Key Takeaways
Accept human-caused warming without pretending every consequence is settled.
The central distinction. Koonin accepts that the planet has warmed by about 1°C since the late nineteenth century and that human emissions contribute to that warming. His challenge is to claims that this evidence makes every forecast, extreme-weather trend, or proposed policy equally certain. Climate science asks three linked questions: what influences the climate, how the climate responds, and what those changes mean for people and ecosystems. Uncertainty grows as the questions build on one another, especially when predicting local impacts decades ahead.
An expert’s change of mind. After leading a 2014 workshop that scrutinized climate evidence and models, Koonin says he found the field less mature than he had assumed. His argument is not that climate is static, but that public discussion often presents qualified findings as settled certainty.
The useful move is to reject a false choice between accepting human influence and accepting every alarming claim. Scientific confidence is claim-specific: evidence for warming can be strong while estimates of regional rainfall or future storm damage remain weak. That distinction is essential in any complex field, from epidemiology to economics. Still, uncertainty is not evenly distributed across all claims. It should not be used to blur the well-established evidence that greenhouse gases warm the planet or to imply that no action is justified until every consequence is known.
Compare decades and regions before calling a weather pattern climate change.
Climate needs context. Weather is what happens day to day; climate describes patterns averaged over decades. The book uses a thirty-year average as the standard reference, while noting that researchers sometimes analyze shorter periods. A hot summer or snowy winter is therefore not, by itself, evidence of a changing climate. Long records matter because year-to-year swings can be much larger than the long-term trend.
Scale changes the story. Global temperature records show roughly 1.1°C of warming from 1900 to 2020, but the trend varied across periods and places. Koonin also argues that local warming near cities can reflect urban growth, not just global greenhouse-gas influences. Temperature anomalies, which compare each place with its own baseline, help researchers compare changes across very different climates.
This is a lesson in measurement design as much as climate. An anomaly makes unlike places comparable, just as standardized scores help compare students taking different exams. But every summary measure hides detail: a global average can obscure regional exposure, while a local record can exaggerate a worldwide trend. Good interpretation requires both scales, plus a clear account of the baseline and time window. Koonin’s warning about cherry-picking applies in every direction: choosing a short interval can create either an alarming spike or a reassuring pause.
Small greenhouse-gas changes can shift the planet's energy balance.
The greenhouse mechanism. Earth warms by absorbing sunlight and cools by radiating heat back to space. Greenhouse gases intercept some of that outgoing heat, much like insulation, before the energy ultimately escapes. Koonin calls the net influence of a factor on this energy balance radiative forcing. Human influences are small compared with the total energy moving through the climate system, but that does not make them irrelevant: the planet’s temperature depends on a finely balanced exchange.
CO2 has a distinct role. Atmospheric carbon dioxide rose from about 280 parts per million before industrialization to about 410 ppm in 2019. Multiple lines of evidence, including the changing carbon-isotope mix, show that the increase is overwhelmingly human-caused. The book also stresses that other influences, including aerosols, natural variability, and changes in sunlight, complicate the total warming picture.
The energy-balance explanation helps correct an intuitive mistake: a small cause can matter when a system is sensitive and the effect accumulates over time. The same logic appears in ecological tipping processes and financial compounding, though those analogies should not be mistaken for identical mechanisms. A useful scientific question is not whether human forcing is small relative to all planetary energy flow, but how it compares with the system’s capacity to shed or redistribute heat. The book is strongest when it asks for that comparison rather than treating smallness as a verdict.
Track cumulative carbon, because modest cuts only slow its buildup.
Emissions are not concentration. Each year, the world emits carbon dioxide, and natural processes absorb roughly half of the added amount. Much of the rest remains in the atmosphere, so the concentration keeps rising as long as emissions continue. Koonin emphasizes that CO2 behaves unlike short-lived air pollution: a cut in emissions slows its accumulation, but does not by itself stabilize the concentration. That requires global emissions to fall to roughly zero, or removal to balance what remains.
The carbon cycle is vast. Fossil-fuel emissions add carbon to a natural system that moves enormous quantities among plants, soils, oceans, and air. The annual human addition is small relative to natural flows, yet it disrupts their balance. Because excess CO2 persists for centuries, today’s choices affect the climate far beyond the year they are made.
