- 1 -
Science and policy characteristics of 1
the Paris Agreement temperature 2
goal 3
Carl-Friedrich Schleussner
1,2,*
, Joeri Rogelj
3,4
, Michiel Schaeffer
1,5
, Tabea Lissner
1,2
, 4
Rachel Licker
6
, Erich Fischer
4
, Reto Knutti
4
, Anders Levermann
2,7,8
, Katja Frieler
2
, 5
William Hare
1,2
6
1
Climate Analytics, Berlin, Germany 7
2
Potsdam Institute for Climate Impact Research, Potsdam, Germany
8
3
Energy Program, International Institute for Applied Systems Analysis, Laxenburg, Austria
9
4
Institute for Atmospheric and Climate Science, ETH Zurich, Zürich, Switzerland 10
5
Environmental Systems Analysis Group, Wageningen University and Research Centre, Wageningen, The 11
Netherlands 12
6
Woodrow Wilson School of Public and International Affairs, Princeton University, Princeton, USA 13
7
Institute of Earth and Environmental Science, University of Potsdam, Potsdam, Germany 14
8
Lamont-Doherty Earth Observatory, Columbia University, New York, USA 15
*email: carl.schleussner@climateanalytics.org
16
17
The Paris Agreement sets a long-term temperature goal of holding global average temperature 18
increase to well below 2°C and pursuing efforts to limit this to 1.5°C above preindustrial levels. Here 19
we present an overview of science and policy aspects related to this goal and analyse implications 20
for mitigation pathways. We show examples of discernible differences in impacts between 1.5°C 21
and 2°C warming. At the same time, most available low emission scenarios at least temporarily 22
exceed the 1.5°C limit before 2100. The legacy of temperature overshoots and the feasibility of 23
limiting warming to 1.5°C or below thus become central elements of a post-Paris science agenda. 24
Countries' near-term mitigation targets for the 2020-2030 period are insufficient to secure the 25
achievement of the temperature goal. An increase in ambition for this period will determine the 26
Agreement's effectiveness. 27
The ultimate objective of the United Nations Framework Convention on Climate Change (UNFCCC) is 28
to “prevent dangerous anthropogenic interference with the climate system”
1
. This objective can be 29
operationalized by, for example, expressing “dangerous anthropogenic interference” (DAI) in terms of 30
a long-term global temperature limit. However, this translation depends on world-views, political, legal 31
and other value judgments and requires continuous evaluation
2
. Scientific assessments, like those by 32
the Intergovernmental Panel on Climate Change (IPCC), are not in the position to recommend specific 33
levels of warming or greenhouse gas concentrations. Nevertheless, they provide essential information 34
for science-based political decisions by outlining the impacts, risks and vulnerabilities, as well as 35
technological, economic and feasibility assessments associated with different goals
2
. 36
Here we reflect on science and policy aspects relating to the temperature and mitigation goals 37
of the Paris Agreement
3
. First, we give a short history of the emergence of 1.5°C and 2°C as limits in 38
- 2 -
the policy debate followed by an overview of the current understanding of the differences in impact 39
indicators between these two warming levels. We then examine the mitigation architecture of the 40
Paris Agreement and how it is designed to progressively increase mitigation ambition to meet its long-41
term global temperature goal (LTTG). In this context, we assess characteristics of emission pathways 42
from the scientific literature. Finally, we outline implications for technological requirements and near-43
term action, and discuss elements of a post-Paris science agenda. 44
A short history of temperature goals 45
The adoption of the LTTG in the Paris Agreement stems from a long-standing climate policy debate. In 46
1996, the European Union Environment Council was first to identify a global mean surface-air 47
temperature (GMT) increase of 2°C above pre-industrial levels as a limit not to be exceeded, based on 48
the IPCC’s Second Assessment Report
4
. This position was subsequently confirmed by EU heads of 49
government in 2005 and 2007
4
, mainly informed by the IPCC’s 2001 Third Assessment Report (TAR)
5
. 50
Eventually, this “not exceed” limit was taken up by the G8 in 2009
