52
Climate brief Politics
A history o action and inaction
The challenge without precedent
The first of six weekly climate briefs looks at the history and politics of attempts
to tackle global warming
n june 1988 scientists, environmental
I activists and politicians gathered in To-
ronto for a “World Conference on the
Changing Atmosphere”. The aspect of its
changing that alarmed them most was the
build-up of carbon dioxide, a greenhouse
gas. In the late 1950s, when systematic
monitoring of the atmosphere’s carbon-di-
oxide level began, it stood at around 315
parts per million (ppm). By that summer, it
had reached 350ppm—and a heatwave was
bringing record temperatures to much of
North America.
The week before the Toronto confer-
ence James Hansen, a climate scientist at
nasa, had pointed to the heatwave when
telling the us Senate that it was time “to
stop waffling…and say that the evidence is
pretty strong that the greenhouse effect is
here”. The Toronto conference took a simi-
lar view, calling for an international effort
to reduce global carbon-dioxide emissions
by 20% by 2005.
A mere four years later a global compact
against climate change had been signed.
Even with a boost from the end of the cold
war, which made global action on shared
concerns seem newly possible and provid-
ed an opening for a new eschatology to re-
place that of nuclear Armageddon, that
seemed like a remarkable political success
on the part of those pressing for action.
Unfortunately, a global agreement to
act is not the same thing as global action.
In this series
1 The politics o climate action
2 Modelling the greenhouse efect
3 The carbon cycle, present and uture
4 The impacts and their timescales
5 Engineering an energy transition
6 The imperative o adaptation
The Economist April 25th 2020
Fossil fuels are the bedrock of industrial
society. Even though the alternative of re-
newable energy has, since 1988, become far
more plausible, a decisive move away from
fossil carbon still means a wrenching and
unprecedented shift.
To many convinced environmentalists
that shift seems self-evidently worthwhile.
It fits with an ideology that commits them
to lives that have less impact on the natural
world. But in the face of climate change, in-
dividual willingness to sacrifice the fruits
of a high-energy lifestyle is not enough.
People, and countries, that do not share
such motivations must act, too.
The challenge of climate politics is to
overcome these differences by negotiating
ways forward that can gain general assent.
It is a challenge that, despite those remark-
able four years, has not been met. Instead
of emissions in 2005 being 20% lower than
they were in 1988, they were 34% higher. By
2017 they were 22% higher still.
Think global, act global
The Toronto attendees’ belief that an inter-
national agreement could bring down car-
bon-dioxide emissions rested in part on an
agreement reached a year before to limit
the production of ozone-destroying chem-
icals, most notable among them the
chlorofluorocarbons (cfcs) used in fridges
and spray-cans. That Montreal protocol
looked like a template in two ways.
The first was that it was global. Since the
1960s the environmental movement had
increasingly taken “saving the planet” as
its rhetorical focus. But practical environ-
mental protections, such as clean-air regu-
lations, almost all worked on a national, or
at most regional, basis. Because the world’s
cfcs are thoroughly mixed together before
they reach the stratosphere’s ozone layer,
the Montreal protocol had to be genuinely
global, and thus balance the needs of devel-
oped and developing countries.
The second was that the Montreal pro-
tocol required remarkable faith in science.
Unlike most pollution controls, which try
to reduce harm already being done, it
called for expensive action to deal with a
problem that, despite the dramatic discov-
ery of the Antarctic ozone hole in 1985, was
not yet hurting people. It was based instead
on the likelihood of future catastrophe.
Climate scientists realised that an emis-
sions-reduction agreement on greenhouse
gases would need a similarly strong con-
sensus on their dangers. This led to the cre-
ation in late 1988 of the Intergovernmental
Panel on Climate Change (ipcc). Including
researchers from governments, academia,
industry and non-governmental organisa-
tions, the processes of the ipcc required
governments to sign off on its conclusions,
so reducing their ability to ignore them.
The ipcc’s first assessment of climate- 1
 The Economist April 25th 2020 Climate brief 53

2 change science, published in 1990, predict-
ed that if greenhouse-gas emissions con-
tinued to rise unchecked, the world would
warm by 0.2-0.5°C (0.4-0.9°F) every decade
over the course of the 21st century, and that
sea-level would rise 3-10cm a decade.
Changes in the three decades since fit with
the low end of both predictions.
Two years later, at an “Earth Summit” in
Rio de Janeiro, the un’s members agreed on
a framework convention on climate
change (unfccc) which committed them
to the “stabilisation of greenhouse-gas
concentrations…at a level that would pre-
vent dangerous anthropogenic interfer-
ence with the climate system”.
Despite the fact that such stabilisation
implied impressive cuts in emissions, the
treaty set no targets along the lines of To-
ronto’s 20% by 2005. They were to be
worked out later. In years to come those ne-
gotiations on emission cuts came to domi-
nate discussions between the parties to the
treaty, sidelining the vital question of how
to help countries, especially poor ones,
adapt to the now inevitable changes. To
talk of such adaptation was equated with
capitulating on emission cuts.
Specific emission cuts were agreed
upon five years after Rio, in Kyoto. They
were not global in extent, applying only to
developed countries, which were responsi-
ble for most of the emissions. They were
not ambitious either. And the Kyoto proto-
col was never ratified by America, then the
largest global emitter.
The un imprimatur gave the unfccc
universal legitimacy. But fashioning a
treaty that all could accept had meant pro-
ducing one with little practical power. The
unfccc lacked any mechanism for making
countries commit to ambitious action, let
alone binding them to such commitments.
If all countries had shared an urgent in-
terest in action, those shortcomings would
not have mattered. But they did not. The
costs of environmental improvements
tend to fall on a few groups—typically,
those doing the polluting. In domestic en-
vironmental politics, progress typically re-
lies on going some way to placate those
groups while increasing the enthusiasm
for action among others and the public.
If emissions had been down to just a few
companies, as with cfcs, or sectors of the
economy, as with the smogs tackled by
clean-air acts, such trade-offs might have
been possible internationally. But fossil-
fuel use permeated rich economies. Those
countries knew the cost of reducing them
could be severe—and that the benefits
would accrue mostly to people in other
countries and future times.
