calcIO does not return ngt or ngcc generation data (only gaschp and x_gas2elec)

[stephenbi]

As a result, calcIO gives <25% of global gas power generation compared against "SE|Electricity|+|Gas (EJ/yr)" from calcEmber. First guess: it's probably related to this commit.

dimSums(seelgen[, 2000:2020, "pegas.seel.gaschp"], 1) / dimSums(ember_nat[, 2000:2020, "SE|Electricity|+|Gas (EJ/yr)"], 1)

[pfuehrlich-pik]

I have really no idea what's going on here, so I'd leave this to you @fbenke-pik :)

[fbenke-pik]

subtype here: output
seelgen <- calcOutput("IO",subtype="output",aggregate = FALSE)[,,"seel"]

Looks to me like this is not part of the mappingfile

i.e. none of the IEA products and flows (column C and D) is mapped to the correct REMIND in, out, tech combination (columns E,F,G).

You will need to create the correct mapping entries for this (maybe @Renato-Rodrigues or @robertpietzcker can help)

[stephenbi]

OK, looks like the behavior is correct, just confusing and poorly documented. The x_gas2elec technology is apparently a catchall for non-CHP gas plants, which then gets assigned completely to ngcc in REMIND (though the comments suggest it was previously assigned completely to ngt). So I guess it's worth asking, @robertpietzcker and @Renato-Rodrigues: is this mapping convention a matter of necessity or tradition at this point?

[robertpietzcker]

I don't get the initial comment.
To me it looks like total gas-based power generation matches pretty well with EMBER data:

so I must somehow be misunderstanding the original issue. Could you try to explain once more, Steve?

2. IEA data doesn't differentiate between ngt and ngcc, so the input data can't split this information - thus it is all assigned to x_gas2elec.

IN REMIND itself, it is then distributed in module 04/datainput from line 375 on according to the share of gas electricity in total power generation, as a low share will likely lead to a mostly peaking role, while a large share means much more plants with high full load hours such as ngcc:
https://github.com/remindmodel/remind/blob/8b9542edfc9d4d86391ccf64181b58ef6761dea5/modules/04_PE_FE_parameters/iea2014/datainput.gms#L375

[stephenbi]

@robertpietzcker No no, you're right, we've since figured out that there's no issue with data or functionality. I just wasn't aware that x_gas2elec is a placeholder and not a tech, and there's no description for it in core/sets.

So, this is just a low priority question on documentation/mapping convention (sorry, only labeled it as such now). Since x_gas2elec is mapped 1-to-1 with ngcc within REMIND, is it really necessary?

[robertpietzcker]

Since x_gas2elec is mapped 1-to-1 with ngcc within REMIND, is it really necessary?

it is not :-)
as written under "IN REMIND itself ..." above, it is a few lines below the original mapping to ngcc split into ngt and ngcc.
it would be great to add a line
!! further below, the values are split between ngt and ngcc - search for "Distribute the initial gas numbers to ngcc and ngt"
in the REMIND code :-)

[fbenke-pik]

Stephen, is this still current? If not, please close. Otherwise, anything to do on my end?