1. The affordability, storage, and power capacity of compact modern recording hardware has evolved passive acoustic monitoring (PAM) of animals and soundscapes into a non-invasive, cost-effective tool for research and ecological management and is particularly effective for bats and toothed whales that consistently echolocate. The use of  PAM at large scales hinges on effective automated detectors and species classifiers which, combined with distance sampling approaches, have enabled species abundance estimation of toothed whales. But standardized, user-friendly, and open-access automated detection and classification workflows are in demand for this key conservation metric to be realized for bats.
2. We used the PAMGuard toolbox including its new deep learning classification module to test the performance of four open-source workflows for automated analyses of acoustic datasets from bats. Each workflow used a different initial detection algorithm followed by the same deep learning classification algorithm and was evaluated against the performance of an expert manual analyst.
3. Workflow performance depended strongly on the signal-to-noise ratio and detection algorithm used: the full deep learning workflow had the best classification accuracy (≤67%) but was computationally too slow for practical large-scale bat PAM. Workflows using PAMGuard’s detection module or triggers onboard an SM4BAT or AudioMoth accurately classified up to 47%, 59% and 34%, respectively, of calls to species. Not all workflows included noise sampling critical to estimating changes in detection probability over time, a vital parameter for abundance estimation. The workflow using PAMGuard’s detection module was 40 times faster than the full deep learning workflow and missed as few calls (recall for both ~0.6), thus balancing computational speed and performance. 
4. We show that complete acoustic detection and classification workflows for bat PAM data can be efficiently automated using open-source software such as PAMGuard and exemplify how detection choices, whether pre- or post-deployment, hardware or software-driven, affect the performance of deep learning classification and the downstream ecological information that can be extracted from acoustic recordings. In particular, understanding, and quantifying detection/classification accuracy and the probability of detection are key to avoid introducing biases that may ultimately affect the quality of data for ecological management.

Brief Overview: These data include raw and processed acoustic recordings of bats at woodland edges at different locations in Denmark, made with two common types of acoustic devices: the proprietary SM4BAT FS running firmware version 2.2.7 [Wildlife Acoustics Inc., Maynard, MA, USA], and the open-source AudioMoth running firmware version 1.4.4 (Hill et al., 2018, openacousticdevices.org).

Study Context: The study focus was to test how different setups for detection of bat calls affect the performance of complete acoustic detection and classification workflows for bat PAM data and to demonstrate such full, open-source workflows using PAMGuard (www.pamguard.org)

Each of the four workflows depended on a different initial detection algorithm: 1) A deep learning detection model based on ANIMAL-SPOT and integrated in PAMGuard, 2) PAMGuard's click detector module, 3) a device-based static trigger on the open source AudioMoth recorder, and 4) a device-based dynamic trigger on the proprietary SM4BAT FS recorder. Data from each workflow was processed in PAMGuard. For the two workflows using the AudioMoth and SM4BAT recorders, detections were imported into PAMGUARD for further processing. Following the detection stage, each workflow used the same deep learning species classifier integrated in PAMGuard from ANIMAL-SPOT. 

The project made extensive use of Raven, PAMGuard to process files and AnimalSpot to develop deep learning models. 

Raven Lite was used to annotate files; it is not open source but can be downloaded free of charge from www.ravensoundsoftware.com. Sonic Visualiser provides an open-source alternative to Raven and can be downloaded from www.sonicvisualiser.org.

PAMGuard was used for automated processing and can be downloaded from www.pamguard.org. 

AnimalSpot is a Python library used to train deep learning models and can be forked from here with a paper describing capabilities here.

The files supplied in the dataset are generally text files and raw sound files - they can be opened with a large number of programs.