An Experimental Evaluation of Electrofishing Catchability and Catch Depletion Abundance Estimates of Larval Lampreys in a Wadeable Stream: Use of a Hierarchical Approach

Abstract

Some lamprey species are in decline, and assessments of local abundance could benefit research and conservation. In wadeable streams, larval lampreys are collected by using specialized backpack electrofishing techniques, although catchability has not been sufficiently evaluated. We assessed removal models for estimating the local abundance of larval lampreys in experimental net-pen enclosures within a wadeable stream. Known numbers of larvae were seeded at densities of 4–130 larvae/m² into 1-m² enclosures that were lined with fine sand and placed into Cedar Creek, Washington (Columbia River basin). Depletion sampling in each enclosure (n = 69) was conducted by three to five electrofishing passes, and abundance was estimated by six removal models that assumed different catchability functions. Catchability averaged 0.28. For the standard removal model, which assumed that catchability varied independently by enclosure but not by pass, the 95% highest posterior density credible intervals (95% HPD-CIs) were very large relative to the abundance estimates. Models assuming that catchability was either equal or a random factor among all enclosures and passes generally produced accurate (mean bias = −0.04) estimates of abundance, and 95% HPD-CIs were much smaller. Based on our data set, the expected bias of abundance estimates for 80% of simulations was less than 20% if five passes were completed from at least four randomly selected quadrats and if catchability was assumed to be a random factor. Additional sampling may be needed at low lamprey densities (especially 2). Our results suggest that local abundance of larval lampreys in wadeable streams can be effectively estimated by depletion sampling at multiple 1-m² quadrats and by use of a hierarchical removal model. Received June 12, 2015; accepted April 20, 2016 Published online August 5, 2016