Drones are an effective tool to assess the impact of feral horses in an alpine riparian environment

Abstract

Protected areas can be impacted by the presence and proliferation of feral species. Effective management of feral species requires reliable tools to monitor their population size and ecological impacts. Here, we used drone-based image analysis to assess evidence of feral horses and horse-specific ecological impacts on alpine riparian habitat. Valleys with low (0), medium (1–16) and high (>16) horse abundances were chosen for drone imagery analysis based on independent aerial counts of horses. Data collection trips were carried out pre- and post-2019/2020 wildfires, which unexpectedly burnt valleys with low horse presence. Drone-based RGB orthomosaic imagery was sufficient to identify seven indicators of horse presence and determine the severity of feral horse impacts. Despite the impact of fire, drone-derived classifications were able to accurately detect a gradient of horse impacts, showing a significant difference in indicators from low presence valleys compared with medium and high presence valleys, which did not differ significantly from each other. The significance of differences between valleys reveals that regions routinely inhabited by feral horses will display significant environmental impacts. Our results clearly indicated significant differences between valleys with low horse presence compared with either medium or high horse presence regions (0.01 for differences between both low and medium and low and high horse presence). This was evident both before and after the 2019/2020 fires, suggesting that wildfires did not significantly impact horse populations or distribution in the sampled region. Overall, it was evident that feral horses have a clear and definable impact on alpine riparian vegetation, and drone surveying can be used to routinely monitor potential spread and the outcome of management actions.