Integrating endocrine biomarkers and landscape structure to advance non-invasive demographic and physiological monitoring of woodland caribou

Abstract

Effective conservation of woodland caribou (Rangifer tarandus caribou) requires monitoring approaches that can provide biological information from wide-ranging, low-density populations where direct observation is difficult. Non-invasive fecal sampling is already used in caribou monitoring for genetic identification, sex determination, and population estimation, but the potential for fecal samples to provide additional information on age-class, pregnancy status, and endocrine variation remains less developed. This thesis evaluated the utility of fecal morphometric and endocrine biomarkers for deriving biological information from winter-collected fecal samples in woodland caribou from the Churchill Range of northwestern Ontario. First, we assessed whether fecal pellet morphometrics could distinguish among age-classes during winter, and whether fecal progesterone concentrations could be used to distinguish pregnant from non-pregnant females. Gaussian mixture modeling revealed no evidence of discrete morphometric size classes within sex after accounting for seasonal variation, indicating that pellet morphometrics provided limited age-class resolution. In contrast, fecal progesterone concentrations exhibited clear bimodality and supported classification of most females as pregnant or non-pregnant. Second, we evaluated whether fecal cortisol concentrations varied across gradients of anthropogenic disturbance and landscape composition. Fecal glucocorticoid concentrations were negatively associated with the proportion of the landscape burned within the previous 40 years and positively associated with the proportion of wetlands, indicating that this endocrine biomarker varied with landscape conditions. Collectively, these findings demonstrate that fecal endocrine biomarkers can strengthen non-invasive monitoring of woodland caribou by supporting pregnancy-status classification and providing insight into landscape-associated variation in fGCM concentrations.