Decide the resolution and way to generate the .geosjon shapefiles of the regions.

[sebastien-lenard]

The website need regularly updated .geojson shapefiles.

Known specs:

• 1 file per region (=subdivision).
• files stored in subfolders, depending on the id of the region and the CRS code.
• shapefile should be in the projection of the root region.
• original shapefiles contain polygons (1 shapefile per type of region) and are in wgs84 (lat lon). Resolution < 0.001 degree.

Known constraints:

• only the root regions and a few countries (e.g. Argentina) have shapefiles delimited by the area of available source data tiles. (e.g. the shapefile for Argentina doesn't occupy southernmost Argentina/Ushuaia). Shapefiles of huc/hydrosheds will not be delimited taking into account the source data tiles (but could be in 2024).
• we can't have .geojson of the same resolution of the original files, otherwise .geojson will have a big Mb size and be long to load. So we need to lower the resolution, in a ~uniform way over the Earth.

  I found a way to have file size in the same order as current, with resolution 500 m and precision = 0 (instead of 15 decimals currently

- New 2023-10-13: Incomplete/controversial regions: at the country and state level, there are currently some territorial gaps in the .geojson (and the shapefiles in the supercomputer backend), notably in eastern Europe and around China. Although this can be corrected for Europe, it's not planned to display country and state dispute boundaries in High Mountain Asia (so potentially, there will be only hydroshed basins).

Temporary specs. I propose the region shapefiles be of a 463 m resolution (same as the resolution of the geotiffs) and be adapted so that no pixel of the data geotiffs be split between inside and outside the region shapefile.

Links

• Python algorithm for simplifying geometry based on the scale of the feature. Expects a geopandas dataframe and returns a geopandas dataframe: https://github.com/nsidc/snow-today-webapp-server/blob/main/snow_today_webapp_ingest/util/simplify_geometry.py
• Code which calls the above algorithm may include other useful functions for e.g. reading and writing: https://github.com/nsidc/snow-today-webapp-server/blob/main/snow_today_webapp_ingest/make_region_shapes_and_index.py

[sebastien-lenard]

@windnagelnsidc and @MattF-NSIDC I have a problem dealing with projection to web mercator (and probably any other projection) from the modis sinusoidal projection for the western part of Alaska, I let you informed of the solution I'll find.

NB: modis sinusoidal is the (weird) original projection of the sensor data we import in supercomputer backend and the projection in which we store all the data in the backend

Update: problem solved, it only concerned a bit of land of extreme east Russia we won't use anyway