These difficulties were exacerbated by
attempts to weaken public support for cli-
mate action. Fossil-fuel companies and
their political allies, understood how im-
damage was to the case for action. The re-
sult was a campaign to make the science
look at best dubious, and at worst fraudu-
lent, which went beyond noting that many
environmental scientists were committed
environmentalists and pointing out truly
open questions (the wide range of the un-
certainties in the first ipcc report has been
slow to narrow). In doing so it helped pro-
duce an environment in which some right-
wing politicians felt able to oppose all cuts
to emissions, with notable successes in
America and Australia.
Future targets beat present action
Another source of resistance to emissions
reduction was the rise of China. Its gdp,
measured at purchasing-power parity and
in real terms, increased sevenfold in the 20
years after Rio. Its carbon-dioxide emis-
sions more than tripled, from 2.7bn to
9.6bn tonnes. China showed no real inter-
est in curbing this world-changing side-
effect, and because it was a developing
country it was not even notionally obliged
to do so by the Kyoto protocol—despite the
fact that, before that protocol was ten years
old, China was a bigger emitter than Ameri-
ca. Resentment over this was one of the
reasons some developed countries became
increasingly unhappy with their commit-
ments. China’s unwillingness to offer real
action contributed to the near collapse of
attempts to move beyond Kyoto at the Co-
→ Changes, fast and slow, in the climate and its politics
Atmospheric CO2, parts per million
0.9 7 8
Strong Eruption
El Niño o Mount
Pinatubo
0.6
→ Temperature anomaly, °C
Relative to 1961-90 average
0.3
0
-0.3
UNFCCC conferences*
199
→ CO2 emissions, gigatonnes Rio
IPCC reports
1 3
1988 Toronto con erence
on the changing atmosphere
197 UN Stockholm 1987 Montreal
con erence protocol negotiated
1970 First 1985 Antarctic ozone
Earth Day hole discovered
7 8
Renewable energy production, GW years 10
penhagen summit of 2009.
Six years after Copenhagen, though, the
un process made its biggest step forward
since Rio: the Paris agreement. This, at last,
set a specific global target. Atmospheric
greenhouse-gas levels were to be stabilised
by the second half of this century at a level
that would see an increase of the average
global temperature over its preindustrial
level well below 2°C, with strenuous efforts
made to keep it down to 1.5°C. All the coun-
tries, developed and developing, that
signed were required to commit to domes-
tic actions towards that aim.
There were several reasons for the suc-
cess: prior talks between America and Chi-
na; skilful French diplomacy; canny nego-
tiation by developing countries. Perhaps
the most important one, though, was that
the cost of renewable energy was tumbling
and investments in the field booming. Re-
ducing emissions while continuing high-
energy lifestyles felt newly possible.
Perhaps it will be. But the reductions the
countries offered in Paris were too small to
meet the 2°C target. That insufficiency has
seen a new generation of climate activists
demand greater ambition at the next big
unfccc meeting, originally to be held this
year in Glasgow but now postponed be-
cause of the covid-19 pandemic. There re-
mains no way for them to force action on
people and countries who do not share
their passion and commitment. 7
7
Strong European Strong El Niño
El Niño heatwave
36.6
35
30
1997 009 015 0 0
Kyoto Copenhagen Paris Glasgow
(post-
poned)
20
4 5
15
10
25 50 100 200 400†

portant a scientific consensus on future Sources: Global Carbon Project; Met Ofce; NOAA; EIA *United Nations Framework Convention on Climate Change †Forecast

Climate brief Science
Computing climate change
Model behaviour
The basic science of the greenhouse effect is quite straightforward. The vital work
of untangling its future effects requires models which are anything but
o imagine earth without greenhouse
T gases in its atmosphere is to turn the fa-
miliar blue marble into a barren lump of
rock and ice on which the average surface
temperature hovers around -18oC. Such a
planet would not receive less of the sun-
light which is the ultimate source of all
Earth’s warmth. But when the energy it ab-
sorbed from the sunlight was re-emitted as
infrared radiation, as the laws of physics
require, it would head unimpeded back out
into space.
Greenhouse gases block that swift exit.
Transparent to incoming sunlight, they ab-
sorb outgoing infrared radiation, thus
warming the atmosphere and, in so doing,
the surface below. The result is an average
surface temperature of some 15oC—warm
enough for open seas and oceans and a vi-
brant biosphere.
In the late 19th century the discovery of
the ice ages led scientists to the conclusion
that climate could change on a global scale.
Svante Arrhenius, a Swedish chemist,
wondered if a weakened greenhouse effect
might be to blame. Carbon dioxide was
known to be a greenhouse gas: Eunice
Foote, an American scientist, had found in
the 1850s that the rate at which a sealed jar
of air warmed up in sunlight depended on
the level of carbon dioxide in that air. So Ar-
rhenius—recently divorced, somewhat
melancholy and in need of a project—be-
gan laboriously to calculate the effects on
In this series
1 The politics o climate action
2 Modelling the greenhouse efect
3 The carbon cycle, present and uture
4 The impacts and their timescales
5 Engineering an energy transition
6 The imperative o adaptation
The Economist May 2nd 2020 55
the climate of halving the atmosphere’s
level of carbon dioxide.
Doing so required him to tackle a pro-
blem of the sort that most frustrates and
most delights scientists who study the
Earth system: a feedback loop through
which a change in one factor affects anoth-
er factor which, in turn, affects the first fac-
tor more.
Because water evaporates more slowly
in cooler climes, the amount of water va-
pour in the atmosphere falls with the tem-
perature. And water vapour, like carbon di-
oxide, is a greenhouse gas. Cooling the
atmosphere dried the atmosphere which
cooled the atmosphere further. Many pen-
cils and thousands of sheets of paper into
his exploration of this, Arrhenius conclud-
ed that halving the carbon-dioxide level
would cool the planet by 5oC (9oF).
