Na u e Clima e Change | Volume 15 | Feb ua y 2025 | 218–226 218
na u e clima e change
Analysis h ps://doi.o g/10.1038/s41558-024-02198-6
Ene gy and socioeconomic sys em
ans o ma ion h ough a decade o
IPCC-assessed scena ios
D. J. an de Ven 1 , S. Mi al 2,3, A. Nikas 4, G. Xexakis 5, A. Gambhi 2,
L. He mwille 6, P. F agkos 7, W. Obe gassel 6, M. Gonzalez-Eguino 1,8,9,
F. Filippidou7, I. Sognnaes 3, L. Cla ke10 & G. P. Pe e s 3
Cha ing u u e emissions pa hways is a cen al ene o IPCC assessmen
epo s (AR), ye i is unclea how unde lying d i e s (including a ound
policy and echnology) ha e in luenced he e olu ion o emissions
pa hways. He e we compa e scena ios in AR5 and AR6 and ind ha
scena ios wi hou speci ic clima e policies en o ced ha e shi ed lowe in
each scena io gene a ion, owing o alling low-ca bon echnology cos s
and educed expec a ions o economic g ow h, educing ossil- uel
sha es in ene gy and indus y. Mi iga ion pa hways consis en wi h 1.5–2 °C
ha e seen inc easing elec i ica ion a es and highe sha es o a iable
enewables in elec ici y in mo e ecen scena io gene a ions, implying
educed eliance on coal, nuclea , bioene gy and ca bon cap u e and
s o age, e lec ing changing cos s. Despi e he sh inking ca bon budge due
o insu icien ecen clima e ac ion, mi iga ion cos s ha e no inc eased
gi en mo e op imis ic low-ca bon echnology cos p ojec ions. Mo ing
o wa d, scena io p oduce s mus con inually ecalib a e o keep ab eas o
echnology, policy and socie al de elopmen s o emain policy ele an .
The decade in be ween he Six h (AR6)1 and Fi h (AR5)2 assessmen
epo s o he IPCC has seen d ama ic cos educ ions and pe o mance
imp o emen s in key clean echnologies such as sola pho o ol aic
(PV), ba e y s o age and wind powe 3–7. Howe e , i has also seen
ela i e s agna ion in long-conside ed mi iga ion ‘game-change s’
8
,
esul ing in conce ns o e he plausibili y o he di usion scales
assumed in modelled mi iga ion pa hways— ega ding ca bon cap-
u e and s o age (CCS)
9
o ca bon dioxide emo al (CDR), pa icula ly
bioene gy wi h CCS (BECCS)
10–13
. The same decade saw he in oduc-
ion o low-ca bon policies and a ge s14 ha ha e helped shi away
om ‘baseline’ (no-clima e-policy) scena ios
15
—alongside inc eased
clima e awa eness among ci izens
16
and dis up i e socioeconomic
e en s amids linge ing e ec s o he G ea Recession ha include a
pandemic, in e na ional con lic and la ge-scale ene gy c isis.
These de elopmen s ha e had a ied implica ions o clima e
ac ion, he eby al oge he o e ing use ul oppo uni ies o assess
he capaci y o in eg a ed assessmen models (IAMs) o keep up17
and o unde s and how he insigh s eme ging om IAM exe cises
ha e changed h oughou he decade. In his pe iod, he scien i ic
ield o IAMs has g own subs an ially in esea ch ou pu 18 and use
in e es
19
, d i en by inc easing dedica ion o he clima e eme gency
20
and imp o emen s in compu a ional powe
21
. This g ow h has allowed
housands o scena ios p ojec ing mi iga ion pa hways wi h p ede ined
emissions o clima e a ge s o be published in he li e a u e, e ealing
insigh s on how emission educ ions migh be p io i ized ac oss
coun ies and sec o s.
To ully e lec on wha can be lea n om ecen scena io
li e a u e and unde s and how po en ial de elopmen s onwa ds can
Recei ed: 20 Feb ua y 2024
Accep ed: 31 Oc obe 2024
Published online: 3 Janua y 2025
Check o upda es
1Basque Cen e o Clima e Change, Leioa, Spain. 2G an ham Ins i u e, Impe ial College, London, UK. 3CICERO Cen e o In e na ional Clima e Resea ch,
Oslo, No way. 4Na ional Technical Uni e si y o A hens, A hens, G eece. 5HOLISTIC, A hens, G eece. 6Wuppe al Ins i u e o Clima e, En i onmen and
Ene gy, Wuppe al, Ge many. 7E3 Modelling, A hens, G eece. 8IKERBASQUE, Basque Founda ion o Science, Bilbao, Spain. 9Uni e si y o he Basque
Coun y (UPV/EHU), Bilbao, Spain. 10Bezos Ea h Fund, Washing on, DC, USA. e-mail: dj. ande en@bc3 esea ch.o g
Na u e Clima e Change | Volume 15 | Feb ua y 2025 | 218–226 219
Analysis h ps://doi.o g/10.1038/s41558-024-02198-6
o sys ema ic compa ison o se e al dimensions o mi iga ion. We
addi ionally include inpu s om he Special Repo on Emissions
Scena ios (SRES), which was used o in o m he AR3 and AR4 ( e . 27).
As baseline-only scena ios, SRES a e compa ed exclusi ely o baseline
scena ios ac oss ensembles. All obse a ions h oughou his sec ion
a e ‘model means’— ha is, he mean indica o alue om all scena ios
o simila clima e ambi ion om he same model (Me hods), gua an ee-
ing esul s a e no biased owa ds speci ic models wi h a la ge numbe
o scena ios in IPCC assessmen s.
E olu ion o baseline scena ios
Baseline scena ios a e an impo an e e ence poin . Fi s , hey a e
uncons ained by ca bon budge s and mos ly d i en by assumed
con inua ion o cu en echnology, economic and socie al ends.
Analysing baseline scena ios, he e o e, is impo an o unde s and
he signi icance o he obse ed di e ences in mo e ambi ious sce-
na ios. Second, he ans o ma i e po en ial o a p ojec o poli ical
ins umen has been a gued o co espond o i s po en ial o shi
om an assumed business-as-usual (BAU)28. BAU scena ios should
e lec he expec ed change unde socioeconomic assump ions,
be add essed
22
, i is ins uc i e o assess wha scena io da abase ensem-
bles say abou he e ol ing ole o echnology and clima e policy
23
.
He e, we p esen such analysis compa ing all scena ios assessed wi hin
he IPCC epo ing p ocess spanning he las decades. The analysis
o scena io ensembles spanning ou IPCC cycles p o ides a mo e
dynamic o e iew o p og ess, allowing o ack he e olu ion o base-
line p ojec ions—as a p oxy o clima e e o e ec i eness, imp o e-
men s in modelling capaci y and ends in low-ca bon echnoeconomic
de elopmen s—and, in u n, o he solu ion space p o ided by IAMs
o achie e p ede ined clima e a ge s. Finally, we aim o unde s and
how mi iga ion po olios ha e e ol ed in IAMs and he d i ing ac o s
behind his e olu ion.
