Wildlife Society Bulletin · 2018 · 10 citations · 36 references
AgricultureAnimal SciencesAnimal NutritionLipid BiochemistryWildlife ManagementBody CompositionLipid AnalysisWildlife NutritionAvian PhysiologyBody Condition IndexBody Size EstimationMorphometric IndicesMorphometric ParameterizationPredictive Equation ValidationProtein Content DeterminationSerum Total Protein
ABSTRACT Assessing nutrient stores in avian species is important for understanding the extent to which body condition influences success or failure in life‐history events. We evaluated predictive models using morphometric characteristics to estimate total body lipids (TBL) and total body protein (TBP), based on traditional proximate analyses, in spring migrating lesser snow geese ( Anser caerulescens caerulescens ) and Ross's geese ( A. rossii ). We also compared performance of our lipid model with a previously derived predictive equation for TBL developed for nesting lesser snow geese. We used external and internal measurements on 612 lesser snow and 125 Ross's geese collected during spring migration in 2015 and 2016 within the Central and Mississippi flyways to derive and evaluate predictive models. Using a validation data set, our best performing lipid model for snow geese better predicted TBL (root mean square error [RMSE] of 23.56) compared with a model derived from nesting individuals (RMSE = 48.60), suggesting the importance of season‐specific models for accurate lipid estimation. Models that included body mass and abdominal fat deposit best predicted TBL determined by proximate analysis in both species (lesser snow goose, R 2 = 0.87, RMSE = 23.56: Ross's geese, R 2 = 0.89, RMSE = 13.75). Models incorporating a combination of external structural measurements in addition to internal muscle and body mass best predicted protein values ( R 2 = 0.85, RMSE = 19.39 and R 2 = 0.85, RMSE = 7.65, lesser snow and Ross's geese, respectively), but protein models including only body mass and body size were also competitive and provided extended utility to our equations for field applications. Therefore, our models indicated the importance of specimen dissection and measurement of the abdominal fat pad to provide the most accurate lipid estimates and provide alternative dissection‐free methods for estimating protein. Published 2018. This article is a U.S. Government work and is in the public domain in the USA.
36
A Simple Test for Heteroscedasticity and Random Coefficient Variation
Trevor Breusch, A. R. Pagan · Econometrica · 1979 · 5.4K citations
Statistical Science · 2010 · 2.4K citations · Full text
RESTITUTION OF MASS–SIZE RESIDUALS: VALIDATING BODY CONDITION INDICES
Albrecht I. Schulte‐Hostedde, Bertram Zinner, John S. Millar et al. · Ecology · 2005 · 1.1K citations · Full text