This is a repository copy of Climate and air-quality benefits of a realistic phase-out of fossil
fuels.
White Rose Research Online URL for this paper:
http://eprints.whiterose.ac.uk/151100/
Version: Accepted Version
Article:
Shindell, D and Smith, CJ orcid.org/0000-0003-0599-4633 (2019) Climate and air-quality
benefits of a realistic phase-out of fossil fuels. Nature, 573 (7774). pp. 408-411. ISSN
0028-0836
https://doi.org/10.1038/s41586-019-1554-z
© The Author(s), under exclusive licence to Springer Nature Limited 2019. This is an
author produced version of a paper published in Nature. Uploaded in accordance with the
publisher's self-archiving policy.
eprints@whiterose.ac.uk
https://eprints.whiterose.ac.uk/
Reuse
Items deposited in White Rose Research Online are protected by copyright, with all rights reserved unless
indicated otherwise. They may be downloaded and/or printed for private study, or other acts as permitted by
national copyright laws. The publisher or other rights holders may allow further reproduction and re-use of
the full text version. This is indicated by the licence information on the White Rose Research Online record
for the item.
Takedown
If you consider content in White Rose Research Online to be in breach of UK law, please notify us by
emailing eprints@whiterose.ac.uk including the URL of the record and the reason for the withdrawal request.
The harmonized climate and air quality benefits of a realistic phase out of fossil 1
fuels 2
3
Drew Shindell
1,2*
and Christopher J. Smith
3
4
5
1
Nicholas School of the Environment and Duke Global Health Initiative, Duke 6
University, Durham, NC, USA 7
2
Porter School of the Environment and Earth Sciences, Tel Aviv University, Tel 8
Aviv, Israel 9
3
Priestley International Centre for Climate, University of Leeds, Leeds, UK 10
11
Fossil fuel combustion produces emissions of the long-lived greenhouse gas carbon 12
dioxide and short-lived pollutants, including sulphur dioxide, that contribute to 13
atmospheric aerosol formation
1
. Atmospheric aerosol can cool the climate, masking some 14
of the warming effect resulting from greenhouse gases emissions
1
. Aerosol particulates 15
are highly toxic when inhaled, however, leading to millions of premature deaths per 16
year
2, 3
. Phasing out unabated fossil fuel combustion will thus provide health benefits but 17
will also reduce aerosol masking of greenhouse gas-induced warming. Given the much 18
more rapid response of aerosols to emissions changes relative to carbon dioxide, there are 19
large near-term increases in the magnitude and rate of climate warming in many idealized 20
studies that typically assume an instantaneous removal of all anthropogenic or fossil fuel-21
related emissions
1, 4, 5, 6, 7, 8, 9
. Here we show that more realistic modelling scenarios do not 22
produce a substantial near-term increase in either the magnitude or rate of warming, and 23
in fact can lead to a decrease in warming rates within two decades of the start of the fossil 24
fuel phaseout. Accounting for the time required to transform power generation, industry 25
and transportation leads to gradually increasing and largely offsetting climate impacts of 26
carbon dioxide and sulphur dioxide, with the rate of warming further slowed by fossil 27
methane emission reductions. Our results indicate that even the most aggressive plausible 28
transition to a clean energy society provides benefits for climate change mitigation and 29
air quality at essentially all decadal to centennial timescales. 30
31
There is a substantial body of literature pointing out that air quality policies, under which 32
cooling aerosol particles are reduced, can be beneficial for human health but lead to 33
‘disbenefits’ for climate change
1, 4, 5, 6, 7, 8, 9
. Such trade-offs clearly exist for some air 34
quality policies, such as flue gas desulfurization of coal-fired power plants, and studies 35
have suggested the alarming possibility that warming rates could accelerate from their 36
current levels of about 0.2C per decade to 0.4 to C were aerosols alone to be rapidly 37
removed
5, 10, 11, 12, 13
. The presence of such trade-offs in response to climate policies is less 38
clear, however. The scientific community has long known that due to the shorter lifetime 39
(days to weeks) of cooling aerosols relative to long-lived greenhouse gases such as 40
carbon dioxide (CO
2
, decades to centuries), cessation of emissions would lead to a near-41
term pulse of warming. This was illustrated most clearly by the Intergovernmental Panel 42
on Climate Change (IPCC) in the Frequently Asked Questions to the Working Group I 43
contribution to the Fifth Assessment Report (AR5)
14
, which showed that ceasing 44
anthropogenic emissions would lead to a spike in warming of about half a degree within a 45
few years, followed by a slow cooling that would require nearly a century to recover to 46
current temperatures. Many studies over the past two decades have found a similar near-47
term warming due to removal of anthropogenic aerosols when all aerosol or all 48
anthropogenic emissions cease
4, 9, 15, 16, 17, 18
. 49
50
Though authors have often framed their work at least in part as an examination of the 51
geophysical commitment to past emissions, such results have also been widely assumed 52
to provide an indication of future behavior were there to be dramatic anthropogenic 53
emission cuts. This has driven a fairly common perception that the broad phasing out of 54
unabated fossil fuel usage required to meet ambitious climate change mitigation targets 55
such as the Paris Climate Agreement also leads to trade-offs, with a near-term increase in 56
both the magnitude and rate of warming as a ‘climate penalty’ (e.g. 57
https://nationalpost.com/news/world/scrubbing-aerosol-particles-from-the-atmosphere-a-58
faustian-bargain-study-finds, ref.
