Data from: Passive-acoustic evidence of lunar, temporal, and weather effects on the vocal activity of four coexisting forest owl species in southern Taiwan
Data files
Apr 22, 2026 version files 3.12 MB
-
GAM.R
3.13 KB
-
README.md
3.42 KB
-
var_h_ALL_MSC02_exp32_2023.txt
713.19 KB
-
var_h_CO_MSC02_exp32_2023.txt
594.68 KB
-
var_h_CSO_MSC02_exp32_2023.txt
595.25 KB
-
var_h_MSO_MSC02_exp32_2023.txt
619.87 KB
-
var_h_NB_MSC02_exp32_2023.txt
592.88 KB
Abstract
Lunar illumination and environmental conditions are known to shape vocal behavior of nocturnal birds, but few studies have simultaneously compared responses across multiple owl species. Using a full-year Passive Acoustic Monitoring (PAM) dataset from Yushan National Park, Taiwan, we investigated how four sympatric Strigidae species—Mountain Scops-Owl Otus spilocephalus, Collared Owlet Taenioptynx brodiei, Collared Scops-Owl O. lettia, and Northern Boobook Ninox japonica—responded to variation in lunar phase, lunar altitude, temporal factors, and weather. We analyzed 1,946 hours of nighttime audio using an AI-based sound recognition model (SILIC ver. exp32), yielding over 700,000 vocalizations. Generalized Additive Models revealed species-specific responses to lunar variables: while Collared Owlet and Northern Boobook increased vocal activity near full moon or high lunar altitude, the other two species remained relatively unaffected. By incorporating both absolute hour and solar-referenced night-phase indices, we further clarified how diel behavior varies across seasons. Weather factors such as wind and rain suppressed vocal activity across all species, but to a varying degree. These results highlight the ecological divergence in nocturnal communication strategies among different owl species, with implications for improving survey timing and species detection in community monitoring.
Dataset DOI: 10.5061/dryad.hmgqnk9zv
Description of the data and file structure
This dataset contains processed passive acoustic monitoring (PAM) data and associated environmental variables used to analyze vocal activity patterns of four coexisting forest owl species in southern Taiwan.
Acoustic recordings were collected in 2023 at a montane forest site in Yushan National Park using autonomous recording units. Vocalizations were automatically detected using an AI-based classifier (SILIC), and results were aggregated into hourly vocal activity rates (VAR).
The dataset integrates three main components:
- Acoustic detection outputs (species-specific vocal activity)
- Temporal variables (hour, month, solar-relative time)
- Environmental covariates (lunar illumination, weather conditions)
The primary response variable, VAR (Vocal Activity Rate), represents the number of detected calls per species per hour and is used to model temporal and environmental influences on owl vocal behavior.
Files and variables
File: GAM.R
Description: R script used for statistical modeling and visualization
File: var_h_CO_MSC02_exp32_2023.txt
Description: Species-specific dataset for Taenioptynx brodiei
Variables
- datetime: Date of observation (YYYY-MM-DD)
- year: Year (2023)
- month: Month (1–12)
- hour: Standardized hour index (1–16) from 16:00 (4 PM) to 07:00 (7 AM of the following day)
- VAR: hourly vocal activity rate (calls/min)
- moon_phase_ratio: Fraction of moon illuminated (0–1)
- moon_altitude_ratio: Relative moon elevation during the recording period
- night_distance: Time relative to sunset/sunrise (continuous variable; negative = night, positive = twilight/day)
- wind_speed: Wind speed (m/s)
- t2m_C: Air temperature (°C)
- sp_hPa: Surface pressure (hPa)
- tcc: Total cloud cover (0–1)
- tp_mm: Precipitation (mm)
File: var_h_CSO_MSC02_exp32_2023.txt
Description: Species-specific dataset for Otus lettia
Variables:
- The same as the file "var_h_CO_MSC02_exp32_2023.txt"
File: var_h_MSO_MSC02_exp32_2023.txt
Description: Species-specific dataset for Otus spilocephalus
Variables
- The same as the file "var_h_CO_MSC02_exp32_2023.txt"
File: var_h_NB_MSC02_exp32_2023.txt
Description: Species-specific dataset for Ninox japonica
Variables
- The same as the file "var_h_CO_MSC02_exp32_2023.txt"
File: var_h_ALL_MSC02_exp32_2023.txt
Description: Combined dataset including all four owl species
Variables
-
The same as the file "var_h_CO_MSC02_exp32_2023.txt"
-
MSO, CO, CSO, NB represent hourly vocal activity rate (VAR) for each species
Variable Species MSO Otus spilocephalus CO Taenioptynx brodiei CSO Otus lettia NB Ninox japonica
Code/software
All analyses were conducted in R (version ≥ 4.2).
Required R packages
mgcv– generalized additive models (GAMs)ggplot2– visualizationcar– variance inflation factor (VIF)lmtest– Durbin–Watson test
