Articles
https://doi.org/10.1038/s41558-018-0091-3
1
International Institute for Applied Systems Analysis (IIASA), Laxenburg, Austria.
2
Institute for Atmospheric and Climate Science, ETH Zurich, Zurich,
Switzerland.
3
Potsdam Institute for Climate Impact Research (PIK), Member of the Leibniz Association, Potsdam, Germany.
4
Joint Global Change
Research Institute, Pacific Northwest National Laboratory, College Park, MD, USA.
5
Fondazione Eni Enrico Mattei, Milan, Italy.
6
Centro Euro-Mediterraneo
sui Cambiamenti Climatici, Milan, Italy.
7
PBL Netherlands Environmental Assessment Agency, The Hague, The Netherlands.
8
Copernicus Institute for
Sustainable Development, Utrecht University, Utrecht, The Netherlands.
9
National Institute for Environmental Studies, Tsukuba, Japan.
10
Department of
Management, Economics and Industrial Engineering, Politecnico di Milano, Milan, Italy. *e-mail: rogelj@iiasa.ac.at
S
cenarios of the energy–economy–land system can facili-
tate the integrated assessment of the impacts and mitigation
of climate change. For the Fifth Assessment Report (AR5)
of the Intergovernmental Panel on Climate Change (IPCC), four
Representative Concentration Pathways
1
(RCPs) have provided cli-
mate researchers with a set of consistent climate forcings
2–4
. More
recently, the Shared Socio-economic Pathways (SSPs) have been
developed
5,6
. SSPs provide a socio-economic dimension to the inte-
grative work started by the RCPs
7
. This framework provides a basis
of internally consistent socio-economic assumptions that represent
development along five distinct storylines
8
: development under
a green-growth paradigm
9
(SSP1); a middle-of-the-road develop-
ment along historical patterns
10
(SSP2); a regionally heterogeneous
development
11
(SSP3); a development that results in both geograph-
ical and social inequalities
12
(SSP4); and a development path that is
dominated by high energy demand supplied by extensive fossil-fuel
use
13
(SSP5).
Prior to 2015, international climate policy under the United
Nations Framework Convention on Climate Change focused on the
goal of keeping the global-mean temperature increase below 2 °C
relative to pre-industrial levels
14
. The Paris Agreement reset this
long-term goal to holding the increase well below 2 °C and pursuing
efforts to limit it to 1.5 °C
15
. In this study, we present a set of strin-
gent climate change mitigation scenarios consistent with an increase
of 1.5 °C in 2100. Six integrated assessment models were included
in this study (AIM, the Asia–pacific Integrated Model
11
; GCAM4,
the Global Change Assessment Model
12
; IMAGE, the Integrated
Model to Assess the Global Environment
9
; MESSAGE-GLOBIOM,
the Model for Energy Supply Strategy Alternatives and their
GeneralEnvironmental Impact combined with the Global Biosphere
Management Model
10
; REMIND-MAgPIE, the Regionalized Model
of Investments and Development combined with the Model of
Agricultural Production and its Impact on the Environment
13;
and
WITCH-GLOBIOM, the World Induced Technical Change Hybrid
model combined with GLOBIOM
16
), with which we attempted
to model scenarios that limit end-of-century radiative forcing
to 1.9 W m
−2
under various SSPs (hereafter called ‘SSPx–1.9’ sce-
narios, with SSPx indicating the specific SSP assumed by the sce-
nario and 1.9 the radiative forcing target in 2100, Methods). This
scenario set allows the structured exploration of climate change
at a level consistent with limiting the global-mean temperature
increase in 2100 to 1.5 °C with approximately 66% probability (see
Fig. 1 and results described below). Overall, all teams were able to
produce 1.9 W m
−2
scenarios in SSP1, and four teams were suc-
cessful in SSP2. Of the three and four modelling frameworks that
attempted to model 1.9 W m
−2
scenarios in SSP4 and SSP5, one and
two were successful, respectively (see Methods, Supplementary
Table 1, Supplementary Fig. 1, Supplementary Text2). From this
set of 1.9 W m
−2
scenarios, a further, stringent climate mitigation
scenario has been selected for inclusion in the Scenario Model
Intercomparison Project
17
(ScenarioMIP) of the Sixth Phase of
the Coupled Model Intercomparison Project
18
(CMIP6), as well as
other CMIP6 MIPs (for example, refs
19,20
, Fig. 1a, Supplementary
Text 1, Methods).
