Incorporation of the simulation of solar-induced chlorophyll fluorescence (and some modifications of radiative transfer)

[RongLi29]

We incorporated simulation of solar-induced chlorophyll fluorescence (SIF), along with some modifications in canopy radiative transfer and some model parameters. The modifications are described in our paper accepted by JAMES https://doi.org/10.1029/2021MS002747.

Scientific background

SIF is faint radiation emitted by chlorophyll as a byproduct of photosynthesis, and global observations of SIF are available from several recent satellites. Due to the link between SIF and photosynthesis, there is a potential to use SIF observation to evaluate or constrain the simulation of photosynthesis. We incorporate the simulation of SIF into CLM, and provide SIF radiance at 740 nm at nadir direction, which can be directly compared to many satellite SIF products. In our simulation, we use a mechanistic way to scale SIF from leaf-level to top-of-canopy at the nadir direction by taking into account canopy scattering, the bidirectional effect, and clumping.

Below are the modifications we made:

1. Simulation of SIF

We added simulations of 1) total leaf-level SIF emission, 2) SIF radiance at nadir direction, and 3) SIF escapes the canopy from the full upper hemisphere. Major modifications are in PhotosynthesisMod.F90.

2. Simulation of reflectance at nadir direction

We added simulation of nadir reflectance to the canopy radiative transfer scheme, providing outputs: 1) refd: Flux scattered at nadir above canopy (per unit direct beam flux), 2)refi: Flux scattered at nadir above canopy (per unit diffuse beam flux). These simulations are needed for simulation of nadir SIF. Major modifications are in SurfaceAlbedoMod.F90.

3. Incorporation of clumping index

We consider clumping for the radiative transfer simulations. An additional field "CI_pft" is added to the input parameter file. PhotosynthesisMod.F90 and SurfaceAlbedoMod.F90 are modified.

4. Some modifications related to the calculation of leaf APAR.
a) modification of the PAR/SR ratio for leaf APAR simulation
In CLM land-only mode, the ratio of visible to total incident solar radiation is set to be 0.5, and the visible radiation is used for calculation of APAR. But the ratio (0.5) is too high according to literature (e.g., Tsubo & Walker, 2005), and it affects simulation of photosynthesis and fluorescence. Based on measurements at 31 Ameriflux sites (text S6 and Table S1 in supplementary information of our paper), we set the ratio between incident PAR and shortwave radiation to 0.435 for simulation of leaf APAR, which is used for simulation of photosynthesis and fluorescence (in SurfaceRadiationMod.F90). But we did not changed the visible or NIR radiation for other parts of CLM.

b) added calculation of PAR absorbed by leaf only
Ideally, PAR absorbed by leaf only should be used for simulation of photosynthesis and fluorescence. But in current CLM, APAR is only calculated as total absorbed PAR (by leaf, stem, and snow) devided by eLAI+eSAI, so there is impact of snow and stem on the calculated APAR. We added calculation of PAR absorbed by leaf only in SurfaceAlbedoMod.F90 and SurfaceRadiationMod.F90, and used this leaf APAR for simulation of photosynthesis and fluorescence. (Details are in text S7 in supplementary information of our paper).

Files we modified: I currently have our code modified based on CESM release cesm2.1.1-rc.05 at https://github.com/RongLi29/CLM5-SIF

PhotosynthesisMod.F90: We added subroutine "fluorescence" to simulate leaf-level fluorescence emission and subroutine "fluorescence_upscaling" to calculate canopy-level SIF. Subroutines "Photosynthesis" and "PhotosynthesisHydraulicStress" are modified to simulate several parameters needed for the "fluorescence" subroutine. Other subroutines are modified for the added variables.

SurfaceAlbedoMod.F90: Major modifications are in the subroutine "TwoStream". We added simulations of nadir reflectance and radiative transfer parameters needed for the simulation of nadir SIF. We also incorporated clumping index and modifications of leaf APAR calculation.

SolarAbsorbedType.F90: added variables needed for the modified leaf APAR calculation

SurfaceAlbedoType.F90: add variables and outputs related to radiative transfer

SurfaceRadiationMod.F90: add variables for PAR absorbed by leaves only (not affected by stem/snow); PAR/SR set to be 0.435 instead of 0.5 for leaf APAR simulation. Note: we only changed the leaf APAR calculation for simulation of photosynthesis and fluorescence. In other parts of CLM, visible and NIR radiation were still derived from the 0.5 ratio.