He also noted that the same relation
would hold the other way round: double
the carbon dioxide and you would get 5oC
of warming. Industry’s coal burning could
thus warm the world—but only, he
thought, very slowly indeed. He never
imagined that the carbon-dioxide level
would increase by a third in just a century.
Around the same time as Arrhenius was
pondering the climate, a Norwegian scien-
tist called Vilhelm Bjerknes was working
on the physics of how heat drives fluid
flow. His students applied these insights to
large scale flows in the atmosphere and the
oceans, laying the foundations of 20th-
century weather forecasting. In 1950 one of
those students’ students, Ragnar Fjørtoft,
was part of the team which first pro-
grammed a computer to forecast the
weather by solving such equations.
The computer models central to today’s
climate research bring together Arrheni-
us’s curiosity and Bjerknes’s techniques.
Programmes developed from weather-
forecasting software calculate how the lev-
el of carbon-dioxide and other greenhouse
gases is likely to affect the world’s flows of
heat, energy and water, and through them
the future climate. To do so they use com-
puters that can be some 25trn times faster
than the one used in 1950.
These climate models do not treat the
atmosphere as a whole. They divide it into
millions of “cells”. The conditions in each
of these cells depend on the conditions in
its neighbours above, below and to the
sides as well as on its own history. The idea
is to calculate how conditions in each cell
change over time. Unlike a weather fore-
cast, which tries to predict how a specific
state of the atmosphere will evolve over a
few days, these climate models simulate
years, even centuries, of weather in order
to discover the averages and probability
distributions that define the climate—the
envelope which constrains the norms and
extremes of future weather. 1
56 Climate brief
2 Dozens of teams at meteorological and
research organisations around the world
run such models, each using different code
to capture the climate’s underlying mecha-
nisms and study everything from future
peak rainfall to the tracks of storms to
shifts in seasonality. Since 1995 the Cou-
pled Model Intercomparison Project, or
cmip, has brought these teams together by
providing standardised tasks for their
models and then looking at the range of re-
sults. Thus, for example, the 56 different
models considered in the fifth of the cmip
projects, which concluded in 2013, found
that doubling the carbon-dioxide level
would, in time, bring about a warming of
between 1.5 oC and 4.5oC. The uncertainty in
what the models suggest at smaller scales
is greater still. Different models can pro-
vide very different pictures of the future of
regional climates.
Rows and floes of angel hair
The wide range of outcomes is, for the most
part, down to the fact that no two models
represent the mechanisms of the climate—
and particularly its feedbacks—in precise-
ly the same way. Some ways of doing things
can be ruled out because the models they
produce fail to capture the behaviour of the
climate as it is, or as it was in the past (stud-
ies of the low-carbon-dioxide ice ages pro-
vide useful calibration, which would have
pleased Arrhenius). But among models
which reproduce past and current climates
reasonably well, there is no clear way to say
which one’s representations are most reli-
able. The differences between the models
represent a basic level of uncertainty, given
the current state of knowledge.
This endemic uncertainty, though, does
not mean the models have nothing useful
to say. Given how long modelling has been
going on, it is now possible to compare pre-
→ Climate models can guide policy even if they are not precise
Global projected temperature change
Relative to 1850-1900 average, °C
1986-2005 average*
Observed temperature change
7 7
Source: IPCC AR5, adjusted to an 1850-1900 baseline
dictions made decades ago with the way
things have turned out. A study published
last year systematically assessed what
models published between the 1970s and
2007 had said about the way the climate
would respond to steady rises in carbon di-
oxide. It found that for 14 out of 17 models
what had happened had been within the
model’s error bars; of the other three, two
had overshot, one had undershot. Taking
the models seriously would have been a
good bet.
The most important source of uncer-
tainty in the models lies in the clouds. As
greenhouse gases warm the atmosphere its
humidity changes, as does the extent to
which it cools with altitude. These changes
affect how clouds develop; the clouds, in
turn, change surface temperature. Most
clouds warm the world; some cool it.
The problem is that the processes which
control a cloud’s thickness, lifetime and
other qualities work on pretty small scales.
The models do not. Even if every layer of
the atmosphere is represented by hun-
dreds of thousands of grid cells, they still
end up being hundreds of kilometres on a
side—much too large to capture the pro-
cesses responsible for individual clouds.
Not all the feedbacks sit squarely within
the atmosphere; some extend beneath it.
Various feedbacks link the atmosphere to
the oceans, which store, move and release
heat in ways that do a great deal to shape
the climate. In the 1960s modellers began
trying to capture these effects by “cou-
pling” models of the ocean to models of the
atmosphere, so that what they saw in the
atmosphere reflected changes in the
oceans and vice versa.
Feedbacks involving the land matter,
too. Cold weather brings snow; snowy
ground, especially under clear skies, re-
flects away more sunlight, cooling things
Model uncertainty
for each pathway
Future scenarios
High emissions
RCP† .
4
Minimal emissions
RCP† .
3 shows emissions increasing at a spectacu-
RCP† 4.
RCP† .
-
*Uncertainties calibrated to 1986-2005, as shown †Representative Concentration Pathway
The Economist May 2nd 2020
further. Biology adds yet more complexity.
A tropical forest pumps water vapour into
the atmosphere with far greater efficiency
than a savannah does. In warmer oceans it
is harder for nutrients to rise to the surface,
which reduces the ability of plankton to
suck carbon dioxide from the atmosphere.
Melting permafrost produces copious mi-
crobial methane—a gas which absorbs in-
frared much more strongly than carbon di-
oxide does. Over the decades modellers
have attempted to build more and more of
these interrelationships into their models,
adding greatly to their complexity.
Unfortunately increasing complexity
does not always reduce uncertainty. A
model which ignores, say, the instability of
ice sheets—as most did until recently—is
clearly missing something important.
However, because there are always differ-
ent ways to incorporate something new,
two models updated to capture ice-sheet
dynamics may diverge more after this “im-
provement” than they did when, unrealis-
tically, they simply ignored the issue. In the
cmip6 process, which is currently winding
up, preliminary results show a wider range
of uncertainties than was seen in cmip5.
