Project

General

Profile

Forums » Support »

EOF analysis

Added by thanu roh 5 months ago

Tried doing EOF analysis for wave height data from ERA5 reanalysis data. this was shown :eigen value computation with one sided jacobi scheme did not converge properly. 10 of 55 pairs of column did not achieve requested orthogonality of 1 e-12. Screenshot of the command used and message displayed is attached.
what does this mean?

Screenshot 2024-07-08 185203.png (69.1 KB) Screenshot 2024-07-08 185203.png

Replies (4)

RE: EOF analysis - Added by Karin Meier-Fleischer 5 months ago

Hi Thanu,

does the file eofdjf.nc have an variable attribute valid_range with (-32768, 32767)? If yes you should delete it.

RE: EOF analysis - Added by thanu roh 5 months ago

Hello Karin. Thanks for the reply. How can i delete that?

RE: EOF analysis - Added by Karin Meier-Fleischer 5 months ago

If you have CDO version 2.4.2 use the new delattribute operator otherwise use NCO's ncatted .

ncatted -a <attribute name>,<variable name>,d,, infile outfile

RE: EOF analysis - Added by thanu roh 5 months ago

tried doing this. still the same error is seen

    (1-4/4)