Multinomial N-mixture models improve the applicability of electrofishing for developing population estimates of stream-dwelling Smallmouth Bass

North American Journal of Fisheries Management
By:  and 

Links

Abstract

Failure to account for variable detection across survey conditions constrains progressive stream ecology and can lead to erroneous stream fish management and conservation decisions. In addition to variable detection’s confounding long-term stream fish population trends, reliable abundance estimates across a wide range of survey conditions are fundamental to establishing species–environment relationships. Despite major advancements in accounting for variable detection when surveying animal populations, these approaches remain largely ignored by stream fish scientists, and CPUE remains the most common metric used by researchers and managers. One notable advancement for addressing the challenges of variable detection is the multinomial N-mixture model. Multinomial N-mixture models use a flexible hierarchical framework to model the detection process across sites as a function of covariates; they also accommodate common fisheries survey methods, such as removal and capture–recapture. Effective monitoring of stream-dwelling Smallmouth Bass Micropterus dolomieu populations has long been challenging; therefore, our objective was to examine the use of multinomial N-mixture models to improve the applicability of electrofishing for estimating absolute abundance. We sampled Smallmouth Bass populations by using tow-barge electrofishing across a range of environmental conditions in streams of the Ozark Highlands ecoregion. Using an information-theoretic approach, we identified effort, water clarity, wetted channel width, and water depth as covariates that were related to variable Smallmouth Bass electrofishing detection. Smallmouth Bass abundance estimates derived from our top model consistently agreed with baseline estimates obtained via snorkel surveys. Additionally, confidence intervals from the multinomial N-mixture models were consistently more precise than those of unbiased Petersen capture–recapture estimates due to the dependency among data sets in the hierarchical framework. We demonstrate the application of this contemporary population estimation method to address a longstanding stream fish management issue. We also detail the advantages and trade-offs of hierarchical population estimation methods relative to CPUE and estimation methods that model each site separately.

Study Area

Additional publication details

Publication type Article
Publication Subtype Journal Article
Title Multinomial N-mixture models improve the applicability of electrofishing for developing population estimates of stream-dwelling Smallmouth Bass
Series title North American Journal of Fisheries Management
DOI 10.1080/02755947.2016.1254127
Volume 37
Issue 1
Year Published 2017
Language English
Publisher American Fisheries Society
Contributing office(s) Coop Res Unit Atlanta
Description 14 p.
First page 211
Last page 224
Country United States
State Missouri, Oklahoma
Other Geospatial Ozark Highlands
Google Analytic Metrics Metrics page
Additional metadata about this publication, not found in other parts of the page is in this table