Unsettled: What Climate Science Tells Us, What It Doesn’t, and Why It Matters

Steven E. Koonin’s Unsettled takes on one of the most consequential—and contentious—issues of our time: climate change. Koonin, a physicist, former Caltech professor, and former U.S. Department of Energy official, does not argue that climate change is a hoax or that humans have no influence on the climate. Instead, he asks a more complicated question: How much do we know with confidence, where does meaningful uncertainty remain, and what happens to those distinctions when climate science reaches the public?

His central argument is that climate science contains both well-established findings and significant areas of uncertainty—and that public discussions do not always distinguish clearly between the two.

TL;DR

Koonin accepts that Earth has warmed, atmospheric carbon dioxide has increased substantially because of human activity, and greenhouse gases influence the climate. His argument is that greater uncertainty emerges when scientists move from those basic conclusions to questions about exactly how much warming will occur, how individual regions will be affected, how particular extreme-weather events will change, and what the economic and societal consequences will be decades from now.

Throughout Unsettled, Koonin urges readers to distinguish between observations, attribution, computer-model projections, economic forecasts, and policy decisions. They are connected, but they are not interchangeable.

The book is also controversial. Many climate scientists argue that Koonin places disproportionate emphasis on uncertainty and understates the strength of the broader scientific evidence concerning human-caused climate change and its risks.

The value of Unsettled, therefore, may not be in accepting every one of Koonin’s conclusions. It is in asking better questions about what we know, how we know it, how certain we are, and how scientific evidence should influence the decisions society makes.

The Data

One of Koonin’s recurring messages is surprisingly simple: look at the actual data and look at it over meaningful periods of time.

Earth’s climate has never been static. Temperatures, rainfall, droughts, hurricanes, sea ice, glaciers, ocean circulation, and other climate indicators fluctuate naturally over years, decades, centuries, and longer periods. Today, human influences—especially increasing greenhouse-gas concentrations—operate alongside those natural processes.

That creates the challenge of attribution: determining how much of an observed change can be attributed to human activity and how much results from natural variability or other factors.

Koonin argues that attribution is considerably easier for some questions than others. The evidence that greenhouse gases contribute to global warming is strong. Connecting climate change to changes in a particular type of extreme event, especially at a specific location or during a particular period, can involve additional uncertainty.

He also cautions against automatically interpreting increasing financial losses from natural disasters as proof that the underlying weather has worsened by the same amount.

Consider hurricanes. A storm striking a heavily populated coastline today can cause dramatically more economic damage than a comparable storm decades ago because there are more people, homes, businesses, roads, and valuable infrastructure in its path. The physical hazard and society’s exposure and vulnerability to that hazard are different variables.

That distinction illustrates one of the book’s most useful habits of thought.

When encountering a climate claim, ask: What exactly is being measured? Over what period? Is there a meaningful trend? What other factors could influence the result? And how confidently can scientists attribute the observed change to human activity?

Going Deeper

One of the most important discussions in Unsettled concerns climate models.

Modern climate models are extraordinary scientific achievements. They simulate interactions among the atmosphere, oceans, land, ice, clouds, greenhouse gases, solar energy, and countless other components of Earth’s climate system.

They are essential tools for studying possible climate futures.

But models are not crystal balls.

The climate system is extraordinarily complicated, and not every physical process can be represented perfectly. Some processes occur at scales too small to model directly and must instead be approximated mathematically. Clouds are a particularly important example because they can both reflect incoming sunlight and influence how heat escapes from Earth.

Different models can therefore produce different projections even though they share the same fundamental physical principles.

There is another source of uncertainty that no climate model can eliminate: human behavior.

How many people will inhabit Earth in 2100? How much energy will they consume? How quickly will developing economies grow? What will energy cost? How widely will nuclear, solar, wind, batteries, carbon capture, or technologies not yet invented be deployed?

Those choices affect future greenhouse-gas emissions.

This means long-term climate projections can contain several layers of uncertainty: future human behavior, future emissions, the climate system’s response, regional consequences, and ultimately the effects on economies and societies.

Koonin’s argument is not that models are useless. Rather, he believes those uncertainties should remain visible when model projections are communicated to the public.

What You Should Know

Perhaps the most important distinction in Unsettled is between science and policy.

