Data from: Tiny patch, big value: A small dry forest patch supports wildlife conservation in Guanacaste, Costa Rica
Data files
Sep 08, 2025 version files 91.14 KB
-
Analysis.R
20.17 KB
-
Full_Camera_Dataset_(Cleaned).csv
67.06 KB
-
README.md
3.75 KB
-
Trapping_Data.csv
161 B
Abstract
Tropical dry forests are among the most threatened tropical ecosystems due to ongoing agricultural expansion and development. Despite this, small forest patches persist within fragmented landscapes, though their conservation value remains poorly understood. To evaluate the role of one such patch in supporting wildlife, we deployed camera traps across three zones (Dense Forest, Grassland, and Sparse Forest) within the Murren Reserve, a small dry forest patch on the southern coastline of Guanacaste, Costa Rica. Camera traps detected 16 vertebrate species, with opportunistic encounters adding 16 more. The most detected species included the white-nosed coati (Nasua narica), raccoon (Procyon lotor), and white-tailed deer (Odocoileus virginianus). Additionally, we recorded nationally endangered species, including the ocelot (Leopardus pardalis), puma (Puma concolor), and mantled howler monkey (Alouatta palliata), as well as nationally threatened species such as the spotted skunk (Spilogale angustifrons) and white-throated magpie-jay (Calocitta formosa). Extrapolated rarefaction curves suggest that further sampling could improve species richness estimates for the Dense Forest and Grassland. Although we did not detect a difference in species composition among zones (ANOSIM, R = 0.14, p = 0.26), a generalized linear mixed model indicated that zone explained 17% of the variation in species richness. These findings highlight the role that small dry forest patches play in conserving both threatened species and broader vertebrate communities. We also provide recommendations for future research to address current data gaps and improve long-term monitoring in fragmented landscapes.
Dataset DOI: 10.5061/dryad.0000000dg
Description of the data and file structure
We evaluated vertebrate biodiversity across distinct zones within a small tropical dry forest by deploying camera traps in the Murren Reserve, Costa Rica. We placed cameras across three zones (Dense Forest, Grassland, and Sparse Forest) during the dry and wet seasons of 2024. Because of incorrect time and date metadata from images captured during the wet season, we focused most analyses on dry season data. We compared daily species richness and overall species composition between zones using generalized linear mixed models, dissimilarity analyses, and rarefaction curves. We also documented opportunistic vertebrate observations to supplement the camera data.
A note about the inclusion of location data of threatened species: The study site covers ~13 ha (about half a square mile), and the publication associated with this dataset already includes its coordinates. In this context, omitting GPS coordinates of the focal species would not meaningfully reduce the risk of illegal hunting. However, the site is privately owned, clearly marked as such, and bordered by a national park managed by government officials. These factors should, in theory, further reduce the risk of illegal hunting.
Files and variables
File: Full_Camera_Dataset_(Cleaned).csv
Description: This is the primary data set used in our article. To clean the data, we first removed all "false captures" (i.e., videos with no animal). Then, we defined a unique capture event (UCE) as any instance in which a species had not been detected by the same camera within the previous 60 minutes. We counted each identifiable individual within a video as a separate UCE and, when multiple videos captured individuals within the same 60-minute window, we used the video with the highest number of identifiable individuals. Each row in this csv file represents one UCE.
Variables
- Video_ID: Unique video ID assigned to a given video (i.e., image)
- Season: Season of camera deployment
- Camera: Unique Camera ID assigned
- GPS_location: Location of camera deployment
- Area: The zone where the camera was deployed (Dense Forest, Grassland, or Sparse Forest)
- Date (MM/DD/YYYY): The date the video occurred
- Time (military): The time the video occurred
- Temperature (degree C): The temperature during the video
- Class: The class organisms in the video
- Species: The common name of the species ID'ed in the video
- Scientific_name: The scientific name of the species in the video
- Species_Code: The first two letters of the genus and the first two letters of the specific epithet, entered as a four-letter code in all caps
- Coder_name: The name of the author to enter the data (although several rows were cross-validated between multiple people)
- Notes: Any misc. notes about the row
File: Trapping_Data.csv
Description: We derived this data from the "Full_Camera_Dataset_(Cleaned).csv" to create our rarefaction curves. The first row represents the total number of UCEs for that zone. Each row represents the number of UCEs for each species identified, in descending order. The cells among each of these rows do not represent a single species.
Variables
- Each of the columns represents one of the zones
File: Analysis.R
Description: The R code we wrote the our data analysis.
Code/software
We used R programming software version 4.4.2 to conduct our analyses. For a detailed account of the data analysis, including packages used, please refer to our published article.