ROMS<kshedstrom version> blowing-up with ' REASON: KEchar = NaN, PEchar = NaN' report

Report or discuss software problems and other woes

Moderators: arango, robertson

Post Reply
Message
Author
yangwu1215
Posts: 10
Joined: Mon Jul 01, 2019 1:22 pm
Location: Nanjing University

ROMS<kshedstrom version> blowing-up with ' REASON: KEchar = NaN, PEchar = NaN' report

#1 Unread post by yangwu1215 »

I'm recently studying the kshedstrom version <https://github.com/kshedstrom/roms/tree/master>, and trying to set the vertical s-coordinate manually.
However, if I set the s-coordinate with an intersified surface layer, the model blow up with the 'REASON: KEchar = NaN, PEchar = NaN' report. The following is the s-coordinate, error report , and running log.
error-report1.png
error-report.png
While, if I set the s-coordinate following the 'vtransform=2, vstretching=4' option, the model works successfully. The following is the s-coordinate, which has a uniform vertical spacing of 0.03.
error-report3.png
Hope to get some help with this problem. Great thanks!
Attachments
run.log
(55.15 KiB) Downloaded 307 times

User avatar
kate
Posts: 4091
Joined: Wed Jul 02, 2003 5:29 pm
Location: CFOS/UAF, USA

Re: ROMS<kshedstrom version> blowing-up with ' REASON: KEchar = NaN, PEchar = NaN' report

#2 Unread post by kate »

That's a very intensified surface layer. Did you try a shorter timestep? There are numerical limitations on the vertical operations which can cause blow-ups like that.

yangwu1215
Posts: 10
Joined: Mon Jul 01, 2019 1:22 pm
Location: Nanjing University

Re: ROMS<kshedstrom version> blowing-up with ' REASON: KEchar = NaN, PEchar = NaN' report

#3 Unread post by yangwu1215 »

Thank for for the suggestion. The model works if I reduce the dt from 300s to 10s.

Post Reply