Resul s
We ha e ga he ed, compiled and compa ed da a om IAM scena ios
included in ou online scena io da abases (Ex ended Da a Table 1).
These include AR5 published in 2014
24
, he Special Repo on Global
Wa ming o 1.5 °C (SR1.5) in 201825 and AR6 in 202226, which all
bene i om a ha monized epo ing empla e o all socioeconomic,
ene gy sys em, emissions and clima e ou comes, he eby allowing
25 30 35 40 45 50 55 60 65
SRES
AR5
SR1.5
AR6
SRES
AR5
SR1.5
AR6
SRES
AR5
SR1.5
AR6
SRES
AR5
SR1.5
AR6
2010202020302050
G ossil CO2
E&IP CO2 emissions in baseline
scena ios and eal-wo ld obse a ions
GCP
EDGAR
a
Lowe
in e qua ile ange
Median Highe
in e qua ile ange
34% 34% 33% 34%
287%
207% 205% 195%
–185%
–158% –163% –158%
–32% 4%
–8% –6%
0
50
100
150
200
250
300
350
SRES AR5 SR1.5 AR6 SRES AR5 SR1.5 AR6 SRES AR5 SR1.5 AR6 SRES AR5 SR1.5 AR6
Popula ion GDP pe capi a Ene gy in ensi y o GDP CO2 in ensi y o
ene gy
E&IP CO2 emissions in 2050 ela i e o 2010 (%)
Kaya decomposi ion, ossil CO2 in 2050 wi h espec o 2010
No-policy baselines
b
Fig. 1 | Analysis o E&IP CO2 emissions in baseline scena ios. a, Medians and
in e qua ile anges o emissions ou comes om no-policy baselines, o
IPCC SRES, AR5, SR1.5 and AR6. Values o 2010 and 2020 include eal-wo ld
obse a ions by Global Ca bon P ojec 53 and EDGAR .8.0 ( e . 54). Fo hese
obse a ions, we ook he a e age o 2009–2011 and o 2019 and 2021 o he
alue in 2010 and 2020, espec i ely, o a oid annual a iabili y (including
Co id-19 impac s) which mos models do no accoun o . b, Kaya iden i y
componen s o mean emission pa hways, sepa a ely o SRES, AR5, SR1.5
and AR6. Values o SRES excluded indus ial p ocess emissions, and ha e
been mul iplied by 1.06 o be compa able wi h AR5, SR1.5 and AR6 scena ios,
which commonly epo he sum o E&IP emissions. S a is ics a e d awn om
model means (Me hods) and SRES n = 15, AR5 n = 53, SR1.5 n = 16, AR6 n = 47.
Supplemen a y Table 1 gi es an o e iew o included scena ios.
Na u e Clima e Change | Volume 15 | Feb ua y 2025 | 218–226 220
Analysis h ps://doi.o g/10.1038/s41558-024-02198-6
whe eas mi iga ion scena ios ou line mo e ambi ious ans o ma ions
o socio echnical and economic sys ems owa ds mee ing ambi ious
clima e goals.
Baseline scena io p ojec ions o CO
2
emissions om ene gy
and indus ial p ocesses (E&IP) in 2030 and 2050 ha e g adually
dec eased om SRES owa ds AR6 (Fig. 1a), e lec ing ans o ma-
ion ha has g adually happened ( o example, due o policy e o s
and low-ca bon echnology deploymen in he his o ical yea s o
models), e olu ion o modelling assump ions and/o capabili ies
( o example, ela i e cos s and enhanced model ep esen a ion o
low-ca bon echnologies) o changes in assumed socioeconomic
ajec o ies. Compa ing scena io emissions in 2010 and 2020 wi h
his o ical alues shows ha 2010 emissions we e gene ally unde -
es ima ed (below median) and 2020 emissions (2019–2021 a e -
age o la gely a oid Co id-19 e ec s) o e es ima ed in all da ase s.
Logically, olde da ase s show highe a iabili y, as es ima ions a e
u he away om model base yea s, allowing ini ial di e ences o
inc ease o e ime. The imp o ed baseline s a ing poin , e en in
he absence o s ong clima e policies, and he clea o e es ima ion
o 2020 emissions, is compa ible wi h ecen li e a u e poin ing o
u u es o mo e han 4–5 °C end-o -cen u y wa ming being inc eas-
ingly unlikely15,29.
Disen angling he 2010–2050 e olu ion o mean emission pa h-
ways o no-clima e-policy baselines (Fig. 1b) in Kaya iden i y compo-
nen s (Me hods) shows how di e ences in mean E&IP CO2 emissions
(Fig. 1) a e d i en by a combina ion o causes. Fi s , ela i e o AR5
and AR6, SRES p ojec ed signi ican ly highe g oss domes ic p oduc
(GDP) g ow h, bu also s onge decline in ene gy in ensi y o GDP and
CO
2
in ensi y o ene gy, e en in he absence o policy. The baseline
emission educ ion om AR5 o SR1.5 and AR6 is d i en by a combina-
ion o all Kaya componen s, wi h again lowe p ojec ed GDP g ow h
playing a signi ican ole, being he main d i e behind AR5 scena ios
o e es ima ing emissions in 2020
30
. O e all, howe e , his o e es i-
ma ion o u u e GDP g ow h h oughou he modelling ensembles
may be d i en by long- e m unce ain y o s uc u al assump ions in
p ojec ions31 (Ex ended Da a Fig. 1).
E olu ion o mi iga ion pa hways
When gi ing IAMs he objec i e o limi global empe a u e inc ease,
adia i e o cing o cumula i e emissions o ce ain le els, hey ypi-
cally es ima e he cos - o wel a e-op imal pa h o e ime and space
owa ds ha le el
32
. Ne e heless, model use s de e mine om which
yea onwa ds IAMs mus look o his pa h. AR5 scena ios assumed
a le el o mi iga ion ac i i y by 2020 ha did no e en ua e, wi h
he SR1.5–AR6 ensembles he e o e de e ing he onse o subs an-
i e mi iga ion ac i i ies un il pos -2020. The ela i ely delayed and
highe peak in emissions esul ed in deepe equi ed emission cu s
in 2050 in SR1.5–AR6 o be compa ible wi h <1.5 °C wa ming (Fig. 2a).
A simila bu mo e mode a e e ec can be obse ed o 2 °C pa h-
ways, a pa e n epea ed o e e y egion (Ex ended Da a Fig. 2),
con i ming ha i is a esul o models a emp ing o ind he op imal
pa hway o speci ied a ge s, a he han unde lying egion-speci ic
de elopmen s.
