An opportunity to trade time for space without compromising data quality in creel surveys

Abstract Objective Fisheries-dependent data (e.g., angler effort, catch, and harvest rates) are important for understanding angler behavior and can be useful for managing some aspects of recreational fisheries. However, the collection of fisheries-dependent data through traditional point-intercept angler creel surveys is expensive and labor-intensive. We evaluated model-based approaches to predict fisheries-dependent metrics, which may be more cost-effective for fisheries management agencies that are trying to balance the need for creel survey information and fisheries-independent population monitoring. Methods We used a Bayesian modeling approach to analyze 33 years of standardized point-intercept creel information from the recreational Walleye Sander vitreus fishery in northern Wisconsin to test whether reduced survey designs could still accurately predict catch, harvest, and exploitation rates. We strategically reduced the creel survey data set in five different ways (i.e., seasonal reductions, removal of survey days) to represent alternative survey designs with reduced creel clerk effort that may be more efficient. Results The analysis of 33 years of creel survey data pooled together generally found no difference between the estimates of fishery-dependent metrics from the current creel design and the five alternative creel designs using Bayesian P-values and root mean square error. An annual analysis of the same metrics also found no difference between the current creel design and the five reduced creel designs for most years (mean of 91.2% across metrics and scenarios). Of note, however, the seasonal reductions in creel effort, while not significantly different from the current creel design, did perform worse than the percentage reductions that were applied evenly across the year. Conclusions The five alternative creel survey designs reproduced key fishery-dependent metrics that were not different from the estimates obtained via the current creel survey design. This suggests that creel effort could be reduced without compromising the critical estimates that are needed for the Walleye management system in northern Wisconsin and allow for more lakes to be surveyed on an annual basis. Although there are few significant differences between reduced-sampling designs and the full design, there are differences in performance among the different reduced designs that should be considered if surveyors are choosing among the different reduced designs for implementation.