Scenarios towards limiting global mean
temperature increase below 1.5 °C
Joeri Rogelj
1,2
*, Alexander Popp
3
, Katherine V. Calvin
4
, Gunnar Luderer
3
, Johannes Emmerling
5,6
,
David Gernaat
7,8
, Shinichiro Fujimori
1,9
, Jessica Strefler
3
, Tomoko Hasegawa
1,9
,
Giacomo Marangoni
5,6
, Volker Krey
1
, Elmar Kriegler
3
, Keywan Riahi
1
, Detlef P. van Vuuren
7,8
,
Jonathan Doelman
7
, Laurent Drouet
5,6
, Jae Edmonds
4
, Oliver Fricko
1
, Mathijs Harmsen
7,8
,
Petr Havlík
1
, Florian Humpenöder
3
, Elke Stehfest
7
and Massimo Tavoni
5,6,10
The 2015 Paris Agreement calls for countries to pursue efforts to limit global-mean temperature rise to 1.5 °C. The transition
pathways that can meet such a target have not, however, been extensively explored. Here we describe scenarios that limit
end-of-century radiative forcing to 1.9 W m
−2
, and consequently restrict median warming in the year 2100 to below 1.5 °C. We
use six integrated assessment models and a simple climate model, under different socio-economic, technological and resource
assumptions from five Shared Socio-economic Pathways (SSPs). Some, but not all, SSPs are amenable to pathways to 1.5 °C.
Successful 1.9 W m
−2
scenarios are characterized by a rapid shift away from traditional fossil-fuel use towards large-scale low-
carbon energy supplies, reduced energy use, and carbon-dioxide removal. However, 1.9 W m
−2
scenarios could not be achieved
in several models under SSPs with strong inequalities, high baseline fossil-fuel use, or scattered short-term climate policy.
Further research can help policy-makers to understand the real-world implications of these scenarios.
© 2018 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
NATURE CLIMATE CHANGE | VOL 8 | APRIL 2018 | 325–332 | www.nature.com/natureclimatechange
325
Articles
Nature Climate ChaNge
Emission and climate-related outcomes
CO
2
and other greenhouse gas (GHG) emissions peak before 2030
and decline rapidly over the next two to three decades in SSPx-
1.9 scenarios (Fig. 1, see Supplementary Figs. 2–6 for other emis
-
sions). By 2050, annual CO
2
and GHG emissions are in the range
of − 9–6 and 1–13 billion tons of CO
2
-equivalent emissions (gigaton
GtCO
2
e yr
−1
, Methods), respectively, across all available scenarios.
Underlying these reductions is a phase-out of industry and energy-
related CO
2
production at a rate of 0.2–7.1% yr
−1
(median: 3.0% yr
−1
,
see Supplementary Tables 2, 3 for a complete overview), combined
with rapid upscaling of carbon capture and storage (CCS) and
carbon-dioxide removal (CDR, see section on system transforma
-
tions below). Near-term emissions vary across the SSPs, because,
in contrast to SSP1, the effectiveness of near-term climate policies
is assumed to be limited in other SSPs (defined by so-called Shared
Policy Assumptions
5,21
). In that case, global mitigation is regionally
scattered and accelerates slower over the next few decades,requiring
it to accelerate faster later on.
All scenarios presented here lead to 1.9 W m
−2
radiative forc-
ing in 2100 within rounding precision (Supplementary Fig. 7), but
they differ in their likelihood of limiting warming below specific
temperature levels. All scenarios keep warming to below 2 °C with
more than 66% probability (Fig. 1d), and maximum (peak) median
temperature estimates vary from 1.5 °C to 1.8 °C. Near-term mitiga
-
tion has a determining role here: higher 2030 emissions come with
a temperature penalty (Supplementary Fig. 8). The probability of
limiting peak warming to below 1.5 °C relative to pre-industrial
levels is approximately halved and peak temperature about 0.2 °C
higher if emissions are at the high (> 45 GtCO
2
e yr
−1
) instead of the
low (< 30 GtCO
2
e yr
−1
) end of the available range in 2030 (Fig. 1e).
