Source attributions of radiative forcing by regions, sectors, and climate forcers
Xuanming Su, Kaoru Tachiiri, Katsumasa Tanaka, Michio Watanabe, Michio Kawamiya
SSource attributions of radiative forcing by regions, sectors,and climate forcers
Xuanming Su ∗ , Kaoru Tachiiri , , Katsumasa Tanaka , , MichioWatanabe & Michio Kawamiya Research Institute for Global Change / Research Center for Environmental Modeling andApplication / Earth System Model Development and Application Group, Japan Agency forMarine-Earth Science and Technology (JAMSTEC), Yokohama, Japan Center for Global Environmental Research, National Institute for Environmental Studies(NIES), Tsukuba, Japan Laboratoire des Sciences du Climat et de lâ ˘A ´ZEnvironnement (LSCE), Commissariat Ã˘alâ ˘A ´ZÃl’nergie atomique et aux Ãl’nergies alternatives (CEA), Gif-sur-Yvette, FranceE-mail: ∗ [email protected] Abstract.
It is important to understand how the emissions of different regions, sectors, orclimate forcers play a role on pathways toward the Paris Agreement temperature targets.There are however methodological challenges for attributing individual contributions dueto complexities associated with a variety of climate forcers affecting the climate system ondifferent spatial and temporal scales. Here, we use the latest historical and future emissionsdata for a comprehensive set of climate forcers as well as land-use datasets and apply thenormalized marginal approach to quantify the forcing contributions of regions, sectors andforcing agents toward the 2 °C and 1.5 °C targets. We show that most of the worldwideregions and sectors need to maintain forcing levels not higher than present levels to attain the1.5 °C target of the Paris Agreement, while slightly higher future forcing levels than presentlevels are allowed for the 2 °C target. Our results illustrate the importance of negative CO emissions, which contribute - . ± . Wm -2 and - . ± . Wm -2 to the 2 °C and 1.5 °Ctargets. Less negative forcings, or more positive forcings are also identified for the land-usealbedo for the 2 °C and 1.5 °C scenarios compared to existing studies. a r X i v : . [ phy s i c s . a o - ph ] S e p . Introduction The Paris Agreement has set goals to limit the global average temperature increase towell below 2 °C and to pursue efforts to limit the temperature increase to 1.5 °C abovethe preindustrial level. The Paris Agreement goals can be translated into the requiredlevels of greenhouse gas (GHG) emission reductions [1–6] for practical implementationpurposes, through calculations of the radiative forcings resulting from various emissionsources. Attaining the radiative forcing of 2.6 Wm -2 and 1.9 Wm -2 are known to be largelyconsistent with achieving the 2 °C and 1.5 °C climate goals, respectively, at an approximately66% probability [1, 3, 7–10]. Here we use the radiative forcing as a benchmark and assesshow individual regions, sectors, and climate forcers can contribute to achieving the ParisAgreement temperature targets based on a latest set of historical and future emission data.Such source attribution cannot be done from emissions data alone because a variety of climateforcers affect the climate system on different temporal and spatial scales in nonlinear ways,requiring dedicated methodologies like the one presented here. Attributing forcing at thelevels of regions, sectors, or climate forcers provides a basis for considering the principle ofcommon but differentiated responsibilities contained in the 1992 United Nations FrameworkConvention on Climate Change (UNFCCC).We identified two issues associated with previous attribution methods. First,comprehensive regional and sectoral assessments should in principle consider a full suiteof anthropogenic sources at the regional and sectoral levels, including GHGs, aerosols andpollutants, as well as land-use albedo. However, not all sources have been considered inprevious studies. For example, aerosols and pollutants were not always examined [11, 12], oronly a subset of aerosol species was considered (such as sulfate aerosols [13]). Also, land-use albedo were sometimes not included [11–13]. This may be due to the lack of datasetsor difficulties in representing regional or sectoral forcings, but the latest datasets provideopportunities to consider a more comprehensive set of GHGs and related agents. Second,among various methods proposed, only the marginal and time-sliced methods are consideredto be useful based on a satisfaction test of eight essential criteria [14, 15]. In principle, anattribution method needs to ensure additivity regarding regions and time. Implementing anon-additive method like the residual method used in [16] may introduce bias to the outcomefor regions with high and low emissions. Some of the methods yield non-zero radiative forcingwhen a region’s concentration has become zero. On the other hand, the two recommendedmethods are computationally expensive, especially when various sources of uncertainties areconsidered.We apply the most up-to-date emissions and land-use datasets (Tables S1 to S3and Figs. S1 to S4), which resolve regions and sectors and contain all pertinent forcingsources, including GHGs, aerosols and pollutants, and land-use albedo, as well as theirassociated uncertainties. Particularly we consider a range of future emissions trajectorieswith various socioeconomic backgrounds and climate mitigation levels [17, 18]. Wecombine the normalized marginal method, which is computationally less expensive thanthe time-sliced method [14], with a simple climate model - the Simple Climate Model2or Optimization version 2 (SCM4OPT v2.0) [19, 20]. SCM4OPT v2.0 is designed tobe lightweight and suitable for performing a large number of simulations required forour study exploring uncertainties, while resolving diverse characteristics of forcing agentsconsidered. Furthermore, based on the premise of previous studies [11, 21, 22], we make theattribution analysis more comprehensive by considering historical and future emissions andproviding perspectives from regions, sectors, and climate forcers. We consider two types ofuncertainties, including those due to our lack of knowledge regarding historical emissions andfuture projections (emission uncertainties), and those due to low confidence in understandingthe climate system (climate uncertainties). With these methodological advances, we quantifythe forcing contributions of regions, sectors, and climate forcers toward the Paris Agreementtemperature goals.
2. Methods
We compile the emissions and land cover datasets at regional and sectoral levels andimplement them to SCM4OPT v2.0 to calculate the marginal forcing effects of forcingagents at regional and sectoral levels. The relative forcing contribution of the forcing agentsat regional and sectoral levels can be distinguished based on the fraction accounting fortheir marginal effects in total marginal effects caused by the forcing agent. The forcingcontributions of the forcing agents at regional and sectoral levels can be therefore attributed.We sum up the associated individual forcings to obtain the radiative forcings resulting fromregional and sectoral sources.
We used historical and future emissions and land cover datasets with both regional andsectoral details. The available datasets are shown in Tables S1 to S3 and Figs. S1 to S4[17, 18, 23–26]. We designated historical sources as those originating from 1850 to 2016,and future projections by 2100 are grouped by the forcing level at 1.9 Wm -2 and 2.6 Wm -2 ,regardless of the underlying socioeconomic development or technological assumptions. Wealso included other scenarios with relatively lower possibilities to achieve the 2 °C and 1.5°C targets, namely, forcing levels of 3.4 Wm -2 , 4.5 Wm -2 , 6.0 Wm -2 , 7.0 Wm -2 and 8.5 Wm -2 (Table S2). Thus, a broad range of forcings can be examined for future climate change.For the future emission datasets, 25 of them are obtained from the Asia-Pacific IntegratedModel/Computable General Equilibrium (AIM/CGE) model [17], and the remaining nine arefrom the Integrated Assessment Modeling Consortium (IAMC) [18] (Table S2). We dividedthe world into eleven regions, including 1) China (CHN), 2) India (IND), 3) Japan (JPN),4) Russia (RUS), 5) the United States of America (USA), 6) sub-Saharan Africa (AFR), 7)Europe (EUR), 8) Latin America and the Caribbean (LAM), 9) the Middle East and NorthAfrica (MEA), 10) other areas in Asia (OAS) and 11) the rest of the world (ROW) (Table S4).For each region, twelve emitting sectors [17, 18, 23] were assessed, namely, 1) agriculture,2) agricultural waste burning, 3) domestic and commercial housing, 4) energy, 5) industry, 6)3ndustrial solvents, 7) surface transportation, 8) waste treatment, 9) open forest burning, 10)open grassland burning, 11) aviation and 12) international shipping (Table S5). In additionto the twelve sectors above, 13) land-use CO emissions and 14) negative CO emissions,namely, through carbon capture and storage (CCS) and bioenergy with CCS (BECCS), wereseparately considered. We compiled the emissions from the available datasets into scenario-,region- and sector-specific emissions and used E n,r,s,e ( t ) to denote scenario ( n )-, region ( r )-and sector ( s )-specific emissions ( e , refer to the emission species) over time t . The emissionsof the same forcing target originating from different Shared Socioeconomic Pathways (SSPs)and integrated assessment models (IAMs) were treated as emission uncertainties. Forexample, we iteratively simulated the 1.9 Wm -2 forcing scenario by using dataset from theAIM/CGE (SSP1-1.9 and SSP2-1.9) and the IAMC (SSP1-1.9), as reported in Table S2. We used the simple climate model SCM4OPT v2.0 to generate the outputs for our analysis.The current model has been updated from the precedent in the following four respects: First,we adopted the ocean carbon cycle of Hector v1.0 [27] and applied the Diffusion OceanEnergy balance CLIMate (DOECLIM) model [28–30] to calculate the temperature change.We calibrated the carbon cycle and temperature modules based on 26 coupled atmosphere-ocean general circulation models (AOGCMs) with outputs for the carbon cycle in the CoupledModel Intercomparison Project, Phase 5 (CMIP5) (Table S6). Second, parameters associatedwith CH , N O and halogenated gases (a total of 37 gases, see Table S7) were tuned against theatmospheric lifetimes and radiative efficiencies in the IPCC Fifth Assessment Report (AR5)[31]. Third, we employed the simple global parameterizations described in OSCAR v2.2 [32]to estimate the radiative forcings resulting from aerosols and pollutants. The radiative forcingof short-lived climate forcers depends on the geographical location of emissions. The spatialdistribution of the radiative forcing of short-lived species is different from that of long-livedspecies [33, 34]. However, these two effects are not considered in our analysis. Fourth,we adopted a simple parameterization scheme [32] to calculate the land-use albedo (seeEq. (67) in the supplementary materials). The equations for the climate model are listedin the supplementary materials.We performed a robustness test over the historical period by using historical emissiondatasets as input and considering the climate uncertainties that were applied in this analysis.The outputs from our model are consistent with those of other models or statistical records(Figs. S5 to S8, S10 and S11). Furthermore, the likelihoods of meeting the 2°C and 1.5°C targets of each of the forcing scenarios obtained from our model agree largely with thecorresponding IPCC ranges (Fig. S12) to limit global warming to 2°C with at least 66%probability and 1.5 °C with 50% [3].
We utilized and expanded the normalized marginal method presented in ref [14, 15, 22] toconduct our analysis. The relative forcing contribution of emission E n,r,s,e (Column 2 in4able S7) to the associated radiative forcing f (Column 4 in Table S7), which is defined as α fn,r,s,e , is proportional to the marginal effect of E n,r,s,e causing the radiative forcing f (seeFig. S13). To calculate α fn,r,s,e , for each E n,r,s,e , we performed two simulations, i.e., 1) onesimulation with all emissions included in the simulation as input, to calculate the associatedradiative forcing termed F all,fn,r,s,e , and 2) another simulation with the emission e reduced by E n,r,s,e · (cid:15) ( (cid:15) = 0 . ) over the evaluation period, that is 1850-2100, to obtain the correspondingradiative forcing named F (cid:15),fn,r,s,e . The relative contribution α fn,r,s,e is obtained by: α fn,r,s,e = F all,fn,r,s,e − F (cid:15),fn,r,s,e (cid:80) r,s,e (cid:16) F all,fn,r,s,e − F (cid:15),fn,r,s,e (cid:17) (1)Therefore, the radiative forcing F fn,r,s,e , which is resulting from E n,r,s,e , is isolated by: F fn,r,s,e = F all,fn,r,s,e · α fn,r,s,e (2)To consider the relevant emission and climate uncertainties, we carried out 200 similarpairs of runs for each forcing-level-specific source E n,r,s,e , with randomized scenarios at thesame forcing level and randomized parameter sets for the climate system, and we call themas one experiment for E n,r,s,e (see the randomization sources of the scenarios (within eachforcing level) and the climate system in Table S8). Here, the value of 200 has been tested toensure that two decimal places of the precision level could be achieved for the mean forcingvalue of F fn,r,s,e under different experiments.We obtain the individual forcing agents by summing the forcings induced by all availableemissions sources. Therefore, a certain forcing agent is probably a mixed effect resulting fromvarious emissions sources. On the other hand, a particular emission may result in differentkinds of radiative forcings, as indicated in Table S7. For example, black carbon (BC) cancause BC forcing, BC on snow and indirect cloud effects. A total of . × runs ( . × for each forcing level) were performed considering all forcing levels, regions, sectors andemissions. To derive the regional forcings, we applied the Monte Carlo approach (n = 20,000)to sum all F fn,r,s,e values belonging to a given region. Here, the value of 20,000 for n was alsotested to ensure the necessary precision for our analysis. The sectoral forcings are similarlyobtained. An overview of all the iterations are contained in Table S8. For each experiment, 200 sample results were acquired. Here, we assumed that the obtainedtemperature increase T over time t followed a normal distribution, and the cumulativedistribution function was defined as: F tT ( τ ) = P t ( T ≤ τ ) (3)We used the exceedance of Eq. (3) to obtain the probability of exceeding a specifiedclimate target τ : 5 tT ( τ ) = P t ( T > τ ) = 1 − F tT ( τ ) (4)Therefore, F tT (2) indicates the probability of exceeding 2°C, and F tT (1 . gives theprobability of exceeding 1.5°C.
