Boreal Wetland Probability Data
The Boreal Wetland probability data set is a raster based product which describes the probability of wetland habitat at a 10 m resolution across Alberta's Boreal Forest Natural Region.
The Boreal Wetland probability dataset was generated using a machine learning framework in R statistical software. The machine learning model was based off of a Digital Elevation Model (DEM), optical satellite data (Sentinel-2), and Synthetic Aperture Radar data (SAR; Sentinel-1). The model was trained with ABMI 3x7 land cover photo-plots which are derived from high resolution 3D image interpretation and give detailed attribution of land cover information.
Overall, the resulting product correctly identifies wetland and upland areas with an accuracy of 83%. This data set is based on fully open source data and processing software. An alternative data set was produced with proprietary LiDAR data but this product is not yet shareable. For more detail on the methodology please refer to the Boreal Wetland probability – technical documentation (available for download below).
Hird, J., DeLancey, E.R., McDermid, G.J., and Kariyeva, J. 2017. “Google Earth Engine, Open-Access Satellite Data, and Machine Learning in Support of Large-Area Probabilistic Wetland Mapping.” Remote Sensing, Vol. 9 (No.12): pp. 1315.
- Species & Habitat Raw Data
- Human Footprint Products
- ABMI GitHub
- Advanced Landcover Prediction & Habitat Assessment (ALPHA) Products
- ABMI Wetland Inventory
- Photoplot Land Cover Data - Training & Validation
- Boreal Surface Water Inventory
- Boreal Wetland Probability Data
- Hydro Temporal Variability
- Other Geospatial Land Surface Data
- GIS Data: Biodiversity
- Species-level Data Sets
- Remote Camera Mammal Data
- Rare Species Data
- Data Archive