Convective updraft and downdraft mass flux values

[CMHood]

I am investigating options for obtaining values of parameterised convective updraft and downdraft mass flux values from WRF for use in the EMEP chemistry-transport model at global scales.

From the registry, I was interested in the UDR_KF, DDR_KF, UER_KF and DER_KF variables for up/downdraft en/detrainment rates, but was not able to obtain output of these with any of the Kain-Fritsch cumulus physics options, although they have the 'h' option by default - is there an additional control setting which affects output of these variables?

From the registry and looking at WRF-Chem code for convective mixing, I have also identified the MFUP_CUP, MFUP_ENT_CUP, MFDN_CUP and MFDN_ENT_CUP variables as corresponding to up/downdraft fluxes, though it is unclear whether they only apply to shallow convective structures? I edited the registry and have obtained values of these variables in WRF output.

A global trial run for a week using cu_physics = 10 (namelist attached) shows some non-zero values of the updraft flux and entrainment variables (MFUP_CUP, MFUP_ENT_CUP) but zero values of the downdraft variables (MFDN_CUP, MFDN_ENT_CUP) throughout. The updraft flux and entrainment values only appear to be non-zero over oceans (map attached for a timestep on the final full day of the run). In the code (module_cu_kfcup.F) a comment says that "if downdraft does not evaporate any water for specified relative humidity, no downdraft is allowed", but it seems unlikely that this would always be the case.

Please can anyone comment on whether this is expected behaviour for these CuP variables (zero downdraft fluxes at all times, zero updraft over land)? Are additional namelist settings or input data required to enable non-zero downdraft flexes?

Or are there any suggestions for a different way of obtaining suitable output values of up/downdraft flux and/or en/detrainment values to represent parameterised convective mixing?

Thanks for your help.

 

[kwerner]

Hi,
It looks like those variables are part of a package and are output when the "kf_edrates" namelist.input parameter is set in the &physics record. Here is some information about this variables from the WRF/run/README.namelist file:

 kf_edrates (max_dom)               !  Add entrainment/detrainment rates and convective timescale output variables for KF-based cumulus              
                                                                   schemes (cu_phys = 1, 11 and 99 only)     
                                                                = 0,  !  no output; = 1, additional output

 

[CMHood]

Thanks kwerner, I had found the kf_edrates parameter and was preparing an update. I have been able to obtain the output en/detrainment rates with this parameter.

I'm now looking for a definition of the physical representation of each of these variables in the papers associated with the Kain-Fritsch convection scheme. In particular it would be helpful to clarify the sign conventions: non-zero output values of DER_KF are negative whereas all of the other three are positive.

I'm also still curious about the significance and behaviour of the CuP flux variables if anyone has comments.

 

[kwerner]

Hi,
Sorry for the delay. As I'm not a physics specialist, I wanted to reach out to a colleague who could provide more insight. They said this this line in the code

DOMGDP(NK)=-(UER(NK)-DER(NK)-UDR(NK)-DDR(NK))*EMSD(NK)

implies that only UER has an opposite sign.

 

[CMHood]

Thanks kwerner for the further response. I had reached a slightly different conclusion from the code:

- Updraft mass flux (UMF) is positive. It is updated based on -UDR (updraft detrainment, positive) +UER (updraft entrainment, positive)

                UPOLD=UMF(NK)-UDR(NK1)
                UPNEW=UPOLD+UER(NK1)
                UMF(NK1)=UPNEW

- Downdraft mass flux (DMF) is negative. It is updated based on +DER (downdraft entrainment, negative) +DDR (downdraft detrainment, positive)

DMF(ND)=DMF(ND1)+DER(ND)
[...]
DMF(ND) = DMF(ND1)+DDR(ND)

This is consistent overall - entrainment increases and detrainment reduces the associated overall mass flux in both cases. However, it is slightly confusing that the sign change is incorporated in the calculation of DDR but applied externally to UDR.

In the DOMGDP calculation you quote, this also makes sense if the expression within the bracket (UER(NK)-DER(NK)-UDR(NK)-DDR(NK)) is the change in the magnitude of mass flux included in the combined up- and down-drafts.