In e ms o Kaya componen s (Me hods), SR1.5 and AR6 include
a much mo e subs an ial educ ion in CO2 in ensi y o ene gy supply,
han does AR5 (Fig. 2b), sugges ing ha hese wo ensembles ely hea -
ily on supply-side echnology inno a ion o compensa e o he lag-
ging mi iga ion esponse o he wo ld. This obse a ion may con i m
c i icisms o IPCC scena ios o en elying on la gely unp eceden ed
echnological change a he han slowe economic g ow h o demand
educ ion o eaching clima e objec i es33, al hough he e is also an
inc emen al ole o slowe economic g ow h p ojec ions (AR6) o
as e declining ene gy in ensi y o GDP (SR1.5) o make up o his
delay ela i e o AR5. In e es ingly, he ole o echnological ca bon
0
5
10
15
20
25
30
35
40
2010 2020 2030
Yea
2040 2050
M CO2 om ene gy and indus y
2 °C AR5
2 °C AR6 (including SR1.5)
1.5 °C AR5
1.5 °C AR6 (including SR1.5)
Obse ed GCP
Obse ed EDGAR
34% 33% 34%
207% 207%
196%
–204% –211% –198%
–41%
–49% –54% –28%
–27% –24%
–100
–50
0
50
100
150
200
250
300 Kaya decomposi ion
2 °C compa ible scena ios
34% 33% 34%
201% 205%
184%
–214%
–224% –199%
–46%
–69% –68% –33%
–30% –30%
–100
–50
0
50
100
150
200
250
300
AR5
SR1.5
AR6
AR5
SR1.5
AR6
AR5
SR1.5
AR6
AR5
SR1.5
AR6
AR5
SR1.5
AR6
Popula ion GDP pe capi a Ene gy in ensi y
o GDP
CO2 in ensi y o
ene gy
CO2 cap u e
and s o age
1.5 °C compa ible scena ios
E&IP CO2 emissions in 2050 ela i e o 2010 (%)
ab
Fig. 2 | Analysis o E&IP CO2 emissions in mi iga ion pa hways. a, Mean
pa hways in AR5 and AR6 om 2010 o 2050 and obse ed emissions in 2010 and
2020 (Fig. 1). A ows in 2020 and 2050 e lec how he delayed mi iga ion in AR6
scena ios ela i e o AR5 scena ios leads o deepe equi ed educ ions in 2050
in AR6 scena ios o be compa ible wi h he same clima e a ge . Values o 2010
and 2020 include eal-wo ld obse a ions by Global Ca bon P ojec 53 and EDGAR
.8.0 ( e . 54). b, Kaya iden i y componen s o 2 and 1.5 °C compa ible scena ios,
sepa a ely o AR5, SR1.5 and AR6. S a is ics a e d awn om model means
(Me hods) and Supplemen a y Table 1 gi es an o e iew o included scena ios
o each indica o .
Na u e Clima e Change | Volume 15 | Feb ua y 2025 | 218–226 221
Analysis h ps://doi.o g/10.1038/s41558-024-02198-6
emo als (such as CCS) sligh ly dec eased om AR5 o AR6, despi e
highe mi iga ion p essu e.
The sec o al s ock ake in mi iga ion e o s in 2010–2050 shows
ha he la ges absolu e con ibu ions come om ene gy supply
(co e ing all ans o ma ion s ages— o example, mining, e ining
and elec ici y), bu addi ional e o in mo ing om 2 o 1.5 °C pa h-
ways comes mos ly om indus y, buildings and anspo (Fig. 3).
Ac oss scena io ensembles, he e is a shi owa ds deepe mi iga ion
in end-use sec o s owa ds 2050, pos -AR5, bo h o mo e modes 2 °C
scena ios and mo e clea ly o 1.5 °C scena ios, e lec ing imp o ed
model inco po a ion o end-use echnologies, such as elec i ica-
ion and hyd ogen in indus ial p ocesses, ba e y-elec ic ehicles
in anspo a ion and elec ic hea pumps in buildings and/o he
ma u i y o hese echnologies ( o example, lowe cos s o g ow h
cons ain s) pos -201534,35.
The e olu ion o CO
2
emissions om Ene gy Supply and AFOLU
a ies less in absolu e quan i y be ween scena io da abases, bu no a-
bly hese sec o s go owa ds ne -nega i e emissions in SR1.5–AR6,
ins ead o ca bon neu ali y (AR5). Ene gy-sec o CO2 emissions
become nega i e p ima ily as a esul o elec ici y gene a ion and uel
supply by BECCS, which i s mus o se esidual sec o al emissions
be o e becoming ne -nega i e. Lowe esidual emissions in elec ici y
gene a ion ( o example, mo e enewables) can acili a e lowe emis-
sions in he ene gy sec o e en i BECCS use is mode a ed. AFOLU CO2
emissions ep esen a balance be ween e o es a ion, a o es a ion,
de o es a ion and—depending on model— o es y. The e is a ade-o
be ween BECCS and AFOLU because o limi ed land a ailabili y and
sa u a ion o AFOLU o e ime36.
Ene gy echnology mix and mac oeconomic cos s
Focusing on mi iga ion scena ios compa ible wi h 2 o 1.5 °C, esul s
indica e g adual e olu ion o mi iga ion po olios. In e ms o p ima y
ene gy mix (Fig. 4a), coal sha es ha e d opped in mo e ecen ensem-
bles: while coal s ill played a non-negligible ole in deep mi iga ion
scena ios in AR5 (~10% o p ima y ene gy by 2050), i s sha e is <10%
in all 2 °C/1.5 °C scena ios in AR6. Reliance on na u al gas was ela-
i ely s able among scena ios in AR5 (~20–25% o p ima y ene gy) bu
dec eased in 1.5 °C scena ios in SR1.5–AR6 ( owa ds ~15%). This d op
is p obably ela ed o he need o deepe mi iga ion ela i e o AR5
(Fig. 2). Reliance on bioene gy o mi iga ion is also sligh ly lowe in
mo e ecen scena ios, whe eas eliance on sola and wind in he mos
ambi ious scena ios inc eased om 10–15% (AR5) o 20–45% o o al
p ima y ene gy demand by 2050 (AR6), in esponse o apidly declin-
ing echnology cos s and enhanced model ep esen a ion o a iable
enewables (VRE) in eg a ion in he elec ici y sys em ( o example,
coupled wi h ba e ies). This las e ec is clea ly ela ed o he elec-
ici y mix and signi icance in inal ene gy demand, due o enhanced
elec i ica ion (Fig. 4b). Fo example, he sha e o enewables in elec-
ici y inc eases om 40–50% (AR5) o 60–80% (SR1.5–AR6) by 2050
h oughou scena ios, while nuclea d ops om 15–20% o 10–15%,
p obably ela ed wi h cos escala ion and limi ed poli ical suppo
a e he 2012 Fukushima e en
37
. The sha e o elec ici y in inal ene gy
demand inc eases wi h deepe mi iga ion, bu he o e all le el is ~10%
highe in AR6 o e AR5, leading o elec i ica ion a es o ~50% a he
han ~40% o he mos ambi ious scena ios. To al p ima y and inal
ene gy use, bo h dec easing wi h inc easing mi iga ion e o , ha e
no signi ican ly e ol ed since AR5 (Ex ended Da a Fig. 3).