3, 7
). Such a view may come from incomplete 59
understanding of scientific studies, or from news and social media reaction from which 60
some may have incorrectly inferred that aerosol removal inevitably leads to accelerated 61
warming regardless of co-emitted greenhouse gases. This perception has led to 62
contentious debates in the policy arena, for example during the approval process for the 63
Summary for Policy Makers of the IPCC Special Report on 1.5°C (hereafter SR1.5) 64
about the role of non-CO
2
emissions reductions. Specifically, some countries with high 65
air pollution burdens pushed for an equal emphasis on the near-term acceleration of 66
warming that would result if they were to shift away from fossil fuels alongside the 67
Report’s presentation of the public health benefits. 68
69
We have studied the pathways included in the recently released SR1.5 (ref.
19
) to 70
investigate whether such a climate penalty exists in realistic scenarios of the transition to 71
clean energy as well as in the idealized ‘zero emissions’ studies. We include 42 pathways 72
classified by the SR1.5 as consistent with 1.5C with no or limited (<0.1C) temporary 73
overshoot of the target (see Methods). These scenarios are least-cost pathways generated 74
by models of the energy-economy-land system and include a rapid phaseout of unabated 75
fossil fuel usage with a median decrease of ~60% by 2050 and 85% by 2100 for all 76
primary energy and a >90% reduction in usage of fossil fuels for electricity generation by 77
2050. The speed at which fossil fuels usage is reduced in these models is based on 78
feasibility assessments of rates of capital turnover, technology switching, socio-economic 79
limits to technological and behavioral shifts, and the requisite financial flows. Rates of 80
change in individual sectors are typically at the high end of those in historical precedents, 81
whereas the scale of the transitions envisioned is substantially larger than any historical 82
precedent for similar rates of change
20
. In other words, although energy-economy-land 83
models have sometimes underpredicted the rates of uptake of specific new technologies
21
, 84
the overall rates of the societal transformation away from fossil fuels in the 1.5C 85
pathways are likely at the upper end of what could be achieved under very ambitious 86
policies. Hence these are likely as close to the ‘zero emissions’ case as is practically 87
possible. These shifts result in rapid and deep cuts in both CO
2
and non-CO
2
emissions, 88
with CO
2
from fossil sources and sulphur dioxide (SO
2
, that is largely co-emitted) 89
decreasing by around 75-85% by 2050 in most scenarios (Figure 1). Some emissions with 90
large non-fossil sources, such as methane (CH
4
), do not necessarily decline by such a 91
large fraction, but typically decrease sharply in the near-term as their fossil portion is 92
eliminated (Extended Data Fig. 1). 93
94
We evaluate the global mean surface temperature response to these emissions changes 95
using the FaIR model that incorporates reduced complexity (relative to Earth System 96
Models) representations of the carbon cycle and the climate system
22, 23
(see Methods). 97
Carbon dioxide removal technologies are excluded to highlight the role of emissions 98
reductions, and some scenarios hence do not stay below 1.5C. Unlike the response to 99
idealized, instantaneous emissions removals, global mean temperatures in realistic 100
pathways do not show a near-term spike in warming (Figure 2). Temperatures continue to 101
increase for at least a decade, and near-term rates of change are highly scenario 102
dependent, but none exhibit an acceleration of warming to 0.3C decade
-1
or higher, and 103
all show a rapid decline in warming rates starting in the 2020s with rates by 2040 ranging 104
from negative (cooling) to less than half the current value (Figure 2). 105
106
We unravel the contributions of individual fossil-related emission decreases to projected 107
temperatures by recalculating changes when holding the fossil portion of individual 108
pollutant emissions constant at 2018 levels while allowing other emissions to follow their 109
specified 1.5C pathways (see Methods). The results show the gradual evolution of 110
temperature responses, with the largest impacts coming from fossil CO
2
, SO
2
and CH
4
111
emissions changes (Figure 3). The pace of change is influenced by the inertia in both the 112
physical climate system and in the socio-economic systems in which fossil fuels are used. 113
For CO
2
, concentrations adjust slowly to emissions changes, leading to a response that is 114