UrbanAlbedoMod.F90: Account for urban landuints for added variables (ftnn, ftin, refd, refi, refd_gr, refi_gr)

CanopyFluxesMod.F90: add input variables when calling PhotosynthesisTotal (for calculation of canopy scattering)

pftconMod.F90: add Clumping index for each pft as an input

An additional field "CI_pft" is added to the input parameter file.

Collaborators: @danicalombardozzi , Mingjie Shi, Christian Frankenberg, Nicholas C. Parazoo , Philipp Köhler, Koong Yi , Kaiyu Guan, and @geoxiyang.

Let me know if there is other information I should provide or if you would like me to open a pull request.

[wwieder]

Thanks for making these contributions, @RongLi29. It seems like you're ready to open a PR for this work?

We have suggestions for how to get started on the CTSM github page that may be helpful, also see the "development guides" on the sidebar of the wiki.

If it's productive to schedule a call with @ekluzek and others to clarify things early in the process we can find a time to make this happen.

Have you been able to evaluate how the code modifications you're proposing modify existing calculations of GPP and the rest of the carbon cycle, or does this just calculate a new diagnostic variable that's more similar to SIF measurements?

Depending on the extent of these changes, I also wonder how difficult it would be to bring these modifications over into FATES once they're implemented here? Would this be helpful for your development @rgknox, @ckoven, @rosiealice, @glemieux, @jkshuman & @adrifoster?

Finally, I wonder if you'd be willing to present at a CLM meeting this spring. I feel like it's been a while since we heard an update from this group?

[wwieder]

Also, I noticed there seem to be a number of options and different parameterizations that it looked like you considered in your github repository. Can these be simplified into a preferred configuration? Are there namelist switches we can introduce if users should have the flexibility to consider different model assumption (e.g. clumping, Vcmax, etc).

Also, I wondered how this code interacts with LUNA, since it looks like you may be hardwiring Vcmax to results from TRY? These are conversations we can also have during a presentation.

[RongLi29]

I think we are close to opening a PR. But I still need to update the code according to the most up-to-date CTSM version. Also, I agree it would be better if we can use namelist to switch between options, and some modifications are needed for that. I am considering an option to switch between the two fluorescence models we considered and an option to chose whether clumping is considered. We can also discuss if we want other namelist options.

It may be good to schedule a call, my schedule is quite open. I am also happy to present at a CLM meeting. Has it been scheduled?

Regarding the impact on carbon cycle, our modifications of canopy radiative transfer affect APAR, so they also affect the simulation of GPP. Here I attach a figure indicating that our modification of PAR/SR ratio for leaf APAR calculation and the incorporation of clumping index can each reduce GPP by ~10% in some regions.

For Vcmax, we hardwrote those and turned off LUNA in the simulations mostly for the purpose of a sensitivity test. I am not very familiar with LUNA, but I think it is not necessary to incorporte our change of Vcmax parameters if we have some confidence in LUNA.

[geoxiyang]

Hi all, We would also love to discuss the possibility to include SIF and canopy reflectance (400-2500 nm) to FATES if that is something you are interested in.

-Xi

On Fri, Mar 4, 2022 at 4:34 PM Rong Li ***@***.***> wrote: I think we are close to opening a PR. But I still need to update the code according to the most up-to-date CTSM version. Also, I agree it would be better if we can use namelist to switch between options, and some modifications are needed for that. I am considering an option to switch between the two fluorescence models we considered and an option to chose whether clumping is considered. We can also discuss if we want other namelist options. It may be good to schedule a call, my schedule is quite open. I am also happy to present at a CLM meeting. Has it been scheduled? Regarding the impact on carbon cycle, our modifications of canopy radiative transfer affect APAR, so they also affect the simulation of GPP. Here I attach a figure indicating that our modification of PAR/SR ratio for leaf APAR calculation and the incorporation of clumping index can each reduce GPP by ~10% in some regions. [image: GPP_PAR_CI] <https://user-images.githubusercontent.com/54550132/156840084-a3ae44ec-a90c-49a1-beb0-73cf39ca1fa1.png> For Vcmax, we hardwrote those and turned off LUNA in the simulations mostly for the purpose of a sensitivity test. I am not very familiar with LUNA, but I think it is not necessary to incorporte our change of Vcmax parameters if we have some confidence in LUNA. — Reply to this email directly, view it on GitHub <#1672 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB3ETLV572TZDH4PISXH3ZTU6J6VXANCNFSM5P3QB37A> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>. You are receiving this because you were mentioned.Message ID: ***@***.***>

-- Sent from my phone; sorry for any typo

[danicalombardozzi]

Thanks for opening this issue @RongLi29! I agree that a recommended configuration and namelist switches will allow you and other users flexibility in using the new code you developed. It seems like submitting a pull request is the next step! @wwieder pointed you to some of the same resources I shared via email, and I'll reiterate that we can set up a meeting to talk through the namelist options, etc. if you'd like to do so before you submit a pull request.

