The effects of variability in catch effort on the precision of statistical population reconstruction

Abstract

Statistical population reconstruction (SPR) models have emerged as a robust and versatile framework for estimating the demographic dynamics of harvested wildlife populations using commonly collected age‐at‐harvest and catch‐effort data. Although numerous studies have suggested that higher interannual variability in catch effort may improve the accuracy and precision of reconstructed estimates, particularly in the absence of auxiliary data on annual abundance or survival, the extent and magnitude of these effects has not been explored. We examined the influence of catch‐effort variability, as measured by the ratio between years of highest and lowest effort, on the relative absolute deviation of reconstructed estimates of population abundance, as well as on the actual percent coverage and width of the corresponding confidence intervals. We used a Monte Carlo simulation to generate catch‐effort data with different levels of variability for populations experiencing a wide range of demographic and harvest conditions. For similar amounts of age‐at‐harvest data, using catch‐effort data with higher interannual variability resulted in reconstructed estimates of annual abundance that had significantly lower deviations from reality, better coverage, and narrower confidence intervals (as measured by the margin of error). These improvements were consistent and linear at low to medium levels of catch‐effort variability, but leveled off and became substantially less pronounced at higher levels. We found that the inclusion of auxiliary data largely mediated this relationship, although higher catch‐effort variability still resulted in more accurate and precise estimates of annual abundance even when these data were included. Our research highlights the need to include a thorough investigation of the available catch‐effort data alongside the established practices of assessing the number of years of available data, the average number of animals harvested each year, and the availability of auxiliary data from radio‐telemetry studies or other sources.