RxnRates diagnostic

[beckyalexander]

Name and Institution (Required)

Name: Becky Alexander Institution: University of Washington

Confirm you have reviewed the following documentation

• [x] Support guidelines
• [x ] User manuals
• [x ] Debugging GEOS-Chem and HEMCO errors
• [x ] Current and past Github issues

Description of your issue or question

I've used the RxnRate diagnostic for the first time. The results don't make sense so I'm wondering if I'm misunderstanding the instructions. I output the sulfate formation rates in both the RxnRate diagnostic and the ProdLoss diagnostic. They should give the same results after I add up the reactions in RxnRates, but they look nothing alike and the RxnRates diagnostic has lots of zeros.

In HISTORY.rc, you list the reaction rates that you want as "RxnRate_EQ001", etc. I'm not sure that I'm finding the reaction numbers that I want.

The reaction rate numbers come from gckpp_Monitor.F90. gckpp_Monitor first lists species, then reactions. The first reaction in the list is: 'SALAAL + O3 + SO2 --> LOx + PSO4 + SO4 - SALAAL ', & ! index1 Does the "index 1" indicate that this is reaction rate RxnRate_EQ001?

Also, are the RxnRates units in s-1? ProdLoss is in molec/cm3/s.

Please provide as much detail as possible. Always include the GCClassic version number and any relevant configuration and log files. I'm using 13.4.1.

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. If there are no updates within 7 days it will be closed. You can add the "never stale" tag to prevent the Stale bot from closing this issue.

[yantosca]

Transferring this to geoschem/geos-chem

[yantosca]

Closing out this issue

[beckyalexander]

I'd like to re-open this. At IGC11 I've learned that I'm not the only one who got unrealistic results with the RxnRates diagnostic. @viral211 suggested that it might be because this diagnostic is written out before the chemistry solver. Perhaps there should be a warning on the website that states that this diagnostic is not yet working so that people do not do long simulations expecting it to work. Currently the alternative is to tag reactions in KPP and turn on the Prod/Loss diagnostic.

[yantosca]

Thanks Becky. We can add this documentation. Tagging @msulprizio @lizziel.

[AlfredMayhew]

Hi, I was planning on using the reaction rate output in my work. After some discussions at IGC, figured I would try and run some tests to see what's going on. So, at risk of confusing the situation, this is what I did:

• I ran a new GCClassic 14.3.1 simulation with reaction rate and rate constant output for the sulfate formation reactions and CO loss reactions (since these are already present in the ProdLoss diagnostics). I also added a new reaction to the mechanism with two isoprene RO2s reacting together (IHOO1 and IHOO4) to mimic the dimer reactions I'm currently working on. I tracked this reaction with a new ProdLoss definition in the kpp file as well as outputting the reaction rate and constant diagnostic for this reaction. I also output species concentrations for the simulation.
• My simulation didn't include any spin-up.
• Then I calculated the rates three different ways:
  • Using the rate constants multiplied by species concentrations. I think this method is actually slightly flawed since I'm calculating using the species concentrations after chemistry.
  • Used the output reaction rates (summing if there is more than one production/ loss reaction).
  • Used the ProdLoss output.
• I made plots of each of these rates in different chemical environments around the globe to check for variations in behaviour.

I found that each of the three methods showed wildly different sulfate production rates, differing by orders of magnitude. It was also inconsistent which method would produce the highest rate (i.e. it wasn't always the case that the ProdLoss rate was higher than the reaction rates, etc.). I wonder if there is something special about the sulfate case, involving its participation in heterogeneous chemistry that confuses the rate calculations?

In contrast, the CO loss rates showed very good agreement between each method. This perhaps makes sense given the slow reaction rates of CO, meaning changes in species concentrations before and after the reaction weren't as severe. This is especially significant given the lack of spin-up.

Finally, my test dimer reaction showed that the difference was not quite as severe as the sulfate case, but did show some disagreement. The agreement was very good when the rate was fast (e.g. a point over the amazon), but in locations where the rate was slower (but not negligible), such as over central US, the ProdLoss output was consistently about 2-5x higher than the other two methods.

I haven't included plots in this post, but I can provide them if it would help to illustrate my points. I'm happy to answer any questions related to my tests.

My question is what to do with this information! I'm looking to get rate output from my simulations to trace which chemical pathways are more important than others. From this post and my discussions with others, it sounds like the preference is to trust the ProdLoss output as an accurate reflection of the rates. Would someone be able to comment on why this is the case?

Thanks!