Opinion|What’s the Least Bad Way to Cool the Planet?
By David Keith
Dr. Keith is a professor of applied physics and of public policy at
Harvard, where he led the development of the university’s solar
engineering research program. He is also a co-host of the podcast
“Energy vs Climate” and the founder and a board member of the company
Carbon Engineering, which provides technology to capture carbon dioxide
from the atmosphere.
How to cool the planet?
The energy infrastructure that powers our civilization must be rebuilt,
replacing fossil fuels with carbon-free sources such as solar or
nuclear. But even then, zeroing out emissions will not cool the planet.
This is a direct consequence of the single most important fact about
climate change: Warming is proportional to the cumulative emissions
over the industrial era.
Eliminating emissions by about 2050 is a difficult but achievable goal.
Suppose it is met. Average temperatures will stop increasing when
emissions stop, but cooling will take thousands of years as greenhouse
gases slowly dissipate from the atmosphere. Because the world will be a
lot hotter by the time emissions reach zero, heat waves and storms will
be worse than they are today. And while the heat will stop getting
worse, sea level will continue to rise for centuries as polar ice melts
in a warmer world. This July was the hottest month ever recorded,
but it is likely to be one of the coolest Julys for centuries after
emissions reach zero.
Stopping emissions stops making the climate worse. But repairing the
damage, insofar as repair is possible, will require more than emissions
To cool the planet in this century, humans must either remove carbon
from the air or use solar geoengineering, a temporary measure that may
reduce peak temperatures, extreme storms and other climatic changes.
Humans might make the planet Earth more reflective by adding tiny
sulfuric acid droplets to the stratosphere from aircraft, whitening
low-level clouds over the ocean by spraying sea salt into the air or by
Yes, this is what it comes down to: carbon removal or solar
geoengineering or both. At least one of them is required to cool the
planet this century. There are no other options.
Carbon removal would no doubt trounce geoengineering in a straw poll of
climate experts. Removal is riding a wave of support among centrist
environmental groups, governments and industry. Solar geoengineering is
seen as such a desperate gamble that it was dropped from the
important “summary for policymakers” in the United Nations’ latest
Yet if I were asked which method could cut midcentury temperatures with
the least environmental risk, I would say geoengineering.
Lest you dismiss me, I founded Carbon Engineering, one of the most
visible companies developing technology to capture carbon directly from
the air and then pump it underground or use it to make products that
contain carbon dioxide. The company’s interests could be hurt if
geoengineering were seen as an acceptable option. I was also an early
proponent for burning biofuels like wood waste, capturing the resulting
carbon at the smokestack and storing it underground. I am proud to be a
part of the community developing carbon removal. These approaches can
help manage hard-to-abate emissions, and they are the only way to
reduce the long-term climate risks that will remain when net emissions
But the problem with these carbon removal technologies is that they are
inherently slow because the carbon that has accumulated in the
atmosphere since the Industrial Revolution must be removed ton by ton.
Still, the technology provides a long-term cure.
Geoengineering, on the other hand, is cheap and acts fast, but it
cannot deflate the carbon bubble. It is a Band-Aid, not a cure.
The trade-off between geoengineering and carbon removal depends on
one’s time horizon. The sooner cooling is pursued, the greater the
environmental and social impacts of carbon removal.
Suppose emissions were under control and you wanted to cool the planet
an additional degree by midcentury. How would removal and
Carbon removal could work. But it will require an enormous industry.
Trees are touted as a natural climate solution, and there are some
opportunities to protect natural systems while capturing carbon by
allowing deforested landscapes to regrow and pull in carbon dioxide as
they do. But cooling this fast cannot be achieved by letting nature run
free. Ecosystems would need to be manipulated using irrigation, fire
suppression or genetically modified plants whose roots are resistant to
rot. This helps to increase the buildup of carbon in soils. To cool a
degree by midcentury, this ecological engineering would need to happen
at a scale comparable to that of global agriculture or forestry,
causing profound disruption of natural ecosystems and the
too-often-marginalized people who depend on them.
Industrial removal methods have a much smaller land footprint; a single
carbon capture facility occupying a square mile of land could remove a
million tons of carbon from the air a year. But building and running
this equipment would require energy, steel and cement from a global
supply chain. And removing the few hundred billion tons required to
cool a degree by midcentury requires a supply chain that might be
smaller than what feeds the construction industry but larger than what
supports the global mining industry.
The challenge is that a carbon removal operation — industrial or
biological — achieves nothing the day it starts, but only cumulatively,
year upon year. So, the faster one seeks that one degree of cooling,
the faster one must build the removal industry, and the higher the
social costs and environmental impacts per degree of cooling.
Geoengineering could also work. The physical scale of intervention is —
in some respects — small. Less than two million tons of sulfur per year
injected into the stratosphere from a fleet of about a hundred
high-flying aircraft would reflect away sunlight and cool the planet by
a degree. The sulfur falls out of the stratosphere in about two years,
so cooling is inherently short term and could be adjusted based on
political decisions about risk and benefit.
