Hello everyone,

I am interested in ERA5 reanalysis data (hourly on pressure levels) for one specific weather station, i.e. one specific grid cell, over 10 years in time. In theory, this should not result in too much data, since it is only for a single location. However, in the CDS store, the item limit prohibits such a download.

As far as I understand, for ERA5 on pressure levels, 1 item= 1 variable * 1 level * 1timestep. Since the sub-area is not taken into account, the number of items is the same as if I tried to download the whole globe over 10 years and therefore greatly exceeds the CDS limit.

Do you know a way to download a timeseries of data for one specific location?

I think, this should be a pretty common task, e.g. to compare model reanalysis data with measurements from instruments located on a site. Nevertheless, I haven`t found this question here in the forum (maybe I have overseen it).

Thank you all,
Paul

5 Comments

  1. I contacted the ECMWF support in this regard, which provided me with the following answer:

    (...)

    Thank you for your reply.

    We indeed suggest that you use the CDS Toolbox to get the timeseries for your location:

    https://cds.climate.copernicus.eu/toolbox/doc/index.html

    There is a sample script in the Toolbox Gallery which will get you started: https://cds.climate.copernicus.eu/toolbox/doc/gallery/3_extract_time_series_and_plot_graph.html

    I hope this helps.

    Kind regards, 

    ECMWF Support

  2. Update: It seems like the CDS Toolbox applies the same limits to the download as the Web Interface. I am still in contact with the support and we are looking for a solution.

  3. Hi Paul - we do this routinely (although for ERA5 single-level data) to make the data available via time-series API. What is the location (lat/lon) that you're looking for? We are planning to make pressure data available later this year and if it's just a single location, we may be able to just process it for you. Please feel free to contact me via email on my profile.

    Cheers,

    Joseph

  4. Thank you, this sounds very interesting! I could not find an email on your profile here, but I retrieved it from an older post of you here (***oikolab.com) and sent you further information.

    Best,
    Paul

  5. So to everyone who might be interested in this issue in the future: Although I haven`t got a final answer from ECMWF yet, currently it seems impossible to get high resolution data for a single location via the ECMWF tools (online API, python API or toolbox editor). You have to move to single level or monthly averaged data.

    Best,
    Paul