HR Model: geopotential (z)

[sergioleonluis]

Dear @wesselb,

I have one question about the high-resolution model.

The geopotential (z) is entered in the batch as a matrix(lat,lon). With this in mind, why do we introduce these two individual values of Z in the model? What latitude/longitude are they associated with?, How were these values calculated/obtained?

model = AuroraHighRes(
# Use manually downloaded and regridded static variables.
surf_stats={"z": (-3.270407e03, 6.540335e04)},
)

Thank you very much for solving this question.

[wesselb]

The value for surf_stats["variable"] is of the form (a, b) and will be used to normalise variable in the following way: (variable - a) / b. During fine-tuning, we specifically normalised the surface-level variable geopotential surface z to the range [0, 1], which corresponds to setting a = min(z) and b = max(z) - min(z). This how the values in the example were computed.