By 2100, this variation disappears and all scenarios limit warming
below 1.5 °C with about 66% probability (Supplementary Figs. 8, 9).
Cumulative CO
2
emissions
2016–2100 (GtCO
2
)
Non-CO
2
radiative forcing
in 2100 (W m
–2
)
00.2 0.
40
.6
–200
0
200
400
600
1.5
2.0 2.5 3.0 4.0
°C
Global mean temperature
increase relative to preindustrial levels
SSP marker implementation
AIM/CGE (A)
GCAM4 (G)
IMAGE (I)
MESSAGE-GLOBIOM (M)
REMIND-MAgPIE (R)
Bold symbols:
SSP marker
implementationWITCH-GLOBIOM (W)
SSP1–1.9 SSP4–1.9
SSP colours:
SSP2–1.9 SSP5–1.9
a
b
c
f
e
d
SSP1–1.9
SSP4–1.9
SSP2–1.9
SSP5–1.9
SSP1–1.9
SSP4–1.9
SSP2–1.9
SSP5–1.9
SSP1–1.9
SSP4–1.9
SSP2–1.9
SSP5–1.9
1.5 °C
in 2100
1.5 °C 2 °C
Exceedance probability (%)
0
20
40
60
80
100
Global Kyoto GHG emissions
(GtCO
2
-eq yr
–1
)
Time (years)
–20
0
20
40
60
2000 2050 2100
Models
Global non-CO
2
emissions
in 2100 (GtCO
2
-eq yr
–1
)
CH
4
N
2
O
F gases
AG IMRWAGMRWGR
0
5
10
Time (years)
SSP5–8.5
Updated CMIP6
ScenarioMIP
Set:
2100 range
SSP3–7.0
SSP4–6.0
SSP2–4.5
SSP1–2.6
SSP1–1.9
SSP4–3.4
Global CO
2
emissions (GtCO
2
yr
–1
)
2000 2020 2040 2060 2080 2100
–40
–20
0
20
40
60
80
100
120
140
Probability of projected
peak (maximum) warming (%)
2030 GHG emissions (GtCO
2
-eq yr
–1
)
2010 30 40 50
0
20
40
60
80
100
Reference scenarios
6.0 W m
–2
4.5 W m
–2
3.4 W m
–2
2.6 W m
–2
1.9 W m
–2
Historical
emissions
Fig. 1 | Emission and temperature characteristics of 1.9 Wm
−2
scenarios under various SSPs. a, Global CO
2
emissions of SSP scenarios with the selected
CMIP6 ScenarioMIP subset highlighted. Historical emission from ref.
52
. All other panels show 1.9 W m
−2
scenario data only. b, Global Kyoto GHG
emissions. Shaded areas show the range per SSP, solid lines the marker scenarios for each SSP and dashed lines single scenarios that are not markers.
Single model detail is provided in Supplementary Fig. 2. c, Non-CO
2
GHGs per scenario in 2100. d, Exceedance probability of various temperature limits
for the 1.9 W m
−2
scenarios with bars showing the full range over all available scenarios per SSP. Except for the first sub-panel, all other panels give
the exceedance probability over the entire twenty-first century. e, Probability of peak warming versus 2030 GHG emissions in 1.9 W m
−2
scenarios. f,
Dependence of cumulative CO
2
emissions on non-CO
2
radiative forcing in 2100.
© 2018 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
NATURE CLIMATE CHANGE | VOL 8 | APRIL 2018 | 325–332 | www.nature.com/natureclimatechange
326
Articles
Nature Climate ChaNge
Whether these pathways provide an acceptable interpretation of the
Paris Agreement long-term temperature goal is not a scientific but a
political question
22,23
, which we do not address here.
