From: nutter@granny.atmo.arizona.edu (Paul Nutter)
Message-Id: <9406271822.AA02279@granny.atmo.arizona.edu>
Subject: estabv.mods
To: ccm-users@ncar.ucar.edu
Date: Mon, 27 Jun 94 11:22:49 MST
Hi ccm-users,
In a posting on Friday, Uma Bhatt said the following:
>I submitted a short test run that called the mod called /ccm/ccm1/r15/estabv.
>mods. The first time I ran it I just copied estabv.mods into my ocean mods.
>The output for the ocean ocean model for 150 days was exactly the same as
>in my original run (where estabv is not called). I also did a test run where
>I incorporated the estabv.mods through a nupdate call (-i option).
>This test run was also exactly the same as my original run. I am wondering if
>I am incorporating the mods correctly since I would expect some changes in
>latent heat flux (since this does need saturation specific humidity at Ts)
>and therefore changes in ocean temperature.
I have also been attempting to incorporate /ccm/ccm1/t42/estabv.mods
into the ccm1 code. To do this, I have used the nupdate call, -i
option. The results are dissapointing. In fact, the results of a
12-day run of the model are exactly identical to that of a forecast
over the same period without the estabv.mods changes. I considered
500mb heights, 850mb temps, prect, psl, and 300mb wind components.
I am surprised there are no changes, especially in the total acc. precip
(prect).
I have done some quick rms error caculations for the above fields over
the northern hemishere to get a feel for the error growth over the
12 day forecast period. The rate of error growth between a forecast
made with today's compiler and one made with the old compiler is
comparable to errors that grow after slightly perturbing the initial
conditions. We were hoping that the modifications to the estabv code
would at least reduce the error growth from unperturbed intitial
conditions.
My question is the same as Uma's: Am I incorporating these changes
properly, or is there something else going on here?
Paul Nutter
nutter@air.atmo.arizona.edu