Open Pakillo opened 3 years ago
If the distinction between datasets is clear and the output is not a mix including different data sources and downscaling methods, it can be a good idea to implement. I think it can be tricky to mix datasets with different downscaling methods and source data unless it is really clear for the user.
By the way, CHELSA has also this product (daily precipitation, 1 km, 2000-2021): https://chelsa-climate.org/daily-precipitation/
The function would produce data from a single dataset, which you can choose using a 'source' argument, for example: get_daily_climate(coords, "tmin", period = "2010-01-01", source = "CHELSA")
This function would be a wrapper for different functions calling each source. Hence adding other datasets would require some refactoring of the code, but wouldn't be too difficult. And having data like CHELSA easily available would be extremely useful!
Being like this, I think it is a good idea.
We could add other climate data sources, particularly if they implement cloud-optimised geotiffs...
One good candidate are climatic layers from OpenLandMap: https://gitlab.com/openlandmap/global-layers. They provide global 1-km resolution daily temperature and precipitation data up to 2020 (which would be nice to complement the current limit of 2017 from the Austrian data set). Further details: https://twitter.com/opengeohub/status/1415401763465793538?s=20 & https://twitter.com/opengeohub/status/1417129100322512898?s=20
The nice thing is that we wouldn't need to change much of the current package. Mostly the destination address, and little else