AR5 2 °C AR5 1.5 °C
SR1.5 2 °C SR1.5 1.5 °C
AR6 2 °C AR6 1.5 °C
–4
–2
0
2
4
6
8
10
12
14
16
18
2010 2030
Yea Yea Yea Yea Yea
2050
G CO2
Ene gy
supply
0
2
4
6
8
10 4
2010 2030 2050
Indus y
(combus ion and
p ocesses) 0
1
2
3
2010 2030 2050
Buildings
0
2
4
6
8
10
2010 2030 2050
T anspo
Obse ed
(EDGAR + GCP)
–2
–1
0
1
2
3
4
5
2010 2030 2050
AFOLU
a
b
–5
–10
0
5
10
15
20
25
30
35
40
45
AR5 SR1.5 AR6 AR5 SR1.5 AR6 AR5 SR1.5 AR6 AR5 SR1.5 AR6
2010* 2020* 2 °C compa ible 1.5 °C compa ible 2 °C compa ible 1.5 °C compa ible
Obse ed emissions Means o 2030 p ojec ions Means o 2050 p ojec ions
AFOLU
Ene gy supply
Buildings
T anspo
Indus y
Fig. 3 | Global CO2 emissions by sec o . Mean alues o sec o al con ibu ion
o CO2 emissions in 2010 o 2050, sepa a ely o scena ios ha a e 2 and
1.5 °C compa ible and o scena ios in IPCC AR5, SR1.5 and AR6. a,b, Absolu e
dis ibu ion pe scena io (a) and e olu ion o e ime by sec o (b). The as e isks
indica e a e ages o obse ed alues om EDGAR .8.0 ( e . 54) (ene gy
consump ion sec o s) and GCB53 (AFOLU sec o ) o 2009–2011 (as es ima e
o 2010) and o 2019 and 2021 (as es ima e o 2020) o a oid annual a iabili y
(including Co id-19 impac s) which mos models do no accoun o . Indus y
es ima es om SR1.5 omi ed because o inconsis en epo ing o emissions
om indus ial p ocesses. S a is ics a e d awn om model means (Me hods)
and Supplemen a y Table 1 gi es an o e iew o included scena ios o each
indica o .
Na u e Clima e Change | Volume 15 | Feb ua y 2025 | 218–226 222
Analysis h ps://doi.o g/10.1038/s41558-024-02198-6
Gene ally, mi iga ion scena ios ha e e ol ed o depend less on
CCS (Fig. 4c). The compa ison o di e en CCS elemen s be ween
scena io ensembles is no s aigh o wa d because o changes in
he epo ing s uc u e o ca bon seques a ion, bu he igu e
shows ha eliance on CCS-powe ed ossil elec ici y signi ican ly
dec eased om AR5 o SR1.5, s aying ela i ely low bu wi h high
model a iance in AR6, which is compa ible wi h he abo emen-
ioned educed eliance on ossil uels and non- enewable elec ic-
i y. Ne e heless, he o e all ole o CDR has no ma kedly educed
since AR5. While he ole o BECCS ma ginally dec eased, he e has
been a ise in new CDR echnologies such as DACCS
38
(Ex ended Da a
Fig. 4). This dis inc ion be ween CDR echnologies is impo an , as
i shows ha , while mi iga ion has shi ed owa ds enewables, he
need o ca bon emo als has no declined wi h he need o deepe
mi iga ion in SR1.5–AR6.
In e es ingly, SR1.5 scena ios appea o coincide mo e wi h AR6
ajec o ies o nea ly all a iables. The ac ha 1.5 °C scena ios in
SR1.5, as in AR6, equi ed deepe mi iga ion compa ed o AR5 (Fig. 2),
may play a ole. Howe e , o many a iables (coal and nuclea sha es,
enewable elec ici y and elec i ica ion), he esemblance clea ly
s a s wi h 2 °C scena ios al eady, and he simila endency o 2 °C
scena ios h oughou he IPCC scena io ensembles (Fig. 2) sugges s
ha he e a e mo e s uc u al causali ies oo; o example, upda ed
model s uc u es and/o assump ions.
Fo decades now, he IPCC has been asked wi h assessing he
li e a u e on mac oeconomic implica ions o educing emissions and
has esponded by publishing bo h es ima es o , and limi a ions inhe -
en in, long- e m mac oeconomic p ojec ions
39
. Acknowledging he
nume ous me hodological issues in hese es ima es, as well as inhe -
en di e ences in how models add ess he economics o mi iga ion,
we ack how assessmen s o mac oeconomic consequences ha e
e ol ed om AR5 o AR6 (Fig. 4d). We obse e ha , o 2 °C scena ios,
es ima es ha e shi ed om educ ions o 2–3% o economic ac i i y
(GDP) in AR5 o 1–2% in AR6. Fo 1.5 °C scena ios, hese di e ences
a e smalle , d opping om 3–5% o 2–4% o o al GDP. Ne e heless,
conside ing he deepe mi iga ion equi emen s in 2050 in AR6 ela-
i e o AR5 (Fig. 2), he obse a ion ha mac oeconomic cos s a e
lowe in AR6 signals ha a b oade sui e o mi iga ion op ions was
0
1
2
3
4
5
6
7
8
9
1,500–1,600
1,400–1,500
1,300–1,400
1,200–1,300
1,100–1,200
1,000–1,100
900–1,000
800–900
700–800
Pe cen age loss in ou pu due o clima e policy
Cumula i e G CO2 2010–2050
Mac o-economic
ou pu loss by 2050
0
10
20
30
40
Pe cen age o o al p ima y ene gy
Coal sha e Bio-ene gy sha e
Gas sha e Wind and sola sha e
0
10
20
30
40
50
60
Sha e o enewables in elec ici y mix (%)
Elec i ica ion o
inal ene gy use
0
10
20
30
40
50
60
70
80
90
100
Sha e o enewables in elec ici y mix (%)
Renewable elec ici y
0
5
10
15
20
25
30
35
Sha e o enewables in elec ici y mix (%)
Nuclea elec ici y
0
2
4
6
8
10
12
14
16
18
20
1,500–1,600
1,400–1,500
1,300–1,400
1,200–1,300
1,100–1,200
1,000–1,100
900–1,000
800–900
700–800
G CO2 cap u ed
Cumula i e G CO2 2010–2050
Ca bon cap u e and
s o age (including DACCS)
P ima y ene gy
sha es in 2050
Elec ici y
indica o s in 2050
2 °C
compa ible
1.5 °C
compa ible
0
10
20
30
40
50
60
70
1,500–1,600
1,400–1,500
1,300–1,400
1,200–1,300
1,100–1,200
1,000–1,100
900–1,000
800–900
700–800
EJ elec ici y gene a ion
Cumula i e G CO2 2010–2050
Fossil elec ici y
wi h CCS
1.5 °C
compa ible
2 °C
compa ible
0
5
10
15
20
25
30
1,500–1,600
1,400–1,500
1,300–1,400
1,200–1,300
1,100–1,200
1,000–1,100
900–1,000
800–900
700–800
EJ elec ici y gene a ion
Cumula i e G CO2 2010–2050
Bio-elec ici y
wi h CCS
2 °C
compa ible
1.5 °C
compa ible
Ca bon cap u e
indica o s in 2050
2 °C
compa ible
1.5 °C
compa ible
a
b
c
d
AR5
SR1.5
AR6
Fig. 4 | Ene gy sys em and mi iga ion indica o s by 2050. a–d, In e qua ile
anges o s uc u al indica o s (p ima y ene gy sha es (a), elec ici y indica o s
(b), ca bon cap u e indica o s (c) and mac oeconomic ou pu loss in (d)) o
mi iga ion pa hways by 2050, plo ed sepa a ely o IPCC AR5, SR1.5 and AR6
as a unc ion o mi iga ion e o in e ms o cumula i e ossil CO2 emissions
om 2010 o 2050, ma ked by 2 °C and 1.5 °C compa ibili y. The me ic o
mac oeconomic ou pu loss (d) di e s be ween epo ed scena ios, and includes
GDP losses, consump ion losses o he a ea unde he MAC cu e. S a is ics
a e d awn om model means (Me hods) and Supplemen a y Table 1 gi es an
o e iew o included scena ios. EJ, exajoule.