3. Results
We performed our analysis based on the available existing scenarios, and the 2 °C or 1.5 °Cresults herein thus reflect the diagnosed compatible scenarios in terms of the 2 °C or 1.5 °Ctargets. The results reveal that the USA, China and the European Union (EU) are three majoremitters, accounting for approximately 45% of all the forcings under the historical, 2 °C and1.5 °C scenarios (see Fig. 1a). China’s share increased from ± % ( . ± . Wm -2 ) by2016 (cf. ± % for Chinese data (1750-2010) in ref [22] with similar methods) to 2 °C’s ± % ( . ± . Wm -2 ) and 1.5 °C’s ± % ( . ± . Wm -2 ), while the share ofthe EU declined, from the historical ± % ( . ± . Wm -2 ) level to the 2 °C level of ± % ( . ± . Wm -2 ) and the 1.5 °C level of ± % ( . ± . Wm -2 ) (for the forcingvalues, see Fig. 2a&Fig. S14). In contrast, the share of the USA exhibited no major changes,contributing to approximately 17% of the total forcings under all three scenarios. However,the absolute values of the forcings varied, i.e., the 2 °C forcing attributed to the USA increasedto . ± . Wm -2 from the current forcing value of . ± . Wm -2 , while the 1.5 °Cforcing value declined to . ± . Wm -2 . Latin America and the Caribbean (as one region)also exhibited a relatively high historical share with ± %; however, the value substantiallydeclined under both target scenarios.CO , including fuel CO , land-use CO , and negative CO , if applicable, is the maincontributor and varies across regions. Among them, China, the USA, and the Middle East andNorth Africa (as one region) exhibited the highest net growth forcings under the 2 °C scenario,with values of . ± . Wm -2 , . ± . Wm -2 and . ± . Wm -2 , respectively. Underthe 1.5 °C scenario, the CO forcings in all regions decreased. The largest decline occurred inLatin America and the Caribbean, from the historical value of . ± . Wm -2 to the 1.5 °Cscenario value of . ± . Wm -2 , which occurred due to the negative CO emissions andthe great decrease in land-use CO emissions.The non-CO forcings described here refer to the forcings induced by sources other thanCO , and these forcings also play an important role in the historical period, although they arealmost adequately controlled under the 2 °C and 1.5 °C scenarios (also shown in Fig. S14).Basically, most of the regions reveal net positive non-CO forcings in the historical period.Particularly in regard to Latin America and the Caribbean, the relatively high net positivenon-CO forcing, combined with the relatively high land-use CO forcing, contributes to acomparatively large forcing share in the historical period, although the fossil-fuel forcing isrelatively smaller. It is worth noting that the regions with nearly zero-sum non-CO forcingscontribute considerable amounts of both positive and negative forcings, such as China and6he rest of the world, in the historical period. For future non-CO forcings, however, sub-Saharan Africa is found to exhibit a reasonable increase in net forcing due to its continuousdevelopment and industrialization and population growth, which requires more biomass forcooking and heating purposes, as well as changes in land cover [17, 18].The total forcing, including the forcings that cannot be assigned to any region, increasesto . ± . Wm -2 under the 2 °C scenario but declines to . ± . Wm -2 under the 1.5°C scenario, which is lower than the current level of . ± . Wm -2 (Fig. 1b). All regionalforcings indicate net warming effects, with positive forcing values. Forcing increases areencountered in most of the regions except in Russia, the EU, Latin America and the Caribbean,and the rest of the world under the 2 °C scenario, while the main increases still occur intwo developing regions, namely, China and the Middle East and North Africa under the1.5 °C scenario (Fig. 2a&Fig. S14). Here, the forcing increases in the Middle East andNorth Africa can mostly be attributed to fossil-fuel CO , sulfate, cloud effects, and land-usealbedo, probably due to industry and energy supply expansions as well as due to the expectedreforestation in this area [17]. The regional nonattributable forcings in the historical periodalso reveal warming effects. These forcings are later notably suppressed under both the 2 °Cand 1.5 °C scenarios (Fig. 1b), and they are mainly attributed to the control of ozone-depletingsubstances (ODSs) under the various scenario assumptions (Fig. S1) [17, 18]. The regional effects are further separated into their sectoral constituents to assess how futurechanges occur (Fig. 2a). First, for the developed regions, relatively large increases areobserved in both industrial and housing sectors under the 2 °C scenario. In regard to energy,the gross forcings related to the USA and EU are considerably high under the 2 °C scenario.However, if combined with the negative CO emissions, the energy forcings decrease to . ± . Wm -2 and . ± . Wm -2 for the USA and EU, respectively, which are lowerthan the current levels. Second, among the developing regions, China’s industry exhibits themost significant forcing increase, with a value of . ± . Wm -2 under the 2 °C scenario.In addition to the industrial sector, prominent increases are found in the agricultural sector,such as in sub-Saharan Africa, and the forcing induced by the agricultural sector increases by . ± . Wm -2 under the 2 °C scenario. In addition, the land-use CO forcings are alleviatedto varying degrees in all regions under the 2 °C scenario. Under the 1.5 °C scenario, most ofthe regions still demonstrate increased forcings in the industrial sector, while in the developedregions, the forcings in the industrial sector decrease. Furthermore, both the negative andland-use CO emissions could result in extensive forcing abatement from the current levels inthe developing regions under the 1.5 °C scenario.Globally, in certain sectors, as shown in Fig. 2b, the forcings still increase to certain levelsunder the 2 °C scenario, except for the land-use CO and other sources that are responsiblefor the main emissions of aerosols and pollutants, such as waste treatment, agricultural wasteburning, forest burning and grass burning. However, under the 1.5 °C scenario, except for theenergy sector and land-use albedo, only a small amount of the forcings is found to increase7n the major emitting sectors, such as domestic and commercial housing, industrial sector,aviation and international shipping. In addition, as also indicated in the analysis of theindividual forcing agents below, the negative CO emissions remove a considerable amountof forcings from the energy sector under the 1.5 °C scenario, and the net forcing level in theenergy sector is lower than the current energy sector level. This result implies that to attainthe 1.5 °C target, efforts need to be implemented to maintain the sectoral forcings below orequal to the current levels. Fig. 3a shows the individual forcing agents for each sectoral source. Fossil-fuel CO dominates the forcings in the housing, energy, industrial and transport sectors, particularlyunder the 2°C and 1.5 °C scenarios when the other GHGs and aerosols and pollutants aresubstantially removed (Figs. S1 to S3) and the resulting impacts are therefore greatly reduced.Among them, first, the negative CO emissions can eliminate considerable amounts offorcings. For example, - . ± . Wm -2 and - . ± . Wm -2 are attributed to the negativeCO emissions under the 2 °C and 1.5 °C scenarios, respectively, (Fig. 3a). It is interestingto note that the reduced amount of the absolute forcing under the 2 °C scenario is even largerthan that under the 1.5 °C scenario. This finding explains the feasibility of the relativelyweaker climate policies adopted under the 2 °C scenario, leading to higher gross fossil CO emissions. However, a fair amount of fossil CO emissions is removed in the later periodwhen the costs related to the negative CO emissions are more reasonable. Under the 1.5 °Cscenario, stronger strategies are implemented after the early period. Thus, the gross fossilCO emissions are relatively lower, and the required negative CO emissions do not needto be as high [17]. Here, the general trend can be simply interpreted as that of emit morebut reduce more. Moreover, if considering the negative CO emissions, the net forcings areactually lower than the current levels in the energy sector under both the 2 °C (0.42 Wm -2 )and the 1.5 °C (0.31 Wm -2 ) scenarios, although gross increases are prominent (Fig. 3a).Second, in regard to agriculture, the major sources are CH and N O. A considerableamount of the forcings induced by CH and N O still remains under the 2 °C and 1.5 °Cscenarios (Fig. 3a), due to difficulties in reducing the CH and N O emissions from agriculture[17, 18] and their relatively long lifetimes, namely, 12.4 years for CH and 121 years for N O(see Table 8.A.1 in ref [8]).Third, the land-use albedo currently exhibits a cooling effect of - . ± . Wm -2 .However, the land-use albedo may reveal warming effects in the future, at . ± . Wm -2 under the 2 °C scenario and . ± . Wm -2 under the 1.5 °C scenario (Fig. 3b). Theforest cover is expected to increase under the 2 °C and 1.5 °C scenarios (Fig. S4), which willlower the surface albedo and reflect less of the incoming solar radiation, which in turn willgenerate lower negative forcings, or more positive forcings, while the current deforestationcauses a negative forcing [35]. Therefore, to achieve the set climate goals, more forcingsneed to be reduced from other sources to compensate for this effect. Here, the land-usealbedo is estimated by a simple parameterization scheme [32, 36] constrained by future land8over changes (see Methods), and the results reveal lower negative forcings, or more positiveforcings, than those by the other estimations (Fig. S9). All potential projected scenarios, including those forcings higher than the 2°C and 1.5 °Cforcings, are shown in Fig. 4a (for the regional contributions) and Fig. 4b (for the sectoralforcings). We translate the forcing levels into probabilities of exceeding 2°C or 1.5°C todemonstrate the likelihood of achieving the climate goals under such conditions. Basically,China, the USA and the EU are still the three major contributors to climate change when highforcings are applied. For example, under the high-forcing scenarios, China may account forapproximately 1.1 Wm -2 , albeit with greater uncertainty, the USA accounts for approximately0.9 Wm -2 , and the EU accounts for approximately 0.6 Wm -2 . All other regions exhibitrelatively lower but still significant radiative forcings under the same circumstances, exceptJapan, where the forcing levels do not greatly change even under the high-forcing scenarios.Sectorally, the energy sector may contribute the highest forcings given its high emissions, upto . Wm -2 , followed by the industry (up to . Wm -2 ) and transport (up to . Wm -2 ), as wellas land-use CO (up to . Wm -2 ). Lower or no negative CO emissions (Fig. S1), relativelyfewer nuclear and renewable energy sources (for example, solar and wind), and more fossil-fueled energy use could bring about extremely high climate-related emissions in the energysector [17, 18, 37, 38]; hence, high forcings are produced. All sectors reveal high radiativeforcing values except for open burning, which remains relatively stable, under the assumedhigh-forcing scenarios (Fig. 4b).
