Wikipedia Deep Dive

Climate sensitivity

14 min read

The Number That Determines Our Future

There's a single number that climate scientists have been hunting for over a century. Get it wrong, and we might spend trillions on problems that aren't as severe as we feared. Get it wrong in the other direction, and we sleepwalk into catastrophe while thinking we have decades to spare.

That number is climate sensitivity.

At its core, climate sensitivity answers a deceptively simple question: if we double the amount of carbon dioxide in the atmosphere, how much warmer will Earth get? The answer turns out to be maddeningly difficult to pin down, and the difference between the low and high estimates isn't academic—it's the difference between a challenging but manageable future and one that requires rebuilding human civilization around a radically different planet.

The Easy Part: Basic Physics

Let's start with what we know for certain. Carbon dioxide is a greenhouse gas, meaning it traps heat that would otherwise escape to space. When sunlight hits Earth, some of that energy bounces back as infrared radiation—heat. Greenhouse gases intercept that heat on its way out, like a blanket draped over the planet.

Before the Industrial Revolution, the atmosphere contained about 280 parts per million of carbon dioxide. By 2020, that number had climbed past 415 parts per million—a fifty percent increase. And we're still burning fossil fuels, so that number keeps rising.

Here's the part scientists can calculate with precision: if you double the carbon dioxide in the atmosphere and nothing else changes, Earth's surface warms by about one degree Celsius, or roughly 1.8 degrees Fahrenheit. This follows directly from the Stefan-Boltzmann law, a foundational equation in physics that describes how objects radiate heat. No controversy there. Every physicist on Earth would get the same answer.

One degree doesn't sound catastrophic. But that number is just the beginning of the story.

Where Things Get Complicated: Feedbacks

The climate system doesn't sit passively while you pump carbon dioxide into it. It responds. And those responses—called feedbacks—can either amplify or dampen the initial warming.

Think of it like pushing someone on a swing. The initial push is the carbon dioxide. But then the swing starts moving on its own, and depending on the timing of your subsequent pushes, you can make the swing go higher and higher, or you can accidentally slow it down.

Unfortunately, most of Earth's major climate feedbacks are self-reinforcing, meaning they amplify warming rather than dampen it.

Consider ice. Ice is white, and white surfaces reflect sunlight back into space—a property scientists call albedo, from the Latin word for whiteness. When the planet warms, ice melts. When ice melts, it exposes darker ocean water or darker land beneath. Those darker surfaces absorb more sunlight instead of reflecting it. That absorption causes more warming, which melts more ice, which exposes more dark surface, which causes more warming. This is the ice-albedo feedback, and it's one of the most powerful amplifying forces in the climate system.

Water vapor provides another crucial feedback. Warmer air holds more moisture—this is why humid summer days feel so oppressive. But water vapor is itself a greenhouse gas, far more abundant than carbon dioxide. So when the planet warms and evaporation increases, more water vapor enters the atmosphere, which traps more heat, which causes more warming, which causes more evaporation. The cycle feeds itself.

Clouds make everything more complicated. Some clouds reflect sunlight and cool the planet. Others trap heat and warm it. The net effect of clouds in a warming world remains one of the biggest uncertainties in climate science. Getting clouds right is enormously difficult because they operate at scales too small for global climate models to resolve directly.

Two Flavors of Sensitivity

Climate scientists actually track two different versions of climate sensitivity, and the distinction matters enormously for policy.

The transient climate response, or TCR, measures how much warming you get while carbon dioxide is still increasing. Imagine a runner who's still accelerating—their current speed isn't their top speed. The transient response captures the temperature while the climate system is still adjusting, still catching up to the extra carbon dioxide we've added.

The Intergovernmental Panel on Climate Change, the world body that synthesizes climate research, estimates that the transient climate response probably falls between one degree and 2.5 degrees Celsius for a doubling of carbon dioxide.

The equilibrium climate sensitivity, or ECS, is the eventual temperature once everything settles down—once the oceans have absorbed all the heat they're going to absorb, once all the fast feedbacks have played out. This is the temperature our descendants will live with long after we've stopped adding carbon dioxide to the atmosphere.

The equilibrium number is higher than the transient number, because some feedbacks take decades or centuries to fully manifest. The deep ocean, for instance, absorbs heat from the surface very slowly. While it's absorbing that heat, it acts as a buffer, keeping surface temperatures lower than they'll eventually become. But eventually the ocean warms all the way through, and that buffering effect disappears.

The most recent major assessment from the Intergovernmental Panel on Climate Change concluded that the equilibrium climate sensitivity most likely falls between 2.5 and four degrees Celsius, with a best estimate of three degrees Celsius. That's a range of roughly 4.5 to 7.2 degrees Fahrenheit, with a best guess around 5.4 degrees Fahrenheit.

