Questions on extreme values of DEM data

[hidethethe]

Hi, thanks for sharing the dataset and code. I noticed that the DEM data contains some extreme outlier values that don’t seem physically reasonable (e.g., -32768).
I observed this when opening the files in QGIS, specifically in the France (324), Honduras (321), and Pakistan (1111009) subsets. Could you please confirm whether this is expected or if there might be an issue with the DEM files?