CMAQ Ideas Repository

Add known issues or new ideas about CMAQ to this page.

I found several issues with CMAQ 4.7.1 that cause crashes - primarily involving bidirectional Hg, but also some general problems. I have developed interim 'work-around' solutions. Modified CMAQ 4.7.1 files are available at the following website:

Issues are the following:

1. Vertical advection (/vadv/yamo): The numerical solver contains infinite iterative loops (in vppm.F,v at statements 66 and 67). The loops appear to generate a solution to a cubic equation. (Can a CMAQ developer verify this? If so, there may be better ways to solve the cubic equation.) It very rarely happens that the loop fails to converge, but once is enough to cause the model to hang up!

My interim solution adds an exit after 100 iterations. Also, it skips vertical advection for the column, rather than terminates - it terminates only after many failed iterations.

2. Dry deposition algorithm chokes on zero leaf area index: I found this in the Hg inline version (/vdiff/acm2_inline_txhgsim) and also in MCIP 3.4 (may be corrected in MCIP 3.6). The problems include divide-by-zero (if LAI=0) and nonconvergence in the related soil solution.

The bidirectional Hg algorithm is supposed to skip the soil calculation if land category = ocean. However, I found a few grids in Hudsons Bay, Canada, that were categorized as land (but with soil type = water) with zero LAI. The work-around skips the soil calculation if LAI=0, and elsewhere sets LAI=1e-25 rather than zero (to prevent divide-by-zero errors).

3. Soil type mismatch between WRF and CMAQ (bidirectional Hg only, /vdiff/acm2_inline_txhgsim): CMAQ sets parameters based on 11 soil types and crashes if soil type is 12 or higher. WRF uses 16 soil types. This is in the CMAQ file HGSIM.F,v. The modified CMAQ file (for WRF input only!) translates between them. For WRF soil types, see

4. Timing roundoff error in bidirectional Hg (/vdiff/acm2_inline_txhgsim): This occurs in WRASX_MEDIA (in file HGSIM.F) - sometimes it tries to write SFHG output and the time is off by one second (65959 instead of 70000, for example) and the write statement crashes.

5. Multipollutant models - missing toxic species: The multipollutant model (files rdemis_aetox.F and rdemis_pm_hg.F) want to read emissions for ALL the toxics - and if any one is missing, they crash. My quick work-around assigns emissions =0 if the emissions input is missing for any listed species.

Please let me know also if any of these apparent problems are really my mistakes!

- Sandy Sillman (ude.hcimu|namllis#ude.hcimu|namllis)

Carbon Bond '05 currently does not apply pressure/temperature sensitivities of quantum yields and cross sections. Previous chemical mechanisms, including CBIV, had pressure/temperature sensitivities for formaldehyde, acetaldehyde, acetone, ozone, methyl glyoxal, and higher ketones. The differences are most important in the upper troposphere. Upper troposphere air, however, does subside and impact surface chemistry.

CMAQv4.7 release is suspended while bugs are fixed in the in-line emissions modules. Improvements are also being made to MCIPv3.4 to maintain consistency between MM5 and WRF grid descriptions.

CMAQv4.7 will not contain the sulfate tracking or carbonaceous aerosol apportionment tools. These utilities will be added into the model and released in an interim version of CMAQv4.7 in 2009.

Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License