Summation errors due to CDR-related energy use

[katarkow]

Due to the assumption that emissions from using fegas to produce heat for DAC are captured, there are summation errors in the reporting. It happens because vm_emiTeMkt accounts for burning all that fegas used for DAC and also synfuels from that captured carbon, so we account for that carbon twice. I thought of the following solutions:

(1) Updating the vm_emiTeMkt in REMIND, so discounting these emissions * share CCU (https://github.com/remindmodel/remind/pull/1589) or (2) Discounting the CCU from gas for DAC in remind2, (a) either from vm_emiTeMkt and vm_emiAllMkt directly where they’re declared #553 or (b) everywhere where they’re used, so Emi|CO2, Emi|CO2|+|Energy, Emi|GHG, …. (potentially painful to remember about this when a new reporting variable uses either of the REMIND variables).

What are your thoughts? @fschreyer @amerfort @robertpietzcker @strefler @LaviniaBaumstark

[katarkow]

I tried (1) here, I thought it would be the cleanest. However, there's a small accounting error in the reporting from using sm_eps to handle cases in which there’s no captured CO2 in a given (t, regi): share = ccs / (captured + sm_eps). Anybody has a better idea to handle this? :)

(2a) Is here: #553

[fschreyer]

Hi Kasia,

thanks for looking into it.

in my understanding it needs both, a change in REMIND and a change in the reporting.

I think, it should come down to the difference between vm_emiCdrTeDetail and vm_ccs_cdr which are not used consistently across REMIND and the reporting. Both variables are used to denote captured carbon from DAC. Looking at q33_DAC_ccsbal, the difference seems to be that the second includes captured carbon from gas heating use for DAC, while the first does not.

I think there are different ways to solve that, but here is my suggestion:

(1) in REMIND: In q_emiAllMkt the vm_emiCdr term should be replaced by vm_ccs_cdr since gas use for DAC causes emissions in q_emiTeDetailMkt here which need to be subtracted from total emissions at some point (These emissions are reported as Emi|CO2|Energy|Demand|+|CDR). As captured CO2 from DAC is subtracted in q_emiAllMkt, I would simply change it there.

(2) in the reporting for total energy emissions summations: I think that vm_emiCdrTeDetail needs to be replaced by vm_ccs_cdr here (or the difference between the two needs to be added) where the energy emissions coming out of REMIND are corrected for the DAC carbon that goes into synfuels.

(3) in the reporting for the carbon capture summations: Same here, vm_emiCdrTeDetail should be replaced by vm_ccs_cdr since q_balCapture which calculates the captured carbon vm_ccs_cdr is used. It is consistent to add the gas heating carbon there as long as the gas use causes emissions in q_emiTeDetailMkt as described in (1).

I hope I do not overlook something. Let me know if you think that causes some other inconsistencies.

There are certainly other ways in which this could be resolved (e.g. redefine vm_emiCdrTeDetail to include captured carbon from gas or remove CDR energy emissions from q_emiTeDetailMkt). But, as it is such a small amount and we only want variable summations to work, I would be in favor of being pragmatic and choose some option with the least intrusive changes.

[LaviniaBaumstark]

what is the status of this issue? Are you fixing it for the REMIND 3.3.0 release?