4
. 51
The gradual adoption of specific warming limits by political bodies can be linked to the evolution 52
of the underlying scientific basis. Although not comprehensive, progress in the understanding of 53
climate impacts and their relation to GMT increase might be best illustrated by the temporal evolution 54
of the IPCC’s “Reasons for Concern” (RFCs), a framework for aggregating impacts, risks, and 55
vulnerabilities that was first developed in 2001 for the TAR
5
. With scientific insights steadily 56
progressing, assessments based on the IPCC’s Fourth Assessment Report (AR4)
6
and in the Fifth 57
Assessment Report (AR5)
7
have identified higher risks for all RFCs at lower temperature levels. 58
By the UNFCCC Copenhagen Conference (COP15) in 2009 and informed by the conclusions of the 59
IPCC AR4
8
, approximately 100 countries were calling for warming to be limited to below 1.5°C relative 60
to pre-industrial levels
4,9
. Although COP15 itself was widely regarded as a failure, two politically 61
durable outcomes from the Copenhagen Accord
10
have ultimately translated into the Paris Agreement: 62
First was the emergence of a long-term goal agreed at head of government level, expressed then as 63
an aim to hold the increase in warming below 2°C, combined with a recognition that deep cuts in global 64
emissions are required "according to science". Second and directly linked was the agreement to review 65
the “hold below 2°C” long-term goal with a view to strengthening it, addressing the 1.5°C limit called 66
for by vulnerable countries. 67
The ‘hold below 2°C’ goal was formally agreed upon in 2010 at the COP16 in Cancun and tied to a 68
review of the adequacy of this limit with a view to examining 1.5°C as an alternative
11
. While this led 69
to little or no reaction in scientific circles, a science-based review process under the UNFCCC was 70
established: the Structured Expert Dialogue (SED). Based principally on IPCC AR5 science, the SED 71
concluded that the “concept, in which up to 2°C of warming is considered safe, is inadequate” and that 72
“limiting global warming to below 1.5 °C would come with several advantages”.
12
At the same time, 73
substantial research gaps with regard to 1.5°C science were identified.
12
74
- 3 -
Climate impacts at 1.5°C and 2°C warming 75
Recent scientific literature allows for an assessment of differences in climate projections and impacts 76
at 1.5°C and 2°C that goes beyond the AR5 assessment. We do not aim at providing a full review, but 77
use a regional- and impact-specific approach to highlight differences for selected impacts
13
(Figure 1). 78
All impacts are assessed at a GMT increase of 1.5°C and 2°C above preindustrial levels. For impact 79
indicators that are expressed relative to the 1986-2005 period (with a GMT increase of about 0.6°C 80
above the 1850-1900 average
14
, our approximation for pre-industrial levels), the results thus relate to 81
a warming of 0.9°C or 1.4°C above the recent past. 82
The occurrence of hot temperature-related extremes has increased robustly over the historical 83
period
15,16
. As temperatures rise, the frequency of hot extremes above a fixed threshold increases non-84
linearly. In a 2°C warmer world, the global occurrence probability of a pre-industrial 1-in-a-1000 day 85
extreme temperature event is projected to be about double compared to 1.5°C (ref. 17, Figure 1a). 86
This is about 27 times higher compared to pre-industrial and more than five times higher than today. 87
Patterns of precipitation-related changes are considerably more uncertain
18
. Nevertheless, the 88
increase of extreme precipitation events due to anthropogenic warming is evident from the 89
observational record
19,20
and the intensity of heavy precipitation events is projected to robustly 90
increase globally between 1.5°C and 2°C warming
21
. These changes are particularly pronounced in 91
specific regions like the high latitudes and South Asia (Figure 1b). Equally, the global occurrence 92
probability of a 1-in-1000 day extreme precipitation event is estimated to increase by about 45% (full 93
model uncertainty range: 28-70%) at 1.5°C compared to pre-industrial levels and by 65% (41-100%) at 94
2°C (ref.
