CMAQ process analysis vertical diffusion and emission

Hi, all!

I’m trying to calculate mass balance of sulfate concentration from Process Analysis with CONC concentration.

I found that EMIS-ASO4 (emissions contribution to sulfate concentration) was too large.
My emission data of SO4 was approximately 900kg/hr but EMIS_ASO4 from Process Analysis was approximately 9000kg/hr.

And also, I found that emission data have been treated in VDIFF routines and saw the routine below from vdiffproc.F:

C read & interpolate emissions data => VDEMIS_DIFF from EMIS_DEFN module

     DO S = 1, N_SPC_DIFF
        DO L = 1, EMLAYS
           DO R = 1, NROWS
              DO C = 1, NCOLS
                 EMIS_PA( C,R,L,S ) = VDEMIS_DIFF( S,L,C,R )
              END DO
           END DO
        END DO
     END DO

So, I think sulfate emission from emission file is treated as ‘VDIF_ASO4’ in process analysis.
Then, where EMIS_ASO4 values come from ??

And I didn’t add VDIF_ASO4 in my process analysis input file.
Then, should I add VDIF_ASO4 in my process analysis input file and VDIF_ASO4 is essential value in process analysis ??

Please let me know any your responses anytime.

Thank you.

HyeonYeong Park

Hello HyeonYeong,

One complication here is that CMAQ automatically remaps SULF emissions to ASO4 (conserving mass while doing so). It is possible that this is making up the difference from what you expect. You can try turning this choice off by navigating to EMIS_VARS.F and setting the scale factor for ASO4 to 0.0.

I would recommend changing the line in your process analysis file to just:
so that all processes are output and you can analyze VDIF, EMIS and others as you like.

Best wishes,

1 Like

Hi, Dr. Ben!

Thank you for your response :slight_smile:

I’ve already tried new run with all values of process analysis and the emission value’s problem was not solved yet.

I think the aerosol conversion factor from pa_update.F scripts might be thought to cause the problem.

C…get conversion factors for aero emissions; as of sep 03 release, incoming
C units are in ppmV/sec for ae species, # aer x 10**6/ # molec air / sec for
C NUM, and m2/mol sec for SRF. Conversion factors convert to
C ug/m3 sec, #/m3 sec, and m2/m3 sec, respectively.

         DO I = 1,N_SPC_DIFF
           IF ( DIFF_MASK_NUM( I ) ) THEN
             PA_EM_CONV( I ) = REFAC * AVO / MGPG
           ELSE IF ( DIFF_MASK_SRF( I ) ) THEN
             PA_EM_CONV( I ) = 1.0E+06 * REFAC / MGPG
             PA_EM_CONV( I ) = REFAC * DIFF_MW( I )
           END IF
         END DO
     END IF


And my SULF emission values was set to zero, and also my aerosol emission unit is ‘g/sec’, not ‘ppmv/sec’. I’m wondering if this problem is occurring to only me…!

If you have any idea related to my problem, please let me know your opinion freely!

Thank you.

Best wishes,
HyeonYeong Park

Thanks for the new information HyeonYeong,

May I ask how you are calculating the emissions from the input file to compare to what process analysis is diagnosing?

Also, what domain and time period are you looking at?


Thank you for your kind response!

I’m not sure that I’m answering what you want and I think the amount of emission value is not the problem, but just emission value’s cycle.

I’m calculating the emissions from the input file just by adding values of each grid point. I think I misunderstand the unit of the results after adding values because the results after adding whole values would be ‘g/sec/m3’.

I’m so sorry to bother you with my misunderstanding… :frowning:

And I found that the cycle of emission input file and the cycle of the results of EMIS_ASO4 were different.

So, I used the PMDIAG file to correct this difference (multiplying values of PMDIAG to PA_VALUES) and it worked.

I want to know if applying PMDIAG file to the results of process analysis is necessary or not.

And also, what unit of aerosol emission (ppmV/sec or g/sec) is considered to calculate PA_EM_CONV.

This is not the case: due to long-standing tradition, emissions are not in MKS-Standard, “interpolatable” units: instead, there is an implicit “per grid cell” in units that are of the general form mass per unit time per grid cell.

1 Like

I’m not exactly sure what you mean here by ‘cycle’. Is this referring to the time step of process analysis output being different than the time step of emissions input? In general, they are both probably hourly. However, it is important to keep in mind that Process Analysis output is aggregated across the entire hour while emissions are instantaneous (Carlie has somewhat recently alerted us to problems with this interpretation, but it is nonetheless currently the interpretation).

I didn’t understand exactly how you are using the PMDIAG data and combining it with Process Analysis output. What variables are you relying on and what equations are you applying? I’ve never used PMDIAG with Process Analysis before.

PA_EM_CONV is defined in order to convert from the units (of aerosols in this case) that are used within VDIFF, and these are somewhat strange units. What version of CMAQ are you using?


Thank you for answering!
I think I make you confused because of using the word ‘cycle’.
I meant that the daily variations of emission input file and EMIS_ASO4 were different.
For example, the daily variation of emission input file is an gradually increasing trend, EMIS_ASO4 from process analysis are decreasing or dramatically increasing trend.
And this may be not solved yet… :frowning:

I just multiplied the values ‘PM25AT’, ‘PM25AC’, ‘PM25CO’ from PMDIAG to ‘EMIS_ASO4I’, ‘EMIS_ASO4J’, ‘EMIS_ASO4K’. (e.g. EMIS_ASO4I = EMIS_ASO4I * PM25AT)
Based on your answer, I think this attempt may be wrong.

I’m using CMAQ version 5.3.1.

I’m so sorry to make the confusion for your understanding.

Thank you for your dedication to answer my question!

HyeonYeong Park

Oh, I got it!!
Thank you for your response :slight_smile:

HyeonYeong Park

Hi, Dr. Murphy!

I’m so sorry to answer too late and made any misunderstanding on my question.

Finally, My advisor professor and I found the reason of the dramatic variation of sulfate emission in our model domain (South Korea).

We did a CMAQ simulation with turning on ‘CTM_OCEAN_CHEM’ to ‘Y’ before and this was the reason of the dramatic variation of sulfate emission because we obtained an agreeable variation on contribution of sulfate emission from the simulation with turning off ‘CTM_OCEAN_CHEM’ to ‘N’.

So, we thought sea spray aerosol emissions were overestimated and affected to calculate final emission contribution on process analysis.

However, we are still getting the contribution of sulfate emission from process analysis more three times than original emission values.

We think this might be because some gaseous sulfur dioxide emission can convert to sulfate aerosol (ASO4I, ASO4J) during emitting in model simulation and affect the process analysis value.

Thank you for your effort to solve my problem.

HyeonYeong Park

1 Like