4. Discussion
In this study, we applied a simple climate model, SCM4OPT v2.0, to determine the forcingcontributions of regions, sectors and climate forcers based on available historical and futureprojection emissions and land-use datasets. This study provides an IAM-based assessmentfrom a forcing perspective at the sectoral and regional levels. The radiative forcings, includingthe forcings resulting from the various sources of GHGs, aerosols and pollutants, and land-usealbedo, are distinguished among the different regions and sectors. The outputs here can beused to inform policy-makers of the relative importance of the forcing levels resulting fromdifferent regional or sectoral sources.The results are interpreted with certain caveats and limitations. First, we analyzeemission datasets that contain regional and sectoral information, while global-scale datasetsare not included. Therefore, our analysis only reflects limited uncertainties that are derivedfrom the available emission estimates. Thus, the analysis here is merely considered as anIAM-based evaluation of the potential future climate, especially under the 2°C and 1.5 °Cscenarios. Second, the outcome of this study is contingent on the set of selected scenariosused, which are treated equally likely. However, the distribution of total range of futureemissions does not necessarily present equal probabilities [39] (Fig. S15). We consider a9urther update for this analysis when the probabilistic emission scenarios are available.The results showed that the 1.5 °C target requires most regions and sectors to maintaintheir forcings not higher than the current levels, while slightly higher future forcing levelsthan present levels are allowed for the 2 °C target. The results here can be used to assessthe gap between the current and targeted climate levels for both regions and sectors interms of the radiative forcing. Furthermore, we found that the negative CO forcing isprojected to contribute - . ± . Wm -2 and - . ± . Wm -2 under the 2 °C and 1.5 °Cscenarios, respectively. Our analysis illustrates the importance of the negative CO emissionsin achieving the climate targets from the perspective of the radiative forcing. By using anew land-use forcing parameterization, we further found that less negative forcings, or morepositive forcings, for the land-use albedo for the 2 °C and 1.5 °C scenarios than those fromexisting studies. A comprehensive consideration of the available forcing sources is importantfor the climate change assessment. 10 R eg i ona l r e l a t i v e c on t r i bu t i on ( % )
12% 8%5% 6%5%15%12% 7% 2% 17%6% (a−1)
Historical01020 01020
8% 9%8% 4%9%12%16% 8% 3% 16%4% (a−2) (cid:176) C01020 01020
9% 9%6% 4%9%13%17% 8% 3% 17%5% (a−3) (cid:176)
C01020 01020
C H N I N D J P N R U S U S A A F R E U R L A M M
E A O A S R O W l l H I S (cid:176) C . (cid:176) C R ad i a t i v e f o r c i ng ( W m - ) b CHNINDJPNRUSUSAAFREURLAMMEAOASROWOther ll Historical2 (cid:176)
C1.5 (cid:176) C Figure 1.
Reginal contributions to climate change. a, Regional relative contributionsto climate change. The regional relative contributions are derived from theelementwise ratio of the regional forcings to the total forcings via the methodsdescribed in ref [22]. Note that the sum of the mean percentages of all regions isnot equal to 100% since regional nonattributable forcings occur (see Fig. 1b). b, Theworld total forcings are divided into regional forcings. The value on top of the barindicates the mean value of the total radiative forcing, and the error bar indicates theassociated uncertainty resulting from the world ensemble. The other forcings in Fig. bare the regional nonattributable climate forcers, including the international shipmentsof . ± . , . ± . , and . ± . Wm -2 and part of the ozone-depletingsubstances (ODSs) of . ± . , . ± . , and . ± . Wm -2 (for the regionalnonattributable forcings, see Table S7), under the historical, 2 °C and 1.5 °C scenarios,as well as mineral dust (Fig. S6) and the effects of solar irradiance and volcanic activity(Fig. S8). The probabilities of reaching the 2 °C and 1.5 °C targets here are 56%and 61%, respectively (see Fig. 4). All uncertainties are represented as one standarddeviation. CHN, China; IND, India; JPN, Japan; RUS, Russia; USA, United Statesof America; AFR, sub-Saharan Africa; EUR, Europe; LAM, Latin America and theCaribbean; MEA, Middle East and North Africa; OAS, other Asian countries; ROW,the rest of the world; Other, regional nonattributable forcings. l ll ll ll ll ll ll ll ll ll ll CHN I ND J P NRU S U SAA F R E UR L A MM EA O AS R O W (cid:176) C2 (cid:176) CHIS1.5 (cid:176) C2 (cid:176) CHIS1.5 (cid:176) C2 (cid:176) CHIS1.5 (cid:176) C2 (cid:176) CHIS1.5 (cid:176) C2 (cid:176) CHIS1.5 (cid:176) C2 (cid:176) CHIS1.5 (cid:176) C2 (cid:176) CHIS1.5 (cid:176) C2 (cid:176) CHIS1.5 (cid:176) C2 (cid:176) CHIS1.5 (cid:176) C2 (cid:176) CHIS1.5 (cid:176) C2 (cid:176) CHIS Radiative forcing ( Wm - ) a ll −10123 H I S (cid:176) C . (cid:176) C R ad i a t i v e f o r c i ng ( W m - ) b AgricultureEnergyIndustrySolventSurface transportAviationIntl' shippingHousingWasteAgricultural wasteForest burningGrass burningNegative CO Land use CO Land albedoOther ll Historical2 (cid:176)
C1.5 (cid:176) C Figure 2.
Sectoral contributions to climate change. a, Sectoral contributions of theeleven regions worldwide. b, The world total forcings are decomposed into sectoralforcings. The value on top of the bar indicates the mean value of the total radiativeforcing, and the error bar indicates the associated uncertainty resulting from the worldensemble. The other forcings are the forcings induced by mineral dust (Fig. S6), solarirradiance and volcanic activity (Fig. S8). The probabilities of reaching the 2 °C and1.5 °C targets here are 56% and 61%, respectively (see Fig. 4). All uncertainties arerepresented as one standard deviation. (a−1) Historical0.25 0.510.16 0.40.530 −0.03 0.23 0.12 −0.400.40.81.2 −0.400.40.81.2 N e g a t i v e C O L U C C O A g r i c u l t u r e O p e n b u r n i n g H o u s i n g E n e r g y I n d u s t r y T r a n s p o r t W a s t e (a−2) (cid:176) C l l l l l l ll l l CO ) −1012 −1012 N e g a t i v e C O L U C C O A g r i c u l t u r e O p e n b u r n i n g H o u s i n g E n e r g y I n d u s t r y T r a n s p o r t W a s t e R ad i a t i v e f o r c i ng ( W m - ) (a−3) (cid:176) C l l l l l l ll l l CO ) −0.501.51 −0.501.51 N e g a t i v e C O L U C C O A g r i c u l t u r e O p e n b u r n i n g H o u s i n g E n e r g y I n d u s t r y T r a n s p o r t W a s t e ll H I S (cid:176) C . (cid:176) C R ad i a t i v e f o r c i ng ( W m - ) b Negative CO LUC CO FF CO CH Strat. H ON OHalogenatedStrat. O Tropo. O SulfateNitratePOA SOABCBC on snowCloud Land albedoDustNatural ll Historical2 (cid:176)
C1.5 (cid:176) C Figure 3.
Contributions of the individual climate forcers. a, Sectoral contributionsdecomposed into individual climate forcers from (a-1) the historical period to 2016, (a-2) the 2 °C climate target by 2100 and (a-3) the 1.5 °C climate target by 2100. Here,open burning sums the agricultural waste burning, forest burning and grass burninglevels. Industry includes industry and solvents, similar to in Fig. 2. Transport totalsthe surface transport, aviation and international shipping values. The annotation valuesunder the energy sector in (a-2) and (a-3) denote the forcing values accounting for thenegative CO emissions. b, The world total forcings are decomposed into individualclimate forcers. The value on top of the bar indicates the mean value of the totalradiative forcing, and the error bar indicates the associated uncertainty resulting fromthe world ensemble. The direct CO emissions are divided into fossil-fuel CO (FFCO ), land-use CO (LUC CO ), and negative CO emissions, if applicable. Thenatural forcings include solar irradiance and volcanic activity (Fig. S8). The land-use albedo, mineral dust (Dust) and natural forcings are applied to Fig. b only.The probabilities of reaching the 2 °C and 1.5 °C targets here are 56% and 61%,respectively (see Fig. 4). All uncertainties are represented as one standard deviation. .5 (cid:176) C 2 (cid:176) C llllllllllllll llllllllll llllllllll ll ll llllll ll ll ll ll llllll ll ll ll ll ll llll ll ll ll llll llll ll ll ll llllllll ll ll ll llllllll ll ll ll llllll
44% 70% 93% 100%
12% 39% 70% 100%
Exceedance probability of (cid:176) CExceedance probability of (cid:176) C R ad i a t i v e f o r c i ng ( W m - ) l l CHNINDJPNRUSUSAAFR EURLAMMEAOASROW a (cid:176) C 2 (cid:176) C lllllllllllllllllllll lllllllllllllll lllllllllllllll lll lll lllllllll lll lll lll lll lllllllll lll lll lll lll lll llllll lll lll lll llllll llllll lll lll lll llllllllllll lll lll lll llllllllllll lll lll lll lllllllll
44% 70% 93% 100%
12% 39% 70% 100%
Exceedance probability of (cid:176) CExceedance probability of (cid:176) C R ad i a t i v e f o r c i ng ( W m - ) lll LUC CO AgricultureOpen burningHousingEnergyIndustryTransportWaste b Figure 4.
Regional and sectoral contributions under the different futureprojections. a, The relationship between the exceedance probability of 2°C or 1.5°Cand the regional forcing contributions. The color points are the regional relativecontributions. The trends are shown by the colored lines obtained via linear regression.b, The relationship between the exceedance probability of 2°C or 1.5°C and thesectoral forcing contributions. The negative CO level is summed into the energysector to simplify the analysis. The color points are the sectoral forcings. The trendsare shown by colored lines obtained through linear regression. The points in Fig. a&bare sampled (from 2020 to 2100, every 10 years) at seven forcing levels, namely, 1.9Wm -2 , 2.6 Wm -2 , 3.4 Wm -2 , 4.5 Wm -2 , 6.0 Wm -2 , 7.0 Wm -2 and 8.5 Wm -2 . Theforcing levels are translated into exceedance probabilities within each sample year.The 1.5°C and 2°C results marked with the vertical lines represent the scenarios withforcing levels of 1.9 Wm -2 and 2.6 Wm -2 , respectively, in 2100. cknowledgments This work was supported by the Integrated Research Program for Advancing Climate Models(TOUGOU), Grant Number JPMXD0717935457, from the Ministry of Education, Culture,Sports, Science and Technology (MEXT), Japan. The computing resources were provided bythe Japan Agency for Marine-Earth Science and Technology (JAMSTEC). We thank M. Abefor providing the data used for model calibration and T. Gasser for sharing the OSCAR v2.2source code.
Author contributions
X.S., K. Tachiiri and K. Tanaka designed the study. X.S. processed the emissions and landcover source data. X.S. developed the model with the help of K. Tanaka and M.W. X.S.performed the calculations and generated the figures. All coauthors contributed to analyzingthe results and writing the paper.
Competing financial interests
The authors declare that they have no competing financial interests.