Na u e Clima e Change | Volume 15 | Feb ua y 2025 | 218–226 223
Analysis h ps://doi.o g/10.1038/s41558-024-02198-6
a ailable in AR6, assump ions o e u u e mi iga ion echnology cos s
ha e become mo e op imis ic and/o s uc u al changes in IAMs ha e
lowe ed he esis ance o majo shi s in ene gy sys ems. In e es ingly,
he mac oeconomic implica ions in SR1.5 la gely esembled hose o
AR6 o 2 °C scena ios, bu hose o AR5 o 1.5 °C scena ios. This may
signal ha assump ions abou majo ene gy ans o ma ions may ha e
been mo e op imis ic in SR1.5 ela i e o AR5, bu model capabili ies
o eaching deep mi iga ion we e no ye as well-de eloped as in AR6.
No e ha hese cos s do no include clima e impac s, and he economic
impac s o clima e inac ion as ou lined in IPCC WGII may well su pass
he IAM- epo ed cos s o mi iga ion39.
The igu es in his s udy show a g adually e ol ing emission/ em-
pe a u e u u e (in no-clima e-policy baselines) and ans o ma ion
pa h o eaching 2 °C and 1.5 °C u u es, as well as inc easing mi iga-
ion s ingency equi ed o nea ly all sec o s and egions o achie e
1.5 oC. We discuss likely echnological d i e s (Box 1) and poli ical
unding- ela ed d i e s (Box 2) o hese changes.
Discussion
IAM esea ch has seen an imp essi e upli du ing he pas wo decades.
Combined wi h consis en scena io epo ing
40
, his allows o a mul i-
dimensional acking o how ans o ma ion pa hways ha e e ol ed.
Ou analysis akes aim a his e olu ion o e he las decade h ough
a de ailed analysis o housands o IAM scena ios in IPCC scena io
da abases.
The e is con inuous e olu ion o mos analysed indica o s
o e scena io ensembles wi h a clea di ec ion owa ds inc easing
BOX 1
The ole o echnology in
e ol ing mi iga ion scena ios
Se e al clean ene gy echnologies ha e been scaled up
signi ican ly o e he pas decade, wi h cos educ ion and
pe o mance imp o emen s un o eseen a his scale in modelled
scena ios, which ha e been g adually implemen ed in mo e ecen
scena ios. The clea es example is sola PV, o which he cos
assump ion o 2020 was h ee o six imes lowe in AR6 han in AR5
(in e qua ile anges), and obse ed cos s a ound 2020 acco ding
o IRENA (2021)55 we e lowe han assumed cos s o 2050 in AR5
and SR1.5 scena ios. Simila ly, o echnologies such as onsho e
and o sho e wind, ba e y s o age, elec ic ehicles, hyd ogen
elec olyse s and elec ic hea pumps, ecen cos educ ions and/
o pe o mance imp o emen s com o ably exceeded modelled
assump ions in ea lie model uns and IPCC epo s. The upda ed
ep esen a ion o cheape - han-expec ed low-ca bon echnologies
in models causes p ojec ed baseline emissions o be lowe , as
clean echnologies compe e agains di y al e na i es e en in he
absence o clima e policy (Fig. 1), and mi iga ion echnologies o
shi adically wi h he same mi iga ion a ge (Fig. 4).
7,000
6,000
5,000
AR5
AR6 AR6
SR1.5 SR1.5
AR5
Capi al cos s—sola PV
US$(2010)/kWe
IRENA (2021)
obse ed cos s
2010, 2015, 2020
Capi al cos s—wind onsho e
4,000
3,000
2,000
2010 2030
Yea Yea
2050 2010 2030 2050
1,000
0
Mo eo e , he ep esen a ion o end-use sec o echnologies
has been subs an ially imp o ed om AR5 owa ds AR6. This
con inuous e olu ion o IAMs also opens new oppo uni ies
o mi iga ion43,47, and can be expec ed o con inue in o u u e
scena io exe cises. Recen indus y-speci ic modelling exe cises
wi h highe sec o al de ail achie e la ge emission educ ions
han IAM scena ios conside ed in he IPCC cycles assessed in his
a icle56,57, o he poin ha some sec o s p e iously conside ed
as ha d o aba e may now ha e he op ion o a much s eepe
mi iga ion ajec o y, such as he i on and s eel sec o wi h new
s eel- ela ed in es men in he Eu opean Union (EU) al eady
domina ed by low-ca bon op ions ( o example, elec ic a c u nace
and hyd ogen-based di ec educ ion o i on). We expec ha hese
ecen de elopmen s will also be e lec ed in he new gene a ion o
IAM scena ios.
BOX 2
Scena ios om Eu opean
modelling eams domina e
IPCC assessed scena io
da abase, shaping ene gy
ansi ion pa e ns
The changing pa e ns we ound in he indica o s examined
in his s udy (Fig. 4) we e especially accen ua ed in scena ios
de eloped by o ganiza ions om Eu ope: AR6 scena ios om
Eu opean o ganiza ions had a highe sha e o sola PV and wind
powe in elec ici y gene a ion and p ima y ene gy as well as
highe elec i ica ion han he es —and less eliance on nuclea
and CCS (Supplemen a y Fig. 1). These di e ences we e less
p onounced in AR5 and SR1.5 and we e e en e e sed o bioene gy.
Simila pa e ns we e ound by compa ing scena ios o di e en
wa ming le els be ween Eu opean and o he o ganiza ions
(Supplemen a y Figs. 2–4), wi h 2 °C scena ios showing he la ges
di e ences. While scena io de elope s om Eu ope we e also
using lowe cos assump ions in AR6 o sola and wind han he
es (Supplemen a y Fig. 5), echnological cos s did no seem o
be he only eason o hese di e ences. Fo ins ance, e en when
con olling o he cos assump ions o sola PV and wind powe ,
he sha e o enewables in p ima y ene gy (Supplemen a y Fig. 1)
was s ill s a is ically signi ican ly highe in scena ios coming om
Eu opean o ganiza ions han he es (median o 25% e sus 10%,
one-way analysis o co a iance, F = 56.49, P < 0.001). Simila ly, e . 3
showed ha p ojec ed sola PV g ow h was highe in AR5 and SR1.5
scena ios om Eu opean o ganiza ions compa ed o scena ios om
Asian o No h Ame ican o ganiza ions.
