Scaling issues in the study of livestock depredation by carnivores

Abstract

Abstract Context Many studies have documented the magnitude and socioecological drivers of livestock depredation, yet few have assessed how observations of depredation might vary with spatiotemporal scale. Understanding scaling relationships may allow for more accurate aggregation of observations collected across diverse extents and resolutions to better inform management actions. Objectives Herein we examine how reported metrics of livestock depredation varied by spatial and temporal scale (extent and resolution) after controlling for other drivers of differences among studies. Methods From 213 published studies we extracted conflict metrics (i.e., number of animals killed, number of attack incidents, and percent annual loss of stock) and regressed each against spatiotemporal extent (i.e., size of study area, duration of study) and resolution (i.e., minimum mapping unit, frequency of observations) while controlling for potential biases. Results The number of attacks or animals killed was positively related to spatial and temporal extent whereas percent annual loss of stock declined with extent. Further, the effects of scaling variables were modified through interactions with other factors (e.g., human density) known to influence human-carnivore conflict. The data available for spatiotemporal resolution were too sparse to draw conclusions. While scaling relationships were generally linear, they became increasingly noisy at broader extents. Conclusions Consistent with other ecological investigations, the scale of observation influenced the observed outcomes in human-carnivore conflict metrics. Authors should report the spatiotemporal dimensions of their observations to improve robust inference in comparative studies. Enhanced understanding of scaling relationships in human-carnivore conflict metrics should improve strategic allocation of resources to better mitigate future conflict.