Data availability
The data used to support the analysis are available from the corresponding author uponreasonable request. 15
1] Rogelj J, Popp A, Calvin K V, Luderer G, Emmerling J, Gernaat D, Fujimori S, Strefler J, Hasegawa T,Marangoni G, Krey V, Kriegler E, Riahi K, Van Vuuren D P, Doelman J, Drouet L, Edmonds J, Fricko O,Harmsen M, Havlík P, Humpenöder F, Stehfest E and Tavoni M 2018
Nature Climate Change Nature Climate Change Mitigation Pathways Compatible with 1.5 ° C inthe Context of Sustainable Development book section 2, pp 93–174 ISBN ISBN[4] Tong D, Zhang Q, Zheng Y, Caldeira K, Shearer C, Hong C, Qin Y and Davis S J 2019
Nature [5] Tachiiri K, Herran D S, Su X and Kawamiya M 2019 Environmental Research Letters Progress in Earth and Planetary Science Nature
Anthropogenicand Natural Radiative Forcing (Cambridge, United Kingdom and New York, NY, USA: CambridgeUniversity Press) book section 8, pp 659–740 ISBN ISBN 978-1-107-66182-0[9] Van Vuuren D P, Stehfest E, Gernaat D E, Van Den Berg M, Bijl D L, De Boer H S, Daioglou V, DoelmanJ C, Edelenbosch O Y, Harmsen M, Hof A F and Van Sluisveld M A 2018
Nature Climate Change Nature
Global Environmental Change Climatic Change
EnvironmentalResearch Letters https://unfccc.int/resource/docs/2002/sbsta/inf14.pdf [15] Trudinger C and Enting I 2005 Climatic Change EnvironmentalResearch Letters ISSN 17489326[17] Fujimori S, Hasegawa T, Ito A, Takahashi K and Masui T 2018
Scientific Data Geoscientific Model Development Earth’s Future Sustainability Science Climatic Change
Nature
Earth SystemScience Data Earth System Science Data Geoscientific Model Development Geoscientific ModelDevelopment Imprecise probability analysis for integrated assessment of climate change doctoralthesisUniversität Potsdam[29] Tanaka K and Kriegler E 2007 Aggregated Carbon cycle, atmospheric chemistry and climate model(ACC2): description of forward and inverse mode Tech. rep. Max Planck Institute for Meteorology[30] Wong T E, Bakker A M R, Ruckert K, Applegate P, Slangen A B A and Keller K 2017
Geoscientific ModelDevelopment Annex II: Climate System Scenario Tables (Cambridge, United Kingdom and New York, NY,USA: Cambridge University Press) book section AII, p 1395â ˘A ¸S1446[32] Gasser T, Ciais P, Boucher O, Quilcaille Y, Tortora M, Bopp L and Hauglustaine D 2017
GeoscientificModel Development Nature ClimateChange Nature Climate Change Journal of Climate Atmospheric Chemistry and Physics Global Environmental Change
169 – 180[38] Kriegler E, Bauer N, Popp A, Humpenöder F, Leimbach M, Strefler J, Baumstark L, Bodirsky B L, HilaireJ, Klein D, Mouratiadou I, Weindl I, Bertram C, Dietrich J P, Luderer G, Pehl M, Pietzcker R, PiontekF, Lotze-Campen H, Biewald A, Bonsch M, Giannousakis A, Kreidenweis U, Müller C, Rolinski S,Schultes A, Schwanitz J, Stevanovic M, Calvin K, Emmerling J, Fujimori S and Edenhofer O 2017
Global Environmental Change Climatic Change upplemental Materials: Source attributions ofradiative forcing by regions, sectors, and climateforcers Xuanming Su ∗ , Kaoru Tachiiri , , Katsumasa Tanaka , , MichioWatanabe & Michio Kawamiya Research Institute for Global Change / Research Center for Environmental Modeling andApplication / Earth System Model Development and Application Group, Japan Agency forMarine-Earth Science and Technology (JAMSTEC), Yokohama, Japan Center for GlobalEnvironmental Research, National Institute for Environmental Studies (NIES), Tsukuba,Japan Laboratoire des Sciences du Climat et de lâ ˘A ´ZEnvironnement (LSCE), CommissariatÃ˘a lâ ˘A ´ZÃl’nergie atomique et aux Ãl’nergies alternatives (CEA), Gif-sur-Yvette, France ∗ To whom correspondence should be addressed; e-mail: [email protected].
Supplementary materials • Model equations for SCM4OPT v2.0 • Table S1 to Table S8 • Fig. S1 to Fig. S15 1 odel equations for SCM4OPT v2.0
Carbon cycle
The perturbation of the atmospheric carbon pool includes the following five components, asdefined in Eq. (1).1 CO emissions from fossil fuels and industrial sources;2 Anthropogenic CO emissions into or removal from the terrestrial biosphere;3 CH oxidation of fossil fuels;4 Carbon fluxes to or from the terrestrial biosphere due to CO fertilization and climatefeedback;5 Carbon uptake by oceans. ∆ C atm ( t ) = E indCO ( t ) + E lndCO ( t ) + E fCH ( t ) − F bio ( t ) − F ocn ( t ) (1) Terrestrial carbon cycle
Both the logarithmic and rectangular hyperbolic forms are adoptedto simulate the CO fertilization effects. First, the logarithmic description is defined as: β log ( t ) = 1 + βln (cid:18) C CO ( t ) C CO (cid:19) (2)Second, the rectangular hyperbolic description is given by Eq. (3-5): β sig − r = 1 + β log (cid:0) /C CO (cid:1) β log (cid:0) /C CO (cid:1) (3) β sig − b = (680 − C b ) − β sig − r (340 − C b )( β sig − r −
1) (680 − C b ) (340 − C b ) (4) β sig ( t ) = 1 / (cid:0) C CO − C b (cid:1) + β sig − b / ( C CO ( t ) − C b ) + β sig − b (5)The CO fertilization coefficient ( β fert ) is given by: β fert ( t ) = (2 − β m ) β log ( t ) + ( β m − β sig ( t ) (6)The NPP ( F NP P ( t ) ) and heterotrophic respiration ( F rsp ( t ) ) are defined as productsof the initial carbon flux and a certain fertilization coefficient, considering an exponentialtemperature feedback effect (Eq. 7 and 8). F NP P ( t ) = F NP P β fert ( t ) exp ( σ NP P ∆ T ( t )) (7) F rsp ( t ) = F rsp β fert ( t ) exp ( σ rsp ∆ T ( t )) (8)2he gross land-use emission levels are defined as the sums of the net land-use emissionsand the corresponding regrowth, as shown in Eq. (9-11): D grossP ( t ) = E lndP ( t ) + G P ( t ) (9) D grossH ( t ) = E lndH ( t ) + G H ( t ) (10) D grossS ( t ) = E lndS ( t ) + G S ( t ) (11)Proportions of the net land-use emission levels are allocated as:1 Living plant pool;2 Detritus pool;3 Soil pool.Please refer to Eq. (12-14). E lndP ( t ) = δ P E lndCO ( t ) (12) E lndH ( t ) = δ H E lndCO ( t ) (13) E lndS ( t ) = δ S E lndCO ( t ) (14)The regrowth here is defined to be linearly related to the relaxation time. G P ( t ) = a P + b P τ P ( t ) (15) G H ( t ) = a H + b H τ H ( t ) (16) G S ( t ) = a S + b S τ S ( t ) (17)The relaxation times are defined as follows: τ P ( t ) = P − ψ (cid:82) t E lndP ( t (cid:48) ) dt (cid:48) dP (18) τ H ( t ) = H − ψ (cid:82) t E lndH ( t (cid:48) ) dt (cid:48) dH (19) τ S ( t ) = S − ψ (cid:82) t E lndS ( t (cid:48) ) dt (cid:48) dS (20)3herefore, the annual decay rates for the living plant pool, detritus pool and soil pool aredefined as shown in Eq. (21-23): dP ( t ) = C P ( t ) 1 τ P ( t ) (21) dH ( t ) = C H ( t ) 1 τ H ( t ) exp ( σ H ∆ T ( t )) (22) dS ( t ) = C S ( t ) 1 τ S ( t ) exp ( σ S ∆ T ( t )) (23)The perturbations of carbon in the living plant pool, detritus pool and soil pool at time tare defined as shown in Eq. (24-26): ∆ P ( t ) = F NP P ( t ) ν P − dP ( t ) − D grossP ( t ) − F rsp ( t ) (24) ∆ H ( t ) = F NP P ( t ) ν H − dH ( t ) − D grossH ( t ) + dP ( t ) ρ p d (25) ∆ S ( t ) = F NP P ( t ) ν S − dS ( t ) − D grossS ( t ) + dP ( t ) ρ p s + dH ( t ) δ d s (26)Therefore, the carbon flux to or from the terrestrial biosphere can be calculated asfollows: F bio ( t ) = ∆ P ( t ) + ∆ H ( t ) + ∆ S ( t ) (27)Here, we fitted the land net primary productivity (NPP), land surface net downwardcarbon flux (NBP), ocean surface downward carbon flux (fgco2) and CO concentration ofSCM4OPT v2.0 to the outputs of three CMIP5 experiments, namely, the historical, RCP26and RCP85 experiments. The calibration procedures were performed in several steps, therebyminimizing the sum of squared errors (SSEs) with the associated variables. Oceanic carbon cycle
We apply the method proposed by [1, 2] to construct the oceaniccarbon cycle (Eq.28-39).
DIC ( obx, t ) · (cid:18) K ( obx, t ) H ( obx, t ) + 2 K ( obx, t ) K ( obx, t ) H ( obx, t ) (cid:19) = (cid:18) ALK ( obx, t ) − K B ( obx, t ) BOR ( obx ) K B ( obx, t ) + H ( obx, t ) − K W ( obx, t ) H ( obx, t ) + H ( obx, t ) (cid:19) · (cid:18) K ( obx, t ) H ( obx, t ) + K ( obx, t ) K ( obx, t ) H ( obx, t ) (cid:19) (28) CO sys ( obx, t ) = DIC ( obx, t )1 + K ( obx,t ) H ( obx,t ) + K ( obx,t ) K ( obx,t ) H ( obx,t ) (29)4 CO ( obx, t ) = CO sys ( obx, t ) K H ( obx, t ) (30) HCO ( obx, t ) = DIC ( obx, t )1 + H ( obx,t ) K ( obx,t ) + K ( obx,t ) H ( obx,t ) (31) CO ( obx, t ) = DIC ( obx, t )1 + H ( obx,t ) K ( obx,t ) + H ( obx,t ) K ( obx,t ) K ( obx,t ) (32) K ( obx, t ) = H ( obx, t ) HCO ( obx, t ) CO sys ( obx, t ) (33) K ( obx, t ) = H ( obx, t ) CO ( obx, t ) HCO ( obx, t ) (34) K B ( obx, t ) = H ( obx, t ) BOH ( obx ) BOH ( obx ) (35) BOR ( obx ) = 416 . · S . BOH ( obx ) + BOH ( obx ) (36) K W ( obx, t ) = H ( obx, t ) OH ( obx, t ) (37) F as ( obx, t ) = κ s α s · ( C CO ( t ) − pCO ( obx, t )) (38) F ocn ( t ) = (cid:88) obx F as ( obx, t ) (39)The carbon in the atmospheric pool is converted into the CO concentration by: C CO ( t ) = C atm ( t ) α ppm gtc (40)The radiative forcing from CO can be obtained as: f CO ( t ) = α CO log C CO ( t ) C CO (41)5 H The change in the CH concentration is directly calculated from the CH emissions fromnatural, industrial and land-use sources and from the CH sinks in the troposphere (based onthe lifetime of OH), stratosphere, and soil. ∆ C CH ( t ) = E natCH + E indCH ( t ) + E lndCH ( t ) θ CH − C CH ( t − τ totCH ( t − (42) τ totCH ( t ) = 1 τ initCH /τ relOH ( t ) + 1 τ soilCH + 1 τ othCH (43)The change in the tropospheric OH abundance relative to the level in 2000 is thusmodeled as: τ relOH ( t ) = S τ CH ∆ T k ( t ) + (cid:18) C CH ( t ) C kCH (cid:19) S OHCH · exp (cid:0) S OHNO x ∆ E NO x ( t ) + S OHCO ∆ E CO ( t ) + S OHV OC ∆ E V OC ( t ) (cid:1) (44) N O The feedback effect of the atmospheric N O concentration on its own lifetime is approximatedas: τ N O ( t ) = τ initN O (cid:18) C N O ( t ) C kN O (cid:19) S τN O (45)The change in the atmospheric N O concentration is calculated as: ∆ C N O ( t ) = E natN O + E indN O ( t ) + E lndN O ( t ) θ N O − C N O ( t − τ N O ( t − (46)The radiative forcings of CH ( f CH ( t ) ) and N O ( f N O ( t ) ) are calculated following thestandard IPCC (2001) methods [3], as shown in Eq. (47-49): f CH ( t ) = α CH (cid:16)(cid:112) C CH ( t ) − (cid:113) C CH (cid:17) − (cid:0) f mn (cid:0) C CH ( t ) , C N O (cid:1) − f mn (cid:0) C CH , C N O (cid:1)(cid:1) (47) f N O ( t ) = α N O (cid:16)(cid:112) C N O ( t ) − (cid:113) C N O (cid:17) − (cid:0) f mn (cid:0) C CH , C N O ( t ) (cid:1) − f mn (cid:0) C CH , C N O (cid:1)(cid:1) (48)6he function f mn ( M, N ) defining the overlap between CH and N O is: f mn ( M, N ) =0 .