The distinction between a flow and a stock is a powerful mental model. Annual emissions are a flow; atmospheric CO2 is a stock built up through cumulative flows. Cutting the flow matters, but the stock does not promptly return to its former level. This resembles debt reduction: spending less can slow the growth of debt without reducing what is already owed. The analogy also reveals a policy question the book raises but cannot settle scientifically: who should bear the cost of reducing a global stock, especially when energy access and historical emissions differ so sharply across countries?
Treat climate models as conditional experiments, not crystal balls.
Models simplify reality. Climate models divide the atmosphere and oceans into grid boxes and calculate how energy, air, and water move between them. Many important processes, including clouds and storms, are smaller than those boxes, so modelers must approximate them. They also tune parameters to make simulations resemble observations. Koonin argues that these choices create substantial disagreement among models, particularly about clouds and climate sensitivity.
Projections depend on assumptions. Representative Concentration Pathways, or RCPs, are scenarios describing possible futures for population, economies, technology, and emissions. They are not predictions. The equilibrium climate sensitivity, or ECS, estimates warming after a hypothetical doubling of preindustrial CO2. The book notes that the IPCC’s likely range remained 1.5°C to 4.5°C from 1979 to 2014, and that some newer models exceeded that range.
A model is best understood as a structured conditional: if these emissions and physical assumptions hold, then these outcomes follow. That makes models useful for comparing scenarios, even when they cannot forecast the exact local future. Koonin is right that model spread and tuning deserve transparency. Yet disagreement does not automatically mean a model has no value; ensemble behavior can expose shared mechanisms and uncertainty ranges. The key is to evaluate each model against observations and its intended use, rather than treating either a single projection or an ensemble average as a direct measurement.
Test extreme-weather claims against records, not vivid anecdotes.
Records can overturn impressions. Koonin examines US temperature records and argues that daily record highs show no significant long-term trend, while record cold nights have declined. He says a government report’s high-to-low record ratio made recent years look more alarming because both kinds of records become rarer over time, but cold records declined faster. His alternative analysis counted absolute records across 725 stations from 1895.
Storm counts need similar care. The book reports no clear long-term human-caused change in hurricanes, while noting that incomplete early observations make the historical record difficult to compare. US tornado reports more than doubled after 1950, but radar began detecting many weaker storms. Looking only at EF1-and-stronger tornadoes shows little trend, and the strongest category, EF3 and above, declined by about 40 percent after 1954.
The central methodological point is robust: a trend in recorded events can reflect a change in observation as well as a change in nature. This is familiar in public health, where improved screening can make a condition appear more common even when underlying risk is stable. Koonin’s examples make that problem vivid, but his selected measures and periods also need scrutiny. A lack of detected trend in one dataset is not proof of no human influence, and findings about the contiguous United States do not automatically describe every region or future climate.
Put rising seas in both historical context and local geography.
The sea has risen before. Since the last ice age, global sea level has risen by about 120 meters as continental ice melted. In the modern record, tide gauges indicate an average rise of roughly 1.8 millimeters per year since 1880, while satellites measured about 3 millimeters per year over the decades before the book’s publication. Koonin notes that comparable rates occurred in some earlier periods, including around 1925 to 1940.
Local exposure matters. Sea level at a particular coast also reflects land sinking or rising, ocean currents, and local water conditions. In Houston-Galveston, groundwater withdrawal lowered the land by as much as three meters. The book argues that global averages and dramatic maps need historical context, but it does not deny that warming, melting land ice, and expanding seawater contribute to rising seas.
A global average is useful for understanding planetary change, but coastal decisions depend on local relative sea level, flood pathways, and the value of exposed infrastructure. This is a classic scale problem in risk analysis: the same global signal can produce very different local consequences. Historical variability is relevant to attribution, but it does not make future exposure harmless. Even a gradual rise can increase the frequency of damaging high-water events, so uncertainty about the pace should inform flexible planning rather than be mistaken for a reason to postpone it.
Separate projected catastrophe from measured human outcomes.
The historical record includes adaptation. Koonin reports that annual deaths from weather-related disasters fell by about 80 times over the past century, largely alongside better warnings, infrastructure, medical care, and economic development. He challenges a projection of 85 additional deaths per 100,000 by 2100 by pointing out that it used the high-emissions RCP8.5 scenario; a more moderate scenario produced a much smaller and highly uncertain estimate.
