On Apr 20, 2012, at 10:35 PM, Phan Van Tan wrote:

Dear Guangshan Chen,
Following your instructions (except Bug 4, which I could not do as you say), I have compiled model successfully. But when running model the following errors occur.
Could you please share your experiences to debug.
Thanks.
-------------------
Successfully initialized clm2 history files
------------------------------------------------------------
Successfully initialized the land model
begin initial run at:
    nstep=            0 year=         2012 month=            1 day=
           9 seconds=            0

************************************************************

Attempting to read monthly vegetation data .....
nstep =            0 month =            1 day =            9
(GETFIL): attempting to find local file SMALL_RCMlai.nc
(GETFIL): using input//SMALL_RCMlai.nc
application called MPI_Abort(MPI_COMM_WORLD, 1) - process 0[cli_0]: aborting job:
application called MPI_Abort(MPI_COMM_WORLD, 1) - process 0
netcdf error from readMonthlyVegetation
netcdf strerror = NetCDF: Index exceeds dimension bound
ENDRUN: called without a message string
rank 0 in job 1 compute-0-0.local_35795   caused collective abort of all ranks
exit status of rank 0: killed by signal 9
-------------------

--- On Wed, 4/18/12, Guangshan Chen <gchen9@gmail.com> wrote:

From: Guangshan Chen <gchen9@gmail.com>
Subject: [RegCNET] Bugs fixed in RegCM4.2 with CLM option
To: regcnet@lists.ictp.it
Cc: "val bennington" <benesh@wisc.edu>, "Steve Vavrus" <sjvavrus@wisc.edu>, "pat behling" <pbehling@wisc.edu>
Date: Wednesday, April 18, 2012, 7:50 AM

Dear all,

The following are the bugs fixed in RegCM4.2 with CLM opiton when I tested it.

Bug 1: the types between lm and lmask are mismatched in mod_lm_interface.F90, line 112
Fix: Below line 106 add the follow line:
integer , pointer ,             dimension(:,:) :: lm_tmp
lm_tmp => lm
Change line 122 to
call assignpnt(lm_tmp,lmask)

Bug 2: SST was the same for every month when using GISST in mod_sst_1deg.F90
Fix: move line 153 “call split_idate(idate,year,month,day,hour)” into the do-loop just below it. (Fixed by Bennington Valerie in CCR UW-Madison)

Bug 3: There is no assigned value for the array “satbrt_clm” in surfrdMod.F90.patch
Fix:
Step 1: Below line 129, define a new variable lndcatf in mod_bats_common.F90
real(dp) , pointer , dimension(:,:) :: lndcatf       ! mddom_io%lndcat chengs++

Step 2: Below line 109, assign the value to lndcatf in mod_lm_interface.F90
call assignpnt(dom1%lndcat,lndcatf) ! chengs++

Step 3: Change line 118 in mod_mtrxclm.F90 to the following:
use clm_varsur,    only : landmask , landfrac, satbrt_clm ! chengs++ satbrt_clm

allocate the array satbrt_clm after line 200:
if ( .not.allocated(satbrt_clm) ) allocate(satbrt_clm(iy,jx)) ! chengs++

Assign the values for satbrt_clm from lndcaft below line 207:
satbrt_clm(i,j) = lndcatf(i,j)

But 4: Monthly vegetation data XXX_RCMlai.nc only has one time slice in it. It should have 12 time slices.

Fix: According to the file of clm2rcm.F90 in RegCM-4.1.1 to change the value of "xhr" in the do-loop at line 458 in clm2rcm.F90 in RegCM-4.2

But 5: Calendar problem in CLM
Note: only Gregorian and noleap calendars can be used in CLM now.

If you define Gregorian calendar in your regcm.in:
Step1: delete "#define NO_LEAP_CALENDAR"    in ESMF.inc
Step 2: change “ESMF_CAL_GREGORIAN” in line 392 and line 410 to “GREGORIAN” in clm_time_manager.F90

Guangshan Chen
Center for Climatic Research
University of Wisconsin-Madison

-----Inline Attachment Follows-----

_______________________________________________
RegCNET mailing list
RegCNET@lists.ictp.it
https://lists.ictp.it/mailman/listinfo.cgi/regcnet