Science can establish physical principles. Measurements can tell us atmospheric carbon dioxide concentrations are increasing. Observations can document changes in temperatures, oceans, glaciers, sea level, precipitation, and other environmental conditions. Models can help researchers investigate possible futures.

But determining what society should do involves another set of questions.

Should governments impose carbon taxes? Expand nuclear energy? Subsidize renewable power? Encourage electric vehicles? Invest in carbon capture? Build stronger flood defenses? Accelerate adaptation?

Scientific evidence can inform these decisions, but it cannot make them for us. Policy also involves economics, technology, acceptable levels of risk, competing priorities, and questions of fairness between countries and generations.

Koonin gives particular attention to adaptation.

Regardless of future emissions, societies will continue experiencing hurricanes, floods, droughts, heat waves, wildfires, and other natural hazards. Better infrastructure, building standards, water management, emergency planning, agricultural practices, forecasting, and flood protection can reduce vulnerability under a variety of climate scenarios.

Energy creates another complication. Billions of people in developing countries seek higher standards of living, and economic development generally requires greater access to affordable and reliable energy.

Any serious discussion about reducing global emissions therefore has to consider not only climate objectives but also the enormous scale of the world's energy system and the aspirations of people seeking greater prosperity.

The Takeaway

Taken together, these themes provide a useful framework for thinking about climate change.

The Data reminds us to look beyond individual events and examine long-term evidence, context, and attribution. Going Deeper demonstrates why models are valuable while showing how assumptions about technology, energy, economics, and human behavior influence long-term projections. And What You Should Know reminds us that scientific findings and policy choices, while connected, are not the same thing.

None of this means the evidence for human-caused climate change is weak. There is strong evidence that greenhouse-gas emissions are warming the planet.

At the same time, confidence is not identical for every climate-related question. Scientists can have high confidence in one conclusion while having greater uncertainty about the magnitude, timing, or regional consequences of another.

This is where Unsettled becomes both useful and controversial.

Koonin encourages readers to question assumptions, examine underlying data, and pay close attention to uncertainty. Many climate scientists counter that his presentation gives disproportionate attention to those uncertainties and can leave readers with an impression of greater scientific disagreement than the broader evidence supports.

The reader therefore does not need to choose between the sweeping statements that “everything is settled” or “climate science is uncertain.”

Reality is more complicated.

Some conclusions are supported with very high confidence. Others involve wider ranges of uncertainty. Still others concern economic, technological, and political choices rather than climate science itself.

That distinction connects directly to the central message of this article: look at the data, understand the assumptions behind projections, recognize the limitations of prediction, and distinguish scientific findings from the policy decisions that follow.

You do not have to agree with every argument in Unsettled to benefit from engaging with it. Its value lies partly in encouraging readers to ask better questions—and then compare Koonin’s answers with broader scientific assessments and evidence before reaching their own conclusions.

The Bottom Line

The title Unsettled can easily be misunderstood. Koonin is not claiming that all climate science is unsettled.

He accepts several foundational conclusions: greenhouse gases affect Earth’s climate, humans have substantially increased atmospheric carbon dioxide, and human activity contributes to warming.

His argument centers on what happens next—when those basic scientific conclusions are translated into detailed predictions about future temperatures, extreme weather, regional impacts, economic damages, and ultimately public policy.

Many climate scientists disagree with aspects of Koonin’s interpretation and contend that he emphasizes uncertainty in ways that understate the strength of the overall evidence. That criticism is an important part of understanding the book.

Perhaps the larger lesson of Unsettled therefore extends beyond climate change.

Good scientific thinking requires being comfortable with two ideas at the same time: we can know important things with considerable confidence while remaining uncertain about others.

Acknowledging uncertainty does not invalidate climate science. Likewise, accepting that humans are warming the planet does not mean every prediction about the climate decades from now carries the same level of confidence.

The challenge is learning to tell the difference.

That means moving beyond headlines, examining evidence, understanding what models can and cannot tell us, separating observations from projections, and recognizing where scientific questions end and policy judgments begin.

In that sense, Unsettled asks readers to do something increasingly valuable in complicated public debates: replace slogans with evidence, certainty with curiosity, and simple answers with better questions.