Across all 13 available scenarios, net zero GHG emissions are
reached around 2055–2075 (rounded to the nearest 5 years). Net
zero CO
2
emissions are reached earlier (Supplementary Table 2).
The year of reaching net zero GHG emissions is inversely correlated
with emissions in 2030. For example, scenarios with 2030 GHG
emissions higher than 40 GtCO
2
e yr
−1
reach global net zero GHG
emissions before 2060 (Supplementary Fig. 10). Cumulative CO
2
emissions over the 2016–2100 period range from − 175 to 475 GtCO
2
(SSP2 median: 250 GtCO
2
, rounded to the nearest 25 GtCO
2
). End-
of-century non-CO
2
radiative forcing strongly influences the varia-
tion across this range
24
(Fig. 1f). These values are consistent with
earlier published estimates
24–27
(Supplementary Text 3) and lead to
2100 atmospheric CO
2
concentrations in the 350–390 p.p.m. range.
Potential feedbacks that are currently not included, such as CO
2
and
CH
4
release from permafrost thawing or changes in other natural
sources, can reduce carbon budgets further
28,29
and therefore alter
the presented climate outcomes.
Even in these very stringent mitigation pathways, sizeable
remaining CH
4
and N
2
O emissions are projected by all models
(Fig. 1c, Supplementary Fig. 6), and in 2100, respectively, 53–85%
and 59–95% of these emissions originate from agriculture. The
uncertainty in CH
4
and N
2
O emissions is large with inter-model
variations dominating inter-SSP variations. High and low estimates
for 2100 differ by a factor of 2–3, mainly owing to uncertainties in
how emissions from agriculture are treated and can be mitigated in
different models
30,31
. Important uncertainties also remain in the CO
2
mitigation contribution of the land-use sector
31
(Supplementary
Fig. 5). Here, emissions decline over the long term, but whether and
to what degree the land-use sector becomes a global sink is very
model-dependent (Supplementary Text 4).
System transformations
Achieving pronounced emission reductions requires a transforma-
tion of the global economy. Previous studies have discussed the
implications of such a global transformation for the energy and
land-use system
32
, highlighting the importance of limiting future
energy demand
32
to keep warming to below 1.5 °C and of changing
consumption patterns
33
combined with sustainable intensification
of agriculture
34
. We here focus on confirming these characteristics
and exploring the extent to which they vary across SSPs.
All 1.9 W m
−2
scenarios in this study strongly limit energy
demand growth (Fig. 2d, Supplementary Fig. 11), with energy
intensity reduction rates of 2–4% yr
−1
from 2020 to 2050 (Fig. 2d).
InSSP2, final energy demand in 2050 is limited to 10–40% above
2010 levels (rounded to the nearest 5%). This compares to 10%
below to 30% above, and 45–75% above 2010 levels in SSP1 and
SSP5, respectively. Energy conservation is therefore a common
strategy in stringent mitigation scenarios, but it also has limits.
Energy supply also has to be transformed to achieve reductions in
deep emissions. This includes upscaling of bioenergy and renewable
energy technologies, shifting away from freely emitting fossil-fuel
use, and the deployment of CDR, such as Bioenergy with Carbon
Capture and Sequestration (BECCS) or large-scale afforestation
(see Supplementary Text 5 for a discussion of CDR in SSPx-1.9 sce
-
narios). Non-biomass renewables (solar, wind, hydro and geother-
mal energy) scale up rapidly over the twenty-first century (Fig. 2a),
reaching mid-century electricity shares of 60–80% and 32–79% in
SSP1 and SSP2, respectively (Supplementary Fig. 12). In the marker
SSP scenarios, these shares are 79%, 60% and 61% in SSP1, SSP2 and
SSP5, respectively. Both solar and wind energy is projected to scale
up consistently across the different SSPs (Supplementary Fig. 13).
Particularly for wind energy, inter-model variations dominate over
differences induced by different SSPs, a feature also present in less
stringent mitigation pathways
35
(Supplementary Table 4). SSP2 and
SSP5 1.9 W m
−2
scenarios see a strong upscaling of nuclear power,
whereas in SSP1, and particularly its marker implementation, the
contribution of nuclear energy use decreases compared to today’s
levels (Supplementary Fig. 13).
