On 25th of February 2026 , as part of regular operational software upgrades, the software handling area extraction for regular lat-lon gridded datasets will be updated to a new version to correct a behaviour that caused an interpolation of the data during the area extraction.
What is the reason for this change?
Until now, area extraction has triggered an interpolation onto a new grid whose south west corner is that given in the supplied area. This interpolation has been identified as undesirable behaviour. In order to preserve the original values it has been decided the software should instead only crop the existing data, returning just those grid points that lie inside the area without modification.
What will be the effect on the data?
Requests that specify a geographical area may now return data on a slightly different grid which will be displaced from the original grid by up to one grid length, and the number of grid points returned may also differ.
Please refer to our Knowledge Base article on Software upgrade for geographical area extraction from data on regular lat-lon grids, which graphically explains the change in behaviour and also provides a workflow to help users identify whether or not this change is affecting their requests.
Which datasets are affected?
The change described above will apply exclusively to data requested for a specific area, and only in cases where the latitude and longitude of the user-provided bounding box do not exist in the original global dataset before area extraction, regardless of how the data was downloaded (i.e. either using the web interface or the API service), and for the following ERA5 and seasonal forecast datasets ONLY:
ERA5 datasets:
- ERA5 hourly data on single levels from 1940 to present (reanalysis-era5-single-levels)
- ERA5 monthly averaged data on single levels from 1940 to present (reanalysis-era5-single-levels-monthly-means)
- ERA5 hourly data on pressure levels from 1940 to present (reanalysis-era5-pressure-levels)
- ERA5 post-processed daily statistics on single levels from 1940 to present (derived-era5-single-levels-daily-statistics)
- ERA5 monthly averaged data on pressure levels from 1940 to present (reanalysis-era5-pressure-levels-monthly-means)
- ERA5 post-processed daily statistics on pressure levels from 1940 to present (derived-era5-pressure-levels-daily-statistics)
- ERA5-Land hourly data from 1950 to present (reanalysis-era5-land)
- ERA5-Land monthly averaged data from 1950 to present (reanalysis-era5-land-monthly-means)
- ERA5-Land post-processed daily statistics from 1950 to present (derived-era5-land-daily-statistics)
- Complete ERA5 global atmospheric reanalysis (reanalysis-era5-complete) (including reanalysis-era5.1-complete)
Seasonal Forecast datasets:
- Seasonal forecast daily and subdaily data on single levels (seasonal-original-single-levels)
- Seasonal forecast subdaily data on pressure levels (seasonal-original-pressure-levels)
- Seasonal forecast monthly statistics on single levels (seasonal-monthly-single-levels)
- Seasonal forecast monthly statistics on pressure levels (seasonal-monthly-pressure-levels)
- Seasonal forecast anomalies on pressure levels (seasonal-postprocessed-pressure-levels)
- Seasonal forecast anomalies on single levels (seasonal-postprocessed-single-levels)
What should users do if they have downloaded sub-area data (as described above) from the listed datasets?
If your requests are affected by this software upgrade, then the recommendation is to download a new data sample (from date of implementation on 25 February 2026) and compare it with the data previously downloaded. This will allow you to decide on whether or not any modifications are needed to your workflows following the software upgrade. Please refer to our Knowledge Base article for additional information.
For any enquiries on this topic, please contact us (ECMWF login required).
ECMWF Support
on behalf of the Data Stores Service