Do wolf kill-sites reflect prey accessibility or abundance in forested landscapes? A LiDAR based assessment of forest and topographic structure at moose mortality sites on Isle Royale National Park

Background/Question/Methods

Environmental heterogeneity plays a potentially important, but poorly understood role in the ability of large predators to obtain prey.  For example, structural heterogeneity may either reduce or enhance the efficiency of a predator’s efforts to search, track, capture, kill, and consume prey. For prey, structural heterogeneity may affect predator detection, avoidance and defense, escape tactics, and the ability to exploit refuges. This study investigates whether and how heterogeneity in topographic and vegetation structure influence the spatial patterns of moose (Alces alces) mortality on Isle Royale National Park using island-wide LiDAR data. We analyzed locations where wolves (Canis lupus) killed moose during the winters between 2002 and 2010 (n=230), and where moose died of starvation between 1996 and 2010 (n=182). Structural metrics derived from LiDAR data were generated and integrated with an algorithm model (Random Forest) to identify, characterize, and classify three-dimensional variables related to mortality classes and random points. We also developed spatial models to predict and assess the likelihood of moose mortality at the landscape scale based on structural variables.  This approach explores whether or not wolf kill-sites or starvation sites relate to habitat structure, prey availability, and moose habitat selection.

 Results/Conclusions

Wolf kill-sites are more likely than malnutrition-sites to occur in areas with greater diversity in vertical forest structure (i.e. canopy layers), more open canopy, and less dense vegetation.  Malnutrition sites are more likely to occur in denser understory vegetation, between 1.5-2.5 meters in height. Female moose are more likely than males to be killed by wolves in denser vegetation, between 1.5–10 meters in canopy height, and die of malnutrition in comparatively less dense forest stands.  Arthritic moose are more likely than healthy moose to die of predation at sites of overall denser and higher vegetation canopy and die of malnutrition in areas of dense understory canopy, between 1.5-2.5 meters.  These results indicate that moose mortality patterns are non-random and highly variable across the landscape for both wolf predation and malnutrition. Important differences exist in site-level topographic and vegetative structure based on moose age, sex and health.   Wolf kill-sites in winter on Isle Royale generally appear more likely to be a result of prey habitat selection, i.e. reflecting broad patterns of prey abundance rather than strong patterns of accessibility.  This result contrasts with patterns observed in other predators, such as large felids, which may be explained by predator hunting mode.