Eu opean scena io de elope s ha e con ibu ed mos scena ios
o he examined epo s (71% in AR6, 75% in SR1.5 and 66% in AR5;
Supplemen a y Table 3), so hey expec edly ha e a la ge impac
on he o e all pa e ns in he IPCC scena io da abases. Since hei
sha e in he da abases has no changed ma kedly om AR5, he
signi ican e olu ion obse ed he e may be due o an inc eased use
o scena io designs ha explo e highe sha es o enewables and/o
easibili y conce ns58.
Na u e Clima e Change | Volume 15 | Feb ua y 2025 | 218–226 224
Analysis h ps://doi.o g/10.1038/s41558-024-02198-6
elec i ica ion a es and highe VRE sha es in elec ici y gene a ion.
Simul aneously, he use o nuclea , bioene gy and CCS wi h ossil
ene gy has declined in mo e ecen scena ios, despi e mo e s in-
gen 2050 a ge s. Since 2010, emaining ca bon budge s ha e sh unk
by ~400 G CO
2
(~0.5 °C), ye he achie abili y and mac oeconomic
implica ions o deli e ing he same clima e a ge ha e no ma kedly
changed. An impo an in e p e a ion o his is ha AR5 po ayed
mo e ambi ious clima e ac ion as ha de and cos lie han nowadays
calcula ed. The SR1.5–AR6 scena io ensembles ea u e mo e ambi-
ious emissions cu s in ene gy end-use sec o s (indus y, buildings
and anspo ), while allowing ene gy supply and AFOLU o each
ne -nega i e CO2 emissions.
The main causes we ound behind he explo ed di e ences
be ween scena io ensembles also imply lessons o u u e scena io
de elopmen . A majo d i e lies in as e - han-expec ed educ ion
in clean echnology cos s and imp o ed ep esen a ion o c i ical
end-use echnologies (Box 1). The de ia ion be ween p ojec ed and
obse ed ends in many key echnologies41,42 poin s o he need o
sus ained e o o u he upda e models. Any ailu e by IPCC sce-
na ios o adequa ely e lec he p ac icabili y/plausibili y o ce ain
changes becomes mo e signi ican wi h declining ca bon budge s. I
is, he e o e, impo an ha modelling communi ies wo k closely wi h
s akeholde s o keep hei c i ical echnology assump ions up- o-da e
amids apidly e ol ing ma ke dynamics. Since CDR le els in mi i-
ga ion scena ios a e dependen on he emission educ ion le e s in
ha d- o-aba e sec o s included in models
43
, such de elopmen s may
educe eliance on CDR.
A second cause we iden i ied was poli ical (Box 2): na a i es
s emming om a Eu opean con ex may be mo e p o- enewables
gi en he apid oll-ou o sola and wind en isioned in he Eu o-
pean G een Deal44 and, pe haps, he EU being a majo unde o such
s udies ( o example, ia H2020 and Ho izon Eu ope; Supplemen-
a y Table 3)18. The dominance o scena ios om Eu opean ins i u es
in scena io ensembles may explain he inc eased ocus on VRE in
mo e ecen scena io ensembles. Eu opean na a i es may, howe e ,
con as wi h na a i es om de eloping economies, such as India,
whe e mi iga ion s a egies a e in e wo en wi h socioeconomic
de elopmen 45.
High economic g ow h p ojec ions ha e his o ically been used
as an a gumen o delay mi iga ion, as u u e gene a ions would be
mo e capable o paying o adap a ion and/o CDR echnologies o
egula e empe a u es46. Al hough mos scena ios in he explo ed
ensembles a e no based on in e gene a ional cos –bene i analysis,
he obse a ion ha p ojec ions o u u e economic g ow h ha e been
con inuously co ec ed downwa ds (Fig. 1 and Ex ended Da a Fig. 1)
may imply educed inno a ion po en ial o b eak h oughs and alue
added o CDR echnologies31. Such echnologies, he e o e, inc ease
u u e cos s o bo h mi iga ion ( h ough la e ac ion) and adap a ion
( h ough empe a u e o e shoo )
47
, while also in ol ing o he isks
o u u e socie ies13. The ack eco d o eal-wo ld echnology cos s
ou pe o ming o ecas s (Box 1) illus a es ha lacking o ecas s o
inpu s o mi iga ion scena ios ha e alsely a ou ed delayed ac ion
h oughou he pas decade.
Howe e emp ing o conclude ha enewables in u u e
low-ca bon elec ici y p ojec ions will con inue o inc ease, owing
o spec acula cos educ ions and deploymen ela i e o CCS and
nuclea , his is no ce ain. G id balancing, social accep ance, mine al
supply chains and so on mus be conside ed as po en ial o se s o
con inued capi al-cos declines. The u u e di ec ion o o he obse ed
ends is also unclea ; no ably, al hough p ojec ed mi iga ion cos s
ha e allen (o , o 1.5 °C pa hways, la gely emained unchanged) wi h
dec easing echnology cos s mo e- han-o se ing sh inking ca bon
budge s, his balance can only las so long. Ine i ably, especially o
1.5 °C pa hways, mi iga ion cos s will ise, as he emaining ca bon
budge e en ually diminishes and emissions emo als become cen al
o keeping below his empe a u e h eshold. A clea e end, om
AR5 o AR6, is he educ ion in coal and gas sha es in u u e p ima y
ene gy demand in he lowes ca bon budge scena ios. This could
well con inue in o AR7 bu , ega dless, nea - e m geopoli ical de el-
opmen s mus be ac o ed in— o example, po en ial lock—in o LNG
in as uc u e in ligh o a wa e o new global in es men 48.
I is likely ha he 1.5 °C empe a u e s abiliza ion goal will
be exceeded wi hin he nex 5–10 yea s49, dic a ing a e aming o
scena ios and IAMs. Ins ead o mo e elabo a e ways o keep below
1.5 °C, ocus may shi owa ds easibili y50, equi ing mo e c i ical
hinking on how di e en echnologies may be deployed unde
di e en policy and egional con ex s. This emphasizes he impo -
ance o using mul iple s eams o e idence when using/in e p e -
ing scena ios, such as sec o al and na ional expe ise. Models and
scena ios a e en icing because o hei abili y o in eg a e hund eds
o housands o assump ions, bu his also makes hem suscep ible
o p o iding a alse sense o con idence in esul s. The su ge o he
‘ne -ze o’ concep , s emming om physical clima e science51, and
inc easingly applied in poli ical pledges and scena io design50, had
no ole in AR5, a ela i ely small ole in AR6, bu will p obably be
ins umen al owa ds AR7, by i sel condi ioning se e al indica o s
explo ed in his s udy.
