Institutional decarbonization scenarios evaluated against the Paris Agreement 1.5 °C goal

Scientifically rigorous guidance to policy makers on mitigation options for meeting the Paris Agreement long-term temperature goal requires an evaluation of long-term global-warming implications of greenhouse gas emissions pathways. Here we employ a uniform and transparent methodology to evaluate Paris Agreement compatibility of influential institutional emission scenarios from the grey literature, including those from Shell, BP, and the International Energy Agency. We compare a selection of these scenarios analysed with this methodology to the Integrated Assessment Model scenarios assessed by the Intergovernmental Panel on Climate Change. We harmonize emissions to a consistent base-year and account for all greenhouse gases and aerosol precursor emissions, ensuring a self-consistent comparison of climate variables. An evaluation of peak and end-of-century temperatures is made, with both being relevant to the Paris Agreement goal. Of the scenarios assessed, we find that only the IEA Net Zero 2050 scenario is aligned with the criteria for Paris Agreement consistency employed here. We investigate root causes for misalignment with these criteria based on the underlying energy system transformation.


Evaluating the Constant Quantile Extension (CQE) method
As described in the Methods section of the main text we truncate each pathway (i.e., model and scenario combination) at 2050 and then use the CQE method to extend the pathways to 2100 and calculate the error as: where is the error, pi,t is the CQE-extended value of pathway i at time t, qi,t is the originally projected value at that time, ( ) is the number of pathways (times) being summed over and is the standard deviation of original projections at that time. Supplementary Table 1 shows the relative errors of using the CQE method on the SR1.5 database. It indicates that the errors are generally low -on average 0.23 (0.30 excluding the HFCs), compared to values above 0.5 when using any of the infilling techniques that infer one species from another 1 .
Supplementary  There is strong uncertainty in the infilled gases for the Equinor Rebalance and IEA NZE scenarios, depending on the method selected to derive the relationship between the lead gas (CO2 emissions from energy and industrial processes) and the infilled gases. In this section, we evaluate the uncertainty for the CH4 and N2O emissions (as an example).
Supplementary Figure 1: Effect of the infilling method for a given lead gas for (a -c) IEA NZE scenario and (d -f) Equinor Rebalance scenario. Thin grey lines represent the scenarios in the SR1.5 database used for infilling.
The RMS pathway selection method can lead to the infilling of relatively extreme emissions that are not necessarily driven by the lead gas. The IEA NZE scenario demonstrates this -while the lead gas emissions drop steeply, the pathway that is selected has relatively high N2O emissions (rms line in Supplementary Figure 1b and c) that indicate a model-specific result. A similar effect is observed for the corresponding emissions for the Equinor Rebalance scenario. On the other hand, the EQW method's inherent assumption of monotonicity can lead to large reductions in the infilled gas without a clear correlation to the lead gas reduction -we see this in the stringent CH4 reductions in the IEA NZE scenario in panel b (for more details see previously published discussion of the methods 1 ). Hence as a default case for the main results, we select the Quantile Rolling Windows (QRW) approach to provide a balanced and consistent approach to infer the missing emissions.

Climate assessment -key characteristics of the pathways
The categorisation of pathways on the basis of their climate impact follows the categorisation scheme adopted in SR1.5 2 , where categories were constructed based on the probability of exceeding a given temperature target. The categories and their respective exceedance probabilities (Pe) are adapted from SR1.5 and presented in Supplementary Table 3.
Supplementary The first climate outcome that is necessary (but not sufficient) to assess the Paris compatibility of the different pathways, is to check whether they achieve a balance between sources and sinks of emissions in the second half of the century. In Supplementary Table 4, we report the total greenhouse gas pathway (infilled using the QRW method) for the institutional scenarios.
Supplementary In Supplementary Table 5 we present the variation of the climate categorization across the different infilling methods assessed in this study. In only one case is there a change in categorization due to infilling method because the scenario results lie near the boundary between two categories. Supplementary

Mitigation lever results
In Supplementary Table 6 we summarise results for two technology mitigation levers, as discussed in the main text. The total final energy demand (Et) and the carbon intensity of final energy (CIt) are given as relative levels compared to 2010 for pathways of various climate outcome categories as derived from the SR1.5, as well as for the scenarios analysed in the present work.

Applying the FaIR model to assess the climate outcome
In this section, we demonstrate the application of the FaIR model 3,4 to assess the climate outcome of the institutional scenarios. For this sensitivity, we perform the climate assessment across all infilling sensitivity cases, and then calculate the average difference between the MAGICC6 and FaIR results. Importantly, we use the same priors (parameter sets) that were used in SR1.5 while performing the FaIR assessment. SR1.5 noted the following, with respect to the comparison between MAGICC6 and FaIR results: "The comparison of these lines of evidence shows high agreement in the relative temperature response of pathways, with medium agreement on the precise absolute magnitude of warming, introducing a level of imprecision in these attributes." Since the purpose of this paper is to assess the climate outcome of institutional pathways, in line with the assessment in SR1.5 (which kept consistency with AR5), we would expect the same outcome to be reflected in our results for the institutional pathways. In Supplementary Table 7, we highlight the difference in peak and end-of-century warming (in both cases we compare the median of the probabilistic temperature distributions) between the MAGICC6, and FaIR setups. A positive value indicates that MAGICC6 has higher values than the corresponding FaIR values. We assess the difference for 18 scenarios (i.e., 6 institutional scenarios, infilled with 3 different methods).

Supplementary Note
The institutions associated with the scenarios considered in this manuscript were contacted for a response to the manuscript. Equinor had no comment, and we received no response from BP or Shell. IEA requested clarifications, which were addressed in the manuscript.