Why the Range Matters

The difference between the low and high ends of that range isn't abstract.

At 2.5 degrees Celsius of warming per doubling, the Paris Agreement goal of limiting warming to well below two degrees Celsius remains achievable—difficult, but achievable—with aggressive emissions reductions. We'd face serious challenges: more extreme weather, rising seas, disrupted agriculture. But the changes would unfold slowly enough that human societies could adapt, albeit painfully.

At four degrees Celsius per doubling, the math becomes brutal. One study estimated that if the equilibrium climate sensitivity exceeds 3.4 degrees Celsius, the Paris Agreement's two-degree target simply cannot be met no matter how fast we cut emissions. We'd blow past that limit, at least temporarily, and face consequences we can only partially anticipate.

The economic stakes are staggering. One analysis concluded that simply halving the uncertainty in the transient climate response—not eliminating it, just cutting the uncertainty in half—could save trillions of dollars by allowing more precisely calibrated climate policy.

How Do We Measure Something We've Never Seen?

Here's the fundamental problem: we've never actually doubled carbon dioxide and waited for the planet to equilibrate. We're running that experiment in real time, on the only planet we have, without the luxury of waiting centuries to see the full result.

So scientists approach the question from multiple angles, hoping that independent lines of evidence will converge on the same answer.

The historical approach looks at temperature and carbon dioxide records since the Industrial Revolution. We know how much carbon dioxide we've added. We know how much the planet has warmed—a little over one degree Celsius so far. In principle, we should be able to calculate climate sensitivity from those numbers.

But there's a catch. Carbon dioxide isn't the only thing we've changed about the atmosphere. We've also added aerosols—tiny particles from burning fossil fuels, especially coal. Aerosols reflect sunlight and have a cooling effect that partially masks greenhouse warming. The problem is that we don't know exactly how strong that cooling effect is. Different assumptions about aerosols lead to different estimates of climate sensitivity.

The paleoclimate approach looks to Earth's distant past. During the Last Glacial Maximum, about twenty thousand years ago, massive ice sheets covered much of the Northern Hemisphere. We can estimate both the temperature and the carbon dioxide concentration during that period. Similar analyses can be applied to other ancient climates, like the Paleocene-Eocene Thermal Maximum about 55 million years ago, when the planet was dramatically warmer than today.

The catch here is that ancient climates differed from modern ones in ways that affect sensitivity. The Last Glacial Maximum had enormous ice sheets that don't exist today. The continents were in slightly different positions. Ocean circulation patterns were different. Translating ancient climate sensitivity to modern conditions requires assumptions that introduce uncertainty.

Climate models provide a third approach. These are sophisticated computer simulations that attempt to capture the physics of the atmosphere, oceans, ice, and land surface. Scientists can run experiments in models that would be impossible in the real world—like instantaneously doubling carbon dioxide and watching what happens over thousands of simulated years.

But models have their own limitations. They can't resolve every physical process, especially small-scale phenomena like individual clouds. They make approximations that may or may not capture reality. And different modeling groups, using different approximations, get different answers.

A Century of Hunting

The quest to determine climate sensitivity has been going on far longer than most people realize.

In 1896, the Swedish chemist Svante Arrhenius published the first quantitative estimate of how much warming a doubling of carbon dioxide would cause. Using pencil, paper, and an enormous amount of patience, he calculated something in the neighborhood of five to six degrees Celsius. Remarkably, despite knowing almost nothing about many of the feedbacks we now recognize, his estimate falls within the modern uncertainty range.

Arrhenius was actually trying to explain ice ages, not predict future warming. He thought the effect would take thousands of years and considered it potentially beneficial for his native Sweden. He couldn't have imagined that humans would double atmospheric carbon dioxide in just a century or two.

Through the twentieth century, estimates bounced around as scientists discovered new feedbacks and developed better tools. A landmark 1979 report chaired by the meteorologist Jule Charney concluded that equilibrium sensitivity probably fell between 1.5 and 4.5 degrees Celsius. That range held remarkably steady for decades—testament either to the robustness of the estimate or to the stubbornness of the underlying uncertainty, depending on your perspective.

The most recent comprehensive assessment, published in 2021, finally narrowed the range slightly, ruling out the lowest and highest values with greater confidence. But the core uncertainty remains. We still can't pin down the answer more precisely than "somewhere between 2.5 and four degrees Celsius."

The Slow Feedbacks We're Not Counting

Even the equilibrium climate sensitivity, the long-term measure, doesn't capture everything. By definition, it excludes feedbacks that take millennia to fully manifest.

Consider the great ice sheets of Greenland and Antarctica. These are not like seasonal ice or mountain glaciers. They're miles thick in places, containing enough frozen water to raise global sea levels by over sixty meters if they fully melted. They respond to warming, but slowly—over centuries and millennia, not decades.

