Data and analysis code for paper entitled 'Passing rail traffic reduces bat activity', appearing in Scientific Reports (2021).
The raw data consists of 38 nights of acoustic, ultrasonic and video recordings made between 15th July - 20th September 2019, at 12 rail-side sites in South-eastern England (see Supplementary Materials Table 1 on article web page for details).
Each night's recordings are contained within a separate zip archive, named by site number, followed by site name, day of recording at that site, and date (e.g. 2. Hassocks - Day 1 - 15.07.19.zip). Zip files are best opened with 7zip (Windows) or The Unarchiver (MacOS).
The folder structure within each night's raw data zip archive is:
XXX represents the detector's individual ID. Acoustic folders contain calibrated acoustic recordings of passing trains (wav format), and if the acoustic data was used in sound pressure level (SPL) calculations, SPL data (Praat v6.1.09 pkf format, also converted to a text file with file extension .Collection for importing into R (see below). Ultrasonic folders contain ultrasonic audio files (wav format) as labelled using Wildlife Acoustics Kaleidoscope Pro v3.1.5, along with Kaleidoscope metadata (csv format), a log (txt format) of files processed, and the settings file (ini format) used by Kaleidoscope. Files identified as noise are stored in a 'NOISE' subfolder. While the individual audio files and Kaleidoscope generated csv files can be opened using standard software, both are best viewed/played using Kaleidoscope v3.1 or higher.
Intermediate processed data which were imported into R (see below), can be found in the zip archive Intermediate Processed Data (R Imports).zip, which has the following folder structure:
|-- Bat Detectors ....|-- Acoustic ....|-- Ultrasonic ........|-- Discarded ........|-- Selected |-- Train Dimensions |-- Video Data |-- Weather Data
The Acoustic folder contains data calibrating times between acoustic and video datasets (csv and xlsx format). Within the Ultrasonic folder, Kaleidoscope metadata files (with additional columns noting feeding buzzes and start/end times, csv format) are separated into Selected and Discarded subfolders depending on whether data from a given detector was used in the analysis (see Methods). The Train Dimensions folder contains the information used to determine average carriage length, and so estimate train speed (xlsx format), the Video Data folder contains timings of train passes manually extracted from the video recordings (csv format), and the Weather Data folder contains Gatwick Airport weather station records (csv format). Detector distance to trains, and site locations /timings (csv format) are provided at the top level of the archive.
All R code (R version 3.6.2) used in further data processing steps, and the final analyses, plus the generated data (as csv files) are provided in a single R Project within the zip archive Data Processing and Analysis.zip. The R project contains a renv database containing all software packages and versions used, allowing the analytical environment to be reproduced using the R command renv::restore(). All paths to external files (outside the project directory) within the R code have been truncated to '/path/to/' before the file name. Consequently, paths to external files should be respecified (e.g. by using a search function to locate the file within the archives) before running affected code.
Abstract: Rail transport is expanding, with a global increase in infrastructure of up to one-third predicted by 2050. Greater reliance on rail is expected to benefit the environment at a planetary level, by mitigating transport-related carbon emissions. However, smaller-scale, more direct consequences for wildlife are unclear, as unlike roads, railway impacts on animal ecology are rarely studied. As a group, bats frequently interact with transport networks due to their broad distribution and landscape-scale movements. Additionally, their nocturnality, and use of echolocation mean bats are likely to be affected by light and noise emitted by trains. To investigate whether passing trains affect bat activity levels, we monitored the two most abundant UK species using ultrasonic detectors at 12 wooded rail-side sites in southern England. Activity fell by ≥ 30-50% each time a train passed, for at least two minutes. Consequently, activity was reduced for no less than one-fifth of the time at sites with median rail traffic, and two-thirds or more of the time at the busiest site. Such activity changes imply repeated evasive action and/or exclusion from otherwise favourable environments, with potential for corresponding opportunity or energetic costs. Hence, disturbance by passing trains may disadvantage bats in most rail-side habitats.