The exclusion of T_GDS10_HYBY in the Grib setup without replacement leads to a silent crash of PrISM, and by extension OASIS3. As a stand-in for actually original data, in places where the temperature above the surface layer is required, model1 formerly used to interpolate it from the surface layer temperature, and the difference in average height between the desired layer and the surface layer with a simple dry adiabatic lapse rate of –0.0098 K/m (Bott 2016).

Because the Grib file setup covers a larger area, the permanent parameter test files did not suffice. There is a substitute though, confront the page ‘Coarse external data’. The 1st substitute lacks the geometrical layer heights. To avoid a crash similar to the previous paragraph, model1 would again produce its own dummy data. This time, it would add the average height of the layer in the test setup to HSURF. Consequently, the dry adiabatic lapse rate scheme just constituted the addition of a constant.