sans said:
Hi,
I want to run CAM3.0, in aquaplanet mode with monthly observed sst that comes alongwith cam input datasets. I want to run the model from 1979 Jan, so, I am using the model generated initial condition file for 1979.
While running with the climatological sst, it runs fine, but with observed sst it shows the following message and stops :
SSTINI: Failed to find dates bracketing ncdate, ncsec= 321 0
ENDRUN: called without a message string
I have made all the necessary changes as suggested in the thread titled : Turning of CLM, posted by brianpm, in the CGD Forum.
I have also set - START_YMD = 19790101, in the namelist.
Any suggestions in this regard, would be of great help.
Thanks,
sans
When running the default model with aqua_planet=.true.,
the model automatically sets the date to March 21 of year
0. *permanently* (except for the history file handler which advances
time just as in a regular default CAM run). The date does
not ever advance, it always remains March 21 throughout
the simulation. So all aspects of the simulation are tied to
March 21 including the reading of SST, ozone and any
other time-dependent boundary datasets.
The modifications I gave to brianpm worked for him because
the climatological dataset has year 0 for its date. This
was good for him if he only wanted to sample the March 21
date because that is all the model would sample since the
model stays at March 21 throughout the run in aqua_planet
mode.
Unfortunately, if you (and he) want to actually sample
the SST dataset in time-advancing mode throughout the run,
then the modifications I gave to him will not work. You
will need to do the opposite (which was his original plan).
You will need to set AQUA_PLANET = .false in the namelist
and make the modifications outlined below.
Since aqua_planet was not designed to run in this manner,
there may be some risk that hidden time-dependencies that
are not relevant for aqua_planet mode will be activated by
making the following mods. In fact, I know that you will
be reading the climatological ozone dataset in time-dependent
mode which may corrupt your results, depending on the purpose
of your research. So you may want to build your own ozone
dataset which you can download from the website I indicated
to brianpm.
Which is to say, proceed at your own risk :-)
1) namelist settings:
AQUA_PLANET = .false.
SSTCYC = .false. (for reading monthly dataset)
Code modifications:
2) reverse *only* the second instance of the
"aqua_planet" logic in
"models/atm/cam/src/control/initext.F90"
3) reverse all instances of the "aqua_planet"
logic in
"models/atm/cam/src/dynamics/eul/inidat.F90"
4) reverse all instances of "aqua_planet" in
"models/atm/cam/src/physics/cam1/physpkg.F90"
5) Also in "inidat.F90", set the variable "landm"
to 0. This is a weighting factor used in
the physics and tied to land.
6) set the following namelist parameters
ECCEN = 0.
OBLIQ = 0.
MVELP = 0.
7) You may also additionally set these namelist parameters to
turn off effects of possible time-dependent aerosols
SULSCL = 0.
CARSCL = 0.
SSLTSCL = 0.
DUSTSCL = 0.
