Change yield of production of secondary organic aerosols from isoprene

[oyvindseland]

What is the feature/what would you like to discuss?

The yield of isprene to SOA should be reduced from 5 % to 0.5 % Impact all chemistry set-up (chem_mech.doc) Should likely be default in NorESM2.3. Sara Blichner should be asked if she still think the 0.5 % is what she recommends. Added her github in the discussion section below

Is there anyone in particular you want to be part of this conversation?

@sarambl

Will this change (regression test) answers?

Yes

Will you be implementing this enhancement yourself?

Yes, but I will need some help

[gold2718]

Completed by #176

[DirkOlivie]

Hi Sara @sarambl (@oyvindseland @gold2718),

we have now reduced the SOA-yield for isoprene from 0.05 to 0.005 in NorESM2.3. We see however that the anthropogenic aerosol ERF becomes considerably stronger: -1.62 W/m2 (+/-0.08 W/m2). In earlier versions (NorESM2), the anthropogenic aerosol ERF was around -1.2 to -1.4 W/m2.

Have you tested the anthropogenic aerosol ERF with these lower SOA-yields? Did you get similar results?

Might the impact of these SOA-yield changes also be different in a simulation containing your NPF code and a simulation without the NPF code (NorESM2.3 currently doesn't contain the NPF code)?

Best regards,
Dirk

[sarambl]

Interesting! No as I mentioned at the time of the recommandation, I didn’t really have time to do a lot of simulations or tests so it was more like ”this makes more sense both theoretically, fits what other models are doing and wrt the measurements I am looking at”.
Note also that the mass yield used to be around 10 percent with the 5 percent yield because the SOA_LV and SOA_SV are around double the mass of isoprene (and the 5 percent was implemented as a molar yield). So I anyway stand by that it should be reduced considerably.

Some quick references that I had in mind:

1. The table below is from the supplement of Blichner et al (2024, nat comms, ) and shows the mass yields in the different models in that study. I think the old yield put

2. Here are some results from sporre et al (2019, https://doi.org/10.5194/acp-20-8953-2020):

3. Finally below is again a plot from the supplement of Blichner et al (2024, nat comms, ) which shows the OA at SMR and ATTO as it is and scaled down proportional to reducing the isoprene yield to zero (this is a scaling, not a re-run).

Some more thoughts
It’s not that surprising that it would influence the ERFaero though, given that NorESM is quite sensitive to condensate (has a large aitken mode) and if there’s less in the PI it makes sense to me that that would make it less sensitive?

I think that it would be perfectly fine to increase it somewhat again, maybe the mass yield could be 3 percent as in EC-Earth? But note that for example ECHAM-SALSA (which has a much higher mass yield for IP SOA, does not overestimate OA that much...
The proper way out would of course be to do a comparison to available composition station data (and Atom data as well) and maybe do a couple of sensitivity tests. But I am assuming that's off the menu due to work load? I thought there was a composition dataset in the making during FORCeS, does anyone know what happened to it?

[oyvindseland]

So even for zero yield the OA mass is too high at ATTO.

So the positive may also come from POM? Or the high POM/POC ratio?