William.Hatheway
Active member
Hi all — I’m seeing a reproducible WRF-Chem/KPP build stop that looks like an internal consistency issue in the CBM4 mechanism files
During a WRF-Chem build (`WRF_CHEM=1`, `WRF_KPP=1`), the `chem/KPP/compile_wkc` step iterates over mechanisms and exits with:
It occurs while processing:
Other mechanisms in the same tree define the third-body species explicitly as `M` (commonly annotated as “third body”), and the coupler logic appears to require that exact symbol name. In my case, CBM4’s `cbm4.spc` does not contain `M`, so the mechanism validation fails and the overall build stops—even if CBM4 is not the mechanism I intend to run.
This suggests a mismatch between:
* the WKC/KPP expectations (“third body must be named `M`”), and
* the shipped CBM4 species file content (missing/renamed third-body definition).
Can NCAR/WRF-Chem maintainers confirm whether:
1. `cbm4.spc` is supposed to include a third-body species named `M`, and/or
2. `compile_wkc` should be more flexible for CBM4 naming, and/or
3. this is a packaging/version regression specific to certain releases?
During a WRF-Chem build (`WRF_CHEM=1`, `WRF_KPP=1`), the `chem/KPP/compile_wkc` step iterates over mechanisms and exits with:
Bash:
> `ERROR: variable name for third body in KPP species file is expected to be M, but was not found in cbm4 species file`It occurs while processing:
Code:
chem/KPP/mechanisms/cbm4/cbm4.spcOther mechanisms in the same tree define the third-body species explicitly as `M` (commonly annotated as “third body”), and the coupler logic appears to require that exact symbol name. In my case, CBM4’s `cbm4.spc` does not contain `M`, so the mechanism validation fails and the overall build stops—even if CBM4 is not the mechanism I intend to run.
This suggests a mismatch between:
* the WKC/KPP expectations (“third body must be named `M`”), and
* the shipped CBM4 species file content (missing/renamed third-body definition).
Can NCAR/WRF-Chem maintainers confirm whether:
1. `cbm4.spc` is supposed to include a third-body species named `M`, and/or
2. `compile_wkc` should be more flexible for CBM4 naming, and/or
3. this is a packaging/version regression specific to certain releases?
Last edited: