Note: this preliminary release is subject to changes.

A global non-overlapping and exhaustive spatial partitioning or segmentation where each segment corresponds to a single Sentinel 2 tile. Properties:

• The segmentation is in WGS84 but can be reprojected to any other coordinate system. No seam artifacts will be introduced by the reprojection, because neighboring segments share polygon vertices.
• For each point within a segment, the corresponding Sentinel 2 tile has data for at least a roughly 9720 m x 9720 m north-oriented square patch centered at that point. This corresponds to a square patch of at least 6873 m x 6873 m in any orientation (45-degree worst case).
• For each point within a segment, the corresponding Sentinel 2 tile generally has a UTM projection that gives the smallest north alignment error. However, some small segments have been eliminated when they can be wholly merged into into a larger segment with another UTM zone projection, somewhat increasing the north alignment error.

Limitations:

• Near the poles the segmentation does not reach all the way to the north or south edges of the north or south-most Sentinel 2 tiles. This means that the segmentation is not usable for visualisations that need to show all the data near the poles, without modification of the segmentation

Technical notes:

• Segments are split at the antimeridian with id's appended by _0 and _1.

Files:

• create_grid.ipynb -- The Jupyter Notebook that creates the segmentation.
• preferred_tiles.geojson -- The segmentation as a GeoJSON. Available after running create_grid.ipynb. The properties:
  • "id": Id of the segment, for example "01CCV" if there is a one-to-one correspondence between the segment and the Sentinel 2 tile of the same name, or "01CDH_1" if multiple segments point to the same tile "01CDH".
  • "tile_id": The Sentinel 2 tile, for example "01CCV"
  • "epsg": The EPSG number of the Sentinel 2 tile, as a string, for example "32701", corresponding to a UTM zone.
  • "utm_area": The area of the segment, calculated with the vertices projected to the UTM coordinate system, as a number in square meters, for example 514730427.83726776.

Python 3.12 or later.

Pip install:

shapely
tqdm
pyproj
pandas
networkx
geopandas
matplotlib