PS 76-163
Developing reliable covariates for survival analyses in snowshoe hares

Thursday, August 13, 2015
Exhibit Hall, Baltimore Convention Center
Dianne Quiroz, Sierra Nevada Research Institute, University of California Merced
L. Scott Mills, Forestry and Environmental Resources, North Carolina State University

Snowshoe hares (Lepus americanus) are one of ecology’s iconic species to study population dynamics, and have recently become a model species to study adaptation to climate change as white coat color molts become mismatched with brown snowless ground.  Factors affecting survival are central to research in these areas. We used a 13-year data set from long-term capture-mark-recapture studies of snowshoe hares in the southern portion of their geographic range to evaluate the validity of hind foot length (HFL) as a covariate in survival analyses.  HFL has been shown to be a compelling predictor of survival in hares, probably both because it is closely correlated with structural body size and because larger feet decrease foot loading on snow.  However, for survival studies using radiotelemetry, HFL would only be a reliable covariate if it did not grow between the time it was last measured and the time of death.  Here we seek to establish a threshold weight above which HFL growth has stabilized, allowing it to be a reliable survival covariate for that animal. 


Initial results for our first site, in the Seeley Lake region of Western Montana, indicate that there were no differences in mean HFL among the four grids sampled within this site (Open Old, Open Young, Closed Old, and Closed Young).  We find that HFL stabilizes at around 1319g for adult males and 1309 g for non-reproductive females.  Although we need to complete analyses at other sites, and test the repeatability and accuracy of HFL measurements, these results provide an initial point of reference for the weights at which HFL can begin to be included as a covariate in survival analyses.