The wo ld now has a mul i ude o mi iga ion policies in place
and, despi e lack o a s ong adi ion o such analysis, hei impac
will complica e he de elopmen o ealis ic ‘ e e ence scena ios’ in
u u e exe cises. Fi s , wi hou eal-wo ld compa ison, e alua ing
no-clima e-policy baselines is no as meaning ul a he global le el15.
Second, despi e a g owing numbe o ele an s udies
14,50,52
, pa hways
e lec ing cu en policies and/o Na ionally De e mined Con ibu-
ions a e also ha d o e alua e, since policy cons an ly changes. Thi d,
backcas ing scena ios ypically pe o med o de e mine how he wo ld
could ans o m and achie e ambi ious clima e a ge s mos ly apply
idealized assump ions, such as globally ha monized ca bon p ices
ac oss sec o s/ egions, a si ua ion inc easingly unlikely o be ealized.
E alua ing scena ios will, hus, equi e an inc easing deg ee o expe
judgemen o elici policy ele an ecommenda ions.
Online con en
Any me hods, addi ional e e ences, Na u e Po olio epo ing sum-
ma ies, sou ce da a, ex ended da a, supplemen a y in o ma ion,
acknowledgemen s, pee e iew in o ma ion; de ails o au ho con i-
bu ions and compe ing in e es s; and s a emen s o da a and code a ail-
abili y a e a ailable a h ps://doi.o g/10.1038/s41558-024-02198-6.
Re e ences
1. IPCC Clima e Change 2023: Syn hesis Repo (eds Co e W i ing
Team, Lee, H. & J. Rome o, J.) (IPCC, 2023).
2. IPCC Clima e Change 2014: Syn hesis Repo (eds Co e W i ing
Team, Pachau i, R. K. & Meye , L. A.) (IPCC, 2014).
3. Jaxa-Rozen, M. & T u ne y e, E. Sou ces o unce ain y in
long- e m global scena ios o sola pho o ol aic echnology.
Na . Clim. Change 11, 266–273 (2021).
4. Bins ed, M. An elec i ied oad o clima e goals. Na . Ene gy 7,
9–10 (2022).
5. Lude e , G. e al. Impac o declining enewable ene gy cos s on
elec i ica ion in low-emission scena ios. Na . Ene gy 7, 32–42
(2021).
6. S auch, Y. Beyond he low-ca bon niche: global ipping poin s
in he ise o wind, sola , and elec ic ehicles o egime scale
sys ems. Ene gy Res. Soc. Sci. 62, 101364 (2020).
7. Baue , C. e al. Cha ging sus ainable ba e ies. Na . Sus ain. 5,
176–178 (2022).
8. Pe dana, S. e al. Expe pe cep ions o game-changing
inno a ions owa ds ne ze o. Ene gy S a eg. Re . 45, 101022
(2023).
Na u e Clima e Change | Volume 15 | Feb ua y 2025 | 218–226 225
Analysis h ps://doi.o g/10.1038/s41558-024-02198-6
9. Lane, J., G eig, C. & Ga ne , A. Unce ain s o age p ospec s
c ea e a conund um o ca bon cap u e and s o age ambi ions.
Na . Clim. Change 11, 925–936 (2021).
10. Ande son, K. & Pe e s, G. The ouble wi h nega i e emissions.
Science 354, 182–183 (2016).
11. Fuss, S. e al. Nega i e emissions—Pa 2: Cos s, po en ials and
side e ec s. En i on. Res. Le . 13, 063002 (2018).
12. F idahl, M. & Leh ee , M. Bioene gy wi h ca bon cap u e and
s o age (BECCS): global po en ial, in es men p e e ences, and
deploymen ba ie s. Ene gy Res. Soc. Sci. 42, 155–165 (2018).
13. Smi h, W. Pando a’s Toolbox: The Hopes and Haza ds o Clima e
In e en ion (Camb idge Uni . P ess, 2022).
14. Sognnaes, I. e al. A mul i-model analysis o long- e m emissions
and wa ming implica ions o cu en mi iga ion e o s. Na . Clim.
Change 11, 1055–1062 (2021).
15. Haus a he , Z. & Pe e s, G. P. Emissions— he ‘business as usual’
s o y is misleading. Na u e 577, 618–620 (2020).
16. Boon-Falleu , M., G andin, A., Bauma d, N. & Che allie , C.
Le e aging social cogni ion o p omo e e ec i e clima e change
mi iga ion. Na . Clim. Change 12, 332–338 (2022).
17. Koasidis, K., Nikas, A. & Doukas, H. Why in eg a ed assessmen
models alone a e insu icien o na iga e us h ough he
polyc isis. One Ea h 6, 205–209 (2023).
18. Nikas, A. e al. Pe spec i e o comp ehensi e and comp ehensible
mul i-model ene gy and clima e science in Eu ope. Ene gy 215,
119153 (2021).
19. Webe , C. e al. Mi iga ion scena ios mus ca e o new use s.
Na . Clim. Change 8, 845–848 (2018).
20. an Beek, L., Haje , M., Pelze , P., an Vuu en, D. & Cassen, C.
An icipa ing u u es h ough models: he ise o In eg a ed
Assessmen Modelling in he clima e science-policy in e ace
since 1970. Glob. En i on. Change 65, 102191 (2020).
21. B aun ei e , L., Ma chand, C. & Blume , Y. Explo ing possible
u u es o ein o cing he s a us-quo? The use o model-based
scena ios in he Swiss ene gy indus y. Renew. Sus ain. Ene gy
T ansi . 3, 100046 (2023).
22. Gui a ch, C. e al. Using la ge ensembles o clima e change
mi iga ion scena ios o obus insigh s. Na . Clim. Change 12,
428–435 (2022).
23. Dekke , M. M. e al. Sp ead in clima e policy scena ios un a elled.
Na u e 624, 309–316 (2023).
24. IPCC Clima e Change 2014: Mi iga ion o Clima e Change
(eds Edenho e , O. e al.) (Camb idge Uni . P ess, 2014).
25. Huppmann, D. e al. IAMC 1.5 °C Scena io Explo e and Da a
Hos ed by IIASA (IIASA, 2018); h ps://doi.o g/10.22022/SR15/
08-2018.15429
26. Bye s, E. e al. AR6 scena ios da abase (1.0) [Da a se ]. Zenodo
h ps://doi.o g/10.5281/zenodo.5886912 (2022).
27. IPCC Special Repo on Emissions Scena ios (eds Nakiceno ic, E. &
Swa , R.) (Camb idge Uni . P ess, 2000).
28. He mwille, L., Obe gassel, W. & A ens, C. The ans o ma i e
po en ial o emissions ading. Ca bon Manag. 6, 261–272
(2015).
29. Pielke, R., Bu gess, M. G. & Ri chie, J. Plausible 2005–2050
emissions scena ios p ojec be ween 2 °C and 3 °C o wa ming by
2100. En i on. Res. Le . 17, 024027 (2022).
30. Bu gess, M. G., Ri chie, J., Shapland, J. & Pielke, R. IPCC baseline
scena ios ha e o e -p ojec ed CO2 emissions and economic
g ow h. En i on. Res. Le . 16, 014016 (2021).
