Julia package used to create DeltaDTM
In src there's the source code of the package. More useful are the files in scripts, which make use of the src files.
They can be called like so:
- julia --project scripts/bias_callto process the bias correction
- julia --project scripts/dtm_callto generate DeltaDTM tiles
- julia --project scripts/make_releaseto generate DeltaDTM zips
To run this as is, one needs at least CopernicusDEM tiles (used both with and without water masked out), the original CopernicusDEM error and mask files. One also need to download ESA-WorldCover 2021 and subtile the dataset into CopernicusDEM tiles. Last but not least, one needs to download all ICESat-2 and GEDI data for their area of interest (can be done with SpaceLiDAR.jl) and save these as GeoParquet (.pq), also tiled to the CopernicusDEM tiles, by saving multiple GeoParquet files (one for each granule) in a folder per tile.
For all these .tif tiles, gdalbuildvrt is run, to include all direct neighbours of such tiles. Each 1 by 1 degree .tif tile thus as a .tif.vrt tile that's 3 by 3 degrees. For the GeoParquet files .pq a similar buffer is made, by making a text (.txt) file containing the paths of all files, including all filepaths in the neighboring tilefolders.
The plots are made in Python (with the exception of the bias plot in src/bias.jl), and can be found in the plots folder.
The validation tables can be generated as .csv by calling scripts/validation.jl, once DeltaDTM has been generated for areas with a validation tile, also tiled to the CopernicusDEM specification.