4 Boundary Datasets
4.1 Administrative Boundaries
The Adaptation Atlas uses GAUL 2024 from FAO as the source for all administrative boundaries. GAUL provides a harmonized, high-resolution global dataset of administrative boundaries. Using GAUL ensures a consistent framework across countries, simplifies data aggregation and analysis, and allows datasets to be easily downloaded and reproduced.
4.1.1 Raw GAUL Boundaries
The raw GAUL boundaries represent the original FAO administrative boundaries with minor adjustments to support analysis. GAUL does not provide country-level (admin 0) boundaries, so both admin 0 and admin 1 boundaries in the Atlas were created by aggregating the admin 2 boundaries. Before aggregation, major water bodies included in GAUL were removed from the admin 2 boundaries. Admin 1 boundaries were quality-checked against the official GAUL admin 1 boundaries to ensure that the aggregation process preserved accuracy. The resulting “raw” boundaries therefore include:
Original FAO geometries (admin 2) without major water bodies.
Aggregated admin 1 and admin 0 boundaries derived from the GAUL 2 geometries.
4.1.2 Analysis Ready Boundaries
Analysis-ready boundaries extend the raw boundaries to better support data aggregation and reporting. These boundaries:
Identify disputed regions within the GAUL boundaries and assign them to all countries that claim them, ensuring that aggregated data matches national representations.
Include additional attributes to facilitate analysis and linking with other datasets.
Preserve the same geometries as the raw boundaries, except where overlapping or unioned boundaries are required to account for disputes.
This approach ensures that the Atlas provides an accurate representation of FAO’s GAUL data while providing countries with data that is representative of how they represent their boundaries.
4.1.3 Rasterized Boundaries
In addition to vector boundaries, the Atlas provides rasterized versions of administrative boundaries for use in zonal extractions. These rasters are available at all three administrative levels (0, 1, and 2) and use the numeric GAUL codes as the raster values.
It is important to note that these raster datasets are based on the raw GAUL boundaries described above and do not account for disputed areas. Overlaps in boundaries cannot be represented in a raster, so any treatment of disputed regions must be performed after the zonal extraction. Depending on the type of data, this can be done using simple summation or a weighted mean. To facilitate this, the Atlas provides a reference table listing each disputed area, the countries claiming it, and the area of each region.
4.1.4 Simplified Boundaries
To ensure efficient rendering while retaining topological accuracy, we applied multi-resolution boundary simplifications using mapshaper. These simplified boundaries are based on the analysis ready boundaries and should be used for visualizations as their file size is significantly smaller than the analysis ready boundaries. The script used for simplification can be found here
4.1.4.1 Boundary Simplification Parameters
Simplification method: Visvalingam / weighted area
Shape preservation: Prevent shape removal = true
| Boundary Level | High Res (%) | Low Res (%) | Very Low Res (%) | Very Very Low Res (%) |
|---|---|---|---|---|
| Admin 2 | 20 | 3 | 0.7 | – |
| Admin 1 | 14 | 4.7 | 1.1 | 0.25 |
| Admin 0 | 13 | 1.4 | 0.5 | 0.18 |
4.2 Watershed Boundaries
The Adaptation Atlas uses watershed boundaries from the WMO Basins and Sub-Basins (WMOBB) dataset, derived from the HydroSHEDS database. This dataset includes one level of watersheds and is provided in two versions: analysis-ready and visualization. The dataset incorporates data from HydroSHEDS and other sources under license.
4.3 Boundary Data Index & Paths
All of the boundary datasets are stored in the Adaptation Atlas S3 bucket and are cataloged in our STAC. You can also use the below table to filter and discover the data mentioned above.