The biggest source of uncertainty,
though, lies not inside the models but out-
side them. Climate change is a problem be-
cause human activity is adding carbon di-
oxide, methane and other greenhouse
gases to the atmosphere at a rate that is
both prodigious and impossible for the
physics, chemistry and biology encoded in
the models to predict.
To estimate how changes in policy
might affect emissions a different family of
models is used—“integrated assessment
models” (iams) which import simplified
results from climate models into models of
the economy.
One of the things that cmip5 asked cli-
mate modellers to look at is the way that
the climate might evolve if emissions fol-
lowed four standardised “pathways” devel-
oped from four particular iams in the
2000s. Three were generated from iams
trying to simulate various types of climate
policy. The fourth, rcp8.5, though often re-
ferred to as “business as usual”, was gener-
ated from an iam run featuring high popu-
lation growth, low technological progress
and very large scale use of coal. As a result it
lar rate, which makes it scary, but not a
helpful baseline.
The uncertainties in what the models
predicted was as striking as ever (see
chart). But they all agreed that only the
pathway embodying the strongest climate
action—much stronger than what is seen
and promised today—might allow the
world to keep the temperature rise since
the 18th century well below 2oC in the 21st,
the target enshrined in the Paris agreement
of 2015. 7

Climate brief Carbon
Carbon cycles
Where nature ends
The scale of human industry’s impact on the planet’s natural flows of carbon and
energy is immense
t is all, in the end, a matter of chemistry.
I Carbon dioxide is a form of what chem-
ists call inorganic carbon—a simple mole-
cule that is pretty inert. Fossil fuels are
made of carbon in its organic form—often
complex molecules that are far from inert.
Combustion turns these organic complex-
ities into inorganic simplicities: carbon di-
oxide, water vapour and heat.
Of the energy that people pay for (as op-
posed to the energy that comes from burn-
ing firewood) 34% comes from burning oil,
27% from coal and 24% from gas. Nuclear
power, hydroelectric power and all other
renewables combined provide just 15%.
The result of all this fossil fuel use is a mod-
ern industrial economy and an annual flow
of 9.5bn tonnes of carbon out of the ground
and into the atmosphere.
Through its effects on the plants, ani-
mals and microbes which make up the bio-
sphere, on the climate and on the oceans,
this industrial flow of carbon links the
Earth’s distant geological past to its future
over millennia to come. It is the single
clearest piece of evidence for the idea that
humans now have a power over the Earth as
great as the forces of nature, and that their
use of this power has opened up a new geo-
logical epoch that some scientists call the
Anthropocene.
To appreciate the importance of this in-
dustrial carbon flow, you have to under-
stand the carbon cycle in which it sits. At
In this series
1 The politics o climate action
2 Modelling the greenhouse efect
3 The carbon cycle, present and uture
4 The impacts and their timescales
5 Engineering an energy transition
6 The imperative o adaptation
The Economist May 9th 2020 47
first, this context seems reassuring. Al-
most all microbes, and all animals, get the
energy that they need for life from breaking
up food made of organic molecules. The
flame-free, internalised form of combus-
tion by which they do so, which biologists
call respiration, produces much more car-
bon dioxide than industry does.
But respiration has a counterpart: pho-
tosynthesis, through which plants, algae
and some bacteria use sunlight to turn in-
organic carbon back into organic mole-
cules. These new molecules are the raw
material from which almost all living
things on Earth are made; the sunlight
stored within them is the source of all the
energy that is released through respiration
when those living things are eaten.
The other great flow of carbon dioxide
into the atmosphere is similarly balanced.
Carbon dioxide dissolved in seawater natu-
rally diffuses into the air above. Carbon di-
oxide in the atmosphere dissolves into sea-
water. Left to themselves, the two flows
balance (see diagram on next page).
These flows create a system in what is
called dynamic equilibrium; if you push it
away from current conditions, it pulls it-
self back. If atmospheric carbon-dioxide
levels go up, the rate at which carbon diox-
ide dissolves into the “sinks” provided by
the oceans and plants will also, all things
being equal, go up. This reduces the sur-
plus, restoring the status quo. Until the
19th century this dynamic equilibrium had
kept atmospheric carbon-dioxide levels
pretty stable for most of the 10,000 years
since the end of the most recent ice age.
The plants-and-food branch of the car-
bon cycle, though, is not quite perfect. Like
the little bit left in the corner of the sardine
can that you can’t get out, not all the organ-
ic matter made through photosynthesis
gets used by creatures that respire. Some
ends up buried in sediments instead.
The amount of carbon which leaks out
of the biosphere this way is tiny compared
with the flow returned to the atmosphere.
But the leak has gone unstopped for a very
long time, and that has allowed the Earth’s
crust to build up a significant store of or-
ganic matter. Now human industry’s use of
the most concentrated and readily avail-
able deposits of these fossil fuels has re-
turned to the carbon cycle in a couple of
centuries a fair fraction of what was
stashed away over hundreds of millions of
years. It is the addition of this new source
with no new sink that has knocked the cy-
cle out of whack.
The world’s seas and plants have tried
their best to keep things in equilibrium, re-
sponding to rising levels of carbon dioxide
by stashing more away in the biosphere
and oceans. They suck up roughly half of all
the extra carbon dioxide that industry puts
into the atmosphere. But that is as much as 1
48 Climate brief
2 they can do. And so the amount in the at-
mosphere grows.
This intensification of the carbon cycle
has side-effects. Plants fed with extra car-
bon dioxide tend to grow more, if circum-
stances allow. Current estimates suggest
the global rate of photosynthesis is 3-7%
higher than it was 30 years ago; satellite im-
ages show the Earth is getting greener.
Such “carbon-dioxide fertilisation” has im-
proved the yields of some crops, and the
growth of some forests and other ecosys-
tems. This is not enough to compensate for
the damage climate change does to agricul-
ture by higher temperatures and altered
rainfall. But, on balance, it is hard to see it
as much of a problem.
