Assessing biological realism of wildlife population estimates in data-poor systems

Large carnivore management is often contentious, particularly in jurisdictions where hunting and conservation efforts collide. Regulated hunting is a common management tool, yet relevant decisions are commonly taken in the absence of reliable population data and are driven by factors other than biological considerations. We used European large carnivore (brown bear Ursus arctos, wolf Canis lupus and Eurasian lynx Lynx lynx) management to evaluate the biological plausibility of reported population estimates used in hunting decisions. We used Romania as a test case as this region is not only data-poor, but the public and private game managers are beneficiaries of revenue from hunting activities. We assessed the following: (i) how population growth rates calculated from reported abundances between 2005 and 2012 compared to published growth rates empirically derived from European and North American populations; (ii) whether biological unrealism compounded through time by testing whether reported estimates fell within the bounds of biologically plausible trajectories; and (iii) the relationship between the occurrence of biologically unrealistic estimates and financial incentives (amount of hunting). For U. arctos, which generates high revenue, estimated annual population growth rates were frequently greater than maximum published growth rates (up to 15 for reported versus 1136 in the literature). Reported estimates were greater than maximum simulated populations in 32% of cases, and the difference was positively correlated with hunting (r(s)=0576). Population growth rates for C. lupus overshot the maximum published growth rate (135) less frequently, reported estimates were within the bounds of biologically plausible estimates (91% of cases), and there was a weak correlation between hunting and biologically unrealistic estimates (r(s)=0182).L. lynx population growth rates derived from reported estimates were lower than minimum simulated populations (60% of cases), and there was a weak correlation between hunting and biologically unrealistic estimates (r(s)=0164).Synthesis and applications. Our study suggests that comparing population estimates used by management agencies to demographic data obtained through rigorous peer-reviewed studies is a useful approach for evaluating the biological plausibility of wildlife data in data-poor systems, especially when management decisions might be influenced by non-scientific incentives.