Using Passive Acoustic Monitoring to Identify Avian Indicators for Reflecting Agricultural Management Practices : A Case Study in Tea Garden of Pinglin

In recent years, the growing global emphasis on biodiversity conservation within Social–Ecological Systems (SES) has catalyzed the development of Long-Term Socio-Ecological Research (LTSER). However, effectively integrating social and ecological data remains a significant challenge. Agroecosystems represent a classic example of human-nature coupled systems, where human agricultural management serves as the core driver. Despite this, most existing research focuses on the broad social or environmental impacts of agriculture, with relatively little attention paid to how specific management practices disturb the activity of local species.

This study focuses on the disturbances caused by agricultural management practices, including pruning, fertilization, pesticide application, and weeding, on avian activities within tea plantations. To achieve high temporal resolution, we utilize Passive Acoustic Monitoring (PAM) to collect soundscape data. These recordings are processed using SILIC, an AI-based biological sound identification and labeling system, to extract precise species and activity information.

To evaluate the short-term impacts of these practices, the research employs Bayesian proportion tests to compare changes in avian habitat occupancy before and after specific management interventions. Furthermore, this study aims to identify bird species that are particularly sensitive to certain agricultural activities and analyze their activity patterns. The findings will serve as a practical reference for conservation and agricultural authorities, enabling the optimization of management schedules to avoid peak avian activity periods and minimize ecological disturbance.

 

Keywords: passive acoustic monitoring, automatic species identification, agricultural management practices, indicator species, tea garden, socio-ecological systems