Low MEGAN emissions in CESM3

[tilmes]

Using the latest beta cases, we are seeing a significant underestimation of biogenic emissions, when running a F-case with -bgc sp:

beta02 was still OK, for example, isoprene emissions were around 460 Tg/yr, which is what we think is reasonable.
However, going to beta04, the biogenic emissions dropped to about 250 Tg/yr.
We think that this is a case when MEGAN changes had been included.

We also run a recent beta05 version with @wwieder changes to the land and we saw some improvement with isoprene emissions going back to around 380 Tg/yr.

We would like to discuss what options there are to increase biogenic emissions.
@shawnh @rrbuchholz @lkemmons

[dlawrenncar]

Can you remind us what land related factors drive isoprene emissions? That is, what is it tied to and what changes with the new MEGAN?

…

On Sat, Mar 15, 2025 at 3:34 PM tilmes ***@***.***> wrote: Using the latest beta cases, we are seeing a significant underestimation of biogenic emissions, when running a F-case with -bgc sp: beta02 was still OK, for example, isoprene emissions were around 460 Tg/yr, which is what we think is reasonable. However, going to beta04, the biogenic emissions dropped to about 250 Tg/yr. We think that this is a case when MEGAN changes had been included. We also run a recent beta05 version with @wwieder <https://github.com/wwieder> changes to the land and we saw some improvement with isoprene emissions going back to around 380 Tg/yr. @shawnh <https://github.com/shawnh> @rrbuchholz <https://github.com/rrbuchholz> @lkemmons <https://github.com/lkemmons> — Reply to this email directly, view it on GitHub <#3016>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AFABYVGHSKKHM7SV4BSAXP32USMHVAVCNFSM6AAAAABZC4DKXCVHI2DSMVQWIX3LMV43ASLTON2WKOZSHEZDENJWGIZTKNQ> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***> [image: tilmes]*tilmes* created an issue (ESCOMP/CTSM#3016) <#3016> Using the latest beta cases, we are seeing a significant underestimation of biogenic emissions, when running a F-case with -bgc sp: beta02 was still OK, for example, isoprene emissions were around 460 Tg/yr, which is what we think is reasonable. However, going to beta04, the biogenic emissions dropped to about 250 Tg/yr. We think that this is a case when MEGAN changes had been included. We also run a recent beta05 version with @wwieder <https://github.com/wwieder> changes to the land and we saw some improvement with isoprene emissions going back to around 380 Tg/yr. @shawnh <https://github.com/shawnh> @rrbuchholz <https://github.com/rrbuchholz> @lkemmons <https://github.com/lkemmons> — Reply to this email directly, view it on GitHub <#3016>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AFABYVGHSKKHM7SV4BSAXP32USMHVAVCNFSM6AAAAABZC4DKXCVHI2DSMVQWIX3LMV43ASLTON2WKOZSHEZDENJWGIZTKNQ> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[rrbuchholz]

Here is a slide I use that is originally from Simone and Louisa on MEGAN2.1.

MEGAN_Info.pdf

[rrbuchholz]

I think some of the updates from MEGAN2.1 moving towards MEGAN3.1 were included, but the updates were frozen -->
See this issue: #1323

I'm not sure of the exact updates that have been added to MEGAN2.1.

[dlawrenncar]

It looks like they are described in this issue #2588

Seems like there is a new drought stress and a heat exposure mechanism. So, I guess maybe this change in emissions could be a consequence of that. The LAI, which is part of the equation, hasn't really changed for SP configurations, so that shouldn't be it. If the change is due to new stress mechanisms, then I am not sure how to resolve this. Maybe we need to go back to the developer. Perhaps emissions factors need to be retuned?

[wwieder]

As we're seeing this in F cases with in SP mode, this is likely related to MEGAN changes that came in with ctsm5.2.020. This made it to CESM with cesm3_0_alpha03a.

@HuiWangWanderInGitHub I wonder if you can help advise on this issue? It seems like global isoprene emissions are much lower than the CAM-CHEM group is expecting? Is there a simple way to consider changing MEGAN emission factors?

Separately, I also still wonder how the parameterization of plant water stress over tropical forests may be interacting with BVOC emissions @linniahawkins, especially compared to the MEGAN calibration that informed @HuiWangWanderInGitHub's parameterization?

[aswann]

FYI we compared isoprene emissions and isoprene in the lower atmosphere in CESM2 with observations over the Amazon in this paper: https://doi.org/10.1029/2023GL107214

Isoprene flux looked ok but concentrations were way too high. Not much data though!

