How to derive cloud water pH from CAMS

Datasets and Usage CAMS - Datasets and Usage
,
1

Dear community,

I would like to obtain cloud water pH values from the CAMS dataset. Since cloud water pH is not among the variables that are archived and can be downloaded, it would be nice if there is a a way to calculate this from the available variables.

This would be nice to have for a PhD project I am supervising, where we look at SO2 chemistry and transport with a simplified model for SO2 chemistry on air mass trajectories. The results are quite dependent on cloud water pH, because of the SO2+O3 reaction in cloud droplets, which is very pH dependent. The SO2 chemistry model does not calculate all of the tropospheric chemistry, but uses many species from CAMS (O3, H2O2, OH...) as fixed input variables. In the moment, pH is set to a globally fixed value in the model, and I would like to improve on this. I also have to say that I am not really an expert for cloud pH.

I have already put some research and effort into this to be able to ask more specific questions. First, since the cloud water pH variable was not available and I also have no access to the CAMS source code and the function calculating the pH values, I used a function from the Geos-Chem model that is calculating pH values as a "makeshift" solution and for first tests (sorry for using source code from the competitors (wink) ). This does work quite well, but my main problem in the moment is to match the input parameters of the Geos-Chem function with the CAMS parameters available for download. Some of the input variables are probably not important, since we are looking at the free troposphere in our project and not directly near the surface, but some of the rest are probably important. Here is the list of input parameters for the Geos-Chem function directly from the documentation of the function and my best guess for the corresponding CAMS input:

  •     SO4nss ! Total nss sulfate mixing ratio [M]: I have no idea what the correct CAMS variable is, maybe "sulphate aerosol mixing ratio" aermr11 (11.210) is the correct one here? I am a little bit confused by "total" and "aerosol" here
  •     HMSc   ! Total HMS mixing ratio [M]: (probably do not need this?)
  •     TNO3   ! Total nitrate (gas+particulate) mixing ratio [v/v] : Probably CAMS HNO3 (6.217), but since this is gas+particulate, I am not sure
  •     TNH3   ! NH3 mixing ratio [v/v]: Probably CAMS "ammonia" NH3 (19.217)
  •     SO2    ! SO2 mixing ratio [v/v]: CAMS SO2 (122.210)
  •     CL     ! Total chloride (gas+particulate) mixing : (probably do not need this?)
  •     TNA    ! Sodium (particulate) [v/v] : (probably do not need this?)
  •     TDCA   ! Total Ca2+ and Mg2+ mixing ratio [M] : (probably do not need this?)
  •     TAA    ! Acetic acid mixing ratio [v/v] : (probably do not need this?)
  •     TFA    ! Formic acid mixing ratio [v/v] : (probably do not need this?)


Any help would be greatly appreciated! In case there is no expert around here, it would also help to point me to someone who can help me on this topic.

Thanks!

Ingo Wohltmann (supervisor) and Chiranjeevi Nalapalu (PhD student)

Alfred Wegener Institute for Polar and Marine Research, Potsdam, Germany

2

Hi Ingo, 

if there is no response in the forum, the next step is to contact the user support portal, to make sure a response is provided. 

Specifically for your question about cloud pH ... 

That variable is not provided because it is not a CAMS product and is not calculated by the operational CAMS systems (o-suite forecasts, reanalysis) , for which data is stored in the ADS. However, we plan to provide cloud pH as part of the upcoming upgrades. 

For the specific questions you have I add Samuel Remy (Samuel Remy). He may be able to advice on the best mapping of available CAMS variables to the GEOS-Chem input.  Please note, Ammonium and Nitrate aerosols are not available from CAMSRA. 

Johannes

3

Hello Ingo,

     As pointed out to Johannes, we are also working in this direction, but this is for the future. We could keep in touch though to compare results and methods, if you are happy with this. As for the CAMS input, this really depends on which CAMS dataset you are using: is it from CAMSRA (in which case, particulate nitrate and ammonium are not available) or CAMS NRT (particulate nitrate/ammonium available after July 2019, update to cycle 46R1). Taking your list:

  •     SO4nss ! Total nss sulfate mixing ratio [M]: I have no idea what the correct CAMS variable is, maybe "sulphate aerosol mixing ratio" aermr11 (11.210) is the correct one here? I am a little bit confused by "total" and "aerosol" here

Yes, here clearly this is 11.210 that shoudl be used

  •     HMSc   ! Total HMS mixing ratio [M]: (probably do not need this?)

To be confirmed, but I don't think this is a CAMS species. MSA (methane sulfonic acid) is a species and is commonly used in pH computations, so maybe it could relate?

