Dear Michael, Gottfried,

thanks for the nice work and the provided updates. The reviewer has suggested minor revisions/review (comments copied here below):

- Figure 8 suggests that the bias correction scheme in increasing positive biases over large parts of the globe (EG, Africa, Australia). This needs to be explained.

- I would still argue that the reader would find information on the typical number of predictors used in a radiance bias correction scheme useful. I think this is typically 6-8.

In addition, I had a look at the manuscript and suggest 2 minor additions:

- please mention briefly what the "real" data is, e.g. UCAR phase data processed at Graz (e.g. for figure 3)?
- it was not clear to me if figure 3 and 8 show the same information, just before and after the bias correction (you do this link for other figures). If they do, please indicate this in the text. And, if they do, I was too wondering why the bias does change quite substantially at all latitudes (even high ones). Isn't that a potential issue? It would be nice if some information can be given why ECMWF should be biased at high latitudes.

Kind regards,

Axel

Hi Michael,

great, thanks for the update.

Cheers,

Axel