Data from: Determinants of seasonal elephant habitat use in small and enclosed conservation area: Mwea national reserve, Kenya
Data files
Jul 20, 2026 version files 38.51 MB
-
BoundaryDistance_FN_tif.ASC.xml
635 B
-
BoundaryDistance_FN_tif.prj
161 B
-
DrySeasonPrediction.dbf
58.54 KB
-
DrySeasonPrediction.prj
408 B
-
DrySeasonPrediction.sbn
18.67 KB
-
DrySeasonPrediction.sbx
1.07 KB
-
DrySeasonPrediction.shp
54.64 KB
-
DrySeasonPrediction.shp.xml
15.91 KB
-
DrySeasonPrediction.shx
15.68 KB
-
DryseasonPrediction.tfw
90 B
-
DryseasonPrediction.tif
7.35 MB
-
DryseasonPrediction.tif.aux.xml
2.16 KB
-
DryseasonPrediction.tif.xml
4.32 KB
-
Elephant_habitat_use_locations__wet_season.xlsx
45.49 KB
-
Elephant_habitat_use_locations_dry_season.xlsx
71.52 KB
-
elevation_mwea_tif.asc
4.43 MB
-
Elevation_Mwea_tif.ASC.xml
625 B
-
Elevation_Mwea_tif.prj
161 B
-
invasivesanthrop_tif.asc
6.19 MB
-
invasivesanthrop_tif.asc.xml
620 B
-
InvasivesAnthrop_tif.prj
161 B
-
mainriverdist.asc
6.18 MB
-
MainRiverDist.ASC.xml
584 B
-
MainRiverDist.prj
161 B
-
ndvimean_mwea_tif.asc
6.64 MB
-
NDVIMean_Mwea_tif.ASC.xml
623 B
-
NDVIMean_Mwea_tif.prj
161 B
-
README.md
10.03 KB
-
WetSeasonPrediction.cpg
5 B
-
WetSeasonPrediction.dbf
19.90 KB
-
WetSeasonPrediction.prj
408 B
-
WetSeasonPrediction.sbn
6.80 KB
-
WetSeasonPrediction.sbx
516 B
-
WetSeasonPrediction.shp
18.58 KB
-
WetSeasonPrediction.shp.xml
15.92 KB
-
WetSeasonPrediction.shx
5.38 KB
-
WetseasonPrediction.tfw
90 B
-
WetseasonPrediction.tif
7.35 MB
-
WetseasonPrediction.tif.aux.xml
2.13 KB
-
WetseasonPrediction.tif.xml
4.32 KB
Abstract
This data is an exploration of factors that influence seasonal habitat-use dynamics of African elephants (Loxodonta africana). In particular, water availability, vegetation productivity, elevation, and invasive species effect on seasonal spatial distributions of elephants in Mwea National Reserve, a small, fenced, and isolated conservation area in Kenya. The data provide novel evidence that elephants preferentially use invasive plant-dominated habitats during resource-scarce dry seasons, especially in areas where they have colonized dry season grazing areas. This underscores the adaptive flexibility of elephants in fragmented landscapes and offers new insights into how biotic pressures interact with resource limitation to determine habitat-use patterns.
Dataset DOI: https://doi.org/10.5061/dryad.x0k6djhz0
Authors
Wambua, J.K., Wato, Y.A., Abdallah, S.A., Mworia, J.K., Lukhoba, C., & Gichuki, N.N.
Summary of Dataset Contents
The dataset contains spatial data of African savanna elephant (Loxodonta africana) habitat use collected during a systematic field surveys in Mwea National Reserve, Kenya. Elephant occurrence was represented by georeferenced dung locations collected during the dry and wet seasons. These occurrence records were analysed together with environmental and landscape predictor variables to model seasonal habitat suitability using the Maximum Entropy (MaxEnt) algorithm.
Components of the dataset:
- Dry season elephant dung occurrence records (Microsoft Excel)
- Wet season elephant dung occurrence records (Microsoft Excel)
- Environmental and landscape predictor variables (ESRI ASCII raster format with associated projection and metadata files)
Description of the Data and File Structure
The dataset consists of two Microsoft Excel spreadsheets containing elephant occurrence records and a collection of ESRI ASCII raster datasets representing environmental and landscape predictor variables used in habitat suitability modelling.
