fish will be caught (catchability) by a particular type of gear and factors that affect catchability. Escape is often not considered in catchability studies, and there is little to no mention of this phenomenon in fisheries textbooks. However, most passive sampling gear, particularly traps, is ‘permeable’, so the number of organisms actually collected will fluctuate based on probability of escapement. Because high or variable escape rates have the potential to bias measures of abundance, the use of passive traps requires recognition and assessment of permeability. To better understand these dynamics, and particularly how they apply to salmonid emergent fry traps, we conducted a series of mesocosm experiments with early-life stage lake trout (Salvelinus namaycush). We assessed entry and exit rates by fitting daily capture data to a hierarchical state-space model. We also evaluated a trap modification and the effect of thiamine deficiency on capture rates; thiamine-deficient fish can be lethargic and unlikely to enter traps, so deficiency in the wild may be underestimated. Results revealed distinct temporal trends, with low initial entry rates that rose around the third week after hatch and remained constant thereafter. Thiamine deficiency decreased entry rates but appeared to have no effect on exit rates. Conversely, a minor trap design change had an unexpectedly large effect on exit rates. We developed a generalizable model for estimating abundance using passive traps based on soak time and entry/exit rates, as well as specific rate estimates for lake trout in emergent fry traps. Such corrections should be performed whenever feasible, as abundance estimates may otherwise be biased by exit-related equilibrium.