Agriculture and economies complicate the alarm. Global yields of wheat, rice, and maize more than doubled from 1961 to 2011, with rising CO2 among several contributing factors. The book cites estimates of modest net economic impacts by 2100, while emphasizing uncertainty in both climate and economic models. These aggregate findings do not mean every region or household benefits, or that food insecurity has disappeared.
Historical improvement is important evidence that vulnerability is not fixed. Wealth, public health, infrastructure, and adaptation can change outcomes even as hazards change. Yet falling disaster mortality does not show that climate risk is absent; it shows that exposure and resilience are powerful parts of the equation. Aggregate economic estimates also conceal distribution: a small global effect can still impose severe losses on a particular community. The best reading therefore keeps two truths together: catastrophe claims need evidence, and modest averages do not erase concentrated harms.
Demand complete evidence when institutions turn research into public messages.
A summary can distort without inventing data. Koonin describes a long information chain from research papers to assessment reports, summaries, press releases, and headlines. At each stage, caveats and context can disappear. He argues that some reports highlighted selected trends while burying contradictory findings, and that media incentives favor dramatic stories over careful uncertainty.
Review the synthesis, too. A Red Team review is an adversarial examination in which an independent group tries to find weaknesses in a report and its conclusions. Koonin proposes this alongside clearer uncertainty language and better scrutiny of graphs. For readers, he recommends checking time periods, units, uncertainty ranges, and whether a story confuses observations with model projections. The goal is not to dismiss experts, but to make their claims easier to judge.
The book’s strongest institutional insight is that trust depends not only on data quality, but also on how evidence is compressed for decision makers. This resembles auditing in finance: independent challenge can reveal errors that routine approval misses. A Red Team is most valuable when it has clear standards, access to the same evidence, and a genuine chance to correct the record. It should not become a ritual of manufacturing symmetry between claims with unequal evidence. Skepticism is a method, not a political identity.
Treat net zero as a global infrastructure transformation, not a slogan.
Energy systems move slowly. Fossil fuels supply about 80 percent of global energy in the book’s account, while population growth and development are expected to raise demand. Power plants, grids, refineries, vehicles, and buildings last for decades and must work reliably together. Koonin argues that replacing this system is not like completing a single technological project: it means coordinating changes across industries, infrastructure, policy, and daily life.
The global arithmetic is difficult. CO2 concentration depends on cumulative emissions, so meaningful stabilization requires very deep global cuts. The Paris pledges described in the book were voluntary and projected to reduce emissions by less than 10 percent by 2030, far short of the pace required for the agreement’s temperature goals. Koonin judges rapid decarbonization unlikely, particularly if affordable, reliable alternatives are not available to developing countries.
The infrastructure lens improves the discussion by replacing a binary choice between ambition and inaction with questions about sequencing, reliability, cost, and adoption. Energy transitions are social systems as well as engineering projects: a technology must be built, financed, connected, maintained, and accepted. Still, difficulty is not impossibility, and forecasts of political failure can become self-fulfilling if treated as destiny. A sound strategy can pair research and deployment with near-term reductions that have clear benefits, while being candid about who pays and who gains.
Build adaptation and contingency plans for a climate that will keep changing.
Adaptation works with uncertainty. Koonin expects societies to adapt because local measures can be scaled to local conditions and adjusted as risks become clearer. Examples include flood protection, resilient infrastructure, emergency preparation, and stronger institutions. He argues that economic development helps poorer countries adapt, since wealth and governance affect their ability to respond to hazards.
Keep other options under study. Solar Radiation Management, or SRM, would reflect a small amount of sunlight to cool the planet. Carbon Dioxide Removal, or CDR, would extract CO2 from the atmosphere. Both face serious limits: SRM could produce uneven side effects and would need to continue, while CDR would require enormous infrastructure, energy, and storage. Koonin presents these as contingency options requiring research and governance, not ready-made fixes.
Adaptation is a practical complement to emissions cuts because it protects people from hazards regardless of their cause. Its local character can also make it easier to implement than a single global bargain. But adaptation has limits: some ecosystems and communities have fewer resources, and not every loss can be reversed or relocated. Research into climate intervention is similarly prudent as contingency planning, but it raises governance problems that cannot be solved by engineering alone. The responsible posture is neither to promise rescue nor to rule out options before understanding their risks.
Analysis
Unsettled is both a climate-science primer and an argument about who gets to speak for science. Its most durable contribution is methodological: follow the measurement, inspect the time window, distinguish observed changes from modeled futures, and keep uncertainty attached to the particular claim it qualifies. That discipline is valuable in a debate where a single dramatic event can be made to carry more explanatory weight than its data supports.