31. Bu gess, M. G. e al. Mul idecadal dynamics p ojec slow
21s -cen u y economic g ow h and income con e gence.
Commun. Ea h En i on. 4, 220 (2023).
32. Gambhi , A., Ganguly, G. & Mi al, S. Clima e change mi iga ion
scena io da abases should inco po a e mo e non-IAM pa hways.
Joule 6, 2663–2667 (2022).
33. Keyße , L. T. & Lenzen, M. 1.5 °C deg ow h scena ios sugges
he need o new mi iga ion pa hways. Na . Commun. 12, 2676
(2021).
34. G uble , A. e al. A low ene gy demand scena io o mee ing
he 1.5 °C a ge and sus ainable de elopmen goals wi hou
nega i e emission echnologies. Na . Ene gy 3, 515–527
(2018).
35. Keppo, I. e al. Explo ing he possibili y space: aking s ock o he
di e se capabili ies and gaps in in eg a ed assessmen models.
En i on. Res. Le . 16, 053006 (2021).
36. Humpenöde , F. e al. In es iga ing a o es a ion and bioene gy
CCS as clima e change mi iga ion s a egies. En i on. Res. Le . 9,
064029 (2014).
37. Nuclea Powe and Secu e Ene gy T ansi ions (IEA, 2022);
www.iea.o g/ epo s/nuclea -powe -and-secu e-ene gy-
ansi ions
38. Realmon e, G. e al. An in e -model assessmen o he ole o
di ec ai cap u e in deep mi iga ion pa hways. Na . Commun. 10,
3277 (2019).
39. Köbe le, A. C. e al. The cos o mi iga ion e isi ed. Na . Clim.
Change 11, 1035–1045 (2021).
40. Coin e, B. The AR6 scena io explo e and he his o y o
IPCC scena ios da abases: e olu ions and challenges o
anspa ency, plu alism and policy- ele ance. npj Clim. Ac ion
3, 3 (2024).
41. He, G. e al. Rapid cos dec ease o enewables and s o age
accele a es he deca boniza ion o China’s powe sys em.
Na . Commun. 11, 2486 (2020).
42. Way, R., I es, M. C., Mealy, P. & Fa me , J. D. Empi ically g ounded
echnology o ecas s and he ene gy ansi ion. Joule 6,
2057–2082 (2022).
43. Edelenbosch, O. Y. e al. Reducing sec o al ha d- o-aba e
emissions o limi eliance on ca bon dioxide emo al. Na . Clim.
Change 14, 715–722 (2024).
44. 2030 Clima e Ta ge Plan (Eu opean Commission, 2021);
h ps://ec.eu opa.eu/clima/euac ion/eu opean-g een-deal/
2030-clima e- a ge -plan_en#documen s
45. Sudha mma Vishwana han, S., F agkos, P., F agkiadakis, K. &
Ga g, A. Assessing enhanced NDC and clima e compa ible
de elopmen pa hways o India. Ene gy S a eg. Re . 49, 101152
(2023).
46. No dhaus, W. The Clima e Casino: Risk, Unce ain y, and
Economics o a Wa ming Wo ld (Yale Uni . P ess, 2013).
47. Fuh man, J. e al. Ambi ious e o s on esidual emissions can
educe CO2 emo al and lowe peak empe a u es in a ne -ze o
u u e. En i on. Res. Le . 19, 064012 (2024).
48. Wo ld Ene gy In es men 2024 (IEA, 2024); www.iea.o g/ epo s/
wo ld-ene gy-in es men -2024
49. Fo s e , P. M. e al. Indica o s o global clima e change 2023:
annual upda e o key indica o s o he s a e o he clima e
sys em and human in luence. Ea h Sys . Sci. Da a 16, 2625–2658
(2024).
50. an de Ven, D.-J. e al. A mul imodel analysis o pos -Glasgow
clima e a ge s and easibili y challenges. Na . Clim. Change 13,
570–578 (2023).
51. Allen, M. R. e al. Ne ze o: science, o igins, and implica ions.
Annu. Re . En i on. Resou . 47, 849–887(2022).
52. Roel sema, M. e al. Taking s ock o na ional clima e policies o
e alua e implemen a ion o he Pa is Ag eemen . Na . Commun.
11, 2096 (2020).
53. F iedlings ein, P. e al. Global ca bon budge 2023. Ea h Sys .
Sci. Da a 15, 5301–5369 (2023).
54. C ippa, M. e al. GHG Emissions o All Wo ld Coun ies
(Publica ions O ice o he Eu opean Union, 2023).
55. Renewable Powe Gene a ion Cos s in 2020 (IRENA, 2021).
Na u e Clima e Change | Volume 15 | Feb ua y 2025 | 218–226 226
Analysis h ps://doi.o g/10.1038/s41558-024-02198-6
56. Ba aille, C. e al. Towa ds ne -ze o emissions conc e e and s eel in
India, B azil and Sou h A ica. Clim. Policy h ps://doi.o g/10.1080/
14693062.2023.2187750 (2023).
57. Low-Ca bon Technologies o he Global S eel T ans o ma ion.
A Guide o he Mos E ec i e Ways o Cu Emissions in S eelmaking
(Ago a Indus y, 2024); www.ago a-indus y.o g/ ileadmin/
P ojek e/2021/2021-06_IND_INT_GlobalS eel/A-IND_324_
Low-Ca bon-Technologies_WEB.pd
58. Riahi, K. e al. Cos and a ainabili y o mee ing s ingen clima e
a ge s wi hou o e shoo . Na . Clim. Change 11, 1063–1069
(2021).
Publishe ’s no e Sp inge Na u e emains neu al wi h ega d o
ju isdic ional claims in published maps and ins i u ional a ilia ions.
Sp inge Na u e o i s licenso (e.g. a socie y o o he pa ne ) holds
exclusi e igh s o his a icle unde a publishing ag eemen wi h
he au ho (s) o o he igh sholde (s); au ho sel -a chi ing o he
accep ed manusc ip e sion o his a icle is solely go e ned by he
e ms o such publishing ag eemen and applicable law.
© The Au ho (s), unde exclusi e licence o Sp inge Na u e Limi ed
2025
Na u e Clima e Change
Analysis h ps://doi.o g/10.1038/s41558-024-02198-6
Ex ended Da a Table 1 | De ails o scena io da abases included in he compa a i e analysis
IPCC epo name
Tempo al co e age o
Global
scena io coun
Global model
coun
Sou ce
SRES
Special Repo on
Emissions Scena ios
~2000
80
10
(IPCC, 2000)
AR5
5 h Assessmen
epo
2010-2014
1184
31
(IPCC, 2014)
SR1.5
Special Repo 1.5°C
2015-2018
416
25
(IPCC, 2018)
AR6
6 h Assessmen
epo
2015-2021 (includes
hose in SR1.5)
2304
95
(Bye s e al.,
2022)
