When ncrcat is applied to packed files (packed as short with add_offset and scale_factor), the add_offset and scale_factor are blindly copied from the first file. The resulting file can thus contain invalid data, as the scale factors may (and many times, do) differ between files.
As I see it, there are two possible fixes:
1) Implement a warning or better, an error, telling the person to use ncpdq first, or, better:
2) Modify the scale and add factors so that the tool produces valid output.
Thank you for your attention to this.
Log in to post a comment.