Under all SSPs, 1.9 W m
−2
scenarios show a clear shift away from
unabated fossil fuels (that is, without CCS, Fig. 2c), and a phase-
out of all fossil fuels. The marker implementations exhibit rapidly
declining contributions of coal until 2040 (less than about 20% of its
2010 contribution in 2040), followed by a phase-out of oil until 2060
(Supplementary Figs. 14, 15). The potential contribution of natural
gas to the primary energy mix is the most uncertain, with mid-cen
-
tury contributions ranging from 22 to 267 exajoules (EJ) yr
−1
across
all scenarios compared to about 100–110 EJ yr
−1
in 2010. Differences
in preferences for gas supply across models here dominate the varia
-
tion in costs and availability assumptions owing to alternative socio-
economic pathways (Supplementary Table 4, Supplementary Fig. 16).
Bioenergy is used in large amounts in all 1.9 W m
−2
scenarios,
and this can raise concerns for food security or biodiversity
36–38
.
These concerns depend both on how and how much bioenergy is
produced. Bioenergy demands can be met through dedicated energy
crops or through residues. The latter option comes with fewer trade-
offs than dedicated bioenergy crops
38
. Models, however, project very
different shares for the use of residues (Supplementary Table 5), and
further research clarifying its potential would be essential. For 2050,
global technical bioenergy potentials (including energy crops and
residues) were identified ranging from < 50 to > 500 EJ yr
−1
. High,
medium and low agreement was attributed to potentials of 100,
300 and > 300 EJ yr
−1
, respectively
36
. Bioenergy use is increased by
1–5% per year between 2020 and 2050 in 1.9 W m
−2
scenarios. Total
bioenergy use in 2050 is kept below about 300 EJ yr
−1
, and in most
cases below 150 EJ yr
−1
(Supplementary Fig. 17). In a green-growth
SSP1 world, markedly lower bioenergy contributions are projected
compared to an SSP2 world that continues the historical experience
(34–112 EJ yr
−1
lower in 2050). Putting this into context, scenar-
ios project approximately 100 EJ yr
−1
of bioenergy use (full range:
38–112, with important variations across SSPs) in baseline scenar
-
ios without any climate policy (Supplementary Fig. 17).
In 1.9 W m
−2
scenarios, land for energy crops and forest area is gen-
erally projected to expand during the twenty-first century, with large
variations across models, and this can impact land for agriculture
and water availability
39,40
(Fig. 2f, Supplementary Fig. 18). However,
in SSP1 the decrease in agricultural land in 1.9 W m
−2
scenarios is
reasonably similar to what is projected in a no-climate-policy base
-
line merely owing to low demand for agricultural commodities and
high agricultural intensification. Pasture is one of the activities most
affected by expanding other land uses and declines robustly across
models and SSPs (Supplementary Fig. 19). In the middle-of-the-road
SSP2 world, pastures decreases by 1–20% in 2050 compared to 2010
levels, and in SSP1, pastures also decrease by 8–16%. In a fossil-fuel
intensive SSP5 scenario, it declines by 15–25%. It is important to note
that SSP1 baseline scenarios already project a pasture-land decrease
of 1–11% due to shifts towards less meat-intensive diets, limited food
waste and a return of the world population to 7 billion people by
2100
5,9,31
. This reaffirms the important role that changes in food con-
sumption in combination with sustainable intensification of agricul-
ture have for stringent mitigation
31,34,41
.
Large-scale afforestation and reforestation can make an impor
-
tant contribution to the overall CDR effort. In the sustainable
SSP1 world, pressure on land is relatively low, and the forest area
in 2050 can therefore expand by 0–24% relative to 2010. However,
in the middle-of-the-road SSP2 scenarios, results are mixed, with
some models projecting forest area to decrease by 2% and oth
-
ers report an increase of up to 18%. SSP5 sees a change of 0–16%
(Supplementary Table 6). Not all models explicitly include affores
-
tation as a mitigation option and ranges therefore span results that
© 2018 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
NATURE CLIMATE CHANGE | VOL 8 | APRIL 2018 | 325–332 | www.nature.com/natureclimatechange
327
Articles
Nature Climate ChaNge
are not fully comparable across models. However, in all 1.9 W m
−2
scenarios climate policy leads to a net forest expansion compared
to no-climate-policy baselines (Fig. 2e). Integrated policy packages
are required that ensure food security is achieved together with
climate change mitigation
42
.