The book also requires the same scrutiny it asks readers to apply to official assessments. Its case is deliberately built around discrepancies between alarming public claims and selected assessment findings. That exposes genuine communication failures, but an inventory of underreported uncertainty is not itself a complete account of climate risk. Broad evidence for human-caused warming and the central role of greenhouse gases is stronger than some passages' rhetoric about unresolved attribution may suggest. Similarly, the absence of a clear historical trend in one extreme-weather measure does not establish that warming cannot affect that hazard, or that future risk is negligible.
Koonin often translates projected economic effects into a percentage of future GDP growth. That is a useful antidote to headlines that omit scale, but aggregate growth can hide losses borne by particular regions, industries, and households. A small net effect can coexist with serious local damage. Conversely, models built on extreme assumptions should not be presented as the likeliest future merely because they describe a severe possibility.
The resulting lesson is not to choose between alarm and reassurance. It is to separate three questions: what evidence supports, what society values, and what implementation can deliver. Scientific uncertainty can change the design and timing of policy without erasing the case for preparation. The book is most persuasive when it demands transparency, and least persuasive when its own emphasis on uncertainty risks sounding like a forecast of safety. Better decisions require both its skepticism toward overstated claims and equal skepticism toward comforting conclusions.
Review Summary
Unsettled receives mostly positive reviews for its balanced, data-driven approach to climate science. Readers appreciate Koonin's credentials and his critique of media sensationalism and scientific integrity. The book challenges common climate change narratives while acknowledging human impact. Some praise its accessibility, while others find it technically challenging. Critics argue Koonin cherry-picks data and downplays climate risks. Overall, reviewers recommend it for those seeking a nuanced perspective on climate science, regardless of their prior beliefs.
People Also Read
Glossary
radiative forcing
Influence on energy balanceA change in the balance between energy entering the climate system and heat leaving it. Greenhouse gases, aerosols, volcanic eruptions, and changes in sunlight can all exert radiative forcing. It is commonly measured in watts per square meter and helps describe how an influence tends to warm or cool the climate.
Representative Concentration Pathways
Possible emissions futuresScenarios used by climate researchers to describe plausible future paths for population, economic activity, technology, emissions, and resulting atmospheric concentrations. Their labels refer to the radiative forcing level associated with each scenario near the end of the century. They are conditional scenarios, not predictions of what will happen.
equilibrium climate sensitivity
Warming after CO2 doublingAn estimate of how much the global average surface temperature would rise after atmospheric carbon dioxide reaches twice its preindustrial concentration and the climate system has had time to adjust. It summarizes the climate system’s response to that change, including feedbacks such as shifts in water vapor, clouds, snow, and ice.
event attribution studies
Human role in an eventAnalyses that estimate whether human influences changed the likelihood or intensity of a particular weather event, such as a flood, heat wave, or hurricane. They commonly compare observations and climate-model simulations under different conditions. The book emphasizes that such conclusions can be difficult to test and depend on the models, records, and assumptions used.
Global Mean Sea Level
Average global ocean heightAn estimate of the average height of the ocean surface across the globe, derived from tide gauges and satellite measurements. It smooths over regional differences caused by currents, gravity, ocean temperature, and other factors. Local relative sea level also depends on land movement, so the global average cannot by itself predict a coastal community’s exposure.
Red Team
Adversarial report reviewAn independent group tasked with rigorously testing a scientific assessment or its conclusions by identifying weak evidence, omissions, assumptions, and possible errors. The authors or a separate review group can respond to those challenges. The method is intended to strengthen or clarify a report through structured criticism, not to presume that its findings are wrong.
Solar Radiation Management
Reflect sunlight to coolA proposed climate intervention that would increase the amount of sunlight reflected back into space, reducing the solar energy absorbed by Earth. One studied possibility is creating reflective particles in the stratosphere, similar in effect to the temporary cooling after large volcanic eruptions. It would not remove carbon dioxide and could produce uneven regional effects.
Carbon Dioxide Removal
Extract CO2 from airA group of proposed methods for taking carbon dioxide directly from the atmosphere and storing it, or using it in products. Approaches include engineered capture systems and increasing carbon uptake by vegetation. The scale required to materially affect atmospheric concentrations would demand substantial energy, infrastructure, and secure storage.
FAQ
What's Unsettled about?