The same cannot be said of the in-
creased flow into the ocean sink. More dis-
solved carbon dioxide makes seawater
more acidic. How bad this acidification
will prove is open to debate. But the process
will probably be very damaging to some
ecosystems, including reefs already
stressed by rising temperatures. Even if
fossil-fuel use were not warming the cli-
mate, this acidification would in itself
count as a frightening global change.
The growth of the two carbon sinks is
also, left to itself, unsustainable. Warm wa-
ter absorbs less carbon dioxide than cold
water. So as the oceans warm their ability to
offset emissions weakens. As to the land
sink, higher temperatures speed up micro-
bial respiration, especially in soils, more
reliably than higher carbon-dioxide levels
speed up photosynthesis.
The Paris agreement of 2015 calls for in-
creases to the atmosphere’s carbon-diox-
ide level caused by fossil fuels to end by the
second half of this century. Even if that
deadline is not met, some mixture of poli-
cy, catastrophe and/or resource depletion
will eventually bring the rise to an end. The
flows of carbon between the atmosphere,
oceans and biosphere will then come back
into balance.
But the equilibrium thus restored will
→ Humans have unbalanced the carbon cycle; could they fix it?
Total carbon stocks and annual flows, gigatonnes of carbon
Past Pre-industrial
Biosphere Atmosphere
109 61 Oceans
600 600
107 60
38,000
Soils
2,000
Geosphere
Sources: California
(solid earth)
Polytechnic; IPCC 12.5m
not be the pre-industrial one. The carbon-
dioxide level will settle down not far short
of whatever the 21st century’s peak level
turns out to be. Which means that tempera-
tures will stay high, too—with all that en-
tails for crops, ice caps and the like.
This plateau will eventually subside.
The erosion of the Earth’s crust exposes sil-
icate minerals that react with carbon diox-
ide, eventually producing solid carbonate
minerals from which the carbon cannot
readily escape. But this “chemical weather-
ing” works on a much longer timescale
than the sinks. Geochemists think it would
take 1,000 years for a post-fossil-fuel car-
bon-dioxide level of around 550 parts per
million to be brought back below today’s
415ppm towards a mid-20th century level
of 315ppm.
Going backwards
What, though, if the Anthropocene transi-
tioned from a past dominated by anthropo-
genic carbon sources to a future character-
ised by anthropogenic sinks? There are two
reasons why this might be appealing. One
is that some fossil-fuel emissions may be
very hard to eliminate from the economy. If
they could be counterbalanced by “nega-
tive emissions” that take carbon dioxide
out of the atmosphere at a similar rate, the
Paris goal of stopping any further increase
to the carbon-dioxide level would be far
easier to meet.
The second attraction of the idea stems
from the other Paris goal, that of keeping
the global temperature increase, compared
to pre-industrial times, well below 2°C. Do-
ing this simply by reducing emissions
would require much steeper cuts than any
seen to date, and they would have to con-
tinue for decades. If the world developed
negative-emission technologies, more
gentle emissions cuts in the near future
could be made up for by negative emis-
sions later on, which would bring the car-
bon-dioxide level back down from its ex-
cessive peak.
Now Early Anthropocene
Biosphere Atmosphere
123 79
600 600
117 80
240
Soils 9.5
2,000
Extracted fossil
carbon
-365
Geosphere
12.5m
The Economist May 9th 2020
Some forms of negative emission look
fairly benign: farming in ways that make
the soil richer in organic carbon; restoring
degraded forests and planting new ones.
More ambitious is the idea of harnessing
photosynthesis to industry; growing plan-
tation crops, burning them to generate
electricity and sequestering the carbon di-
oxide given off underground, rather than
letting it out into the atmosphere, an ap-
proach called bioenergy with carbon cap-
ture and storage, or beccs.
Then there is the idea of stripping car-
bon dioxide out of the atmosphere with re-
newably powered open-air chemical engi-
neering: “direct air capture”, or dac. And
there is also the possibility of helping
along the chemical weathering process by
grinding up silicate rocks into fine dusts,
thus speeding up the reactions that store
carbon dioxide away in stable minerals.
There are two big problems with these
ideas. One is the scale at which they need to
operate to make a difference. Imagine that
in 2060 the world had, through a vast ef-
fort, renounced 90% of its fossil-fuel use.
To offset the remaining recalcitrant 10%
would still require a sink capable of soak-
ing up about 1bn tonnes of carbon a year.
The industrial systems for taking carbon
dioxide from the air currently on the draw-
ing board operate at barely a thousandth of
that scale. Creating such a flow through
photosynthesis would require a plantation
about the size of Mexico.
This leads to the second problem. Imag-
inary backstops are dangerous. If countries
build negative emissions into their think-
ing, they will cut emissions more slowly on
the basis that any overshoot can be
mopped up later. But they will not neces-
sarily undertake the huge efforts required
to make those negative emissions a reality.
The Anthropocene fact that humans are
now integral to the processes of the planet
does not mean that they can change those
processes without great effort—let alone
just through wishful thinking. 7
Negative emissions: 1 enhanced weathering
2 reforestation 3 soil improvement 4 BECCS* 5 DAC*
Net-zero future Later Anthropocene
Biosphere Atmosphere
1
2
Oceans
Oceans
38,000
3
Soils ?
4 5
Geosphere
*Explained in text

Climate brief Impacts
From seconds to centuries
Bad times
The damage done by climate change will be severe, widespread
and sometimes surprising
n november 21st 2016, a line of thun-
O derstorms passed through the Austra-
lian state of Victoria. By the end of the fol-
lowing day, it had sent 3,000 people to
hospital. Storms typically hurt people by
blowing down buildings, flooding streets
or setting fires. In this case, though, the ca-
sualties were caused by asthma. Late that
afternoon a peculiarly powerful downdraft
generated by the storm front pushed a layer
of cold air thick with pollen, dust and other
particles through Melbourne. The city’s
ambulance service was swamped within
hours. At least ten people died.
The risks that weather and climate pose
to human life are not always as specific to
the peculiar circumstances of time and
place as that sudden-onset asthma epi-
demic. But they are complex functions of
what, where and who, and their mecha-
nisms are not always easily discerned.
