Currently, missing values associated with NaN data within datasets are not handled in DetectNodes (Priority) and DetectBlobs. It would be best to automatically look for the missing_values or _Fillvalues (if any) in NetCDF files and ignore/skip those values in the computation.