[RongLi29]

Thank you for the support @wwieder @danicalombardozzi! I will start preparing for the pull request. A meeting discussing some details such as the namelist options would be helpful. Can we schedule one in the coming weeks?

[wwieder]

Yes, I'll work on scheduling a meeting. Would it be helpful to have any of your collaborators outside of NCAR on the call @RongLi29 ? We'll discuss who should be involved on our end too, both for CLM and potentially FATES?

[RongLi29]

Thank you! I will check will our collaborators. We can have discussions for both CLM and FATES.

[dlawrenncar]

It seems like it would be useful to have a science presentation on this at a CLM science meeting and then a separate technical discussion about the code. Not sure what order would be best for this, but I kind of think the science meeting would be a higher priority. That will help us understand what the benefits are and what the impacts on the simulations are. Might be worth trying to run with LUNA prior to that meeting just to confirm that the behavior is similar.

…

On Sat, Mar 5, 2022 at 8:56 AM Rong Li ***@***.***> wrote: Thank you! I will check will our collaborators. We can have discussions for both CLM and FATES. — Reply to this email directly, view it on GitHub <#1672 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AFABYVFN2EWT4Y622EJJWLLU6N72BANCNFSM5P3QB37A> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>. You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[RongLi29]

Sounds good. I am happy to do a science presentation and can run a few simulations with LUNA next week.

[wwieder]

Can we tentatively plan on a March 17 or March 31 presentation at 1:00 MST? @dlawrenncar this would be in lieu of a PPE meeting, Is this OK?

[RongLi29]

Either works for me.

[danicalombardozzi]

I won't be able to attend either of those meetings.

…

On Sun, Mar 6, 2022 at 8:39 AM Rong Li ***@***.***> wrote: Either works for me. — Reply to this email directly, view it on GitHub <#1672 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AGHW2QKCB36Z346JB3DZP7DU6TGRJANCNFSM5P3QB37A> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>. You are receiving this because you were mentioned.Message ID: ***@***.***>

-- Dr. Danica Lombardozzi she/her/hers Terrestrial Sciences Section Climate and Global Dynamics National Center for Atmospheric Research Boulder, CO 80305 email: ***@***.*** office: (303) 497-1777

[ekluzek]

A subgroup met, and @RongLi29 is going to add sif_method as a namelist item in PhotosynthesisMod to control which dataset is used. There's a few other changes, and then we'll start working on turning this into a branch off of CTSM that will based off the release-clm5.0 version the work is based off of. We'll then make a PR based off the release-clm5.0 version that we may make other changes to. Then we'll figure out a plan for updating to the latest CTSM version and make a new PR for that, which will be brought into our usual development cycle.

[wwieder]

👍 thanks for this update

On Fri, Apr 15, 2022 at 5:33 PM Erik Kluzek ***@***.***> wrote: A subgroup met, and @RongLi29 <https://github.com/RongLi29> is going to add sif_method as a namelist item in PhotosynthesisMod to control which dataset is used. There's a few other changes, and then we'll start working on turning this into a branch off of CTSM that will based off the release-clm5.0 version the work is based off of. We'll then make a PR based off the release-clm5.0 version that we may make other changes to. Then we'll figure out a plan for updating to the latest CTSM version and make a new PR for that, which will be brought into our usual development cycle. — Reply to this email directly, view it on GitHub <#1672 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB5IWJF7N6ZNMMO42PFFZOTVFH4FTANCNFSM5P3QB37A> . You are receiving this because you were mentioned.Message ID: ***@***.***>

-- Will Wieder (he/his) Project Scientist & Land Model Working Group Co-Chair CGD, NCAR 303-497-1352

[wwieder]

@RongLi29 we haven't heard any updates on this thread, which makes me think you have other priorities right now, but let us know if this is work you're still interested in developing and on what timeline. Thanks.

[RongLi29]

Hi Will, I'm really sorry for not providing updates. I currently have a few things going on and cannot work on it very soon. But I'm definitely still interested in developing. I'll aim at working on this end of the year if that's okay.

[wwieder]

That's fine. Thanks for following up. I just wanted to follow up because it's been a while since we heard from you.