Abstract
Finding methods to reduce livestock depredations is important for conserving recovering populations of large carnivores and mitigating impacts on ranching livelihoods. This is especially true for ensuring the successful reintroductions of endangered carnivores, such as the Mexican wolf ( Canis lupus baileyi ), an endangered subspecies of gray wolf being reintroduced to their native range in the Southwestern United States and Mexico. Diversionary feeding—providing food caches to divert predators away from preying on livestock—is a strategy to reduce depredations by Mexican wolves but has not been evaluated for its effectiveness. We used data from the Mexican wolf recovery program from 2014 to 2021 in a Bayesian hierarchical model to evaluate whether diversionary feeding reduced livestock depredations by wolf packs and what factors correlated with depredations. Our model accounted for the non‐detection of depredation events, given that some depredations are unencountered or unreported on extensive rangelands. We found that diversionary feeding reduced depredations by an average of 0.78 depredations (43.9%) per pack in the 3 months after application. Prey density was negatively correlated to depredations before diversionary feeding. Minimum pack size and annual livestock density were negatively correlated with depredations after diversionary feeding, while prey density was positively correlated. We estimated a mean of 63% ± 5.4% of depredations were detected with high variation between packs (40.4% ± 7.9% – 74.0% ± 5.3%). Because detections were only two‐thirds of the model‐estimated depredations in our study, our model could improve compensation and targeting of nonlethal tools to mitigate the financial burden of co‐occurrence with wolves by elucidating factors that lead to lower detection and adjusting payment multipliers. Our results indicate diversionary feeding can reduce depredations by wolves on large landscapes in the Western United States but is not a panacea for conflict reduction.

This publication has 65 references indexed in Scilit: