### rho_phy and unit of tracers into WRF atmosphere

Posted:

**Mon Jul 27, 2020 6:43 pm**Hi,

I'm using WRF to compute transport of passive greenhouse gas tracers (chem_opt=17, emiss_opt=17). Those are usually expressed as dry air mole fractions, and I assumed WRF does the same. A student's question prompted me to look into this and now I'm not sure whether emissions are actually emitted as wet air mole fractions.

The lines in question are in module_ghg_fluxes.F, lines 77f (in WRF 4.1.1):

The question boils down to whether rho_phy is the density of dry air or the density of wet air. I traced the definition of rho_phy via emissions_driver.F to subroutine chem_prep in module_chem_utilities.F:

I didn't track down 'alt' (the input to chem_prep is grid%alt), but this looks like density of wet air to me. Comments in e.g. module_mosaic_cloudchem.F and module_sorgam_vbs_aqchem.F suggest that, too. On the other hand, a comment in module_chem_plumerise_scalar.F says it's dry air density ('dn0' in subroutine plumerise). And if it's wet air, wouldn't the equation be inconsistent because 8.0461e-6 is computed with the molar mass of dry air...?

In emissions_driver, conversion factors for some species (called 'conv'; analogous to 'conv_rho' above) contain the factor 1/rho_phy just as conv_rho, while others contain the factor alt.

Can someone explain which density rho_phy actually is?

Best regards,

Friedemann

I'm using WRF to compute transport of passive greenhouse gas tracers (chem_opt=17, emiss_opt=17). Those are usually expressed as dry air mole fractions, and I assumed WRF does the same. A student's question prompted me to look into this and now I'm not sure whether emissions are actually emitted as wet air mole fractions.

The lines in question are in module_ghg_fluxes.F, lines 77f (in WRF 4.1.1):

Code: Select all

```
conv_rho=8.0461e-6/rho_phy(i,k,j)*dtstep/dz8w(i,k,j) ! 8.0461e-6=molar_mass(air)/3600, [g/mol/s]
chem(i,k,j,p_co2_ant)= chem(i,k,j,p_co2_ant) + conv_rho* emis_ant(i,k,j,p_e_co2)
```

Code: Select all

`rho(i,k,j) = 1./alt(i,k,j)*(1.+moist(i,k,j,P_QV))`

In emissions_driver, conversion factors for some species (called 'conv'; analogous to 'conv_rho' above) contain the factor 1/rho_phy just as conv_rho, while others contain the factor alt.

Can someone explain which density rho_phy actually is?

Best regards,

Friedemann