search

CliMA / ClimaLand.jl

Clima's Land Model
Apache License 2.0
36 stars 9 forks source link

O2.6.3 Add SIF to Big Leaf model #595

Open kmdeck opened 5 months ago

kmdeck commented 5 months ago

We would like to calibrate the big leaf model with SIF data. To do so, we need a model predicting SIF.

Refer to discussion in https://github.com/CliMA/ClimaLand.jl/issues/152 for notes on SIF modeling

braghiere commented 5 months ago

We will have the first meeting about moving @Yujie-W's RT version to current CliMA Land next week on Tuesday at 4pm. @AlexisRenchon is leading this, so I am adding him as an assignee.

kmdeck commented 5 months ago

We will have the first meeting about moving @Yujie-W's RT version to current CliMA Land next week on Tuesday at 4pm. @AlexisRenchon is leading this, so I am adding him as an assignee.

ok! That's great. I think that is a separate effort, though (and not currently tracked on our milestones or OKRs). By this I mean that we should add a parameterization for SIF (even if simple, not ideal) to the existing model, and making the canopy model more realistic/complex is a parallel effort

braghiere commented 5 months ago

We will have the first meeting about moving @Yujie-W's RT version to current CliMA Land next week on Tuesday at 4pm. @AlexisRenchon is leading this, so I am adding him as an assignee.

ok! That's great. I think that is a separate effort, though (and not currently tracked on our milestones or OKRs). By this I mean that we should add a parameterization for SIF (even if simple, not ideal) to the existing model, and making the canopy model more realistic/complex is a parallel effort

On the contrary, this is highly related! Especially because @Yujie-W's RT code is the best SIF model that exists out there and it has been thoroughly tested and showcased in @eviatarbach's paper. Of course a few adaptations (simplifications) will be needed for this particular task, but these efforts are completely aligned. Thanks for bringing this up! :)

kmdeck commented 5 months ago

ok, maybe we can discuss more what the work entails! @AlexisRenchon already has several tasks on his plate related to clima land milestones due this year (diagnostics, fluxnet site calibrations,...). Do you mean this would be only adding the RT model as another option within the existing big leaf model (in addition to Beer-Lambert, and the 2 stream scheme)? Or do you mean that Alexis is going to be leading the integration of Yujie's entire model into ClimaLand? The latter will be a large effort and is not something we budgeted resources for. The former is smaller, but I think we should still assess the amount of time it will take and where it lies in terms of priorities.

AlexisRenchon commented 5 months ago

Thanks all, I think this is a good discussion to have. I will talk more about it with Tapio. For now my intention is just to have regular meeting (30 min to 1 hour once a week to start, maybe bi-weekly eventually), with Christian Yujie Renato and sometimes other, where we actively work (pair coding) to integrate Yujie's model as an alternate canopy module. I am not sure yet how feasible it is - we will see during the meetings. My intention is for nobody to work on this outside of these meetings, for now.

braghiere commented 5 months ago

Just to make it clear that both models are not that different, apart from the RT code. So, my suggestion would be to start to implement a 3rd RT scheme in CliMALand.jl based on Yujie's model. This would already unlock so many new scientific possibilities, including SIF estimates, which is a requirement. Happy to discuss further and help out as needed.