Affiliation:
1. Ohio Division of Wildlife, Fairport Harbor Fisheries Research Station Fairport Harbor Ohio USA
2. Oklahoma State University Stillwater Oklahoma USA
Abstract
AbstractObjectiveYellow Perch Perca flavescens are popular sport fish; however, obtaining adequate length data can be problematic during low‐abundance years. Using fish from the sport fishery cleaning stations provides a possible source of data, but fish from this source have already been filleted, making length data questionable, and may not have intact backbones, so lengths cannot be determined. Therefore, we tested if Yellow Perch total length measured from filleted fish are similar to the total length measured before filleting and we also developed regression equations to predict total length from filleted fish length, head length, or mandible lengths and compared their accuracy to known total lengths.MethodsYellow Perch were collected from the Ohio Department of Natural Resources standardized bottom trawl survey. Each fish was measured for total length using a measuring board, and head and mandible lengths were measured with digital calipers. A subset of fish (N = 46) was filleted, and total length was again measured to see if filleting altered length measurements. We used linear regression with 10‐fold cross validation to estimate the total length of Yellow Perch from filleted fish length, head length, and mandible length.ResultOur results show that all three measurements were good predictors (R2 > 0.98) of fish total length, with precision being greatest for filleted length, followed by head length (mandible length had notably lower precision). Filleted fish lengths were significantly longer than intact total length, but we provide a regression equation that can be used to estimate unfilleted length using filleted fish length.ConclusionWe recommend estimating total length from filleted length regression or head length regression in cases when total length cannot be directly measured. Our mandible length regression can also be used, but it was slightly less precise than the head length regression, which should be used instead when practical.
Reference14 articles.
1. Building Predictive Models inRUsing thecaretPackage
2. lmerTest Package: Tests in Linear Mixed Effects Models
3. Lenth R.(2024).emmeans: Estimated marginal means aka least‐squares means. R package version 1.10.0.https://CRAN.R‐project.org/package=emmeans.