RE: problems with continuous run of ccm3

Subject: RE: problems with continuous run of ccm3
From: Ghan, Steven J (
Date: Wed Aug 22 2001 - 16:45:08 MDT


The sensitive dependence on parameter values may reflect the chaotic nature of
even SCM simulations. See the paper by Hack and Pedretti, Assessment of Solution
Uncertainties in Single-Column Modeling Frameworks, Journal of Climate: Vol. 13,
No. 2, pp. 352ĘC365.


-----Original Message-----
From: Yuqiong LIU []
Sent: Wednesday, August 22, 2001 11:59 AM
Subject: problems with continuous run of ccm3

Dear CCM3 users,

Recently I hooked up ccm3 (actually a single column) with our atomatic
parameter optimization algorithm. The coupled system seems to be working
now, however, I've got the following two specific problems:

1) When I called ccm3 in a row continuously in my algorithm, I've got
different lsm flux simulations even if the parameters are exactly the
same for each call. This would not happen if I call the model
separately, one call for each run. Also I got no such a problem with the
same but off-line land surface model. To my knowledge, this could be
caused by the random number generator (RNG) used in the model. So I'm
wondering if there is any RNG used in the atmospheric part of the model,
like cloud fraction estimation. Or there may be other sources, please
let me know if you have any comments or suggestions.

2) The second problem is related to the stability of the model
simulations on land surface part. When I change one of the lsm parameter
a little bit(e.g. changing porosity from 0.43482 to 0.43483), I get
about 20 w/m^2 of RMSE in latent heat flux simulation (This is the
results of every 20 mins time step and I'm not averageing it). Does
anybody know why this has happened? Your suggestions will be highly


Yuqiong Liu
Dept of Hydrology & Water Reources, U of A

This archive was generated by hypermail 2b27 : Tue Jan 22 2002 - 11:15:58 MST