Even just a linear interpolation may be very useful to avoid boundary effects. Apparently there are almost no changes in elevation, so simplistic linear interpolation may be more viable as a result.
Does this apply to all ERA5 variables or just e.g. temperature?
In torchGeo we think there is little/no choice of interpolation method? Check this. We could plot some samples to see the type of interpolation going on
Even just a linear interpolation may be very useful to avoid boundary effects. Apparently there are almost no changes in elevation, so simplistic linear interpolation may be more viable as a result.
Does this apply to all ERA5 variables or just e.g. temperature?
In torchGeo we think there is little/no choice of interpolation method? Check this. We could plot some samples to see the type of interpolation going on
Possible paper for help: https://iopscience.iop.org/article/10.1088/1748-9326/ac48b3