The elephant occurrence records contain georeferenced GPS locations of elephant dung recorded during systematic field surveys and represent indirect evidence of seasonal elephant habitat use. These occurrence points served as the response (dependent) variables in Maximum Entropy (MaxEnt) modelling.
Environmental predictor variables were derived from field surveys, remotely sensed imagery, Digital Elevation Models (DEMs), and GIS analyses. All predictor layers were projected to the WGS 84 / UTM Zone 37 South coordinate system, clipped to the Mwea National Reserve boundary, aligned to a common spatial extent and cell size, and exported to ESRI ASCII (.asc) format for compatibility with MaxEnt.
Associated projection (.prj) and metadata (.xml) files accompany several raster datasets to preserve coordinate reference system information and processing history.
Study Area
The study was conducted in Mwea National Reserve, a fenced and relatively small protected area located in central Kenya. The reserve comprises a tropical savanna ecosystem supporting a resident population of African savanna elephants and other herbivores.
Project Description
The dataset was developed to investigate the environmental and landscape determinants influencing seasonal elephant habitat use within a small enclosed conservation landscape.
The data support analyses of:
- Seasonal elephant habitat use
- Spatial distribution patterns
- Habitat suitability modelling
- Habitat preference
- Ecological and anthropogenic drivers of habitat selection
- Conservation planning and protected area management
Dataset Description:
File 1: Elephant_habitat_use_locations_dry_season.xlsx
Contains georeferenced elephant dung occurrence records collected during the dry season.
File 2: Elephant_habitat_use_locations__wet_season.xlsx
Contains georeferenced elephant dung occurrence records collected during the wet season.
File 3: Environmental Variables Dataset
The environmental dataset contains raster predictor variables representing ecological and landscape factors used in Maximum Entropy (MaxEnt) habitat suitability modelling. Raster datasets are supplied as ESRI ASCII grids together with associated projection and metadata files where applicable.
Environmental Dataset File Inventory
File: boundarydistance_fn_tif.asc
Description: ESRI ASCII raster representing Euclidean distance from the reserve boundary (metres). Used to quantify the influence of enclosure and edge effects on elephant habitat use.
File: BoundaryDistance_FN_tif.prj
Description: Projection file containing the coordinate reference system for the boundary distance raster.
File: BoundaryDistance_FN_tif.ASC.xml
Description: ArcGIS metadata documenting raster processing history and conversion to ASCII format.
File: elevation_mwea_tif.asc
Description: Digital Elevation Model (DEM) representing terrain elevation across Mwea National Reserve.
File: Elevation_Mwea_tif.prj
Description: Projection information associated with the elevation raster.
File: Elevation_Mwea_tif.ASC.xml
Description: ArcGIS metadata documenting raster processing history.
File: invasivesanthrop_tif.asc
Description: Raster representing the spatial distribution of invasive plant species used as an environmental predictor in habitat suitability modelling.
Additional environmental raster layers (e.g., NDVI, distance to roads, distance to water sources) follow the same file structure and are described below.
Elephant Occurrence Dataset:
The occurrence data consist of GPS locations of elephant dung recorded during systematic field surveys.
Variables
- X: Easting coordinate (UTM)
- Y: Northing coordinate (UTM)
- Type: Observation type (Elephant dung)
- Species: African savanna elephant (Loxodonta africana)
- Age Class: Estimated dung decomposition class
Dung Age Classification
Only dung classes A, B, C1 and C2 were used in habitat suitability modelling because they represent recent elephant presence.
- A: Very fresh dung
- B: Fresh dung
- C1: Moderately decomposed dung
- C2: Advanced decomposition
- D: Old/highly decomposed dung
Environmental and Landscape Variables:
Environmental predictor variables were selected based on their ecological relevance to elephant habitat selection and space use.
Variables Include:
- Elevation: Digital Elevation Model (DEM) representing terrain elevation (metres above sea level).
- Purpose: Influences topography, drainage, vegetation composition and elephant movement.
- Distance to Permanent Water Sources: Euclidean distance (metres) from permanent rivers and water sources.
- Purpose: Represents accessibility to water, particularly during the dry season.
