It sounds like you are wondering about the aerosol extinction? There are several things that might be output as extinction, but I don't see any that appear as "extinction coefficient". For example, searching for outfld calls in the CAM src directory:
Code:
$grep -Iinr "outfld.*extinct" .
./physics/cam/modal_aer_opt.F90:1070: call outfld('EXTINCTdn'//diag(list_idx), extinct, pcols, lchnk)
./physics/cam/modal_aer_opt.F90:1086: call outfld('EXTINCT'//diag(list_idx), extinct, pcols, lchnk)
./physics/cam/modal_aer_opt.F90:1106: call outfld('EXTINCTUVdn', extinctuv, pcols, lchnk)
./physics/cam/modal_aer_opt.F90:1107: call outfld('EXTINCTNIRdn', extinctnir, pcols, lchnk)
./physics/cam/modal_aer_opt.F90:1161: call outfld('EXTINCTUV', extinctuv, pcols, lchnk)
./physics/cam/modal_aer_opt.F90:1162: call outfld('EXTINCTNIR', extinctnir, pcols, lchnk)
Assuming you're interested in aerosol extinction, I think the relevant code is in the file:
$CESM/components/cam/src/physics/cam/modal_aer_opt.F90
The variable tauxar is the layer extinction optical depth, and has dimensions of columns, levels, [shortwave, longwave] bands. For both shortwave and longwave, there's a loop where tauxar is computed like tauxar(i,k,ilw) = tauxar(i,k,ilw) + dopaer(i).
That dopaer is the "aerosol optical depth in layer". It is computed as dopaer(i) = pext(i)*mass(i,k) (shortwave) or dopaer(i) = pabs(i)*mass(i,k) (longwave).
Mass is just the layer mass.
pext is "parameterized specific extinction (m2/kg)" This is determined based on "cext" and some conditional logic that appears to depend on the aerosol size. "cext" appears to be derived from interpolation of the specific extinction... but I could be misinterpreting exactly what is going on there.
Please take a look at that code to see if it helps alleviate your doubt.
Maybe more useful is the reference at the top of that code:
Ghan, S. J., and Zaveri, R. A. (2007), Parameterization of optical properties for hydrated internally mixed aerosol,
J. Geophys. Res., 112, D10201, doi:
10.1029/2006JD007927.