Estimating fat content in barred owls (Strix varia) with predictive models developed from direct measures of proximate body composition

Body condition indices and related metrics can help assess habitat quality and other ecological processes, and ideally, these metrics are based on measures of lipids directly extracted from the species of interest. In recent decades, barred owls (Strix varia) have become a species of conservation concern as they invaded older forests of the US Pacific Northwest, and caused population declines of the closely related and federally threatened northern spotted owl (Strix occidentalis caurina). A simple and effective measure of barred owl body condition could help to understand how habitat quality varies within their new range, which in turn can inform their management and other aspects of their ecology. Using 77 barred owl carcasses collected during experimental removals in Washington and Oregon, USA, we measured the amount of lipid in each specimen with proximate body composition analysis. We then fit and compared (with adjusted R(2) values) alternative linear regression models to estimate the percent lipids in dry mass of the owls based on morphometric body condition indices, a qualitative fat score of subcutaneous breast fat, sex and the time of year females were collected (relative to egg production). Adjusted R(2) values for all models ranged from 0.49 to 0.87, with the best model including mass divided by foot-pad length, fat score, sex and the time of year a female was collected. Most models generated comparable estimates of percent lipids at a population level and we provided correction factors to apply these models when used with live barred owls, allowing for site-specific comparisons of body condition among individuals inhabiting a diversity of environmental conditions.