47 log (cid:32) . (cid:18) M N (cid:19) . + 0 . M (cid:18) M N (cid:19) . (cid:33) (49) Halogenated gases
All the available halogenated gases are treated separately with regard to their concentrations[1, 4]: C hc ( t + 1 , hc ) = τ hc E ( t, hc ) µ hc ρ atm m atm · (cid:18) − exp( − τ hc ) (cid:19) + C hc ( t, hc ) (cid:18) − exp( − τ hc ) (cid:19) (50)The radiative forcing from each halogenated gas is given by: f hc ( t, hc ) = α hc (cid:0) C hc ( t, hc ) − C hc (cid:1) (51) Direct effect of aerosols
We update the estimation of the direct effects from aerosols based on [5]. The change inthe sulfate burden is assessed to capture the radiative forcing impacts resulting from sulfateaerosols. C SO ( t ) = C SO + α SO τ SO (cid:0) E indSO ( t ) + E lndSO ( t ) (cid:1) + α SO τ dms E dms ( t ) + Γ SO ∆ T as ( t ) (52)Similarly, the concentration of primary organic aerosols (POAs) is defined as: C P OA ( t ) = C P OA + τ indOM α P OM E indOC ( t ) + τ lndOM α P OM E lndOC ( t )+Γ P OA ∆ T as ( t ) (53)The black carbon (BC) concentration is: C BC ( t ) = C BC + τ indBC E indBC ( t ) + τ lndBC E lndBC ( t )+Γ BC ∆ T as ( t ) (54)The concentration of nitrate aerosols is: 7 NO ( t ) = C NO + τ NO x (cid:0) E indNO x ( t ) + E lndNO x ( t ) (cid:1) + τ NH (cid:0) E indNH ( t ) + E lndNH ( t ) (cid:1) + Γ NO ∆ T as ( t ) (55)The concentration of secondary organic aerosols (SOAs) is: C SOA ( t ) = C SOA + τ V OC (cid:0) E indV OC ( t ) + E lndV OC ( t ) (cid:1) + τ BV OC E BV OC ( t )+Γ SOA ∆ T as ( t ) (56)Thus, the direct radiative forcing caused by aerosols and pollutants is: f aero ( t ) = α rfaero δC aero ( t ) (57) Mineral dust aerosols
The historical radiative forcing from mineral dust aerosols is obtained from MAGICC 6.0 [4].The future forcing level is assumed to remain at a constant value of -0.1 Wm -2 after 2005. f mindust ( t ) = − . (58) Cloud effects
The tropospheric burden of soluble aerosols can be obtained by: C solu ( t ) = C solu + (cid:88) aero ∈ SO ,P OA,BC,NO ,SOA α aerosolu (cid:0) C aero ( t ) − C aero (cid:1) (59)The cloud forcing effects are estimated by: f cloud ( t ) = f BC ( t ) κ BCadj + φ solu ln (cid:18) C solu ( t ) C solu (cid:19) (60) Stratospheric ozone
The equivalent effective stratospheric chlorine (EESC) concentration is calculated as: C EESC ( t ) = a EESC (cid:32)(cid:88) Cl n Cl f Cl C hc ( t, Cl ) + α br (cid:88) Br n Br f Br C hc ( t, Br ) (cid:33) (61)The concentration of stratospheric ozone is:8 O s ( t ) = C O s + ξ O sEESC (cid:0) C EESC ( t ) − C EESC (cid:1) + ξ O sN O (cid:18) − C EESC ( t ) − C EESC C XEESC (cid:19) ∆ C lagN O ( t ) + Γ O s ∆ T as ( t ) (62)Thus, the forcing effect of the stratospheric ozone burden can be obtained by: f O s ( t ) = α rfO s (cid:0) C O s ( t ) − C O s (cid:1) (63) Tropospheric ozone
The tropospheric ozone concentration is estimated to be: C O t ( t ) = C O t + ξ O tCH ln (cid:18) C CH ( t ) C CH (cid:19) + Γ O t ∆ T as ( t )+ (cid:88) aero ∈ NO x ,CO,V OC ξ O taero (cid:0) E indaero ( t ) + E lndaero ( t ) (cid:1) (64)The radiative forcing from the tropospheric ozone is then calculated as: f O t ( t ) = α rfO t (cid:0) C O t ( t ) − C O t (cid:1) (65) Stratospheric water vapor from CH oxidation The forcing effect of the stratospheric water vapor from CH oxidation f H O ( t ) is calculatedby: f H O ( t ) = α rfH O (cid:113) C CH (cid:115) C lagCH ( t ) C CH − (66) Land-use albedo
The forcing effect from the land-use albedo is estimated according to the annual mean albedoat the biome and regional scales, using the changes in regional land cover as input followingthe methods described in ref [5]. f LCC ( t ) = − π trans φ rsds (cid:88) bio α bioLCC ∆ A bioLCC ( t )∆ A Earth (67)9
C on snow
The forcing effect of BC on snow is determined as a linear function of the BC emission level: f BCSnow ( t ) = a BC + b BC (cid:0) E indBC ( t ) + E lndBC ( t ) (cid:1) (68) Natural sources
Regarding the various natural sources, the volcanic and solar forcings are assumed to be thenatural forcing inputs for CMIP6. f volc ( t ) = f CMIP volc ( t ) (69) f solar ( t ) = f CMIP solar ( t ) (70) Global mean temperature
The estimation of the global mean temperature is based on the Diffusion Ocean Energybalance CLIMate (DOECLIM) model by using the total radiative forcing as input [6, 7].Here, we reestimated the climate sensitivity, vertical ocean diffusivity and radiative forcingcoefficient for CO doubling based on the CMIP5 outputs related to each available GCM. Thedetailed descriptions and equations are contained in the references [6, 7].For the simple climate module, the time step was calibrated to be 1/6 year for SCM4OPTv2.0 to avoid possible convergence problems when calculating the ocean carbon cycle [1]. Thecalibrated results are shown in Figs. S5 to S11. We also included the results produced by othermodels or associated statistical records for comparison purposes. NomenclatureAerosol and pollutants α rfaero Radiative efficiencie for aerosol aeroα
P OM
Conversion of POM from Tg(OC) to Tg(OM) α SO Conversion of SO from TgS to Tg(SO4) α aerosolu Soluble fraction for aerosol aeroδC aero ( t ) Aerosol aero concentration in time t ∆ T as ( t ) Global mean temperature relative to 1850 in time t Γ BC BC sensitivity to global mean temperature Γ NO N O x sensitivity to global mean temperature Γ P OA
Primary organic aerosol sensitivity to global mean temperature Γ SO Sulfate sensitivity to global mean temperature10
SOA
NMVOCs sensitivity to global mean temperature κ BCadj
Adjustment coefficient of BC radiative forcing to cloud forcing effect φ solu Intensity effect coefficient for soluble aerosols τ indBC Lifetime of industrial BC τ lndBC Lifetime of land use BC τ BV OC
Lifetime of biogenic NMVOCs τ dms Lifetime of dimethyl sulfide τ NH Lifetime of
N H τ NO x Lifetime of
N O x τ indOM Lifetime of industrial primary organic aerosols τ lndOM Lifetime of land use primary organic aerosols τ SO Lifetime of SO τ V OC
Lifetime of NMVOCs C aero ( t ) Aerosol concentration in time tC aero Initial aerosol concentration C BC ( t ) Concentration of BC in time tC BC Initial concentration of BC C NO ( t ) Concentration of nitrate aerosols in time tC NO Initial concentration of nitrate aerosols C P OA ( t ) Concentration of primary organic aerosols in time tC P OA
Initial concentration of primary organic aerosols C SO ( t ) Sulfate concentration in time tC SO Initial sulfate concentration C SOA ( t ) Concentration of SOA in time tC SOA
Initial concentration of SOA C solu ( t ) Number concentrations for soluble aerosol in time tC solu Initial number concentrations for soluble aerosol E indBC ( t ) Industrial BC emissions in time tE lndBC ( t ) Land use BC emissions in time tE BV OC ( t ) Biogenic NMVOC emissions in time tE dms ( t ) Dimethyl sulfide emissions E indNH ( t ) Industrial
N H emissions in time tE lndNH ( t ) Land use
N H emissions in time tE indNO x ( t ) Industrial
N O x emissions in time tE lndNO x ( t ) Land use
N O x emissions in time tE indOC ( t ) Industrial OC emissions in time t lndOC ( t ) Land use OC emissions in time tE indSO ( t ) Industrial SO emissions in time tE lndSO ( t ) Land use SO emissions in time tE indV OC ( t ) Industrial NMVOC emissions in time tE lndV OC ( t ) land use NMVOC emissions in time tf aero ( t ) Direct radiative forcing for aerosol aero in time tf BC ( t ) BC radiative forcing in time tf cloud ( t ) Cloud forcing effects in time tf mindust ( t ) Radiative forcing from mineral dust CO α CO Forcing scaling parameter, = . =5.35 Wm -2 [8] α ppm gtc Unit conversion factor from ppm to GtC, = 2.123 GtC ppm -1 α s Solubility of CO in seawater β CO fertilization factor β fert CO fertilization coefficient β log Fertilization coefficient β m Allocation coefficient between the two descriptions of the CO fertilization effects β sig ( t ) Effective CO fertilization factor at time t ∆ C atm ( t ) Atmospheric carbon pool in time t ∆ P ( t ) , ∆ H ( t ) and ∆ S ( t ) Total changes in the carbon levels for the living plant pool, thedetritus pool and the soil pool δ d s Fraction of dH ( t ) that goes to the soil pool δ i land-use emission distribution factors κ s CO transfer velocity ν P , ν H and ν S =1- ν P - ν H NPP partition factors for the living plant pool, the detritus pool andthe soil pool ψ Fraction of gross deforestation without regrowth ρ p d and ρ p s =1- ρ p d Fractions of dP ( t ) that are distributed to the detritus and soil pools,respectively σ H and σ S Temperature feedback coefficients for detritus pool and soil pool σ rsp Sensitivity to changes in temperature τ i ( t ) Regrowth relaxation time, a i and b i are parameters that are estimated based the CMIP5outputs ALK ( obx, t ) Total alkalinity for ocean box obx and time tBOH ( obx ) Ocean boric acid