As these ice sheets shrink, they reduce Earth's albedo, allowing the planet to absorb more sunlight. They also change ocean circulation patterns and atmospheric dynamics in complex ways. These effects are not included in the standard equilibrium sensitivity calculation.

Scientists have another measure, called Earth system sensitivity, that attempts to capture these very slow feedbacks. Estimates suggest Earth system sensitivity could be roughly double the equilibrium sensitivity—meaning the ultimate warming from doubled carbon dioxide, if you wait long enough, might be twice what the standard estimates suggest.

This matters for thinking about the very long-term future. The carbon dioxide we emit today will affect climate for thousands of years. Our great-great-great-grandchildren, and their descendants for hundreds of generations, will live with the consequences of our choices.

Tipping Points and Non-Linearity

There's something else lurking in the climate system that the concept of sensitivity struggles to capture: the possibility of sudden, irreversible shifts.

Climate sensitivity, as conventionally defined, assumes the system responds smoothly and proportionally to forcing. Double the carbon dioxide, double the temperature increase. That's a linear relationship.

But the Earth system may not behave linearly. It may have tipping points—thresholds beyond which the system lurches into a fundamentally different state.

The ice sheets are a potential tipping point. Once they start to collapse, the process may become self-sustaining and irreversible on any human timescale. The Amazon rainforest might cross a threshold beyond which it can no longer sustain itself and converts to savanna, releasing enormous amounts of stored carbon. Permafrost in the Arctic contains vast quantities of frozen organic matter; if it thaws, microbes will decompose that organic matter and release carbon dioxide and methane, driving further warming.

These tipping points, if they exist and if we cross them, would effectively increase climate sensitivity in ways that our current estimates don't fully capture. The system might be relatively stable up to a point, then become dramatically more sensitive once certain thresholds are crossed.

What the Numbers Mean for Us

Let's make this concrete. We've already increased atmospheric carbon dioxide by about fifty percent since preindustrial times, from 280 to over 415 parts per million. If we continue on our current trajectory, we'll likely reach double the preindustrial concentration—560 parts per million—somewhere around mid-century.

If the equilibrium climate sensitivity is at the low end of estimates, around 2.5 degrees Celsius, that doubling would eventually produce warming that, while serious, falls within what human societies could probably adapt to with sufficient investment and foresight.

If it's at the high end, around four degrees Celsius, we're looking at a transformed planet. Large parts of the tropics could become essentially uninhabitable during heat waves. Agriculture would shift dramatically, with some regions becoming more productive and others failing. Sea level rise would accelerate, threatening coastal cities that are home to hundreds of millions of people. Ecosystems would be under severe stress, with mass extinctions likely.

And here's the uncomfortable truth: we won't know which end of the range we're dealing with until we've already committed to the outcome. Climate change has a profound time lag. The warming we experience today results from emissions decades ago. By the time we know with certainty whether sensitivity is at the high or low end, we'll have already emitted the carbon dioxide that locks in that future.

The Prudent Approach

Faced with this uncertainty, how should we act?

One perspective holds that we should plan for the high end of the range. If sensitivity turns out to be low and we've reduced emissions aggressively, the worst outcome is that we transitioned to clean energy somewhat faster than strictly necessary—and reaped the co-benefits of reduced air pollution, enhanced energy security, and technological leadership in the industries of the future.

But if sensitivity turns out to be high and we planned for the low end, the consequences are irreversible on any timescale that matters to human civilization. We can always slow down emissions reductions if the evidence shifts toward lower sensitivity. We cannot undo a century of emissions if the evidence shifts toward higher sensitivity.

This asymmetry—the irreversibility of climate change versus the reversibility of climate policy—suggests that uncertainty itself is an argument for more aggressive action, not less. We're not flipping a coin with equal upsides and downsides. We're facing a situation where the bad outcomes are potentially catastrophic and the good outcomes are merely inconvenient.

The Search Continues

Climate scientists haven't given up on narrowing the range. New satellite observations provide better data on clouds and energy flows at the top of the atmosphere. Ice cores and ocean sediments reveal more details about ancient climates. Computing power increases, allowing more detailed simulations.

But there's a fundamental limit to how precise we can get. The climate system is chaotic, meaning small differences in initial conditions can produce large differences in outcomes. And we have only one planet, one history, one experiment. We can't rerun Earth's climate with slightly different parameters to see what happens.

Climate sensitivity will likely remain somewhat uncertain for as long as the question matters. And it will matter for the rest of this century and beyond—as long as humans are making decisions about energy, land use, and the atmosphere we all share.

The uncertainty isn't a reason for paralysis. It's a reason for humility about our predictions and wisdom in our actions. The single number that determines our future remains elusive, but what we do know is more than enough to act on.