Adding two million tons of sulfur to the atmosphere sounds reckless,
yet this is only about one-twentieth of the annual sulfur pollution
from today’s fossil fuels. Geoengineering might worsen air pollution or
damage the global ozone layer, and it will certainly exacerbate some
climate changes, making some regions wetter or drier even as it cools
the world. While limited, the science so far suggests that the harms
that would result from shaving a degree off global temperatures would
be small compared with the benefits. Air pollution deaths from the
added sulfur in the air would be more than offset by declines in the
number of deaths from extreme heat, which would be 10 to 100 times
Geoengineering’s grand challenge is geopolitical: Which country or
countries get to decide to inject aerosols into the atmosphere, on what
scale and for how long? There is no easy path to a stable and
legitimate governance process for a cheap, high-leverage technology in
an unstable world.
Which is better? Carbon removal is doubtless the safest path to
permanent cooling, but solar geoengineering may well be able to cool
the world this century with fewer environmental impacts and less social
and economic disruption. Yet no one knows, because the question is not
being asked. Geoengineering research budgets are minuscule, and much of
the work is accomplished after hours by scientists acting outside their
The United Nations Intergovernmental Panel on Climate Change assumes
enormous use of carbon removal to meet the Paris Agreement target of
1.5 degrees Celsius (2.7 degrees Fahrenheit), but not because
scientists carefully compared removal and geoengineering. This was a
glaring omission in the I.P.C.C. report, given that one of the very
few areas of agreement about geoengineering is that it could lower
Research is minimal because geoengineering has influential opponents.
The strongest opposition to geoengineering research stems from fear
that the technology will be exploited by the powerful to maintain the
status quo. Why cut emissions if we can seed the atmosphere with sulfur
and keep the planet cool? This is geoengineering’s moral hazard.
This threat is real, but I don’t find it a convincing basis to forgo
research, particularly given evidence that support for
geoengineering research is stronger in regions that are poorer and more
vulnerable to climate change, regions that would benefit most from
Some will no doubt exaggerate the benefits of solar geoengineering to
protect the fossil fuel industry. But this threat is not unique to
geoengineering. Carbon removal may pose a stronger moral hazard today.
Activists like Al Gore once opposed adaptive measures such as flood
protection, out of fear it would distract from emission cuts. They now
embrace such measures, yet support for emissions cuts has never been
higher, proving that support for one method of limiting climate risks
need not reduce support for others.
Emissions cuts are necessary. But pretending that climate change can be
solved with emissions cuts alone is a dangerous fantasy. If you want to
reduce risks from the emissions already in the atmosphere — whether
that’s to prevent forest fires in Algeria, heat waves in British
Columbia or floods in Germany — you must look to carbon removal, solar
geoengineering and local adaptation.
Emissions monomania is not an ethical climate policy because those
three approaches together do what emissions cuts cannot: They reduce
the future harms caused by historical emissions and provide a reason to
hope that collective action can begin repairing Earth’s climate within
a human lifetime.
Perhaps the best reason to take cooling seriously is that benefits seem
likely to go to the poorest countries. Heat reduces intellectual
and physical productivity with economywide consequences. Hotter regions
are more sensitive to extra degrees of warming, while some cool regions
may even benefit. A year that’s a degree warmer than normal will see
economic growth in India reduced by about 17 percent, while Sweden
will see growth increased by about 22 percent.
Poor people tend to live in hot places. This, combined with the fact
that an added degree causes more harm in warmer climates, explains why
the costs of climate change fall heaviest on the poor — and why the
benefits of cooling will be felt the most in the hottest regions.
This dynamic explains why the one study to quantitatively examine
the consequences of geoengineering for global inequality found that it
might reduce economic inequality by about 25 percent, similar to the
impressive reduction the United States achieved in the four decades
following the New Deal.
Cooling the planet to reduce human suffering in this century will
require carbon removal or solar geoengineering or both. The trade-offs
between them are uncertain because little comparative research has been
done. The fact that one or both are taboo in some green circles is a
dreadful misstep of contemporary environmentalism. Climate justice
demands fast action to cut emissions and serious exploration of
pathways to a cooler future.
David Keith is a professor of applied physics and of public policy
at Harvard, where he led the development of the university’s solar
engineering research program. He is also a co-host of the podcast
“Energy vs Climate” and the founder and a board member of the
company Carbon Engineering, which provides technology to capture
carbon dioxide from the atmosphere.
The Times is committed to publishing a diversity of letters to the
editor. We’d like to hear what you think about this or any of our
articles. Here are some tips. And here’s our email:
Follow The New York Times Opinion section on Facebook, Twitter
(@NYTopinion) and Instagram.