some strange behavior of Energy Service|Transportation

[orichters]

Dear colleagues, in the NGFS scenarios, we see in some regions shocks in the variable Energy Service|Transportation|Freight|Railways that do not translate into emissions, so the emissions intensity per billion_tkm is affected by this shock. An example for China is attached.

Similarly, for some countries, very pronounced shocks are observed either at the beginning of the series (scenarios d_strain and d_delfrag, both are delayed transition scenarios with cm_startyear = 2030) or at the end (o_1p5c = 1.5°C and o_2c = Below 2°C). In the first case, for example, the number of billion_tkm is doubled in Latin America or the Middle East & Africa between 2035 and 2040 in these scenarios for Navigation or for road in the case of reforming economies. This pattern is also observed at the global level for Navigation.

In the case of o_1p5c and o_2c, high intensity compensatory patterns between sectors (navigation and road in Japan, for example) of are observed at the end of the series. On the one hand, there is an intertemporal compensation between Navigation and Road between 2090 and 2100. Furthermore, the relevance of the series change for each scenario, but I don't understand why in similar scenarios like 1.5C or 2C there is so much disparity between such diverse modes of transportation in the last part of the time series. In the case of EU-28, there is intertemporal compensation but no reversal of the importance of the series.

I would expect more stable trends or for them to be consistent in emissions and usage assuming similar technology between the periods. Comments warmly welcome, @johannah-pik, @jmuessel, @robertpietzcker

[jmuessel]

Could you provide a graph on FE demand or a path to the runs, so that I can dig deeper into the transport outputs? What are the underlying scenarios for transport, could you provide a path to the scenario config?

[orichters]

Sure, sorry, the data is at /p/tmp/oliverr/NGFS_v5/remind-2024-03-27/output, always the -rem-4 runs. See /p/tmp/oliverr/NGFS_v5/remind-2024-03-27/compScen-NGFS-rem-4-2024-04-05_08.52.34-REMIND-MAgPIE.pdf for a compare scenarios. The config is /p/tmp/oliverr/NGFS_v5/remind-2024-03-27/config/scenario_config_NGFS.csv.

[jmuessel]

Thank you! I will have a brief look. We are currently testing the refactored EDGET version, it will be released soon. The new version might already solve issues like this one. So I would wait until the release - if that's ok for you. I added getting back to you about this issue on my to-do list for you after the release.

[orichters]

We have to submit mid May. Don't know whether your "soon" is within the next two weeks? You can also tell me how to use the new EDGE-T version and I rerun and see whether the problems persist.

[jmuessel]

We do our best to provide the new version asap, getting back to you then

[johannah-pik]

Two weeks is maybe a bit tight for submission and bugsearch if the problem persists.. Especially, when its likely to be related to the delayed transition with cm_startyear = 2030. Do you think these variables are really necessary to be reported in the submission? Can you quickly state which transport scenarios do you use?

[orichters]

Sure: Mix2 for h_cpol and d_strain, Mix3 for h_ndc and d_delfrag, and Mix4 for o_2c, o_1p5, o_lowdem.

[orichters]

So the uptick in Navigation seems to be a Mix 4 issue.

The 2040 peak might be related to the delayed transition with cm_startyear = 2035, but I don't really understand why.

[johannah-pik]

We will look into the new Mix4 pathways with the standalone version.. If it occurs there as well, its easy to fix. If not, I fear this will take us longer than your deadline would require..

[orichters]

Ok, in case the new release isn't ready then, I would just take out the variables that look fishy to me.

[johannah-pik]

Maybe thats the best. Sorry! Hope to bee soon in another position