What is more, they can interact with each
other. For example, if the southern spring
of 2016 had not brought weather particular-
ly well suited to the growth of allergenic
grasses, would that stormy afternoon have
been so catastrophic? Such complexities
mean that a gradual change to the climate
can lead to sudden changes in the impacts
on human beings when things pass a cer-
tain threshold. And that threshold will not
necessarily be discernible in advance.
Not all the ways in which today’s weath-
er harms people will be exacerbated by cli-
In this series
1 The politics o climate action
2 Modelling the greenhouse efect
3 The carbon cycle, present and uture
4 The impacts and their timescales
5 Engineering an energy transition
6 The imperative o adaptation
The Economist May 16th 2020 49
mate change. But research suggests that
many of them will. Most of the problems
people have with weather and climate
come from extremes. When means shift a
little, extremes can shift a lot (see chart
overleaf). Today’s rare extremes become to-
morrow’s regular disturbances; tomor-
row’s extremes are completely new.
How damaging these impacts will be to
the economic and physical welfare of hu-
mankind depends on how much warming
takes place and how well people adapt—
both of which are currently unknowable.
But it is possible to get a qualitative sense
of what they could mean by looking at the
range of timescales over which they oper-
ate. At one end, a thunderstorm’s pollen
surge, sweeping by in minutes; at the other,
sea-level rise which could last longer than
any civilisation in human history.
In terms of short-lived events, the worst
sort of bad day that the world’s weather can
offer is generally taken to be the one on
which you get hit by a tropical cyclone,
which is why hurricanes (as they are
known in the Atlantic) and typhoons (as
they are known in some other places) have
become so heated a part of the arguments
about climate change. A single hurricane
can do more than $100bn in damage, as
Harvey did when it hit Houston in August
2017, or kill thousands, as Maria did the fol-
lowing month in Puerto Rico.
Tropical cyclones can only form over a
sea or ocean with a surface temperature of
27°C or more. The area where such tem-
peratures are possible will definitely in-
crease with warming. But that does not
mean hurricanes will become more com-
mon. Their formation also requires that
the wind be blowing at a similar speed
close to the surface and at greater alti-
tudes—and this condition, models say, will
become less common in future over many
of the places where hurricanes spawn.
Thus models do not predict a great increase
in the number of tropical cyclones; Atlan-
tic hurricanes may well become more rare.
But more heat in the oceans means that
those tropical cyclones which do get going
are more likely to become intense. There is
thus broad agreement among experts that
the proportion of hurricanes which reach
category four or five looks set to increase.
So, too, does the rainfall associated with
them, because warmer air holds more
moisture. Studies of the flooding caused by
Hurricane Harvey suggest that warming
due to climate change increased its rainfall
by about 15%. Extreme rainfall events of
many sorts increase in warmer worlds.
The heat which powers hurricanes at
sea can, on land, kill directly. Humans cool
themselves by sweating, a process that be-
comes less effective the more humid the at-
mosphere. Combining the heat and the hu-
midity into something called the wet-bulb 1
50 Climate brief
2 temperature (wbt) allows scientists to
measure temperatures in a way that re-
flects that difficulty (similar measures in
America are called the heat index). wbts of
35°C and above are lethal.
Until recently it was thought that wbts
that high would not be seen until warming
had continued for decades. A review of
weather-station data from 1979 on, how-
ever, shows that for very brief periods local
wbts almost that high are already being ex-
perienced occasionally in South-East Asia,
the Persian Gulf and the coastal south-west
of America, and that their frequency had
doubled since 1979. With 2.5°C (4.5°F) of
global warming above pre-industrial lev-
els, which is quite possible in the second
half of this century if action on emissions
is not significantly increased, these unlive-
able conditions will become a regular oc-
curence in parts of the humid subtropics.
Another recent study defines climates
which people find liveable according to
where, historically, they have lived, and
then sees which such areas move beyond
those climatic bounds as the world warms.
Temperature rises quite plausible by 2070
would see many areas where people live to-
day develop climates unlike any that peo-
ple have lived in before (see map). Some
econometric analyses based on inter-
annual differences suggests that, in gen-
eral, higher temperatures lead to lower la-
bour productivity and more violence.
→ A hotter planet is a more extreme one; some regions become unliveable
Effect of changes in global temperature
Increase in mean Increase in variance
→
Fewer cold More hot More cold
extremes extremes extremes
↙ ↘
Cold Average Hot Cold Average
Projected change in suitability* for human habitation in 2070
With warming of between 2°C and 2.5°C above pre-industrial levels
Sources: IPCC AR5; “Future of the human climate niche”, by Chi Xu et al., 2020
In the nearer term, there is an increased
likelihood of heatwaves. Between August
3rd and 16th 2003, Europe saw 39,000 more
deaths than would have been expected on
the basis of previous years. The excess mor-
tality was due to a summer that was hotter,
by some estimates, than any for the previ-
ous 500 years. Modelling suggests that,
even in 2003, climate change had made
such a heatwave at least twice as likely.
Extreme heatwaves are becoming more
frequent not only because temperatures
are climbing. Warming-induced changes
in the climate system can weaken the pro-
cesses that normally move weather around
the world, allowing conditions to get stuck.
Such stalling can be the difference between
a hot week and a lethal month, or in winter
a cold snap and a deep freeze.
Springtime and harvest
Hot summers can also harm crops, both di-
rectly—many important crops are very
sensitive to temperatures above a certain
threshold—and through water stress. Mil-
der winters can also do harm by allowing
pests to survive, hurting yields.
When unusually hot and dry conditions
suck the moisture off the land, the subse-
quent droughts do not just exacerbate the
problems for farmers. They also increase
the risk and severity of fires—which an in-
crease in the amount of lightning will, in
some regions, spark off more frequently.
Increase in mean and variance
→ The big unknown, though, once you get
More hot Less impact on
extremes cold extremes
↙ ↘
Hot Cold Average
More suitable/
no change
*Based on temperature and precipitation levels
The Economist May 16th 2020
This is an issue not just in warm, fire-prone
places such as Australia. For several
months in the summer of 2019, large
swathes of northern Russian and Canadian
forest—and even some of Greenland’s few
woodlands—went up in flames.
