Q: About the model coordinate and elevation that should be contained in the score files, in our case, the observation equivalent is a weighted average of 3 surrounding grid points. I could use the coordinates of the grid point with the highest weight. What is your approach?
A: Model coordinate and elevation: ideally the nearest grid point should be used, without any interpolation. If that is not possible for technical reasons, you could put the weighted average location and elevation, using the same weights as for the actual forecast value.
Q: Is there a list of stations to use, as in case of radiosondes?
A:There is no list of SYNOP stations to use, every centre is free to choose their own set. This is the advantage of exchanging errors for individual stations. In practice, the more the better, so that the intersection between centres (for intercomparison of areal averages of scores) becomes as large as possible.
Q: How should the SEEPS scores be exchanged?
A: SEEPS is not part of the planned exchange of surface scores. The SEEPS computation requires a station climatology, which was considered too complicated for a basic exchange of surface scores.