BOH ( obx ) Ocean borate 12 OR ( obx ) Total boron for ocean box obxC b Concentration at which the NPP is zero, which is taken to be 31 ppm [9] C CO ( t ) Atmospheric CO concentration in time t C CO ( t ) CO concentration C CO Pre-industrial CO concentration (278 ppm) C P ( t ) , C H ( t ) and C S ( t ) Amounts of carbon remaining in the living plant pool, detritus pooland soil pool CO sys ( obx, t ) Dissolved inorganic (DIC) of the system for ocean box obx and time tCO ( obx, t ) Concentration of ocean carbonate CO − for ocean box obx and time tD grossi ( t ) Gross land-use emission level, i ∈ { P, H, S } denote the living plant pool, thedetritus pool and the soil pool, respectively DIC ( obx, t ) Dissolved inorganic for ocean box obx and time tdP , dH and dS Initial decay rates E indCO ( t ) CO emissions from fossil fuels and industrial sources E lndCO ( t ) Anthropogenic CO from or removal to the terrestrial biosphere E lndH ( t ) Detritus pool E lndi ( t ) Net land-use emission level, i ∈ { P, H, S } denote the living plant pool, the detrituspool and the soil pool, respectively E lndP ( t ) Living plant pool E lndS ( t ) Soil pool E fCH ( t ) CH oxidation of fossil fuels F rsp Pre-industrial heterotrophic respiration F as ( obx, t ) Carbon fluxes between the atmosphere and surface ocean box for ocean box obx and time t , if applicable F bio ( t ) Carbon flux to or from the terrestrial biosphere F bio ( t ) Carbon fluxes to or from the terrestrial biosphere due to CO fertilization and climatefeedback f CO ( t ) CO radiative forcing in time tF NP P ( t ) Net primary productivity (NPP) in time t F ocn ( t ) Carbon uptake by the ocean in time t F rsp ( t ) Heterotrophic respiration in time t G i ( t ) Carbon flux originating from regrowth, i ∈ { P, H, S } denote the living plant pool, thedetritus pool and the soil pool, respectively G i ( t ) Land use regrowth, a i and b i are parameters that are estimated based the CMIP5outputs H ( obx, t ) Concentration of [H + ] for ocean box obx and time tHCO ( obx, t ) Concentration of ocean bicarbonate
HCO − for ocean box obx and time t ( obx, t ) First acidity constant of carbonic acid for ocean box obx and time tK ( obx, t ) Second acidity constant of carbonic acid for ocean box obx and time tK B ( obx, t ) Dissociation constant of boric acid for ocean box obx and time tK H ( obx, t ) Henryâ ˘A ´Zs constant for ocean box obx and time tK W ( obx, t ) Dissociation constant of water for ocean box obx and time tOH ( obx, t ) Concentration of OH − P , H and S Initial states of the living plant pool, the detritus pool and the soil pool pCO ( obx, t ) Sea surface partial pressure for ocean box obx and time t Other GHG emissions α CH CH scaling factors, =0.036 α hc Halogenated gas radiative efficiency α N O N O scaling factors, =0.12 ∆ C CH ( t ) Change in the CH concentration in time t ∆ C N O ( t ) N O concentration change in time t ∆ T k ( t ) Temperature change above the 2000 level µ hc Molar mass of halogenated gas hcρ atm
Average density of air τ initCH Initial lifetime of OH, =9.6 years τ initCH /τ relOH ( t ) CH lifetimes in the troposphere τ initN O Initial N O lifetime, =120 years τ othCH CH lifetimes in stratosphere, =120 years τ soilCH CH lifetimes in soil, =160 years τ totCH ( t ) CH lifetime in time tτ hc Lifetime of halogenated gas hcτ N O ( t ) N O lifetime in time tθ CH CH conversion factor, 2.78 Tg ppb -1 θ N O N O conversion factor, =4.81 Tg ppb -1 C CH CH pre-industrial concentration, =721.9 ppb C hc Halogenated gas pre-industrial atmospheric concentration C N O N O pre-industrial concentration, =273.0 ppb C kCH CH concentration in the year 2000 C kN O N O concentration in the year 2000 C hc ( t + 1 , hc ) Concentration (in ppt) of halogenated gas hc in year t + 1 C N O ( t ) N O concentration E ( t, hc ) Halogenated gas emission level of hc in kt yr -1 E indCH ( t ) Industrial CH emissions in time t indN O ( t ) Industrial N O emissions in time tE lndCH ( t ) Land-use sources CH emissions in time tE lndN O ( t ) Land-use sources N O emissions in time tE natCH Natural CH emissions, =274.5 MtCH yr -1 E natN O Natural N O emissions, =8.4 MtN O-N yr -1 f hc ( t, hc ) Halogenated gas radiative forcing M and N CH and N O concentration inputs m atm Total mass of the atmosphere S OHx
Sensitivities of the tropospheric OH to CH , NO x , CO and VOC, with values of -0.32,+0.0042, -1.05E-4 and -3.15E-4, respectively S τ CH CH temperature sensitivity coefficient of tropospheric chemical reactions, =0.0316°C -1 [4] S τ N O N O sensitivity coefficient, =-0.05
Ozone α br Ratio of effectiveness in ozone depletion between bromine and chlorine α O s Stratospheric ozone radiative efficiency α O t Tropospheric ozone radiative efficiency ∆ C lagN O ( t ) N O concentration with time-lag in time t Γ O s Stratospheric ozone sensitivity to global mean temperature Γ O t Tropospheric ozone sensitivity to global mean temperature ξ O taero Tropospheric ozone sensitivity of aerosol aeroξ O tCH Tropospheric ozone sensitivity of CH effect ξ O sEESC Stratospheric ozone sensitivity to EESC ξ O sN O Stratospheric ozone sensitivity to N Oa EESC
A fractional release factor of the EESC C EESC ( t ) EESC concentration in time tC EESC
Initial EESC concentration C XEESC
Non-linear interaction parameter between chlorine and nitrogen chemistries C hc ( t, Cl ) and C hc ( t, Br ) Gas mixing rates in the stratosphere for chlorine and bromine C O s ( t ) Stratospheric ozone concentration in timen tC O s Initial stratospheric ozone concentration C O t ( t ) Tropospheric ozone concentration in time tC O t Initial tropospheric ozone concentration f Cl and f Br Release efficiencies of stratospheric halogens for chlorine and bromine f O s ( t ) Forcing effect of stratospheric ozone burden in time tf O t ( t ) Radiative forcing of tropospheric ozone in time t Cl and n Br Numbers of chlorine and bromine atoms, respectively
Surface albedo α bioLCC Yearly averaged albedo at the for biome ∆ A Earth
Surface area of the Earth ∆ A bioLCC ( t ) Surface area change for biome in time tφ rsds Radiative short-wave and downward flux at the surface π trans Global short-wave and upward transmittance f LCC ( t ) Land-use albedo forcing in time t Other α rfH O Stratospheric water vapor radiative efficiency ∆ C lagCH ( t ) CH concentration with time lag in time ta BC and b BC Forcing scaling parameters of the BC on snow f BCSnow ( t ) Forcing effect of the BC on snow f H O ( t ) Forcing effect of the stratospheric water vapor from CH oxidation in time tf solar ( t ) Solar irradiance forcing effects in time tf CMIP solar ( t ) Solar irradiance forcing effects for CMIP6 in time tf volc ( t ) Volcanic forcing effects in time tf CMIP volc ( t ) Volcanic forcing effects for CMIP6 in time t able S1. Datasets of historical emissions
Source Period Emission Format ReferenceCEDS 1750-2014 CO , CH , BC, CO, NH ,NMVOC, NO x , OC, SO Spatial (sectoral) Ref [10]EDGAR v4.3.2 1970-2012 CO , CH , N O, BC, CO, NH ,NMVOC, NO x , OC, SO Regional and sectoral/Spatial (sectoral) Ref [11]EDGAR v4.2 (*) 1970-2008 CO , CH , N O, CO, NH , F-gases,NF3, SF6, NMVOC, NO x , SO Regional and sectoral/Spatial (sectoral) Ref [12]PRIMAP v2.0 (**) 1850-2016 CO , CH , N O, F-gases, HFCs,PFCs, NF3, SF6 Spatial (sectoral) Ref [13]RCP historical 1850-2000 CH , BC, CO, NH , NO x , OC,SO , VOC Spatial (sectoral) Ref [14](*) Halogenated gas emissions are used in EDGAR v4.3.2 since these emissions are not includedin EDGAR v4.3.2.(**) N O is employed in the other datasets when not included.
Table S2.
Datasets of the future scenarios at the various forcing levels
Forcing levels(Wm -2 ) Source Scenario Reference1.9 AIM/CGE SSP1-1.9, SSP2-1.9 Ref [15]1.9 IAMC SSP1-1.9 Ref [16]2.6 AIM/CGE SSP1-2.6, SSP2-2.6, SSP3-2.6(*), SSP4-2.6, SSP5-2.6 Ref [15]2.6 IAMC SSP1-2.6,Ref [16]3.4 AIM/CGE SSP1-3.4, SSP2-3.4, SSP3-3.4, SSP4-3.4, SSP5-3.4 Ref [15]3.4 IAMC SSP4-3.4, SSP5-3.4-OS Ref [16]4.5 AIM/CGE SSP1-4.5, SSP2-4.5, SSP3-4.5, SSP4-4.5, SSP5-4.5 Ref [15]4.5 IAMC SSP2-4.5 Ref [16]6.0 AIM/CGE SSP1-Baseline, SSP2-6.0, SSP3-6.0, SSP4-Baseline, SSP5-6.0 Ref [15]6.0 IAMC SSP3-LowNTCF(**), SSP4-6.0 Ref [16]7.0 AIM/CGE SSP2-Baseline, SSP3-Baseline Ref [15]7.0 IAMC SSP3-7.0 Ref [16]8.5 AIM/CGE SSP5-Baseline Ref [15]8.5 IAMC SSP5-8.5 Ref [16](*) The SSP3-2.6 scenario was not available in Table 2 in ref [15], however, the dataset was provided inhttps://doi.org/10.7910/DVN/4NVGWA. We retained SSP3-2.6 in our analysis.(**) The target forcing level of SSP3-LowNTCF was 6.3 Wm -2 (Table 1 in ref [16]). We classified it tothe closest forcing level of 6.0 Wm -2 . Table S3.
Datasets of CO emissions from land-use change Source Period Format ReferenceHoughton et al. (2012) (*) 1960-2010 Regional Ref [17, 18]MPIMET 1850-2005 Spatial grid Ref [19]PRIMAP v1.2 1850-2015 Regional Ref [13]Smith and Rothwell (2013) 1850-2010 Regional Ref [20] able S4. Please refer to the spreadsheet in the supplementary tables. Atmospheric driversand radiative forcings. Note: This table is compiled based on Figure SPM.5 in IPCC (2013)and references Gasser et al. (2016) and Su et al. (2017). All emissions from internationalshipping activities are regional nonattributable.
Table S5. able S6. Equilibrium climate sensitivity (ECS) used in this study compared to other referencesModel This study Ref [21] Ref [22] Ref [23] Ref [24] Ref [25] Ref [26] Ref [27]ACCESS1-0 3.88 - 3.83 3.8 3.79 3.45 3.76 3.8ACCESS1-3 3.59 - - - 3.45 2.8 3.22 -bcc-csm1-1 2.80 - 2.82 2.8 2.88 - 2.73 2.8bcc-csm1-1-m 2.79 - 2.87 2.9 - - 3.1 -BNU-ESM 4.11(*) - - 4.1 4.11 - 4.08 4.1CanESM2 3.66 3.69 3.69 3.7 3.68 3.6 3.63 3.7CCSM4 2.90 - 2.89 2.9 2.92 - 2.8 2.9CNRM-CM5 3.27 3.25 3.25 3.3 3.25 3.16 3.07 3.3CNRM-CM5-2 3.46 - - - - - - -CSIRO-Mk3-6-0 4.24 4.08 4.08 4.1 3.99 2.96 3.55 4.1FGOALS-g2 3.45(*) - - - 3.45 - 2.46 3.45FGOALS-s2 4.16(*) - 4.17 - 4.16 - 4.14 4.16GFDL-CM3 3.97 3.97 3.97 4 3.96 3.2 3.85 4GFDL-ESM2G 2.57 2.39 2.39 2.4 2.38 - 1.81 -GFDL-ESM2M 2.71 2.44 2.44 2.4 2.41 - 2.23 2.4HadGEM2-ES 4.58 4.59 4.59 4.6 4.55 4.32 4.6 4.6IPSL-CM5A-LR 4.05 4.13 4.13 4.1 4.1 3.46 3.92 4.1IPSL-CM5A-MR 4.11 - - - - 3.4 - -IPSL-CM5B-LR 2.64 - 2.61 2.6 2.59 - 2.43 2.6MIROC5 2.70 2.72 2.72 2.7 2.71 2.12 2.22 2.7MIROC-ESM 4.67 4.67 4.67 4.7 4.65 3.47 3.88 4.7MPI-ESM-LR 3.64 3.63 3.63 3.6 3.6 3.08 3.27 -MPI-ESM-MR 3.48 - - - 3.44 2.94 3.14 3.4MPI-ESM-P 3.47 3.45 3.45 - 3.42 - 3.07 -MRI-CGCM3 2.60 2.6 2.6 2.6 2.59 2.19 2.52 2.6NorESM1-M 2.82 2.8 2.8 2.8 2.83 2.11 2.48 2.8 (*) The ECS values for BNU-ESM, FGOALS-g2 and FGOALS-s2 are retrieved from ref [24]. Allother values in this study are estimated by using the standard regression method [22, 28] based onthe available CMIP5 experiments of the preindustrial control (piControl) and abrupt 4xCO scenario(abrupt4xCO2). (**) Based on Table 9.5 in IPCC AR5-WG1 [23]. Table S7.
