Results 12 June - cars

[Hertwich]

The GHG emissions are now in the right order of magnitude, several hundred Mt per year. There are some surprising results, however:

• Reduced material content (downsizing) leads to an increase both in the material-related and total emissions in SSP1 and SSP2 but not LED.
• EoL RR improvement leads to increased emissions of MaterialGHG incl. Recycling Credit in 2050.
• CaS reduces material-related emissions by ca. 1%, system emissions by 25% (SSP1 and LED).

[stefanpauliuk]

downsizing emissions go up for some scens!

Two things went wrong here. First, a mistake in scenario target table (broken links): Only SUVs for Other_ASIA, LAM, and SSA region, that was fixed now. Second, 3_SHA_DownSizing_Vehicles actually is a segment shift parameter, not a downsizing parameter. For mostly high-income regions, the segment shift we model results in downsizing (shift to smaller segments on average) but for others, notable India, it is the other way round: shift to larger segments for all scenarios, incl. LED (light trucks go up to 1%). That means, for regions like the US the no-downsizing case are the 2016 values left constant but for countries like India, it is the other way round: The target table values represent no downsizing and the 2016 values represent the downsizing. That swapping of values is done by ODYM-RECC according to a list supplied, and this list was now updated.

[stefanpauliuk]

EoL RR improvement leads to increased emissions of MaterialGHG incl. Recycling Credit in 2050.

There is still a mistake in the recycling calculation that causes this problem and that I need to fix asap. Am working on it.

CaS reduces material-related emissions by ca. 1%, system emissions by 25% (SSP1 and LED).

Model code error: Under CaS, total km driven were also reduced, but this is wrong!!! Under a car-sharing-only scenario, the vehicle-km travelled stay constant! -> mistake in model equation: wrong parameter value used. Fixed now. The total new vehicles and total EoL vehicles (sum over all types) are now also exported.

[stefanpauliuk]

Another discussion item: The now quite small impact of material substitution: This is because low-carbon el. aluminium is ca. on par with the newly introduced low-carbon-hydrogen steel. We need to discuss: Which one is more likely to work out in the midterm future? Price-wise? Large-scale H2-based steel may arrive in 2030 but this is higly uncertain, whereas low-carbon Al already exists today. Need to think about how to model deployment of low-carbon hydrogen realistically.

[Hertwich]

CaS reduces material-related emissions by ca. 1%, system emissions by 25% (SSP1 and LED).

Model code error: Under CaS, total km driven were also reduced, but this is wrong!!! Under a car-sharing-only scenario, the vehicle-km travelled stay constant! -> mistake in model equation: wrong parameter value used. Fixed now. The total new vehicles and total EoL vehicles (sum over all types) are now also exported.

@stefanpauliuk Car sharing leads participants to drive less given that the ownership costs are now felt at the margin. It reduces vehicle lifetime km because car sharers bulk the cars more than owners do. I think our data for the US was that driving went down and number of cars also went down, but not as much. @PaulWolfram confirm please. I think we can expect more robust/easily repaired vehicles, like the Citi Bikes, that will extend lifetime, once car firms make dedicated vehicles for car sharing, as they have started to.

[Hertwich]

Another discussion item: The now quite small impact of material substitution: This is because low-carbon el. aluminium is ca. on par with the newly introduced low-carbon-hydrogen steel. We need to discuss: Which one is more likely to work out in the midterm future? Price-wise? Large-scale H2-based steel may arrive in 2030 but this is higly uncertain, whereas low-carbon Al already exists today. Need to think about how to model deployment of low-carbon hydrogen realistically.

I would prefer a later introduction of H2-steel. I am not sure what the price implications are, but would think that it makes Al more competitive compared to today. In any case, Al was just a stand-in for all lightweight materials and each has its own issues, so I think we should not over-interpret this. Seeing an effect that diminishes over time seems about right.

[stefanpauliuk]

I agree @Hertwich ! H2-based steel is in its infancy plus that it may double-triple the steel price. For the RCP2.6 scenario, I now assume a linear ramp-up of H2-based steel from 0% in 2030 to 100% in 2070, instead of 0% in 2020 to 100% in 2050.

[stefanpauliuk]

@stefanpauliuk Car sharing leads participants to drive less given that the ownership costs are now felt at the margin. It reduces vehicle lifetime km because car sharers bulk the cars more than owners do. I think our data for the US was that driving went down and number of cars also went down, but not as much. @PaulWolfram confirm please. I think we can expect more robust/easily repaired vehicles, like the Citi Bikes, that will extend lifetime, once car firms make dedicated vehicles for car sharing, as they have started to.

Yes but during our scenario workshops we used the model with an exogenously specified and fixed kilometrage that does not change under CaS. Only the fleet size changes in the modelling approach, which has been in use since August last year. Re-opening this issue would mean that we have to redo and re-programm the entire vehicle demand section.

