grassland, forest module and soil land use processes for a closed land carbon cycle

[jnnsbrr]

No description provided.

[chrisdwells]

Thanks Jannes! I started making notes on this before I realised it was a draft version, so thought may as well send over anyway though I'm sure you're aware of the below- cheers!

--

Undefined units; think it'd be good to get those in so it's more readily understandable to the group.

In "soil carbon decay" there's "initial surface temperature anomaly", which is a set value and the units aren't degC. With impacts generally we've been locking things to pre-industrial levels so there's variation between members in 1980 and to make it less sensitive to the 1980 temperature anomaly. If there's good reason to change from that for some impacts then we definitely can, but good to chat through if that's OK. Those units aren't right for the temperature anomaly, and we should be careful as that anomaly (ie 1980 value) is dependent on the run. Also in this module, the decay rates are 1/year for fast and slow, but dmnl for litter - should all be 1/year?

In "cropland soil carbon" there's "soil carbon emissions cropland GtC" which is forced to be positive; it hits zero in a lot of the run - what's the rationale for not going negative?

Just in terms of the naming, there's a "soil carbon emissions" module which has soil carbon emissions and litter emissions - should the module or one of those be renamed to emphasise the sum of soil and letter (unless I'm misunderstanding the process)?

[jnnsbrr]

Dear @wasbridge,
on the issues you mentioned:

• I added units (I guess almost everywhere)
• Together with @chrisdwells I added a fix for initial surface temperature in the climate module
• Now also emissions could also be negative
• For the soil carbon emissions module - we can rename it, I am quite flexible here.

I still have one big problem which I tried to solve the whole day but failed ...
I added a terrestrial carbon balance module which works as a check that I am not missing any carbon throughout the system.
Terrestrial carbon balance over time should be zero, but it is not. It accumulates carbon which I must have missed somewhere but I cannot find the source (or sink) ... maybe you have a clue? Could also be that I made a mistake when doing the balance calculation ...

This is the major problem left, everything else should be almost ready to merge!

Thanks in advance :-)

[wasbridge]

I'm on vacation till Tuesday. I'll look then. My first instinct is a missing land use change flow. Thanks for all the hard work!

…

On Thu, Sep 19, 2024, 12:10 PM Jannes Breier ***@***.***> wrote: Dear @wasbridge <https://github.com/wasbridge>, on the issues you mentioned: - I added units (I guess almost everywhere) - Together with @chrisdwells <https://github.com/chrisdwells> I added a fix for initial surface temperature in the climate module - Now also emissions could also be negative - For the soil carbon emissions module - we can rename it, I am quite flexible here. I still have one big problem which I tried to solve the whole day but failed ... I added a terrestrial carbon balance module which is a check that I am not missing any carbon throughout the system. Terrestrial carbon balance over time should be zero, but it is not. It accumulates carbon which I must have missed somewhere but I cannot find the reason ... maybe you have a clue? This is the major problem left, everything else should be almost ready to merge! Thanks in advance :-) — Reply to this email directly, view it on GitHub <#34 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAGMI2VOVGPNTM24O7NXZPDZXLZRBAVCNFSM6AAAAABOCXZQVSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGNRRGQ2DAOBTHA> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[chrisdwells]

Jannes I'd say first thing is to sort out the units; needs doing anyway and that might show up the issue. Just conceptually though I'm not sure I see how the balance works, though it's completely not my field (ha ha). Would we expect the soil emissions to balance at the same timestep as the production? The emissions are driven by the carbon stocks, so don't directly match up with the production inputs?

[jnnsbrr]

Dear @chrisdwells,
thanks a lot for the first quick review!
I fixed the missing units.
I implemented the balance as a test to make sure no carbon gets lost along the way or no carbon is introduced out of nowhere into the system - these kind of tests are a sort of standard in larger land (carbon) cycle models. Still I am not sure if the test is valid in the way it has been implemented in a SD way. @wasbridge would be really helpful to talk to you on that next week. I think with this amount of carbon processes we have to make sure the model/module is consistent. Its easy to mess it up 😬

[lnnrtrmm]

Dear @jnnsbrr,

I had a look at the implementation this morning, but I also couldn't find the error in the accounting.

There is, however, another issue that worries me even more. The sum of the carbon stocks actually looses carbon over time (~24 GtC between 1980 and 2024), which is the opposite of what it should do, right? The land carbon sink took up ~25% of annual total co2 emissios during the last decades, whereas the land-use change emissions make up 10% of all emissions currently (20% in 1980). Hence, there should be a net increase of the total land carbon pool, which we don't have here. Neither when looking at the evolution of the stocks, nor in the "Cumulative terrestrial carbon balance". Is there generally something missing or is this because all this is not calibrated yet?

[jnnsbrr]

Dear @lnnrtrmm thanks a lot for your review!
And good though that you also did not find any major issues.
Yes right - that is what we meant. I am still not sure if we are missing something or if the approach itself is not valid. I really hope Billy can shed some light on the matter 😬

[wasbridge]

@jnnsbrr So I just fixed all the unit errors, and got the existing structure roughly in line with history again. With that fixed its easier to look at the problem.

The thing that sticks out to me here is the spikes from CO2 in the future in Net Primary Production are TOTALLY missing from "change in stocks". That points to just one source of the missing balance... those CO2 effects impact net primary production, and terrestrial carbon loss, but not soil carbon balance cropland, soil carbon balance grassland etc etc.

What I would do is take what I have commited and in your Terrestrial carbon balance module make balances for each land use type. i.e. for cropland I would have a cumulative terrestrial cropland carbon balance, which adds in the difference between cropland net primary production and crop production in GTC and all of the other changes (soil, production, residue etc). By looking at each land type individually you should be able to find out either what things you're missing in your balance calculation, or what things are making the balance go wacky. Going land use type by land use type should also help identify the issue raised by @lnnrtrmm