17
). Changes in the water cycle may be experienced by half of the world's population at a 95
warming of 2°C (ref. 22). 96
Contrary to an almost global increase in heavy precipitation events, only about 25% of the 97
global land area is projected to experience substantial changes in dry spell length at 2°C warming
21
. 98
Projected changes are most pronounced for subtropical regions, in particular the Mediterranean. 99
Patterns of change in water availability emerge similar to changes in water-cycle extremes
23
. While 100
global changes are not significant, an increase in total water availability is projected for high-latitude 101
regions, as well as the South-Asian monsoon regions for both a 1.5°C and 2°C warming
21
. At the same 102
time, water availability is projected to decrease in subtropical regions, and most prominently in the 103
Mediterranean, where the projected median reduction of at least 9% at 1.5°C over 50% of the region’s 104
land area nearly doubles to at least 17% at 2°C (ref. 21, Figure 1c). 105
Crop yield projections are subject to considerable uncertainty arising both from uncertainty in 106
projections of climate as well as crop models
24
. In addition, many factors affecting crop yield 107
projections – such as elevated CO
2
and ozone concentrations
25,26
as well as management options
27,28
, 108
nitrogen limitations
29
and the impact of heat extremes
30
– are not well-constrained and are 109
represented very differently across agricultural models
31
. Observational evidence indicates that 110
substantial impacts of extremes on crop yields are already evident in the second half of the 20
th
111
century
32
. 112
These substantial uncertainties, particularly regarding the efficiency of the CO
2
fertilization 113
effect, render a robust differentiation between climate impacts on crop yields between 1.5°C and 2°C 114
difficult. However, using multi-model inter-comparison data, some general patterns can be identified 115
for the four most common global crop types: maize, wheat, rice and soy (Figure 1f). While high 116
- 4 -
latitudes may gain, local wheat and maize crop yield reductions are projected over the tropical land 117
area (30°S-30°N) at 1.5°C warming, with more significant reductions projected at 2°C (ref. 21). The 118
vulnerability of wheat and maize production in tropical regions is also evident from a meta-analysis of 119
crop yield projections
27
. Using multi-model ensemble projections that resemble observed yields, it is 120
estimated that global wheat production will decrease by about 6% per °C of warming
33
. At 1.5°C and 121
2°C warming, local rice and soy yields are projected to increase in the tropics compared to present-day 122
yields, as the positive effect of CO
2
-fertilization counterbalances detrimental impacts of climate change 123
in the model projections. However, additional gains for warming above 1.5°C are limited (c.f. Figure 124
1f), and reductions are evident for all crop types for projections that exclude the effects of CO
2
-125
fertilization
21,24
. 126
Also relevant for global food security, ocean ecosystems appear to be particularly vulnerable 127
to anthropogenic greenhouse gas emissions due to warming, deoxygenation, and ocean acidification
34
. 128
As a result, several key risks for oceanic ecosystems are assessed as severe and widespread due to 129
climate change at warming levels below 1.5°C (ref. 35). In particular, global coral reef systems are 130
projected to be threatened by both ocean acidification and thermal stress
36,37
. Projections indicate that 131
nearly all global warm-water coral reef systems will be at risk of long-term degradation at 2°C due to 132
temperature induced bleaching, unless very optimistic scenarios of coral reef adaptation are 133
assumed
37
. Limiting warming to 1.5°C reduces the fraction of ocean grid cells with reefs under risk of 134
long-term degradation based on the frequency of bleaching events (Figure 1d). Similarly, Arctic 135
ecosystems and traditional livelihoods are under substantial pressure as sea-ice vanishes – only a 136
warming of well below 2°C may ensure that significant areas of end-of-summer sea-ice remain in the 137
Arctic
38
. A recent extrapolation of observed sensitivities of nations’ economic growth to temperature 138
fluctuation indicates substantial negative effects of climate change on the global economy (ref. 39). 139
The study identifies differences in impacts between 1.5°C and 2°C on economic production globally, 140
with tropical countries affected most
39
. 141
Many impacts of climate change, in particular those related to large-scale systems such as ice-142
sheets and the global oceans, will not materialize fully until 2100 even under 1.5°C scenarios, but 143
rather over centuries and millennia. Prominent examples include ocean acidification
40
, glacier melt, 144
sea-level rise
41,42
and loss of permafrost
43
. Model projections, as well as evidence from the paleo-145