Unusual infernos have plagued Califor-
nia for many years now, again as a result of
parched conditions, which are drying out
rivers, lakes and underground aquifers
across the entire south-west of the state.
This is no regular drought. It is 19 years in
the making, enough for it to be classed as a
“megadrought”.
Tree-ring records show only four such
in the region over the past 1,200 years, and
suggest that this could be as bad as the
worst of them, which took place in the 17th
century. Such droughts are linked to chang-
ing patterns of circulation in the ocean.
Models suggest that such patterns are
themselves altered by warming, which can
thus change the frequency of other large-
scale regional shifts in the climate.
And then there is the longest term
change: sea level. The sea’s rise comes from
three different mechanisms—the expan-
sion of the oceans as they absorb more
heat, the addition of meltwater from
shrinking glaciers on land, and the physi-
cal break down of ice sheets such as those
on Antarctica and Greenland. The first two
factors are currently driving an increase of
about 1cm every three years, and are set to
do so at a similar rate well into the 21st cen-
tury even if global warming is held well be-
low 2°C; the time it takes seawater to warm
up gives the process a significant inertia.
Such rises will erode coasts and increase
flooding—especially when pushed inland
by the surges intense storms produce.
More hot
extremes
to the century time scale, is the stability of
the great ice sheets. It is widely believed
that there are points of no return after
which such sheets are doomed slowly to
Hot
collapse, thus increasing sea levels by
many metres. Where these points of no re-
turn are is not clear. It is possible that they
Less suitable might be passed even if warming is kept to
1.5°C above the pre-industrial.
Outside historical A high likelihood of drought and crop
human niche failures; changes to regional climate that
upset whole economies; storms more de-
structive in both their winds and their
rains; seawater submerging beaches and
infiltrating aquifers: what is known about
the impacts of climate change is already
worrying enough. The known unknowns
add to the anxiety. It is not just the question
of the ice sheets, an uncertainty massive
enough to weigh down a continent. There
are other tipping points, too, which could
see ocean currents shift, or deserts spread.
And in the spaces between all these trou-
bles are the unknown unknowns, as sur-
prising, and deadly, as a thunderstorm that
kills through pollen. 7

Climate brief Energy
Once more, with renewables
Not-so-slow burn
The world’s energy system has been transformed before—but never as quickly or
completely as must happen now
or more than 100,000 years humans
F derived all their energy from what they
hunted, gathered and grazed on or grew for
themselves. Their own energy for moving
things came from what they ate. Energy for
light and heat came from burning the rest.
In recent millennia they added energy
from the flow of water and, later, air to the
repertoire. But, important as water- and
windmills were, they did little to change
the overall energy picture. Global energy
use broadly tracked the size of a population
fed by farms and warmed by wood.
The combination of fossil fuels and ma-
chinery changed everything. According to
calculations by Vaclav Smil, a scholar of en-
ergy systems at the University of Manitoba,
between 1850 and 2000 the human world’s
energy use increased by a factor of 15 or so.
The expansion was not homogeneous;
over its course the mixture of fossil fuels
used changed quite dramatically. These are
the monumental shifts historians call “en-
ergy transitions”. They require huge
amounts of infrastructure; they change the
way the economy works; and they take
place quite slowly.
James Watt patented his steam engine
in 1769; coal did not exceed the share of to-
tal energy provided by “traditional bio-
mass”—wood, peat, dung and the like—un-
til the 1900s (see chart overleaf). It was not
until the 1950s, a century after the first
commercial oil well was drilled in Titus-
In this series
1 The politics o climate action
2 Modelling the greenhouse efect
3 The carbon cycle, present and uture
4 The impacts and their timescales
5 Engineering an energy transition
6 The imperative o adaptation
The Economist May 23rd 2020 49
ville, Pennsylvania, that crude oil came to
represent 25% of humankind’s total prim-
ary energy. Energy transitions were slow
largely because the growth in total energy
use was fast. In the century it took oil to
capture a quarter of the total, that total in-
creased. They are also always incomplete.
New fuels may reduce the share of the pie
that old fuels control, but they rarely re-
duce the total energy those fuels supply.
Much more “traditional biomass” is
burned by the world’s poor today than was
burned by the whole world in 1900.
To give the world a good chance of keep-
ing global warming, measured against the
temperature pre-coal, well below 2°C
(3.6°F) will require an energy transition far
larger and quicker than any before it. In the
next 30-50 years 90% or more of the share
of the world’s energy now being produced
from fossil fuels will need to be provided by
renewable-energy sources, nuclear power
or fossil-fuel plants that bury their waste
rather than exhaling it.
During this time, the pie will keep grow-
ing—but not necessarily as fast as it used
to. The direct relationship between gdp
and energy use, which held tight until the
1970s, has weakened over the past half cen-
tury. It is possible for growth per person to
continue without energy use per person in-
creasing. Though the population is no lon-
ger growing as fast as it did at the 20th-cen-
tury peak of its increase, it will still be the
best part of 2bn higher by mid-century. And
all those people should be able to aspire to
modern energy services. Today more than
800m people still lack electricity—hence
all that burning of traditional biomass.
The good news, however, is that govern-
ments say they are willing to push through
the change. Previous transitions, though
shaped by government policy at national
levels, were mostly caused by the demand
for new services that only a specific fuel
could provide, such as petrol for engines.
The growth in renewable-generation
capacity is the exception. It has not been
driven by the fact that renewable electrons
allow you to do things of which those from
coal are not capable. It has largely been dri-
ven by government policy. This has not al-
ways had the near-term effects for which
such policy should aim. Germany’s roll-out
of renewables has been offset by its retreat
from nuclear, and its emissions have risen.
But subsidies there and elsewhere have
boosted supply chains and lowered the
cost of renewable technologies.
