South-central British Columbia cattle habitat use intensity data from GPS eartag tracking across three pastures during two growing seasons (April-September 2024 and 2025). The dataset was created by Ellen Wu for modeling with ten environmental predictors, including elevation, vegetation indices, and distance to water. It supports a dual-model framework comparing Random Forest and Generalized Additive Model performance.
Use Cases
- Modeling cattle distribution patterns based on GPS fix counts per grid cell.
- Comparing machine learning and statistical habitat models based on environmental predictors like NDVI and topographic position index.
- Identifying high-use versus low-use grazing areas based on relative habitat rankings.
- Assessing spatial and interannual transferability of habitat selection models.
Strengths
- Data covers two consecutive growing seasons (2024 and 2025), allowing for interannual validation.
- Analysis employed ten environmental predictors, including monthly median NDVI and EVI.
- Models were validated using spatial cross-validation and independent holdout validation.
Limitations
- Row count and file formats are unknown, which may limit suitability assessment.
- Column-level documentation is absent; field semantics must be inferred after download.
- Models exhibited limited spatial extrapolation across pastures and reduced interannual transferability.
Provenance
- Source
- Borealis Harvested Dataverse
- Collection Method
- GPS eartag tracking data collected from cattle across three fenced pastures.
- Time Range
- April-September 2024 and April-September 2025
- Freshness
- Last updated 2026-05-02 04:10:09; freshness should be verified.
- Geography
- South-central British Columbia, Canada