Using morphological measurements to predict subspecies of Midcontinent sandhill cranes

Abstract

The Midcontinent population of sandhill cranes (Antigone canadensis) has historically been classified into 3 putative subspecies, but genetic analyses have identified only 2 genetically distinct subspecies. Previous studies have successfully used morphometrics in combination with an individual's sex to differentiate subspecies of sandhill cranes that had been inferred based on breeding area, but no study has used a sample of genetically determined subspecies to discriminate and develop predictive models. Using measurements from 843 adult sandhill cranes captured throughout their range and annual cycle (in 4 States and 1 Canadian province during 1998–2007), we used linear discriminant analysis to classify genetically identified A. c. canadensis (lesser) and A. c. tabida (greater) sandhill crane subspecies, and developed a field-ready tool to predict subspecies using common morphometric measurements without determination of an individual's sex. Our top-ranked model was 89.5% accurate overall, and used flattened wing chord, total culmen, and tarsometatarsus lengths to correctly identify 93.1% of A. c. canadensis and 82.8% of A. c. tabida subspecies. Additionally, we identified measurement thresholds based on posterior probabilities of correct classification to aid in subspecies determination when the linear discriminant procedure provided equivocal results. We also investigated whether sex determination could increase accuracy of our top-ranked model, and found that accuracy increased <1% when including this information. We suggest collection of the morphometric measurements used in our top-ranked model to determine subspecies of adult Midcontinent sandhill cranes. Our method does not require determining sex of the individual to correctly classify subspecies, allows for accurate and rapid subspecies determination, and can largely avoid additional costs and time associated with genetic analyses to determine subspecies. © 2019 The Wildlife Society.