Please refer to the spreadsheet in the supplementary tables, Sector mapping. Note:(*) The AIM/CGE negative CO2 and land-use CO2 emissions are extracted from the regionaldataset rather than from the spatial dataset. (**) Forest burning and grassland burning levelsare adjusted based on the percentage share in 2012 in EDGAR v4.3.2. able S8. An overview of the iterations regarding the climate system, scenarios, regions,sectors and emissions
Sources Quantity CompositionClimate system 63 Terrestrial carbon cycle, ocean carbon cycle, aerosols andpollutants, climate influences, cloud effects, climate systemScenarios 7 1.9 Wm -2 , 2.6 Wm -2 , 3.4 Wm -2 , 4.5 Wm -2 , 6.0 Wm -2 , 7.0 Wm -2 and 8.5 Wm -2 Regions 11 CHN, IND, JPN, RUS, USA, AFR, EUR, LAM, MEA, OAS, ROWSectors 12 Agriculture, agricultural waste burning, domestic housingand commercial, energy, industry, industrial solvents, surfacetransportation, waste treatment, open forest burning, open grasslandburning, aviation and international shippingEmissions 48 Industrial CO , land-use CO , CH , N O, BC, CO, NH , NO x , OC,SO , VOCs and halogenated gases (a total of 37 gases includingHFC-23, HFC-32, HFC-125, HFC-134a, HFC-143a, HFC-152a,HFC-227ea, HFC-236fa, HFC-245fa, HFC-365mfc, HFC-43-10mee,CF , C F , C F , c-C F , C F , C F , C F , C F ; SF , NF ,CFC-11, CFC-12, CFC-113, CFC-114, CFC-115, CCl , CH CCl ,HCFC-22, HCFC-141b, HCFC-142b, Halon-1211, Halon-1202,Halon-1301, Halon-2402, CH Br and CH Cl) llllll ( G t C / y r) a Fossil fuel CO lllllll −10123 −101231850 1900 1950 2000 2050 2100 ( G t C / y r) b Land−use change CO lllllll ( M t CH / y r) c Anthropogenic CH (excl. open burning) lllllll ( M t CH / y r) d Open burning CH lllllll ( M t N O − N / y r) e Anthropogenic N O (excl. open burning) lllllll ( M t N O − N / y r) f Open burning N O lllllll ( G t C O − eq / y r) g Kyoto Protocol fluorinated gases lllllll ( G t C O − eq / y r) h Montreal Protocol ozone−depleting substances (ODS)
Forcing 1.9 Wm - Wm - Wm - Wm - Wm - Wm - Wm - Figure S1.
Historical and future GHG emissions.
The future projections includeseven forcing levels, namely, 1.9 Wm -2 , 2.6 Wm -2 , 3.4 Wm -2 , 4.5 Wm -2 , 6.0 Wm -2 , 7.0Wm -2 and 8.5 Wm -2 . The uncertainty ranges denote the upper and lower trends. Theerror bars to the right show the upper and lower trends in 2100 at each forcing level.Open burning includes the emissions from agricultural waste burning, forest fires andgrassland fires. Sources: the historical emissions stem from ref [10, 11, 13, 14]; thefuture trends stem come ref [15, 16]; land-use CO originates from ref [13, 19, 20, 29];open burning is from ref [30]. llllll ( M t B C / y r) a Anthropogenic BC (excl. open burning) lllllll
12 121850 1900 1950 2000 2050 2100 ( M t B C / y r) b Open burning BC lllllll ( M t O C / y r) c Anthropogenic OC (excl. open burning) lllllll ( M t O C / y r) d Open burning OC lllllll ( M t C O / y r) e Anthropogenic CO (excl. open burning) lllllll ( M t C O / y r) f Open burning CO lllllll ( M t N / y r) g Anthropogenic nitrate (excl. open burning) lllllll ( M t N / y r) h Open burning nitrateForcing 1.9 Wm - Wm - Wm - Wm - Wm - Wm - Wm - Figure S2.
Historical and future aerosol and pollutant emissions (a-h).
The futureprojections include seven forcing levels, namely, 1.9 Wm -2 , 2.6 Wm -2 , 3.4 Wm -2 , 4.5Wm -2 , 6.0 Wm -2 , 7.0 Wm -2 and 8.5 Wm -2 . The uncertainty ranges denote the upperand lower trends. The error bars to the right show the upper and lower trends in 2100at each forcing level. Open burning includes the emissions from agricultural wasteburning, forest fires and grassland fires. Sources: the historical emissions stem fromref [10, 11, 13, 14]; future trends come from ref [15, 16]; open burning originates fromref [30]. llllll ( M t S / y r) i Anthropogenic sulfate (excl. open burning) lllllll ( M t S / y r) j Open burning sulfate lllllll ( M t NH − N / y r) k Anthropogenic NH (excl. open burning) lllllll ( M t NH − N / y r) l Open burning NH lllllll ( M t V O C / y r) m Anthropogenic
VOC (excl. open burning) lllllll ( M t V O C / y r) n Open burning
VOCForcing 1.9 Wm - Wm - Wm - Wm - Wm - Wm - Wm - Figure S3.
Historical and future aerosol and pollutant emissions (i-n).
The futureprojections include seven forcing levels, namely, 1.9 Wm -2 , 2.6 Wm -2 , 3.4 Wm -2 , 4.5Wm -2 , 6.0 Wm -2 , 7.0 Wm -2 and 8.5 Wm -2 . The uncertainty ranges denote the upperand lower trends. The error bars to the right show the upper and lower trends in 2100at each forcing level. Open burning includes the emissions from agricultural wasteburning, forest fires and grassland fires. Sources: the historical emissions are fromref [10, 11, 13, 14]; future trends come from ref [15, 16]; open burning stems fromref [30]. ( m illi on ha ) Desert & urbanForestGrassland & shrublandCroplandPasturea Historical land cover changes b Future land cover projection lllllll −1600−1400−1200−1000 2020 2040 2060 2080 2100 ( m illi on ha ) b−1 Desert & urban lllllll −1200−800−4000 2020 2040 2060 2080 2100 ( m illi on ha ) b−2 Forest lllllll −1900−1800−1700−1600−1500 2020 2040 2060 2080 2100 ( m illi on ha ) b−3 Grassland & shrubland lllllll ( m illi on ha ) b−4 Cropland lllllll ( m illi on ha ) b−5 Pasture Forcing 1.9 Wm - Wm - Wm - Wm - Wm - Wm - Wm - Figure S4.
Historical and future land cover changes, compared to the values in1700.
The future projections include seven forcing levels, namely, 1.9 Wm -2 , 2.6Wm -2 , 3.4 Wm -2 , 4.5 Wm -2 , 6.0 Wm -2 , 7.0 Wm -2 and 8.5 Wm -2 . The uncertaintyranges denote the upper and lower trends. The error bars to the right show the upperand lower trends in 2100 at each forcing level. Sources: LUH2 v2h [31]; LUH2v2f [32]; AIM-SSP/RCP gridded emission and land-use data [15]. ll l l l l l l l l l l l l l l l l l l l l l l l l l l C O ( W m - ) lll SCM4OPT v2.0MAGICC6 IPCC AR5OSCAR v2.2 b ll l l l l l l l l l l l l l l l l l l l l l l l l l l CH ( W m - ) lll SCM4OPT v2.0MAGICC6 OSCAR v2.2 c ll l l l l l l l l l l l l l l l l l l l l l l l l l l N O ( W m - ) lll SCM4OPT v2.0MAGICC6 OSCAR v2.2 d l l l l l l l l l l l l l l l l l l l l l l l l l l l F − G a s e s ( W m - ) ll SCM4OPT v2.0 MAGICC6 e ll l l l l l l l l l l l l l l l l l l l l l l l l l l O D S ( W m - ) ll SCM4OPT v2.0 MAGICC6
Figure S5.
Simulation of the radiative forcings induced by greenhouse gases(GHGs) compared to existing studies (IPCC AR5 [33], MAGICC6 [4] andOSCAR v2.2 [5]).
The uncertainties in SCM4OPT v2.0 indicate the 17th and 83rdpercentiles. The MAGICC6 time series are extracted from RCP calculations [34]. TheOSCAR v2.2 uncertainties are produced by 500 runs, accounting for the 17th and 83rdpercentiles, downloaded from https://github.com/tgasser/OSCARv2. The error bars in2011 denote the forcing values over the period of 1750-2011 in IPCC AR5 (Table 8.2). ll l l l l l l l l l l l l l l l l l l l l l l l l l l B C ( W m - ) lll SCM4OPT v2.0MAGICC6 OSCAR v2.2 b ll l l l l l l l l l l l l l l l l l l l l l l l l l l −0.8−0.6−0.4−0.20 −0.8−0.6−0.4−0.201850 1900 1950 2000 S O ( W m - ) lll SCM4OPT v2.0MAGICC6 OSCAR v2.2 c ll l l l l l l l l l l l l l l l l l l l l l l l l l l −0.3−0.2−0.10 −0.3−0.2−0.101850 1900 1950 2000 N O x ( W m - ) lll SCM4OPT v2.0MAGICC6 OSCAR v2.2 d ll l l l l l l l l l l l l l l l l l l l l l l l l l l −0.3−0.2−0.100.1 −0.3−0.2−0.100.11850 1900 1950 2000 D u s t ( W m - ) ll SCM4OPT v2.0 MAGICC6 e ll l l l l l l l l l l l l l l l l l l l l l l l l l l −1.5−1−0.50 −1.5−1−0.501850 1900 1950 2000 C l oud ( W m - ) lll SCM4OPT v2.0MAGICC6 OSCAR v2.2 f ll l l l l l l l l l l l l l l l l l l l l l l l l l l −0.5−0.4−0.3−0.2−0.10 −0.5−0.4−0.3−0.2−0.101850 1900 1950 2000 P O A ( W m - ) lll SCM4OPT v2.0 OSCAR v2.2 g ll l l l l l l l l l l l l l l l l l l l l l l l l l l −0.3−0.2−0.100.10.2 −0.3−0.2−0.100.10.21850 1900 1950 2000 S O A ( W m - ) lll SCM4OPT v2.0 OSCAR v2.2
Figure S6.
Simulation of the radiative forcings induced by aerosols and pollutants,compared to existing studies (IPCC AR5 [33, 35], MAGICC6 [4] and OSCARv2.2 [5]).
The uncertainties in SCM4OPT v2.0 indicate the 17th and 83rd percentiles.The MAGICC6 time series are extracted from RCP calculations [34]. The OSCARv2.2 uncertainties are produced by 500 runs, accounting for the 17th and 83rdpercentiles. The error bars in 2011 denote the forcing values over the period of 1750-2011 in IPCC AR5 (Table 8.4 and Figure SPM.5). ll l l l l l l l l l l l l l l l l l l l l l l l l l l −0.16−0.12−0.08−0.0400.04 −0.16−0.12−0.08−0.0400.041850 1900 1950 2000 S t r a t o s phe r i c o z one ( W m - ) lll SCM4OPT v2.0MAGICC6 IPCC AR5OSCAR v2.2 b ll l l l l l l l l l l l l l l l l l l l l l l l l l l T r opo s phe r i c o z one ( W m - ) lll SCM4OPT v2.0MAGICC6 IPCC AR5OSCAR v2.2 c ll l l l l l l l l l l l l l l l l l l l l l l l l l l S t r a t o s phe r i c w a t e r v apou r ( W m - ) lll SCM4OPT v2.0MAGICC6 IPCC AR5OSCAR v2.2 d ll l l l l l l l l l l l l l l l l l l l l l l l l l l −0.3−0.2−0.10 −0.3−0.2−0.101850 1900 1950 2000 Land − u s e a l bedo ( W m - ) lll SCM4OPT v2.0MAGICC6 IPCC AR5OSCAR v2.2 e ll l l l l l l l l l l l l l l l l l l l l l l l l l l B l a ck c a r bon on s no w ( W m - ) lll SCM4OPT v2.0MAGICC6 IPCC AR5OSCAR v2.2
Figure S7.
