hc precision with NETCDF_CHECK_VAR

Report or discuss software problems and other woes

Moderators: arango, robertson

Post Reply
Message
Author
yongjinxiao
Posts: 2
Joined: Wed Feb 25, 2009 5:32 pm
Location: Virginia Institute of Marine Science

hc precision with NETCDF_CHECK_VAR

#1 Post by yongjinxiao » Thu Jan 14, 2010 4:18 pm

Hello all,

I had a very strange error message while running biotoy in my own study location:

ETCDF_CHECK_VAR - inconsistent value of variable: hc 4.941826E+01 4.941826E+01
in file: /in/biotoy_ini_108_194.nc

This looks very much a precision issue, and as some previous post mentioned, NETCDF_CHECK_VAR seems too strict.

From what I read from the wiki and threads in this forum, the way ROMS calculate is hc=min(hmin,Tcline) according to
viewtopic.php?f=17&t=1389&p=5150&hilit= ... _VAR#p5150

My initial file was generated with MatLab 7.7, and the I wrote a matlab script to pull off data from a larger domain ROMS output netcdf files.
I had Tcline=50, both in ocean_bio_toy.in and in my initial file.
Bathymetry in initial file: h = 49.418258666992188 m
My MatLab script defined hc=min(h,Tcline), which in this case, equal to h.

I thought both ROMS and MatLab use double precision (DOUBLE_PRECISION is activated in cpp), and they should not have different prevision, should they? I'm using ROMS version 366 by the way.

Any comment or advice would be appreciated.

Yongjin

Post Reply