During the 2010s the levelised cost (that
is the average lifetime cost of equipment,
per megawatt hour of electricity generated)
of solar, offshore wind and onshore wind
fell by 87%, 62% and 56%, respectively, ac-
cording to Bloombergnef, an energy-data
outfit (see chart overleaf). This has allowed
deployments that were unthinkable in the 1
50 Climate brief
2 2000s. Britain now has 10,000 offshore
wind turbines. They are built by developers
chosen based on how low a price they are
willing to take for their electricity (the gov-
ernment pledges to make the cost up if the
market price falls below it).
In 2015 winning bids were well over
£100 ($123) per mwh, far higher than the
cost of fossil-fuel electricity. Thanks to pre-
dictable policy, fierce competition and
technical progress, a recent auction
brought a bid as low as £39.65 per mwh,
roughly the level of average wholesale
power prices. Solar and onshore wind are
even less expensive. About two-thirds of
the world’s population live in countries
where renewables represent the cheapest
source of new power generation, says
Bloombergnef.
Solar power is the really spectacular
achiever, outstripping the expectations of
its most fervent boosters. Ramez Naam, a
bullish solar investor, recently recalibrated
his expectations to foresee a future of “in-
sanely cheap” solar power. By 2030, he
reckons, in sunny parts of the world, build-
ing large new solar installations from
scratch will be a cheaper way of getting
electricity than operating fully depreciated
fossil-fuel plants, let alone building new
ones. Michael Liebreich, a consultant on
renewable energies, speculates about a “re-
newable singularity” in which cheap re-
newable electricity opens up new markets
that demand new capacity which makes
electricity cheaper still.
Even without such speculative won-
ders, the effect of renewables is apprecia-
ble. Together with natural gas, which
America’s fracking revolution has made
cheaper, solar and wind are already squeez-
ing coal, the energy sector’s biggest emitter
(a megawatt of coal produces a stream of
emissions twice the size of that given off by
a megawatt of gas). In 2018 coal’s share of
global energy supply fell to 27%, the lowest
in 15 years. The pressure that they can apply
→ The phenomenal growth in energy use brought about by fossil fuels will be hard to decarbonise
Global energy consumption, by source, terawatts
Log scale
Fossil uels
Biomass
. Ofshore wind
Human labour
. Coal
Animal labour
. Coal
7 5 5 5
Sources: Vaclav Smil; BP Statistical Review of World Energy; BloombergNEF
to oil is not yet as great, because oil mostly
drives cars, and electric cars are still rare.
But as that changes, renewables will come
for oil, as they are already coming for gas.
There are stumbling blocks. Neither the
sun nor the wind produces energy consis-
tently. Germany’s solar-power installa-
tions produce five times more electricity in
the summer than they do in the winter,
when demand for electricity is at its peak.
Wind strengths vary not just from day to
day but from season to season and, to some
extent, year to year. This amounts to a
speed bump for renewables, not a block-
ade. Long transmission lines that keep
losses low by working at very high voltages
can move electricity from oversupplied ar-
eas to those where demand is surging. Lith-
ium-ion batteries can store extra energy
and release it as needed. The economic
stimulus China announced in March in-
cludes both ultra-high-voltage grids and
electric-vehicle-charging infrastructure.
Thou orb aloft, somewhat dazzling
As the sun and wind account for a larger
share of power, renewables might store
power by splitting water to create hydrogen
to be burned later. More ambitiously, if
technologies for pulling carbon dioxide
back out of the air improve, such hydrogen
could be combined with that scavenged
carbon to make fossil-free fuels.
In doing so, they might help remedy the
other problem with renewables. There are
some emissions which even very cheap
electricity cannot replace. Lithium-ion
batteries are too bulky to power big planes
on long flights, which is where artificial fu-
els might come in. Some industrial pro-
cesses, such as cement-making, give out
carbon dioxide by their very nature. They
may require technology that intercepts the
carbon dioxide before it gets into the atmo-
sphere and squirrels it away underground.
When emissions cannot be avoided—as
may be the case with some of those from
Global energy consumption, by source, % of total
Solar Nuclear Other
Hydropower Wind renewables
Traditional
biomass
5 5
*Average of fixed and tracking systems †Estimated using battery-pack prices before 2018 ‡Combined Cycle Gas Turbines
The Economist May 23rd 2020
farmland—they will need to be offset by re-
moving carbon dioxide from the atmo-
sphere either with trees or technology.
None of this happens, though, without
investment. The International Renewable
Energy Agency, an advisory group, esti-
mates that $800bn of investment in renew-
ables is needed each year until 2050 for the
world to be on course for less than 2°C of
warming, with more than twice that need-
ed for electric infrastructure and efficien-
cy. In 2019 investment in renewables was
$250bn. The big oil and gas firms invested
twice as much in fossil-fuel extraction.  
If governments want to limit climate
change, therefore, they must do more. They
do not have to do everything. If policy
choices show that the road away from fos-
sil fuels is right, private capital will follow.
Investors are already wary of fossil-fuel
companies, eyeing meagre returns and the
possibility that action on climate change
will leave firms with depreciating assets.
But governments need to make the sig-
nals clear. Around the world, they cur-
rently provide more than $400bn a year in
direct support for fossil-fuel consumption,
more than twice what they spend subsidis-
ing renewable production. A price on car-
bon, which hastens the day when new re-
newables are cheaper than old fossil-fuel
plants, is another crucial step. So is re-
search spending aimed at those emissions
which are hard to electrify away. Govern-
ments have played a large role in the devel-
opment of solar panels, wind turbines and
fracking. There is a lot more to do.
However much they do, though, and
however well they do it, they will not stop
the climate change at today’s temperature
of 1°C above the pre-industrial. Indeed,
they will need to expand their efforts great-
ly to meet the 2°C target; on today’s poli-
cies, the rise by the end of the century looks
closer to 3°C. This means that as well as try-
ing to limit climate change, the world also
needs to learn how to adapt to it. 7
Levelised cost of energy, $ per MWh
2019 prices
5
Battery
storage†
4
Gas Solar*
6 3
Oil
4
Onshore wind
CCGT‡
4 6