[MITgcm-devel] heat fluxes

[Martin Losch]

Hi,
I found a problem in the forcing data that may concern the suite of 
experiments that are based on global_ocean.90x40x15, but I am not sure:

The ncep forcing fields in global_ocean.90x40x15 are derived from NCEP 
data that I (or someone else) downloaded from the LDEO climate data 
server in 2001 or 2002 (can't remember). I can no longer find them 
there, but that's a different story. The download data is in  
fjord.mit.edu:/data3/mlosch/ncep_data
as netcdf files. There's also a matlab-script that reads the data and 
create flux files such as ncep_qnet.bin for a global_ocean.90x40x15 
configuration (or whatever configuration you prescribe in the structure 
grd). I defined qnet as the sum of solar radiation, long wave flux, 
latent heat flux and sensible heat flux.

When you integrate the temporal mean of qnet zonally and cumulatively 
integrate it meridionally you should get the implied ocean heat flux as 
a function of latitude, that you than can compare to Trenberth+Caron 
2001, J.Clim, for example. Unfortunately this curve hardly gives any 
poleward heat flux in the southern hemisphere (other than 
Trenberth+Caron etc. find), so I assume, something went wrong there. I 
repeated the computations with the "raw" data from the netcdf files in 
/data3/mlosch/ncep_data, but the result is the same: hardly any 
poleward heat flux in the Southern hemisphere. (I am attaching a figure 
for the computation on the 90x40x15 grid to give you an idea what I 
mean, never mind the additional lines on that plot). Where do I go 
wrong?

Martin
-------------- next part --------------
A non-text attachment was scrubbed...
Name: hflux.ps.gz
Type: application/x-gzip
Size: 3767 bytes
Desc: not available
URL: <http://mitgcm.org/pipermail/mitgcm-devel/attachments/20050601/1747097d/attachment.gz>
-------------- next part --------------