Exact calculation of u_10 and v_10 wind components on ERA5 single levels

Hi all,

I am trying to find out, how u_10 and v_10 are retrieved on the single levels in ERA5. Comparing to the model levels indicates an overestimation of the velocities at 10m. I assumed that they are interpolated from the model levels, however, i cannot reconstruct the same values from the model levels.

For the CARRA dataset, some information is given regarding the 10 m winds calculation, but I am missing something similar for ERA5.
Copernicus Arctic Regional Reanalysis (CARRA): Data User Guide - Copernicus Knowledge Base - ECMWF Confluence Wiki

The same applies for the 100 m wind components, although they are much closer to the model-level interpolated values.

Has anyone experienced something similar?

1 Like

Hi Paul, in ERA5 the 10m is a diagnostic parameter computed using a roughness length of 3cm. It does not necessarily correspond to the model 10m wind. See Section 3.10.2 in IFS Documentation CY41R2 - Part IV: Physical Processes | ECMWF. All the best from Denmark. Rémi, C2Wind.