[HuiWangWanderInGitHub]

I think the low isoprene issue could be related to the drought stress algorithm implemented in PR #2588. That algorithm reduces isoprene emissions when drought conditions occur. One thing you could try is adding "GAMMAS" to the output list to check the impact of drought. As mentioned in my JAMES paper (https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2022MS003174), the CLM model can exaggerate the severity of drought, which can significantly reduce isoprene emissions. So, I would suggest validating the drought simulations for these cases at the same time. If CLM has difficulty simulating water stress accurately, it's also reasonable to turn off drought stress, especially if it's not the focus of your study.

I haven't integrated the MEGAN3 features into the current CLM model yet, since it's not simply a matter of updating the emission factors. It’s still using MEGANv2.1 with the same emission factors. So I’ve only made some scientific representation changes for the Arctic region and drought stress.

Regarding tropical regions, in this simulated tropical forest environment study (https://www.nature.com/articles/s41586-022-05020-5), isoprene also increases under mild drought and decreases under severe drought, which is consistent with observations in temperate forests. So I assume it would be the same, but we definitely need more flux-level data to validate this for tropical forests.

[wwieder]

Thanks for this suggestion. Can you suggest how to turn off the drought stress for a sensitivity test we can try?

[HuiWangWanderInGitHub]

Sure, it is quite simple. Just change line 568 in clm/src/biogeochem/VOCEmissionMod.F90 from gamma_sm = get_gamma_SM(btran(p)) to gamma_sm = 1.0_r8. This will turn off the impact of water stress.

[tilmes]

@wwieder is this a namelist parameter we can change? Should we do this on top of the changes you suggested earlier for beta05? Thanks!

[wwieder]

I think for a sensitivity test you'd want to change this line of code in sourceMods, Simone, so gamma_sm = 1.0_r8, as suggested by @HuiWangWanderInGitHub.

CTSM/src/biogeochem/VOCEmissionMod.F90

Line 568 in 3fe643a

gamma_sm = get_gamma_SM(btran(p))

For now, I suggest repeating your F-case in SP mode using the beta05 tag & namelist changes + adding the source mod in VOCEmissionMod to see if that provides VOC emissions that you're hoping to see.

Down the road, if turning off the soil moisture control over isoprene emissions is a feature that's helpful to have for the CESM3, we can make this a namelist flag so users can more easily this change.

[ekluzek]

@tilmes @wwieder does this issue affect the current set of coupled simulations that are going on? Or is it just for certain CAM-Chem configurations? We are now asking this question in CSEG meetings, and I wanted to make sure we knew how important these impacts are. Thanks

[tilmes]

Hi Erik, yes it would impact the coupled simulations as well, I am not sure how much. We are also running now a run to looking at specified oxidant chemistry to compare with the T4S chemistry run, Simone

…

On Tue, Mar 25, 2025 at 2:01 PM Erik Kluzek ***@***.***> wrote: @tilmes <https://github.com/tilmes> @wwieder <https://github.com/wwieder> does this issue affect the current set of coupled simulations that are going on? Or is it just for certain CAM-Chem configurations? We are now asking this question in CSEG meetings, and I wanted to make sure we knew how important these impacts are. Thanks — Reply to this email directly, view it on GitHub <#3016 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AFJRPITL7K7PVQO74DZRYID2WGYYFAVCNFSM6AAAAABZC4DKXCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDONJSGM4DCNJWHE> . You are receiving this because you were mentioned.Message ID: ***@***.***> [image: ekluzek]*ekluzek* left a comment (ESCOMP/CTSM#3016) <#3016 (comment)> @tilmes <https://github.com/tilmes> @wwieder <https://github.com/wwieder> does this issue affect the current set of coupled simulations that are going on? Or is it just for certain CAM-Chem configurations? We are now asking this question in CSEG meetings, and I wanted to make sure we knew how important these impacts are. Thanks — Reply to this email directly, view it on GitHub <#3016 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AFJRPITL7K7PVQO74DZRYID2WGYYFAVCNFSM6AAAAABZC4DKXCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDONJSGM4DCNJWHE> . You are receiving this because you were mentioned.Message ID: ***@***.***>

-- *Simone Tilmes,* *Atmospheric Chemistry, Observations & Modeling LabNational Center for Atmospheric Research* * PO Box 3000* *Boulder, Colorado 80307-3000* *303-497-1445* *303-497-1400 (fax) * ***@***.*** ***@***.***>*

[rosiealice]

Looping myself into this conversation as we currently have isoprene emissions working in NorESM as well (linked up to FATES, which changes a bit about how they are calculated but not that much).

[rosiealice]

Also would one want to turn gammas off altogether? Or just make it less intense (i.e. by taking it's square root?) . Turning if off altogether seems like an extreme thing to do...