Cloud-native geospatial Sentinel-2 grid

An update of scottyhq/mgrs to convert the Sentinel-2 grid from KML to the cloud-native file format GeoParquet. Two versions are available in this repository:

• sentinel-2-grid.parquet - the full grid
• sentinel-2-grid_LAND.parquet - only tiles that intersect with land areas based on the high resolution land mask by Natural Earth

Access examples

The GeoParquet files can be accessed directly from this repository without having to download the files. Here are some examples to get you started:

GeoPandas

import geopandas as gpd

gdf = gpd.read_file("https://github.com/maawoo/sentinel-2-grid-geoparquet/raw/main/sentinel-2-grid.parquet", engine="pyogrio")

GDAL / OGR

Extract individual tile and save as GeoJSON

ogr2ogr -f "GeoJSON" 32TNT.geojson /vsicurl/https://github.com/maawoo/sentinel-2-grid-geoparquet/raw/main/sentinel-2-grid.parquet -sql "SELECT * FROM \"sentinel-2-grid\" WHERE tile = '32TNT'"

Print file info

ogrinfo -al -so /vsicurl/https://github.com/maawoo/sentinel-2-grid-geoparquet/raw/main/sentinel-2-grid.parquet

This should print the metadata of the file:

INFO: Open of `/vsicurl/https://github.com/maawoo/sentinel-2-grid-geoparquet/raw/main/sentinel-2-grid.parquet'
      using driver `Parquet' successful.

Layer name: sentinel-2-grid
Geometry: Polygon
Feature Count: 56686
Extent: (-180.000000, -83.835951) - (180.000000, 84.644279)
Layer SRS WKT:
GEOGCRS["WGS 84",
    ENSEMBLE["World Geodetic System 1984 ensemble",
        MEMBER["World Geodetic System 1984 (Transit)"],
        MEMBER["World Geodetic System 1984 (G730)"],
        MEMBER["World Geodetic System 1984 (G873)"],
        MEMBER["World Geodetic System 1984 (G1150)"],
        MEMBER["World Geodetic System 1984 (G1674)"],
        MEMBER["World Geodetic System 1984 (G1762)"],
        MEMBER["World Geodetic System 1984 (G2139)"],
        ELLIPSOID["WGS 84",6378137,298.257223563,
            LENGTHUNIT["metre",1]],
        ENSEMBLEACCURACY[2.0]],
    PRIMEM["Greenwich",0,
        ANGLEUNIT["degree",0.0174532925199433]],
    CS[ellipsoidal,2],
        AXIS["geodetic latitude (Lat)",north,
            ORDER[1],
            ANGLEUNIT["degree",0.0174532925199433]],
        AXIS["geodetic longitude (Lon)",east,
            ORDER[2],
            ANGLEUNIT["degree",0.0174532925199433]],
    USAGE[
        SCOPE["Horizontal component of 3D system."],
        AREA["World."],
        BBOX[-90,-180,90,180]],
    ID["EPSG",4326]]
Data axis to CRS axis mapping: 2,1
Geometry Column = geometry
tile: String (0.0)
epsg: Integer64 (0.0)
utm_wkt: String (0.0)
utm_bounds: String (0.0)

Is the Sentinel-2 grid a good solution for my use case?

I will not be able to answer this of course, and it depends on your specific use case. However, I wanted to leave this publication by Bauer-Marschallinger & Falkner (2023) here to advocate for alternative grid systems such as the Equi7 grid. With more and more Earth Observation data being available, we as a community need to move forward and innovate in the way we access and process this huge volume of data.