Data from: Butterflies respond to habitat disturbance in tropical forests through activity shifts
Data files
Apr 22, 2025 version files 342.76 KB
-
glmmfinal.csv
7.20 KB
-
microclimate_data.csv
332.04 KB
-
README.md
3.53 KB
Abstract
Habitat disturbance can alter the dynamics of the forest microclimate by disrupting the canopy structure, particularly in structurally complex tropical forests. These changes may impact ectotherms, whose performance and fitness are highly sensitive to climatic conditions. Behavioural responses, such as changes in activities, may help buffer forest ectotherms like butterflies from microclimate changes in disturbed tropical forests. Using field surveys from four tropical forest sites in Asia, we compared flight activity peaks, durations, and intensity for populations of 21 forest-associated butterfly species between open-canopy and closed-canopy forests. We then compared the temperature and illumination that each species experienced during its activity period between the two forest types. Although butterfly populations began their activity earlier and reached peak levels sooner in open-canopy forests compared to closed-canopy forests, the duration and intensity of activity remained similar across populations. Despite these shifts in activity timing between forest types, butterflies experienced comparable temperature conditions in both forest types but were exposed to higher illumination levels in open-canopy forests. Overall, we demonstrate that tropical butterflies can compensate for microclimate changes in tropical forests by shifting their activity patterns. This may help butterflies buffer against temperature increases but not against higher illumination levels following forest canopy opening due to habitat disturbance. Our results emphasize the importance of understanding how animal activity responds to habitat disturbance.
Dataset DOI: 10.5061/dryad.prr4xgxzc
Description of the data and file structure
We surveyed butterflies in each location when adult butterflies were most abundant. We conducted surveys on sunny and windless days. In each study site, two experienced observers recorded butterfly species and abundance using adjusted Pollard walks (30 min at a constant pace, recording butterflies encountered within 5 m of established transects; Basset et al., 2013) at each transect every hour from morning (about half to one hour after sunrise) to dusk (about half to one hour before sunset). At each site, butterfly observers were the same and were randomly switched between forest types or surveys to minimize bias caused by different observers. We recorded the presence of active (displayed flying, feeding, mud puddling, territorial behaviour, mating and ovipositing) butterfly individuals during the sampling.
To measure the microclimate difference between forest types, we placed 1-2 data loggers (Onset HOBO UA-002-64) in each habitat suspended 1m above ground under vegetation (for avoiding direct solar radiation). These data loggers recorded air temperature (°C) and illumination (Lux) at every 5-minute interval in TPK and KHC, at 1-minute intervals in XTBG and MBCA. We aggregated (averaged across) 1-minute interval recordings to 5-minute intervals to make recordings in different sites comparable and used these 5-minute interval recordings in the subsequent analyses.
Files and variables
The dataset included two csv files designated for the two analysis used in the manuscript.
The microclimate dataset contains 30min-interval (transformed to numerical numbers 0 to 23.5) illumination (in Lux) and temperature (in celsius) readings collected from our four sampling sites in two forest types (open and closed). In the same forest type and sites, the microclimate readings contains multiple days.
The glmmfinal dataset contains activity start, end, duration, and intensity (calculated from activity curves estimated by kernel estimation) for each population in each forest type in four sites (TPK – Tai Po Kau Nature Reserve, Hong Kong;KHC – Khao Chong Botanical Garden, Thailand; XTBG – Xishuangbanna Tropical Botanical Garden, Yunnan, China; MBCA – Maliau Basin Conservation Area, Sabah, Malaysian Borneo). Variable "intensity_10_2" indicate activity intensity from 10 am to 2 pm. "stad_accu_temp" indicate standardized accumulated temperature profiles in closed forest, calculated by dividing the cummulative temperature during activity by activity duration. Similar calculations apply to the variable "stad_accu_light", and "stad_o_accu_temp" (in open forest). "habi" is short for habitat of open and closed canopy forests. "paired" labelled same letters to a same species in one site. This variable is used for a random factor in glmm, to make sure the comparison is within species. "Scenario" is for hypothetical changes in butterfly activity. When the value is "real_", it means the data in different "stad_accu_" variables are from observed dataset. When the value is "fixed_", it means the different "stad_accu_" metrics in open canopy forests are caclulated within the time periods of activity calculated from the closed canopy species population (assumming butterflies do not shift activity pattern).
Code/software
Microsoft excel can be used to view the data file.