Project Objectives:

1. Develop a machine learning model capable of automatically identifying selected Philippine bird species based on their calls

2. Evaluate the performance of the classifier model in reliably detecting and accurately identifying species calls across varied field recordings

3. Analyze diel (within-day) and seasonal (across the year) vocal activity patterns of selected Philippine bird species using the trained model

Abstract:
Passive acoustic monitoring (PAM) and species recognition models provide a promising approach for large-scale, rapid biodiversity monitoring, but their performance remains largely understudied for Philippine birds. As part of a broader effort to develop and evaluate acoustic monitoring tools for select bird species in the Philippines, our study assesses the performance of BirdNET Analyzer v2.4.0 model in detecting the Philippine Nightjar (Caprimulgus manillensis). We collected acoustic recordings across the University of the Philippines Diliman campus from January to May 2026 and analyzed them using BirdNET, with a subset of detections validated through manual review. Across 36 recording locations surveyed, Philippine Nightjars were detected at 78% of sites. Our preliminary findings indicate a high model accuracy, with only one false positive identified among 1,213 BirdNET detections. Ongoing analyses are assessing model recall and precision through additional manual verification of recordings. Philippine Nightjars vocalized throughout the night and are most frequently detected in areas characterized by layered open forest vegetation. Our initial results demonstrate the utility of BirdNET for detecting Philippine Nightjars calls and provide preliminary insights into the species' spatio-temporal  activity patterns. This work further contributes to the growing application of automated bioacoustic monitoring in the Philippines and the development of future species-specific acoustic classifiers for Philippine avifauna.

Project Objectives:

1. To co-develop a research plan, an acoustic monitoring protocol, and specific conservation objectives related to migratory birds in the EAF.

2. To create a robust set of acoustic training data of Southeast Asian migratory bird species to improve automated detection and classification models (e.g., BirdNET).

3. To use the data and proof of concept developed in this project to apply for larger funding sources that can help support long-term locally-led monitoring within the EAAF

Abstract: 

The Las Piñas–Parañaque Wetland Park (LPPWP) is a protected wetland ecosystem located along the southern coast of Manila Bay, Philippines. Covering approximately 181 hectares, it consists of mudflats and mangrove forests that provide crucial habitat for both migratory and resident bird species. To address these challenges, enhanced monitoring techniques are essential for assessing habitat health and species diversity. Passive acoustic monitoring (PAM) offers a non-invasive method that uses stationary recorders deployed in the field to collect acoustic data from a site’s surrounding environment over extended periods, enabling long-term monitoring of wildlife and the environment. The Locally-Led East Asian Flyway Acoustics (LEAFA) Migratory Program, supported by the Cornell Lab of Ornithology, aims to establish a collaborative, regional conservation framework for migratory birds along the EAAF by promoting participatory research. Given LPPWP’s ecological importance and its designation as a Critical Habitat, Ramsar Site, and Protected Area, this research will provide valuable insights as part of the LEAFA Migratory Bird Program’s regional effort into the spatiotemporal distribution and acoustic behavior of migratory species in Southeast Asia.