Hi there,
I followed the WRF-CMAq installation document, and modified the config_cmaq.csh. Then run bldit_cctm.csh script. The following error occurred, anyone can help me, thanks.
if ( ! 1 ) then
if ( 0 != 0 ) then
if ( -e /home/lth/WRF-CMAQ/CMAQ_Project/CCTM/scripts/BLD_CCTM_v533_gcc/CCTM_v533.cfg ) then
mv CCTM_v533.cfg.bld /home/lth/WRF-CMAQ/CMAQ_Project/CCTM/scripts/BLD_CCTM_v533_gcc/CCTM_v533.cfg
if ( 1 ) then
Illegal variable name.
You might consider the manual way to build the coupled model.
-- download WRF 4.4
-- build CMAQ by executing bldit_cctm.csh with build_twoway turns on
-- move the built CMAQ code, BLD* into WRF direction with the name cmaq
-- setenv WRF_CMAQ 1
-- setenv IOAPI /home/wdx/lib/x86_64/gcc-9.1/ioapi_3.2 (this is an example for gcc)
-- setenv WRFIO_NCD_LARGE_FILE_SUPPORT 1
(you can put these three setenv lines in your .bashrc or .cshrc and are activated every time you login in)
-- configure
-- compile em_real
Hi david,
Thank you for your reply. I think some wrong with my git. If I only turn on build_twoway , the same problem(‘Illegal variable name’) occurs. How should I modify the bldit_cctm.csh to only build the CMAQ portion without going through downloading WRF process.
Hi David,
The BLD_CCTM_v533_gcc has created, but there was no CCTM_v533.exe in the directory. Only when I turn off the build_twoway, the CCTM_v533.exe can be created in the directory.
You don't need CCTM.exe since you are building a WRF-CMAQ coupled model where CMAQ is considered a set of subroutines in the WRF model. The very last step "compile em_real >& mylog" in the instructions I have provided will produce wrf.exe which is the WRF-CMAQ coupled model.
Yes seems like there is an issue with the git command and starting a new shell (maybe when you start the new shell your environment changes?). For future uses, you may want to investigate that but simply making a short script with only the git --version command. But for now as David has suggested, please manually try to build WRF-CMAQ
Please also note, that if you are trying to couple WRFv4.3-CMAQv5.3.3 to asses short-wave radiative aerosol feedback, that version contains a bug that impacts this feedback. This bug will be fixed in the next release of CMAQ and we recommend waiting, if you can. If you cannot, I’m sure we can assist you in patching this bug yourself.
It looks like you have created an executable, wrf.exe, for the WRF-CMAQ coupled model. You can double check it by typing the following and see something return:
I have mentioned before that we have uncovered a bug in the WRf-CMAQ coupled model with CMAQ version 5.3 or later. Here is the way to fix the bug in twoway_feedback.F90:
– vi twoway_feedback.F90
– goto subroutine feedback_read
– add a new variable declaration
integer, save :: loc_nlays
– add the following line right before “firstime = .false.”
loc_nlays = nlays3d
– replace all occurance of NLAYS3D with loc_nlays after the
line "firstime = .false.
If you have any questions, please feel free to contact me (wong.david-c@epa.gov).
Thanks Andshrew… I shall take the question to the sccm forum, see if I can get some traction there. To answer your query, We are a 3000 node, single site with just 2 DPs, a single MP, we don’t do anything complicated with it at all which is why it is so frustrating that the simple stuff is failing. Thanks for the input. I might look at running the upgrade, just in case there’s corruption in the install files as the upgrade might update them in \server\sms_Xxx\client… long shot…