[zerateltu]

I compared the vehicle stocks between "recycling only" and "MIU: CaS & RiS" scenarios. In both scenarios, the total vehicle stock increases in 2060. Please refer to column EP in 'backgr_calc' tab of "test_RECC_vehi_results_analysis_template_June 12, 2020.xlsx" in Dropbox (Yale_FES)/G7 RECC/Result Analysis.

[stefanpauliuk]

EoL RR improvement leads to increased emissions of MaterialGHG incl. Recycling Credit in 2050.

There is still a mistake in the recycling calculation that causes this problem and that I need to fix asap. Am working on it.

I spent more than a full working day chasing a bug in the mass balance calculation that got introduced when switching to time and age-cohort dependent material composition. The model is now fully mass-balanced again and global results will re-run overnight.

[stefanpauliuk]

@zerateltu can you please update your analysis to the latest results (\Dropbox\G7 RECC\Results\RECC_Global_Germany_June_19_2020) and be more specific? Which scenario folders did you compare, and what is the reference for your comparison? 2015?

[zerateltu]

@stefanpauliuk I updated the results analysis using the "RECC_Global_Germany_June_19_2020" data (now saved as "/G7 RECC/Result Analysis/test_RECC_vehi_results_analysis_June 19, 2020.xlsx") The issue I was referring to is that in-use vehicle stock (e.g., "In-use stock, pass. vehicles; LED;  RCP2.6") is higher in 2060 compared to 2015 when ALL ME strategies are applied (while in last year's model run, we see reduction in in-use stock in 2060 compared to 2015)

[Hertwich]

@stefanpauliuk Looking at the global annual vehicle emissions in 2050, currently our emissions reductions are 21% in ssp1. In the IRP report, we had 30% for G7 and 35% for India plus China (Fig. 7 SPM). Ride sharing contributed one half of the savings in cumulative emissions, or 6 Gt, in the G7 (2016-2060, Fig 8 SPM). In the current model run, savings from ride sharing are 6 Gt for the entire world. Is this smaller impact of ride sharing a result of a change in the calculation of vehicle ownership? @stefanpauliuk @zerateltu @PaulWolfram

[stefanpauliuk]

@Hertwich yes. I checked with the vehicle service calculator and with the sensitivity analysis for the USA, model version v2.4.: Here, a RiS share of 30% (LED) achieves a stock reduction of 8.5 % both in the vehicle service calculator and in the actual model results, which is correct for our parameter choices. For v.2.2 (IRP-version), the stock reduction was 23%.

[stefanpauliuk]

Just to double-check @zerateltu and @PaulWolfram : The parameter 6_MIP_RideSharing_Occupancy_V1.0 is at 1.4, which is a multiplier in the model, meaning that 1.4 corresponds to a 40% increase of the occupancy rate (as intermediate range of the 25....75% found in the literature). Correct?

[PaulWolfram]

Correct, Stefan!

On Fri, Jun 26, 2020 at 3:52 PM Stefan Pauliuk notifications@github.com wrote:

Just to double-check @zerateltu https://github.com/zerateltu and @PaulWolfram https://github.com/PaulWolfram : The parameter 6_MIP_RideSharing_Occupancy_V1.0 is at 1.4, which is a multiplier in the model, meaning that 1.4 corresponds to a 40% increase of the occupancy rate (as intermediate range of the 25....75% found in the literature). Correct?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/YaleCIE/RECC-ODYM/issues/27#issuecomment-650368552, or unsubscribe https://github.com/notifications/unsubscribe-auth/AI3X5EBYMROTAE6SHT6RPWDRYT4ADANCNFSM4N4D6EZQ .

-- Paul Wolfram, PhD cand. Yale Center for Industrial Ecology 195 Prospect St. New Haven, CT 06511 Web http://environment.yale.edu/profile/paul-wolfram Google Scholar https://scholar.google.co.uk/citations?user=wx1tTHQAAAAJ&hl=en * LinkedIn https://de.linkedin.com/in/pwolfram

[stefanpauliuk]

@zerateltu yes because of the large room for development of pass. vehicle transport in the RoW regions. We have as global fleet in million pavs: 2015: 1036. No ME: Global2020_6_1822_57_33pav LED: 1247 // SSP1: 2069 // SSP2: 3582

Full ME: Global2020_6_1822_53_16pavFYI_FSD_EoL_MSU_ULD_RUS_LTE_CaS_RiS LED: 928 // SSP1: 1681 // SSP2: 3172

Thus a fleet size reduction from CaS and RiS of ca. 25 % (LED), 19% (SSP1), and 11% (SSP2). A good overall question is thus: are these overall reduction what we think is suitable for these three socioeconomc scenarios?

[zerateltu]

@stefanpauliuk thanks for the explanation. The values you showed are the same as what I used for result analysis, but I was comparing 2060 values with 2015 values in each scenario (that's why I only see in-use stock reduction in Full ME, LED).

[stefanpauliuk]

@zerateltu is there still need for discussion on this issue?

[zerateltu]

not anymore, thanks @stefanpauliuk