Real-time bioacoustic monitoring for eastern black rails at Fort Drum Wildlife Management Area

The federally listed threatened eastern black rail (Laterallus jamaicensis jamaicensis) has experienced a significant population decline over the past 30 years, primarily due to habitat loss. Florida, which has the largest remaining population along the Atlantic coast (~1,400 pairs), is vital for the conservation of this species. Thus, accurate survey data is critical for understanding the populations and distribution of this species in Florida to assess their conservation status and to make informed habitat management decisions.

Traditional bird survey methods (e.g., call response surveys) have limitations due to the cryptic coloration and secretive behavior of eastern black rails. In addition, these survey methods are generally costly, limited in space and time, time-consuming, and, most importantly, do not include a permanent record. Consequently, there is no way to validate the data. In contrast, passive acoustic monitoring (PAM) systems are based on the deployment of autonomous recording units (ARUs) in the field to record soundscapes. PAMs can monitor animal populations 24 hours a day, 365 days a year, across a variety of habitats simultaneously, without requiring human attention for extended periods, and all recordings can be permanently stored. Thus, the use of ARUs may prove to be an important survey method for monitoring eastern black rails with minimal field effort.

In April and May 2024, Earthology conducted a study lead by Jeff Howe to 1) identify the presence of eastern black rails at the Fort Drum Wildlife Management Area (FDWMA), Indian River County, Florida, and 2) test the Automated Remote Biodiversity Monitoring Network (ARBIMON) platform’s ability to detect the absence/presence of the eastern black rail. To date, the presence of eastern black rails at FDMCA had only been determined using call-response surveys. ARBIMON is an acoustic analysis platform that provides an efficient way to upload, store, and analyze large acoustic datasets. If successful, this web-based application would allow researchers to rapidly analyze acoustic files for the absence/presence of eastern black rails compared to the labor-intensive process of aurally reviewing acoustic files or conducting call-response surveys.

Acoustic recordings were collected along the eastern portion of the FDWMA (Figure 1). The FDWMA encompasses nearly 21,000 acres of wetland and upland communities, including pine flatwoods, dry prairie, hardwood swamp and floodplain swamp. The marsh areas represent the southernmost portion of the St. Johns River’s headwaters. The FDWMA is managed as a wildlife management area through the Florida Fish and Wildlife Conservation Commission.

This study found that PAM systems can be deployed with minimal effort and environmental disturbance and can efficiently collect a large volume of audio recordings. ARBIMON was used to identify the absence/presence of the eastern black rail using 132 hours of acoustic recordings. A field survey conducted over the same number of days by a researcher would be logistically and economically impractical. However, it is also impractical to aurally review 132 hours of recordings. Consequently, the use of automated machine learning systems is essential, especially with considerably larger datasets; however, they present their own set of difficulties, particularly in developing a mathematical model that detects the call of interest with a high degree of precision.

Although this was a small study consisting of just two sites, the ARBIMON platform was able to process a moderate-sized dataset and detect eastern black rail calls. That said, the algorithm integrated into the ARBIMON Pattern Matching platform appears to be quite simple, hence producing results that varied widely and produced a high number of false positives. Consequently, as machine learning analysis data processing tools continue to improve, the precision of detecting targeted species will improve. More in-depth results were recently published in the journal Florida Scientist volume 88.

Fort Drum Marsh Conservation Area. Photo credit to St. Johns River Water Management District.