BECCS contributes the largest part of CDR in 1.9 W m
−2
scenar-
ios (Supplementary Fig. 20). Between 150–1,200 GtCO
2
(rounded
to nearest 25 GtCO
2
), equivalent to about 4–30 years of current
annual emissions, is removed from the atmosphere via BECCS
during the twenty-first century, with important variation between
models and across SSPs (Fig. 3a, d). SSP1 shows the lowest BECCS
deployment over the twenty-first century (150–700 GtCO
2
) owing
to its lower final energy demand and baseline emissions, compared
to SSP2 (400–975 GtCO
2
) and SSP5 (950–1,200 GtCO
2
). None of
the SSPx-1.9 scenarios explicitly attempted to limit the contribu
-
tion from BECCS. The numbers reported here therefore represent
projections of estimated cost-effective BECCS deployment in 1.9
W m
−2
scenarios, but do not represent minimum BECCS require-
ments in a strict sense.
Abated fossil fuels—that is, fossil fuels combined with CCS (fos
-
sil–CCS)—are often used in models as a bridging solution. However,
fossil–CCS still results in residual CH
4
emissions from coal mining
or gas handling, and CO
2
emissions due to imperfect capture and
leakage. These emissions can become too substantial for very strin
-
gent mitigation transitions. Indeed, almost all 1.9 W m
−2
scenarios
deploy less cumulative fossil–CCS than weaker mitigation scenar
-
ios (Fig. 3c). Optimal 1.9 W m
−2
strategies are therefore not merely
‘more of the same’. Overall, the BECCS share of total CCS increases
(Supplementary Fig. 20). CDR is thus preferred over fossil–CCS in
very stringent mitigation scenarios.
Differential mitigation
A previous study
43
has identified characteristics of 1.5 °C pathways in
comparison to 2 °C pathways. These characteristics were (i) greater
mitigation efforts on the demand side; (ii) energy efficiency improve
-
ments; (iii) CO
2
reductions beyond global net zero; (iv) additional
GHG reductions mainly from CO
2
; (v) rapid and profound near-term
decarbonization of energy supply; (vi) higher mitigation costs; and
(vii) comprehensive emission reductions implemented in the com
-
ing decade. Using our 1.9 W m
−2
and 2.6 W m
−2
scenarios as prox-
ies for 1.5 °C and 2 °C pathways, these characteristics still hold when
assessed with four additional models and varying socio-economic
assumptions (Fig. 4, Supplementary Text 6, and results above). None
of the 1.9 W m
−2
scenarios show a peak of emissions after 2020, and
82–98% of additional cumulative mitigation over the 2020–2100
period is achieved through CO
2
reductions (Supplementary Fig. 21).
Fig. 4 further illustrates the relatively stronger demand-side mitiga
-
tion efforts in 1.9 W m
−2
scenarios, particularly in the transport and
building sectors (see also Supplementary Figs. 22–24).
