[gradsusr] Streamfunction (fish_psi)

[Hart, Robert]

Hi James,


When the GFS went to 0.25 degree output last year, the fishpak routines became unreliable for the GFS:


http://opengrads.org/doc/udxt/fish/fish.html


"Notice that fish() becomes numerically unstable for horizontal resolutions finer that 1/2 degrees or so. In such cases use the spherical harmonic based sh_fish given in extension shfilt."


Indeed, I have found sh_fish is a very reliable replacement at these resolutions.


Bob Hart


________________________________
From: gradsusr-bounces at gradsusr.org <gradsusr-bounces at gradsusr.org> on behalf of James Spinks <jspin881 at hotmail.com>
Sent: Wednesday, March 23, 2016 4:22 PM
To: gradsusr at gradsusr.org
Subject: [gradsusr] Streamfunction (fish_psi)

Hello,

I'm having some trouble using the fish_psi function to plot streamfunction for the GFS data.

'open gfs.ctl'
'd fish_psi(ugrdprs,vgrdprs)'

The field is returning constant value 0 or -inf.

Interestingly enough, I use to be able to plot streamfunction using fish_psi with no problem for the GFS. So this 'bug' has occurred over the last few months or so.

Anyone have any ideas what may be causing the problem with the GFS data? I have use the function fish_psi just fine using the ECMWF-interim data.

I'm using grads-2.1.a2.oga.1

Thanks,
James Spinks
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://gradsusr.org/pipermail/gradsusr/attachments/20160324/5048b14d/attachment.html