Simulation of the radiative forcings induced by human activities,other than the GHGs and aerosols and pollutants above, compared to existingstudies (IPCC AR5 [33], MAGICC6 [4] and OSCAR v2.2 [5]).
The uncertainties inSCM4OPT v2.0 indicate the 17th and 83rd percentiles. The MAGICC6 time series areextracted from RCP calculations [34]. The OSCAR v2.2 uncertainties are producedby 500 runs, accounting for the 17th and 83rd percentiles. The error bars in 2011denote the forcing values over the period of 1750-2011 in IPCC AR5 (Table 8.6). l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l −4−20 −4−201850 1900 1950 2000 2050 2100 V o l c an i c ( W m - ) lll SCM4OPT v2.0 MAGICC6 IPCC AR5 OSCAR v2.2 b l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l −0.100.10.20.3 −0.100.10.20.31850 1900 1950 2000 2050 2100 S o l a r i rr ad i an c e ( W m - ) lll SCM4OPT v2.0 MAGICC6 IPCC AR5 OSCAR v2.2
Figure S8. Assumptions of the radiative forcings induced by the natural sources ofvolcanic activity and solar irradiance compared to existing studies.
The volcanic andsolar irradiance forcings used in SCM4OPT v2.0 are assumed in accordance with volcanicactivity [36] and solar irradiance [37] forcing inputs for CMIP6, and the volcanic forcing isnormalized to zero in 1850. l l l l l l l l l l l l l l l l −0.3−0.2−0.100.10.2 −0.3−0.2−0.100.10.21850 1900 1950 2000 2050 2100 Land − u s e a l bedo c hange ( W m - ) Forcing 1.9 Wm - Wm - Wm - Wm - Wm - Wm - Wm - Reference llll
CMIP5IPCC AR5RCP3PD (MAGICC6)RCP45 (MAGICC6) RCP60 (MAGICC6)RCP85 (MAGICC6)REMIND 1.7
Figure S9. Land-use albedo forcings estimated by SCM4OPT v2.0 compared to existingstudies. l l l l l l l l l l l l l l l l l l l l l l l l l l l −10123 −101231850 1900 1950 2000 T o t a l r ad i a c t i v e f o r c i ng ( W m - ) lll SCM4OPT v2.0MAGICC6IPCC AR5OSCAR v2.2
Figure S10. Total radiative forcing simulated by SCM4OPT v2.0 compared to existingstudies (IPCC AR5 [33], MAGICC6 [4] and OSCAR v2.2 [5]).
The uncertainties inSCM4OPT v2.0 indicate the 17th and 83rd percentiles. The MAGICC6 time series areextracted from RCP calculations [34]. The OSCAR v2.2 uncertainties are produced by 500runs, accounting for the 17th and 83rd percentiles. The error bars in 2011 denote the totalanthropogenic radiative forcing relative to 1750 (Figure SPM.5). G l oba l m ean t e m pe r a t u r e r e l a t i v e t o − ( (cid:176) C ) SCM4OPT v2.0GISTEMP v4HadCRUT 4.6Japan Meteorological Agency (JMA)
Figure S11. Historical global mean temperature increase above the preindustrial level,generated by SCM4OPT v2.0 and compared to existing statistical records.
The anomaliesdeviate from the average over 1890-1910. The SCM4OPT v2.0 uncertainties result from theemission source- (CEDS [10], EDGAR v4.3.2 [11] and RCP historical [34]) and climateuncertainties described in this paper. The uncertainties in HadCRUT 4.6 indicate the 95%confidence interval of the combined effects of all the uncertainties described in the HadCRUT4error model. GISTEMP v4 from ref [38]; HadCRUT 4.6 from ref [39]; Japan MeteorologicalAgency (JMA) from ref [40]. .8Wm - - - - - - - radiactive forcing (Wm - ) a (cid:176) C 1.8 (cid:176)
C 2.3 (cid:176)
C2.9 (cid:176)
C 3.4 (cid:176)
C4.1 (cid:176)
C 4.6 (cid:176) C mean temperature relative to ( (cid:176) C) b Forcing 1.9 Wm - Wm - Wm - Wm - Wm - Wm - Wm - Figure S12. Probability distributions of the total radiative forcing and global meantemperature at forcing levels of 1.9 Wm -2 , 2.6 Wm -2 , 3.4 Wm -2 , 4.5 Wm -2 , 6.0 Wm -2 ,7.0 Wm -2 and 8.5 Wm -2 , estimated by SCM4OPT v2.0. a, Total radiative forcing in 2100.b, Global mean temperature increase relative to 1850 in 2100. The color values indicate themean value at each forcing level. mission (E)Radiative forcing (F) E A E B F A F B Tangent
Figure S13. The normalized marginal method for the attributions of radiative forcings.
The figure is plotted based on Figure 5 in ref [41]. ad i a t i v e f o r c i ng ( W m - ) CHN IND JPN RUS USA AFR EUR LAM MEA OAS ROW C O O t he r C O O t he r C O O t he r C O O t he r C O O t he r C O O t he r C O O t he r C O O t he r C O O t he r C O O t he r C O O t he r −0.4−0.200.20.4 −0.4−0.200.20.4 a Historical l l l l l l l l l l l l l l l l l l l l l l
CHN IND JPN RUS USA AFR EUR LAM MEA OAS ROW C O O t he r C O O t he r C O O t he r C O O t he r C O O t he r C O O t he r C O O t he r C O O t he r C O O t he r C O O t he r C O O t he r −0.200.20.4 −0.200.20.4 b (cid:176) C l l l l l l l l l l l l l l l l l l l l l l CHN IND JPN RUS USA AFR EUR LAM MEA OAS ROW C O O t he r C O O t he r C O O t he r C O O t he r C O O t he r C O O t he r C O O t he r C O O t he r C O O t he r C O O t he r C O O t he r −0.200.20.4 −0.200.20.4 c (cid:176) C Negative CO LUC CO FF CO CH Strat. H ON OHalogenatedStrat. O Tropo. O SulfateNitratePOA SOABCBC on snowCloud Land albedo Figure S14.
Regional forcings are decomposed into CO -induced forcings andthose not directly related to CO . a, Historical period (1850-2016); b, 2 °C (1850-2100); c, 1.5 °C (1850-2100). The direct CO emissions are separated into fossil-fuelCO (FF CO ), land-use CO (LUC CO ), and negative CO emissions, if applicable.The value on top of the bar indicates the mean value summing all components of theleft CO bar. The value at the bottom of the bar indicates the mean value summingall components of the right bar. All uncertainties are represented as one standarddeviation. .9 Wm - Wm - Wm - Wm - Wm - Wm - Wm - A I M / C G E I A M C B o t h A I M / C G E I A M C B o t h A I M / C G E I A M C B o t h A I M / C G E I A M C B o t h A I M / C G E I A M C B o t h A I M / C G E I A M C B o t h A I M / C G E I A M C B o t h C u m u l a t i v e C O e m i ss i on s s i n c e ( G t C ) Figure S15.
Cumulative CO emissions projected by AIM/CGE and IAMC. Theuncertainties are represented as one standard deviation. All the AIM/CGE projectionsare slightly higher than those by the IAMC. This figure shows only the cumulative CO emissions, representing part of the systematic deviations, and variations regardingaerosols and pollutants also occur.
1] Hartin C A, Patel P, Schwarber A, Link R P and Bond-Lamberty B P 2015
Geoscientific ModelDevelopment Biogeosciences Radiative Forcing of Climate Change (Cambridge, United Kingdom and New York,NY, USA: Cambridge University Press) chap 6, pp 349–416[4] Meinshausen M, Raper S C B and Wigley T M L 2011
Atmospheric Chemistry and Physics GeoscientificModel Development Geoscientific ModelDevelopment Geophysical Research Letters Implications of CO2 Effects on Vegetation for the Global Carbon Budget (Berlin,Heidelberg: Springer Berlin Heidelberg) pp 159–199 ISBN 978-3-642-84608-3[10] Hoesly R M, Smith S J, Feng L, Klimont Z, Janssens-Maenhout G, Pitkanen T, Seibert J J, Vu L, AndresR J, Bolt R M, Bond T C, Dawidowski L, Kholod N, Kurokawa J I, Li M, Liu L, Lu Z, Moura M C P,O’Rourke P R and Zhang Q 2018
Geoscientific Model Development Earth System Science Data Earth SystemScience Data Scientific Data Geoscientific Model Development Biogeosciences Global Biogeochemical Cycles emissions[20] Smith S J and Rothwell A 2013 Biogeosciences Geophysical Research Letters [22] Forster P M, Andrews T, Good P, Gregory J M, Jackson L S and Zelinka M 2013 Journal of GeophysicalResearch: Atmospheres
Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to theFifth Assessment Report of the Intergovernmental Panel on Climate Change
ISBN 9789291691432[24] Sherwood S C, Bony S and Dufresne J L 2014
Nature
Philosophical Transactions of the Royal Society of London A:Mathematical, Physical and Engineering Sciences [26] Tsutsui J 2017
Climatic Change
Tellus Series a-Dynamic Meteorology and Oceanography Geophysical Research Letters [29] Le Quéré C, Andrew R M, Friedlingstein P, Sitch S, Hauck J, Pongratz J, Pickers P A, Korsbakken J I, eters G P, Canadell J G, Arneth A, Arora V K, Barbero L, Bastos A, Bopp L, Chevallier F, Chini L P,Ciais P, Doney S C, Gkritzalis T, Goll D S, Harris I, Haverd V, Hoffman F M, Hoppema M, HoughtonR A, Hurtt G, Ilyina T, Jain A K, Johannessen T, Jones C D, Kato E, Keeling R F, Goldewijk K K,Landschützer P, Lefèvre N, Lienert S, Liu Z, Lombardozzi D, Metzl N, Munro D R, Nabel J E M S,Nakaoka S I, Neill C, Olsen A, Ono T, Patra P, Peregon A, Peters W, Peylin P, Pfeil B, Pierrot D, PoulterB, Rehder G, Resplandy L, Robertson E, Rocher M, Rödenbeck C, Schuster U, Schwinger J, SéférianR, Skjelvan I, Steinhoff T, Sutton A, Tans P P, Tian H, Tilbrook B, Tubiello F N, van der Laan-LuijkxI T, van der Werf G R, Viovy N, Walker A P, Wiltshire A J, Wright R, Zaehle S and Zheng B 2018 EarthSystem Science Data Geoscientific Model Development Anthropogenicand Natural Radiative Forcing (Cambridge, United Kingdom and New York, NY, USA: CambridgeUniversity Press) book section 8, pp 659–740 ISBN ISBN 978-1-107-66182-0[34] Meinshausen M, Smith S J, Calvin K, Daniel J S, Kainuma M L T, Lamarque J F, Matsumoto K, MontzkaS A, Raper S C B, Riahi K, Thomson A, Velders G J M and van Vuuren D P 2011
Climatic Change
Summary for Policymakers (Cambridge, United Kingdom and New York, NY, USA: CambridgeUniversity Press) book section SPM, pp 1–30 ISBN ISBN 978-1-107-66182-0[36] Zanchettin D, Khodri M, Timmreck C, Toohey M, Schmidt A, Gerber E P, Hegerl G, Robock A, PausataF S, Ball W T, Bauer S E, Bekki S, Dhomse S S, Le Grande A N, Mann G W, Marshall L, Mills M,Marchand M, Niemeier U, Poulain V, Rozanov E, Rubino A, Stenke A, Tsigaridis K and Tummon F2016
Geoscientific Model Development Geoscientific Model Development Journal of Geophysical Research: Atmospheres [40] (JMA) J M A 2019 Annual Anomalies of Global Average Surface Temperature (° C)[41] UNFCCC 2002 Methodological Issues: Scientific and Methodological Assessment of Contributions toClimate Change, Report of the Expert Meeting, Note by the Secretariat Tech. rep. URL https://unfccc.int/resource/docs/2002/sbsta/inf14.pdfhttps://unfccc.int/resource/docs/2002/sbsta/inf14.pdf