Mitigation costs increase substantially between 1.9 and
2.6 W m
−2
scenarios reflecting higher marginal abatement costs
(Figs. 4,5). The relative carbon price increase is largest in SSP2
(Fig. 4) and also SSP1 sees large relative increases across all models
(Supplementary Figs. 22–24). However, in absolute terms, carbon
f
Baseline
e
a
d
bc
SSP1
SSP colours:
SSP4
SSP2
SSP3
SSP5
SSP1
SSP4
SSP2
SSP3
SSP5
SSP1
SSP4
SSP2
SSP5
SSP1
SSP4
SSP2
SSP5
SSP1
SSP4
SSP2
SSP5
SSP1
SSP4
SSP2
SSP3
SSP5
SSP1
SSP4
SSP2
SSP3
SSP5
SSP1
SSP4
SSP2
SSP3
SSP5
SSP1
SSP4
SSP2
SSP3
SSP5
SSP marker
implementation
Final energy reduction in 1.9 W m
–2
scenarios
rel. to baseline over 2020–2100 period (%)
Global final energy in 1.9 W m
–2
scenarios
in 2050 as a factor of 2010 levels
Global annual final energy intensity rate of change
in 1.9 W m
–2
scenarios between 2020 and 2050 (%)
15
20
25
30
35
40
0.8
0.9
1.0
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
–4.5
–4.0
–3.5
–3.0
–2.5
–2.0
–1.5
AIM/CGE (A)
GCAM4 (G)
IMAGE (I)
MESSAGE-GLOBIOM (M)
REMIND-MAgPIE (R)
WITCH-GLOBIOM (W)
Spain
SSP1
SSP2
SSP3
SSP4
SSP5
Colombia
Democratic
Republic of
the Congo
Australia
Change in global cropland for agriculture
in 2100 relative to 2010 (Mha)
Primary energy
non-biomass renewables (EJ yr
–1
)
Time (years)
SSP1–1.9 SSP4–1.9
SSP colours:
SSP2–1.9 SSP5–1.9
2000 2050 2100
Time (years)
2000 2050 2100
Time (years)
2000 2050 2100
0
100
200
300
400
500
600
Primary energy
biomass with CCS (EJ yr
–1
)
0
100
200
300
400
500
Primary energy
coal without CCS (EJ yr
–1
)
0
50
100
150
200
2.6 W m
–2
2.6 W m
–2
1.9 W m
–2
2.6 W m
–2
Baseline
2010
2050 2100
Afforestation and reforestation
rel. to 2010 (%)
0
5
10
15
20
25
–5
0
5
10
15
20
25
30
35
40
45
Global baseline forest cover in 2010 (Mha)
0
500
1,000
1,500
2,000
2,500
3,000
3,500
4,000
4,500
–800 –600 –400 –200 0 200 400 600 800
1.9 W m
–2
1.9 W m
–2
Fig. 2 | Overview of key decarbonization characteristics in 1.9 Wm
−2
scenarios. a, Primary energy from non-biomass renewables (wind, solar, hydro
and geothermal energy). b, Primary energy from biomass with CCS (BECCS). c, Primary energy from coal without CCS. Shaded areas in a–c show the
range per SSP, solid lines the marker scenarios for each SSP and dashed lines single scenarios that are not markers. d, Three illustrations of global final
energy demand in 1.9 W m
−2
scenarios showing, from left to right, the average reduction from baseline over the 2020–2100 period, the change in 2050
compared to 2010 levels, and the annual rate of final energy intensity change. e, Global forest cover, and change relative to 2010 due to afforestation and
reforestation in 2.6 and 1.9 W m
−2
scenarios. f, Change in global cropland for agriculture in 2100 relative to 2010 in ‘Baseline’ scenarios in the absence of
climate change mitigation, as well as in 2.6 and 1.9 W m
−2
scenarios. Results are grouped per SSP (coloured lines with black symbols). rel., relative. Mha,
million hectares.
© 2018 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
NATURE CLIMATE CHANGE | VOL 8 | APRIL 2018 | 325–332 | www.nature.com/natureclimatechange
328
Articles
Nature Climate ChaNge
prices (Fig. 5), consumption losses and energy supply mitigation
investments (Supplementary Fig. 26) are highest when assuming the
less favourable socio-economic conditions of SSP2, SSP4 and SSP5.
For instance, the average discounted carbon prices (discounted to
2010 over the 2020–2100 period; Fig. 5) are estimated to be about
50–165 US$ per tCO
2
e in SSP2 (rounded to the nearest 5). They
are approximately 35–65% lower in SSP1, and for the two reported
SSP5 scenarios the change is − 30% and + 5%, respectively. The large
range of carbon prices is mainly driven by model uncertainties,
which were already identified for 2.6 W m
−2
scenarios
5
, but are more
pronounced here owing to the more stringent target.
