COS 238-1 - Acoustic monitoring reveals shifts in avian vocalization and biodiversity in response to forest management practices

Abstract: Forested landscapes around the world face unprecedented impacts from land-use conversion and climate change, creating an urgent need to robustly and rapidly evaluate effects on wildlife. Passive acoustic monitoring has emerged as a powerful approach to assess the consequences of these impacts on sonant animals by revealing changes in vocalization activity and biodiversity across space and time. Despite this, traditional approaches to acoustic species identification can be time and resource intensive, severely limiting the application of this approach at scale. Here we use acoustic indices and machine learning techniques to automatically quantify vocalizations of avian species in a temperate rainforest, providing insight into patterns of community composition and bioacoustic activity along a gradient of developmental stages including previously-logged stands and in response to extreme climatic conditions.

Our objective was to evaluate how avian community composition and vocal activity respond to forest management practices in a changing climate. Passive acoustic recordings were made during the years of 2020-2022 at 217 sites distributed among four forest developmental stages (stand initiation, stem exclusion, commercially thinned, and mature forest) in 16 watersheds of the Olympic Experimental State Forest, managed by the Washington State Department of Natural Resources. During the study period, the Pacific Northwest region experienced an extreme heat wave event, resulting in different peak temperatures among stand types. With these data we calculated alpha and beta acoustic indices and used machine learning classification techniques to identify species-specific vocalizations, yielding quantitative measures of acoustic activity and community composition across stand types over time.

Results suggest significant differences in community composition between developmental stages and management practices, reflecting differences in habitat function. Variation in the abundance and timing of vocalizations in response to the heat wave event indicated changes in avian behavior. Comparison of the magnitude of these changes in acoustic activity between developmental stages suggests a range in the severity of wildlife impacts due to the intersection of temperature and habitat structure. These findings illustrate the utility of monitoring bioacoustic communities as indicators of wildlife response to disturbance that can inform future forest management in a changing climate.