Error in "module_ecpp_stats.F90" due to new cloud physics scheme in CESM model

[abhisheklodh]

Dear All,

I am trying to modify the microphysics of the cesm model: "spcam_m2005" by including a new scheme which takes into account secondary aerosols. Hence, I have modified the "crm_physics.F90" and "crmx_crm_module.F90" file. After initializing the new variables I am getting error in the module "crmx_module_ecpp_stats".:
*******************************************************************************************************************
[5] nre,rho_cloudice,rhoa,D_ice(j),D_snow(i_ac) is 289.6582 800.0000
[5] 0.1292351 9.3572657E-04 5.6023085E-03
[5] nre_large,rhog_small,xr1_small,rhoa,xr2_large,K_ns is 289.6582
[5] 800.0000 4.6786328E-04 0.1292351 2.8011543E-03 5557.976
[5] nre,rho_cloudice,rhoa,D_ice(j),D_snow(i_ac) is 289.6582 800.0000
[5] 0.1292351 1.1382556E-03 5.6023085E-03
[5] nre_large,rhog_small,xr1_small,rhoa,xr2_large,K_ns is 289.6582
[5] 800.000[6] forrtl: error (72): floating overflow
[6] Image PC Routine Line Source
[6] cesm.exe 0000000005C074FB Unknown Unknown Unknown
[6] libpthread-2.31.s 000014E0D4464420 Unknown Unknown Unknown
[6] cesm.exe 0000000002EBC8CB crmx_module_ecpp_ 160 crmx_module_ecpp_stats.F90
[6] cesm.exe 0000000002E7EECA crmx_module_ecpp_ 568 crmx_module_ecpp_crm_driver.F90
[6] cesm.exe 0000000002D623AE crmx_crm_module_m 2823 crmx_crm_module.F90
[6] cesm.exe 0000000002B34BA3 crm_physics_mp_cr 3481 crm_physics.F90
[6] cesm.exe 0000000001FEAB1B spcam_drivers_mp_ 489 spcam_drivers.F90
[6] cesm.exe 0000000001169D06 physpkg_mp_phys_r 1074 physpkg.F90
[6] cesm.exe 000000000086B583 cam_comp_mp_cam_r 259 cam_comp.F90
[6] cesm.exe 000000000082F7B7 atm_comp_mct_mp_a 521 atm_comp_mct.F90
[6] cesm.exe 0000000000477604 component_mod_mp_ 737 component_mod.F90
[6] cesm.exe 0000000000441179 cime_comp_mod_mp_ 2823 cime_comp_mod.F90
[6] cesm.exe 000000000045F89A MAIN__ 133 cime_driver.F90
[6] libc-2.31.so 000014E0D4282083 __libc_start_main Unknown Unknown
[6] cesm.exe 000000000041852E Unknown Unknown Unknown
*******************************************************************************************************************

At line 160 of "crmx_module_ecpp_stats.F90" : I am printing rh,

rhsum1 = rhsum1 + rh
print *,"rhsum1 is", rhsum1,"rh is",rh

I am also not modifying rh in the whole suite. Don't understad why this error is coming.

Any suggestion is this regard will be really helpful.

Regards
Abhishek

 

[peverley]

Hi Abishek,

I'm afraid I can't really tell what's going on without more details of what changes you have made to the code. It looks like some modification you made has had unintentional consequences on the relative humidity.

Also, when you're printing out rh and rhsum1, I would recommend not trying to print the entirety of those three-dimensional arrays.

If you are unable to get to the source of the problem, I may be able to help if you provide me with more details on what modifications you have made.

Sorry I can't be of more help,
Courtney

 

[abhisheklodh]

Hi Courtney,

I have introduced my own microphysics module - "module_mp_lin_double.F90" which has new cloud physics scheme to replace the m2005 spcam.
In the scheme: crm_physics.F90, there is a call to crm at line number 1591(approx.) :

call crm (lchnk, i, state_loc%t(i,:), state_loc%q(i,:,1), state_loc%q(i,:,ixcldliq), state_loc%q(i,:,ixcldice), &
..................... taux_crm(i), tauy_crm(i), z0m(i), timing_factor(i), qtotcrm(i, :) )

i modified the argument list of call crm to introduce my new variables. The model build was successfully done.The values: crm_nx, crm_ny and crm_nz and nzm is 32, 1, 28

While giving a CESM model run with new microphysics at the 1st time step, in the file: crmx_module_ecpp_stats.F90, module - "crmx_module_ecpp_stats" is giving "floating overflow" error for the variables- qcloud_bfsum1, rhsum1 and qlsink_bfsum1 in the subroutine "rsums1":

! qcloud_bfsum1 = qcloud_bfsum1 + qcloud_bf
! print *,"qcloud_bfsum1 is", qcloud_bfsum1
! rhsum1 = rhsum1 + rh
! print *,"rhsum1 is", rhsum1,"rh is",rh

The error file are attached. Kindly suggest. Note this is used for converting back in rsum2ToAvg.

Thanks and Regards
Abhishek

 

[peverley]

Hi Abhishek,

Thanks for the additional information. It looks like maybe you attached the wrong log file? That looks like the build log. Can you attach your cesm.log.* and atm.log* files if you're still getting this error?

Also, is the module module_mp_lin_double.F90 somewhere I can look at it? If not, can you attach that as well. I'd like to see what's happening in there that is different from m2005.

Thanks!
Courtney