search

remindmodel / remind

REMIND - REgional Model of INvestments and Development
Other
99 stars 129 forks source link

Exposing SE electricity tax rate tech specific ramp up logistic function parameters as a switch #1668

Closed Renato-Rodrigues closed 4 months ago

Renato-Rodrigues commented 5 months ago

Purpose of this PR

Checklist:

Further information (optional):

/p/projects/ecemf/REMIND/2040_scenarios/v05_2024_04_09_rev1/compScen-h2Tax-2024-05-10_10.40.15-EU27.pdf

Renato-Rodrigues commented 4 months ago

Here are the results of different parametrizations of the b parameter:

run name type se_trade b
05_Nzero_57_bio7p5_EU27 old reference no 10
05_Nzero_57_bio7p5_seTrade_EU27 new reference yes 10
05_Nzero_57_bio7p5_seTrade_EU27_off par test yes disabled
05_Nzero_57_bio7p5_seTrade_EU27_b30 par test yes 30
05_Nzero_57_bio7p5_seTrade_EU27_b20 par test yes 20
05_Nzero_57_bio7p5_seTrade_EU27 par test yes 10

/p/projects/ecemf/REMIND/2040_scenarios/v05_2024_04_09_rev1/compScen-h2Tax-2024-05-10_10.40.15-EU27.pdf

image

@fschreyer and @robertpietzcker Should we disable the elh2 tax rate (05_Nzero_57_bio7p5_seTrade_EU27_off scenario) or increase its default value (05_Nzero_57_bio7p5_seTrade_EU27_b20 or 05_Nzero_57_bio7p5_seTrade_EU27_b30)?

there is not much effect on bio based hydrogen although.

robertpietzcker commented 4 months ago

This seems very different from the effects that Felix observed (if I remember correctly), where adding the tax strongly decreased electrolysis-based hydrogen.

or do I remember your results wrongly, @fschreyer ?

actually, I don't get the results - shouldn't the second bar (triangle) be with the current default tax, and the third bar (plus) with the tax turned off? why then does the one with the tax turned off have less electrolysis and more biomass hydrogen?

Renato-Rodrigues commented 4 months ago

actually, I don't get the results - shouldn't the second bar (triangle) be with the current default tax, and the third bar (plus) with the tax turned off?

Yes.

why then does the one with the tax turned off have less electrolysis and more biomass hydrogen?

I see the opposite of what you describe in EUR by 2050. The plus bar (tax turned off) has more electrolysis (middle tone green) than the default parametrization (second bar, triangle shape). The biomass (light green) values are stable, and there is no gas anymore with the tax turned off or with the reduced tax effects.

In Germany, more or less the same happens but at a lower degree due to the imports.

robertpietzcker commented 4 months ago

I see the opposite of what you describe in EUR by 2050.

haha, funny, I didn't look at the legend but somehow from memory associated the light green with electrolysis. Sorry :-)

ok, I would put the default value of b to 30. with b=30, the share of electricity used for hydrogen is ~10-20% in most regions in 2050 (although only 4% in FRA and DEU), which I think is in a reasonable range for a high-VRE power system where electrolysis still has a stabilizing impact on the power system and is thus likely treated favorably by the tax system.

Renato-Rodrigues commented 4 months ago

ok, I would put the default value of b to 30.

https://github.com/remindmodel/remind/pull/1668/commits/1b86906386c55ecbf44005eb1b1f2266379f771d

fschreyer commented 4 months ago

ok, I would put the default value of b to 30.

I agree for EU that the policy scenarios look ok. I am not sure whether we should do that in general. The thing is if we have so much tax exemptions for hydrogen globally, Npi scenarios will probably see a lot of e-fuel production in MEA etc. again, which we deem unrealistic (see here). I would rather only change it to 30 for EU and leave the rest of the world as it is. We can also make it a tax policy which is only on in policy runs, but I would not like to decide that here.

there is not much effect on bio based hydrogen although.

Actually, you do not see so much difference across your scenarios because the demand-side/CCS/bioenergy assumptions are more important than the supply price. If you do a low CCS/bio run that needs to phase-out fossil fuels even more, then higher hydrogen penetrations are to be expected and I would presume the model goes to higher hydrogen prices/quantities even with the same tax parameters.

On the biogenic H2: The levels in those runs on global level are comparable to what we had before the H2 update. They actually did not change that much. However, in Ariadne we used to have lower biogenic H2 because we had low Bio/CCS assumptions that forced a fossil phase-out. In your net-zero runs, quite some fossils are still used in 2050 and after, which frees more bioenergy to be used elsewhere.

Renato-Rodrigues commented 4 months ago

ok, I would put the default value of b to 30.

I agree for EU that the policy scenarios look ok. I am not sure whether we should do that in general. The thing is if we have so much tax exemptions for hydrogen globally, Npi scenarios will probably see a lot of e-fuel production in MEA etc. again, which we deem unrealistic (see here). I would rather only change it to 30 for EU and leave the rest of the world as it is. We can also make it a tax policy which is only on in policy runs, but I would not like to decide that here.

Do you want me to set a NPi run to double check that effect?

I can also just change the default values to be b=10 globally, and b=30 only to EUR regions if you two agree with that. Just let me know.

robertpietzcker commented 4 months ago

well, if you are not using the cluster otherwise, I think having an NPi to see if the synfuel-problem appears again at b=30 (and goes away at b=10) would be useful.

In Renatos NetZero runs, the synfuel prices in MEA are around 40-60$/GJ, while oil prices stay around 15$/GJ - so I would expect that very little synfuel comes in NPi scenarios

(if time is scarce, I am also fine with having b=30 only in EU and 10 elsewhere, but this may bring critical questions in the future)

Renato-Rodrigues commented 4 months ago

Below you can find a current NPi run with variations of the above mentioned parameter. These results should change a bit after a new calibration, as they assume the model calibrated under the previous parametrization (b=10 if I am not mistaken, and not b=30 as we are proposing).

image

I wouldn't say that the values under b=30 are that crazy for the end of the century. Opinions? @fschreyer @robertpietzcker

Renato-Rodrigues commented 4 months ago

you can find the test runs and comparison pdfs for the NPi scenario here: /p/projects/ecemf/REMIND/trunk/remind_14_05_2024_h2TaxTests

robertpietzcker commented 4 months ago

looks good to me (given that the NPi doesn't represent a "no policy" world but rather "starting from today, with weak policies" - I think it is not unrealistic that in the long term (after 2050), more expensive oil resources may be replaced by synfuels...

it might be interesting to get a comment by @nicobauer on this :-)