  •     TNO3   ! Total nitrate (gas+particulate) mixing ratio [v/v] : Probably CAMS HNO3 (6.217), but since this is gas+particulate, I am not sure

This should be HNO3+NO3: if you are using CAMS NRT post July 2019, then you should also include the two nitrate species (from gas-particle partitioning and from heterogeneous reactions): 247.210 and 248.210

  •     TNH3   ! NH3 mixing ratio [v/v]: Probably CAMS "ammonia" NH3 (19.217)

Maybe you could confirm if this refers to NH3 only or NH3+NH4: "TNH3" often refers to NH3+NH4. If so, and if you are using CAMS NRT post July 2019, then you should use particulate ammonium: 249.210

  •     SO2    ! SO2 mixing ratio [v/v]: CAMS SO2 (122.210)

Yes

  •     CL     ! Total chloride (gas+particulate) mixing : (probably do not need this?)

On the particle side, Cl is part of the sea-salt aerosol species, and could be derived from the three bins of sea-salt aerosol. In our tests, we are using a 0.128 fraction to derive Cl- from NaCl (+water).

  •     TNA    ! Sodium (particulate) [v/v] : (probably do not need this?)

This could be derived from sea-salt aerosol and desert dust (3 bins each): we are using a 0.07 scaling factor for Na+ from NaCl, and a 0.012 scaling factor for Na+ from desert dust

  •     TDCA   ! Total Ca2+ and Mg2+ mixing ratio [M] : (probably do not need this?)

For Ca2+, it can be derived from desert dust, using a 0.03 or 0.05 scaling factor (depending on the litterature).

  •     TAA    ! Acetic acid mixing ratio [v/v] : (probably do not need this?)

To be confirmed, but not available in CAMS datasets

  •     TFA    ! Formic acid mixing ratio [v/v] : (probably do not need this?)

Formic acid is available in CAMSRA and CAMS NRT.


I hope this helps. If you wish to compare results, you can contact me at sr@hygeos.com.

Samuel Remy

4

Dear Samuel,

many thanks for your reply! That helps a lot. I think I have still some questions (and you had some questions for me that I will try to answer).

Unfortunately, I don't know what the difference between CAMSRA and CAMS NRT is. A colleague of mine is downloading the data for us, I think it is the CAMS reanalysis at https://www.ecmwf.int/en/research/climate-reanalysis/cams-reanalysis (that page links to a variable list on the upper right). Is this the same as CAMSRA?

I think for the question what species some of the variables actually include, it is probably helpful to look into the paper that describes the Geos-Chem Ph model that I use. You can find it at doi:10.5194/acp-20-12223-2020 . There is e.g. Table 3. Does that give you sufficient information?

From the table, I have the impression that you are right and that should be the totals, i.e. TNO3=HNO3+NO3 and TNH3=NH3+NH4.

What can I do in case I am not using CAMS NRT (which I suspect) and still need NO3 and NH4 etc. There seems to be NH4 (21.217) and NO3 (is it 32.217 or 51.217?) in the variable list of the CAMS reanalysis. But that seems to contradict what you say in your reply ("particulate nitrate and ammonium are not available in CAMSRA").

I have to admit, I am quite confused about all the CAMSRA, CAMS reanalysis and CAMS NRT stuff and how that relates to all of these different variables. But I hope you can shed some light on this.

Best wishes, Ingo Wohltmann

5

Dear Ingo,

   Sorry for the late answer, days are busy before the end of they year. Yes, CAMSRA is the acronym of the CAMS Reanalysis. I had a look to the reference you sent, indeed, this is a well know reference paper that we used already. The inputs of the H+ computation are quite similar to what we are doing (wink) Clearly, TNO3 is HNO3 + NO3 and TNH3  is NH3 + NH4. For the global CAMS, there are basically two products: the reanalysis, which uses a frozen model version to reprocess a long period of time (2003 to 2022) with a consistent chemistry/aerosol model, and the CAMS near real time (NRT), which uses a model version that is frequently updated, usually at a annual or twice a year frequency. So the model version used for the near real time is more recent: in your case this is why nitrate aerosol is included in the NRT and not in the reanalysis. There are plans to produce a newer reanalysis in the coming years.

   NO3/NH4 in CAMSRA refer to nitrate/ammonium as computed then in the chemistry scheme, before nitrate and ammonium were included in the aerosol scheme. These are unvalidated products, unlike nitrate and ammonium in the aerosol species. However, it is probably better than nothing...Otherwise, I would suggest using the more recent CAMS NRT dataset if you can?

Samuel

6

Many thanks!

I will probably don't have more time to look into that this year, since I am going on Christmas vacation tomorrow (smile)

I will have a look next year if I get this running with the recommended input parameters.

NRT is probably not an option. If I understand it correctly, this is not available for many years. Our model runs are e.g. for the years 2009 and 2010 in the moment. I suppose there is no NRT output for these years.

One thought that just came to my mind is: If there would be a possibility to use the function from the source code that CAMS uses for pH calculation directly, that would simplify things so much (e.g. we would know exactly what input parameters we need, and it would be consistent with what we do in the rest of our chemistry model, which bis based on CAMS background fields). I am aware that there might be legal reasons or a general ECMWF policy that prevent this (e.g. you may not want this to be openly distributed). But maybe there is a way to be allowed to use this small part of the source code for our PhD project and to have a sort of collaboration there.

Best wishes, and Merry Christmas and a Happy New Year in advance

Ingo