search

dklinges9 / mcera5

mcera5
12 stars 9 forks source link

Change to monthly ERA5 requests and added ERA5-Land support #29

Closed Paul-Haimerl closed 2 months ago

Paul-Haimerl commented 3 months ago

Recently the Climate Data Store (CDS) introduced a more restrictive cap on the size of data requests. As a consequence, all ERA5-data requests submitted through build_era5_request and request_era5 get rejected.

This commit changes build_era5_request as to requests data packages at monthly intervals, thus remaining below the new CDF file size limit. The output file now follows the pattern <outfile_name>_<year>_<month>.nc as opposed to <outfile_name>_<year>.nc.

Furthermore there is an efficiency improvement for the internal function uni_dates.

dklinges9 commented 3 months ago

Thanks for the notice and contribution Paul! I'll review your code shortly – for now just acknowledging that this has been received.

Paul-Haimerl commented 2 months ago

In addition, two new functions, build_era5_land_request and extract_land_clim, modelled after build_era5_request extract_clim, enable easy download of ERA5-Land data.

Since the catalogue of climate variables is not congruent between ERA5 and ERA5-Land, extract_land_clim returns fewer data points than extract_clim. Specifically, emissivity, cloud cover, net longwave radiation, upward longwave radiation, downward longwave radiation, direct normal irradiance, and diffuse normal irradiance are not included. Furthermore, diurnal temperature correction is not possible due to the lack of a land-sea-mask for ERA5-Land.

Going forward, it may be possible to employ ERA5-reanalysis data to complement ERA5-Land and impute e.g. a diurnal correction. Furthermore, there are probably additional useful climate variables that are part of the ERA5-Land dataset and currently not taken into account byextract_land_clim (that is where my lack of ecological understanding comes in I am afraid).

Hope the commits are helpful as e.g. a draft for further updates.