record, indicate multi-millennial global sea-level rise of several meters per degree (°C) of warming, 146
with contributions from the Greenland and partly disintegrated West-Antarctic ice-sheets
44,45
. 147
The presented collection of differences in impact indicators points towards hot spots of change 148
between 1.5°C and 2°C, particularly in tropical regions. These findings tend to support the earlier 149
assessments by vulnerable countries that 1.5°C is a less risky limit than 2°C
9
. The following section will 150
examine the mitigation components included in the Paris Agreement and their implications for 151
reaching the LTTG. 152
The Paris mitigation architecture 153
An inter-locking set of articles of the Paris Agreement provides the legally binding mitigation ambition 154
architecture of the agreement
46
. This includes the LTTG in Article 2.1(a), the linked long-term 155
mitigation goals expressed in Article 4.1, and obligations on Parties to progressively increase their 156
(mitigation) ambition to collectively achieve these goals. Such an increase over time should be realized 157
- 5 -
through successive five-yearly nationally determined contributions (NDCs), which are progressively 158
more ambitious and informed by science assessments. Table 1 outlines key elements of this legally 159
binding mitigation ambition architecture and our respective interpretation. 160
The LTTG in Article 2.1(a) is operationalized by means of the long-term global mitigation goals in 161
Article 4.1. This includes a peaking of global emissions as soon as possible with rapid reductions 162
thereafter in accordance with the best available science so as to achieve “a balance between 163
anthropogenic emissions by sources and removals by sinks of greenhouse gases in the second half of 164
this century”
3
. Details of these emission goals are to be determined consistent with the LTTG. The 165
term “balance” here essentially means that globally aggregated anthropogenic greenhouse gas (GHG) 166
emission are required to reach zero. 167
Individual Party’s ambition levels as expressed in each successive NDC are required to “represent 168
a progression beyond the Party’s then current nationally determined contribution” (Article 4.3)
3
. This 169
is part of the ambition mechanism under the Paris Agreement that additionally includes a five yearly 170
“global stocktake” of the globally aggregated effects of countries’ actions and ambitions with regard 171
to the purpose of the Paris Agreement and its long-term goals. The first stocktake is to be held in 172
2023 and is a crucial milestone for the Agreement's effectiveness. 173
Agreed as part of the enabling COP decisions to commence the implementation of the Paris 174
Agreement, a “facilitative dialogue” will be held in 2018. It is mandated to assess how the global 175
aggregate effect of Parties’ commitments (NDCs) and actions are tracking with respect to the long-176
term global mitigation goals, and to inform the preparation and updating of NDCs for 2025 and 2030. 177
The IPCC was invited, and has since decided
47
, to prepare a Special Report on 1.5°C by 2018, specifically 178
including an assessment of this issue, as well as on the impacts of warming of 1.5°C. 179
The content and timing of the facilitative dialogue and the IPCC Special Report on 1.5°C will be 180
significant in the context of the request for Parties to the Paris Agreement to communicate or update 181
their NDCs by 2020 so as to increase the level of ambition, both individually and in aggregate terms. 182
The results of this process may well be the first substantive test of the Paris Agreement’s effectiveness. 183
Finally, Decision 10/CP.21 encourages the scientific community to conduct research on the gaps 184
identified in the aforementioned SED, including on 1.5°C impacts at regional and local scales and 185
scenarios that limit warming below 1.5°C above pre-industrial. 186
Implications for mitigation pathways 187
Mitigation pathways in line with different interpretations of the Paris Agreement and the respective 188
solution space have not been systematically explored so far. The following analysis is based upon an 189
ensemble of opportunity of scenarios available in the current literature. These scenarios are derived 190
by Integrated Assessment Models (IAMs) and any conclusions drawn should be mindful of the 191
assumptions underlying these models
48
. Furthermore, the classification of scenarios applied by us 192
should not be seen as an interpretation of the Paris Agreement. 193
Emissions reduction pathways meeting a specific temperature limit are most commonly 194
interpreted in terms of the probability they imply for staying below such a limit. For example, the ‘hold 195
below 2°C’ limit has often been assessed in the scientific literature, including the IPCC AR5 Synthesis 196
Report
49
, assuming a greater than 66% (or ‘‘likely’’
50
) probability of holding the increase in GMT below 197