substantially extended in time in comparison with the response to SO
2
given that both are 115
largely phased out in the first half of the century (Figure 1). However, the response to 116
CO
2
is also clearly visible in the near-term. For SO
2
, the temperature response is limited 117
only by the response of the climate system, but the emissions changes are gradual as the 118
models include the reality that it takes substantial time to transform energy, transportation 119
and industrial systems under least-cost pathways. Hence roughly 2-3 decades are required 120
to reach 2/3 of the 2100 temperature response to SO
2
changes under these scenarios 121
despite their assumption of systemic rates of change that are faster and broader than any 122
historical precedent
20
. This gradual response to aerosol changes in plausible 1.5C 123
scenarios is consistent with findings using an intermediate complexity model
18
. 124
125
These results differ greatly from the idealized picture of a near-instantaneous response to 126
the removal of aerosol cooling followed by a slow transition to dominance by the effects 127
of CO
2
. In these more plausible cases, the temperature effects of CO
2
, SO
2
and CH
4
128
reductions roughly balance one another through about 2040, after which the cooling 129
effects of reduced CO
2
continue to grow whereas the SO
2
reduction-induced warming and 130
CH
4
reduction-induced cooling effects taper off so that CO
2
reduction-induced cooling 131
dominates (Figure 3). Examining the impact of CO
2
and SO
2
alone (Figure 3d), the faster 132
response of SO
2
means that the net effect of these two pollutants would indeed be a short-133
term warming, but a very small one of between 0.02 and 0.10C in the ensemble mean 134
temperature response (up to 0.30C for the 95
th
percentile across pathways). Accounting 135
for all fossil-related emissions (Figure 3e), any brief ‘climate penalty’ decreases to no 136
more than 0.05C (0.19C at the 95
th
percentile), with the smaller value largely due to the 137
additional near-term cooling from methane reductions. Nearly all the warming in the 138
2020s and 2030s (Figure 2) is thus attributable to the impact of the residual emissions 139
(mainly of CO
2
) during the gradual fossil phase out as well as response to historical 140
emissions
17
. 141
142
What explains the difference in our results in comparison with perception of a climate 143
penalty due to the rapid removal of aerosol cooling? In large part, the difference between 144
the response times of aerosols and CO
2
is smoothed out when both emissions are reduced 145
gradually compared with idealized zero emissions simulations. Note also that aerosol-146
cloud interactions are highly non-linear, with a substantial fraction of the forcing 147
remaining even at low aerosol precursor emissions. In addition, the perception of a 148
climate penalty may also reflect results from earlier work on transitioning away from 149
fossil fuels suggesting that the effects of sulfate could substantially outweigh those of 150
CO
2
. That was likely true in the past, as the ratio of SO
2
to CO
2
emissions (in tonnes of 151
S/C) was ~1/100 in 1980, roughly double the ~1/200 value in 2019 (using SR1.5 scenario 152
data). This stems from an increase in CO
2
emissions of ~70% along with a reduction in 153
SO
2
emissions of ~20% due to air pollution controls in many regions. Hence over the past 154
40 years, the world’s success in curbing SO
2
emissions along with its failure to curb CO
2
155
emissions have led the world to a state where aerosols mask a substantially smaller 156
portion of the effect of CO
2
, greatly diminishing any ‘climate penalty’ resulting from 157
simultaneously phasing out emissions of both pollutants. Prominent analyses showing 158
that aerosol reductions owing to clean air policies have likely led to observed increases in 159
warming
24, 25
and could cause rapid acceleration in future warming
5, 10, 11, 12, 13
may have 160
also left such a strong impression that the same is presumed to be the impact of any 161
future reductions in SO
2
, even when accompanied by CO
2
reductions. Finally, studies 162
have shown that complete cessation of CO
2
emissions leads to fairly constant global 163
temperatures
15
, which has implied to some that CO
2
reductions can be neglected in 164
determining the climate impact of a fossil-fuel phaseout
7
. On the contrary, when 165
compared to continued present-day emissions, the phaseout of CO
2
is more important 166
than concurrent air pollution reductions for climate over the long term, and no less 167
important in the short term (Figure 3). Hence the misperception may stem from 168