Enabling and disabling factors
Our results show that some socio-economic developments and
assumptions about policy effectiveness preclude achieving strin
-
gent mitigation futures (Fig. 5). Such failures were anticipated
for SSP3, in which very heterogeneous regional development and
debilitating policy assumptions already rendered limiting end-of-
century radiative forcing to 2.6 W m
−2
unachievable in the models
5
(Supplementary Text 2). However, in SSP4 and SSP5 limiting radia
-
tive forcing to 1.9 W m
−2
proved difficult too. In SSP4, a world that
promotes both geographical and social inequalities, only one out of
three models attempting a 1.9 W m
−2
scenario was successful. Weak
mitigation is achieved rather easily in SSP4
5,12
. However, the lack
of control over land-related emissions in developing countries and
lower acceptability of CCS in developed countries in SSP4 make
very low emissions pathways unachievable
12
. Also in SSP5, a world
dominated by high economic growth and fossil-fuel development,
challenges to mitigation are high
13
. Finally, under a middle-of-the-
road development (SSP2) and under a green-growth paradigm
(SSP1) four and six models, respectively, were able to produce a
1.9 W m
−2
scenario (Supplementary Table 1).
AIM/CGE
60 45 34 26 19
0
11
GCAM4
60 45 34 26 19
0
IMAGE
60 45 34 26 19
0
0.5
MESSAGE-GLOBIOM
60 45 34 26 19
0
0.5
REMIND-MAgPIE
60 45 34 26 19
0
1
WITCH-GLOBIOM
60 45 34 26 19
0
1
2000 2020 2040 2060
Time (years)
2080 2100
0
5
10
15
20
25
30
35
Annual CO
2
sequestered by CCS
across 1.9 W m
–2
scenarios (GtCO
2
yr
–1
)
a
SSP1
SSP colours:
SSP4
SSP2
SSP5
Cumulative CO
2
stored by bioenergy with CCS
(BECCS) in twenty-first century (in 1,000 GtCO
2
)
d
Cumulative CO
2
stored by fossil fuels with CCS
(Fossil–CCS) in twenty-first century (in 1,000 GtCO
2
)
c
Cumulative CO
2
stored by CCS
in twenty-first century (1000 GtCO
2
)
b
Baseline
AIM/CGE
60 45 34 26 19
0
1
2
GCAM4
60 45 34 26 19
0
1
2
IMAGE
60 45 34 26 19
0
1
MESSAGE-GLOBIOM
60 45 34 26 19
0
1
REMIND-MAgPIE
60 45 34 26 19
0
1
WITCH-GLOBIOM
60 45 34 26 19
0
1
AIM/CGE
60 45 34 26 19
0
1
GCAM4
60 45 34 26 19
0
1
IMAGE
60 45 34 26 19
0
0.6
MESSAGE-GLOBIOM
60 45 34 26 19
0
1
1
REMIND-MAgPIE
60 45 34 26 19
0
WITCH-GLOBIOM
60 45 34 26 19
0
0.6
SSP5
SSP3
SSP2
SSP4
SSP1
Baseline
Baseline
BaselineBaselineBaseline
Baseline Baseline Baseline
BaselineBaselineBaseline
Baseline
Baseline
BaselineBaseline
Baseline
Baseline
Fig. 3 | BECCS, fossil–CCS and CCS across SSPs and across climate targets. a, Annual amount of CO
2
stored by CCS in 1.9 W m
−2
scenarios. Shaded areas
show the range per SSP, solid lines the marker scenarios for each SSP and dashed lines single scenarios that are not markers. b, Variation per modelling
framework and per SSP of cumulative CO
2
stored by CCS during the twenty-first century when moving from a world in the absence of climate policy
(baseline) to increasingly more stringent climate targets (6.0, 4.5, 3.4, 2.6 and 1.9 W m
−2
) c,d, As b but for fossil–CCS and BECCS, respectively. Note that
axis limits vary across models.
© 2018 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
NATURE CLIMATE CHANGE | VOL 8 | APRIL 2018 | 325–332 | www.nature.com/natureclimatechange
329