- Exploration of Climate Science: Unsettled by Steven E. Koonin examines the complexities of climate science, highlighting what is known and what remains uncertain.
- Challenges Popular Perceptions: The book critiques the oversimplified narrative of climate change, arguing that the science is not as settled as often portrayed.
- Two-Part Structure: It is divided into two parts, focusing on the science of climate change and societal responses, providing a comprehensive view of the issues.
Why should I read Unsettled?
- Informed Decision-Making: Koonin emphasizes understanding climate science uncertainties to make informed policy decisions.
- Expert Perspective: As a former undersecretary for science, Koonin offers insights based on his extensive experience in academia and government.
- Critical Examination: The book encourages readers to critically evaluate media and policy narratives about climate change.
What are the key takeaways of Unsettled?
- Science is Not Settled: Koonin asserts that significant uncertainties exist in climate science, affecting projections and policy decisions.
- Miscommunication in Climate Science: The book discusses how climate information often gets distorted, leading to public misconceptions.
- Pragmatic Policy Approach: Koonin advocates for a balanced approach to climate policy, considering both costs and benefits.
What are the best quotes from Unsettled and what do they mean?
- “I do not think ‘The Science’ says what you think it says.”: This quote invites readers to question prevailing climate science narratives.
- “Uncertainty is a prime mover and motivator of science.”: It highlights that acknowledging uncertainties drives scientific progress.
- “The science of climate is neither settled nor sufficient to dictate policy.”: This underscores the need for a balanced approach to climate action.
How does Steven E. Koonin define climate change in Unsettled?
- Human and Natural Influences: Koonin explains that climate change results from both human activities and natural variability.
- Complex Interactions: He emphasizes the complexity of the climate system, making accurate predictions challenging.
- Historical Context: The book provides a historical perspective, noting significant natural variations before substantial human influence.
What are the main uncertainties in climate science discussed in Unsettled?
- Data Limitations: Many climate data sets are incomplete or of low quality, complicating trend analysis.
- Model Discrepancies: Different climate models often produce conflicting results, leading to uncertainty in projections.
- Communication Failures: The way climate science is communicated often leads to misunderstandings among the public.
How does Unsettled address the topic of extreme weather events?
- Skepticism About Trends: Koonin argues that many extreme weather events do not show significant trends attributable to human influences.
- Event Attribution Studies: The book critiques the reliability of studies linking specific weather events to climate change.
- Historical Context: Koonin emphasizes the importance of historical data in understanding extreme weather patterns.
What does Steven E. Koonin say about sea level rise in Unsettled?
- Gradual Increase: Sea levels have been rising at a slow and steady rate, with historical data showing an average increase of about 30 cm per century.
- Human Influence Uncertainty: Challenges exist in attributing recent sea level rise to human activities.
- Tide Gauge and Satellite Data: Both types of data are crucial for understanding sea level changes.
How does Unsettled view the media's role in climate discussions?
- Sensationalism Critique: Koonin criticizes the media for sensationalizing climate issues, leading to public confusion.
- Need for Accurate Reporting: He emphasizes the importance of accurate and nuanced reporting on climate science.
- Responsibility of Journalists: Journalists should strive for a deeper understanding of climate science to avoid perpetuating misinformation.
What are Koonin's recommendations for addressing climate change in Unsettled?
- Focus on Adaptation: Koonin advocates for prioritizing adaptation strategies to manage climate impacts effectively.
- Research and Monitoring: Increased investment in climate research and monitoring is necessary to inform policy decisions.
- Balanced Approach: A balanced approach that considers both mitigation and adaptation is encouraged.
What is the significance of the term "The Science" in Unsettled?
- Critique of Consensus: "The Science" refers to the oversimplified narrative that climate science is settled.
- Call for Open Debate: Koonin advocates for open debate and discussion within the scientific community.
- Encouragement of Critical Thinking: The term serves as a reminder to think critically about "settled science" narratives.
How does Steven E. Koonin suggest improving climate science communication in Unsettled?
- Transparency and Clarity: Clear and transparent communication of climate science, including uncertainties, is essential.
- Engagement with Non-Experts: Scientists should engage with non-experts to foster understanding and informed discussions.
- Red Team Reviews: Implementing Red Team reviews can critically assess climate assessments and improve communication robustness.
Download PDF
Download EPUB
.epub digital book format is ideal for reading ebooks on phones, tablets, and e-readers.