- Distance to Roads: Euclidean distance (metres) from roads within the reserve.
- Purpose: Represents potential anthropogenic disturbance and accessibility.
- Distance to Reserve Boundary: Euclidean distance (metres) from the reserve boundary.
- Purpose: Represents enclosure effects and evaluates the influence of reserve boundaries on seasonal habitat use.
Associated files:
File: boundarydistance_fn_tif.asc
File: BoundaryDistance_FN_tif.prj
File: BoundaryDistance_FN_tif.ASC.xml
Vegetation Productivity (NDVI)
- Normalized Difference Vegetation Index (NDVI) derived from Landsat imagery.
- Purpose: Represents vegetation greenness, primary productivity and forage availability.
Ivasive Plant Distribution
- Raster layer representing the spatial distribution of invasive plant species.
- Purpose: Evaluates the influence of invasive vegetation on seasonal elephant habitat selection.
Associated file: invasivesanthrop_tif.asc
Environmental Data Processing
Environmental variables were prepared using ArcGIS Pro (ESRI).
Processing included:
- Projection to WGS 84 / UTM Zone 37 South
- Conversion of vector datasets to raster where required
- Clipping to the Mwea National Reserve boundary
- Alignment to a common spatial extent and raster resolution
- Export to ESRI ASCII (.asc) format using the ArcGIS Raster to ASCII conversion tool
- Inclusion of accompanying projection (.prj) and metadata (.xml) files
Coordinate Reference System
The spatial datasets use the following coordinate reference system:
- Parameter
- Description
- Projection
- Universal Transverse Mercator (UTM)
- UTM Zone
- 37 South
- Datum
- WGS 84
- Units
- Metres
- EPSG Code
- 32737
Geographic Extent
The environmental layers cover Mwea National Reserve and the surrounding landscape included in the habitat suitability modelling.
Data Sources
Environmental predictor variables were derived from:
- Field surveys
- Landsat satellite imagery
- Digital Elevation Models (DEMs)
- Hydrological datasets
- Road network GIS layers
- Reserve boundary shapefiles
- Existing land-cover datasets
Recommended Software
The dataset can be viewed and analysed using:
- MaxEnt
- ArcGIS Pro
- QGIS
- R
- Python GIS libraries (Rasterio, GDAL, GeoPandas)
File Formats
- .xlsx : Elephant occurrence data
- .asc : Environmental raster layers
- .prj: Coordinate reference system information
- .xml: ArcGIS metadata and processing history
- .tif: Intermediate raster datasets (where applicable)
- .shp: Shapefile geometry
- .dbf: Attribute table
- .shx: Shape index
Intended Uses
These datasets are suitable for:
- Species distribution modelling
- Habitat suitability modelling
- Spatial ecology
- Landscape ecology
- Conservation planning
- Protected area management
- Ecological modelling
- GIS analyses
- Reproducible research
Citation
When using these data, we invite you you to cite both the dataset and the associated publication.
Dataset
Wambua, J.K., Wato, Y.A., Abdallah, S.A., Mworia, J.K., Lukhoba, C., & Gichuki, N.N. Determinants of Seasonal Elephant Habitat Use in a Small and Enclosed Conservation Area: Mwea National Reserve, Kenya. Dryad Digital Repository. https://doi.org/10.5061/dryad.x0k6djhz0
Associated Publication
Wambua, J.K., Wato, Y.A., Abdallah, S.A., Mworia, J.K., Lukhoba, C., & Gichuki, N.N. Determinants of Seasonal Elephant Habitat Use in a Small and Enclosed Conservation Area: Mwea National Reserve, Kenya.
Data Collection
Elephant habitat use was assessed using indirect evidence of elephant presence. GPS coordinates of elephant dung were recorded during systematic field surveys conducted in both the dry and wet seasons. Each dung observation was assigned a decomposition age class based on field assessment.
Environmental predictor variables were developed from field surveys, remotely sensed imagery, Digital Elevation Models (DEMs), and GIS-derived distance surfaces. All predictor layers were projected into the Universal Transverse Mercator (UTM) coordinate system, clipped to the reserve boundary, resampled to a common spatial resolution, and converted to ASCII raster format for compatibility with MaxEnt.