Abstract
Managed grazing is the most widespread and economically significant form of grassland utilization worldwide. Accurate quantification of the spatiotemporal distribution of grazing intensity (GI) is crucial for promoting sustainable management of livestock-grassland ecosystems. However, a reliable method for dynamically monitoring GI and quantifying key proxies under real-world grazing conditions is still lacking. In this study, we developed a practical approach to estimate GI using sequential unmanned aerial vehicle (UAV) monitoring and evaluated its feasibility in a typical household pasture on the Qinghai-Tibetan Plateau, China. Our findings show that: (1) yak dung is clearly identifiable in UAV image, although detection accuracy decreases with increasing flight altitude (from 100% at 2 m to 93.16% at 20 m); (2) yak dung density serves as a feasible proxy for GI, effectively capturing its temporal and spatial variability; (3) yak dung density reflects cumulative GI from May to September, and its representativeness increases with the length of accumulation. The proposed approach is characterized by high frequency, accuracy, and efficiency. It is well-suited for studying animal behavior and evaluating livestock-resource relationships, thereby providing valuable insights for sustainable grassland ecosystem management.