Book summary “Unsettled”: summary

Introduction

You are bombarded daily with media reports claiming that climate catastrophe is imminent. Headlines warn of apocalyptic heatwaves, vanishing ice, and unprecedented droughts, creating a pervasive sense of climate emergency. Yet you may wonder how well these alarming declarations match the underlying scientific research.

Steven E. Koonin, a distinguished physicist and former Department of Energy undersecretary, demonstrates that public perception is largely disconnected from actual scientific reports. The author argues that official assessment documents actually contain reassuring data that directly contradict popular alarmism.

The book examines the state of climate science by addressing several fundamental questions:

• How much influence do humans actually exert on global temperatures?
• Why do complex computer models struggle to project regional climates accurately?
• What do official observational records reveal about extreme weather trends?
• How vulnerable is society to projected economic and environmental impacts?

1. What We Know About Warming

Global temperatures have risen modestly since 1850 through natural and human influences, but distinguishing long-term climate trends from short-term weather anomalies and urban heat effects requires careful data analysis.

Temperature Anomalies and Oceans

Weather describes constant, day-to-day variations in temperature, precipitation, and wind. Climate represents a statistical average over decades, typically defined by the World Meteorological Organization as a thirty-year span. Conflating short-term weather fluctuations with long-term climate trends distorts public understanding of environmental data. Moving from Wisconsin to Arizona means buying air conditioning based on climate, but checking Tuesday's forecast for an umbrella on Thursday relies on weather.

Climatologists track global temperature anomalies rather than absolute temperatures to compare changes across different geographic regions and eliminate baseline distortions. Anomalies measure how much a location's temperature deviates from a long-term baseline average, usually calculated over a thirty-year period. This method allows researchers to place shifts in the Arctic on the same analytical footing as changes in the Tropics. Four independent analyses show that the global average surface temperature anomaly has risen by about 1 degree Celsius since 1850.

Local temperature spikes reported by the media often reflect urban expansion rather than global climate change. Concentrated human activity, buildings, and roads cause cities to warm significantly faster than surrounding rural areas. Professional researchers smooth temperature anomaly maps over distances of 1,200 kilometers or 750 miles to filter out these localized effects. A Washington Post front-page story on August 13, 2019, featured county-by-county maps displaying dramatic warming that was largely driven by local urbanization and energy development rather than global greenhouse gases.

Oceans hold more than 90 percent of the climate system's heat and change over decades to centuries. Because deep ocean data from past centuries is limited, modern analysis relies heavily on robotic floats that descend up to 2 kilometers to measure water columns. The international Argo program deployed a fleet of more than 3,900 robotic drifters that provide yearly temperature and salinity profiles across roughly 60 percent of the world's oceans.

Paleoclimate and Greenhouse Basics

Direct thermometer measurements only became widespread in the mid-nineteenth century. To understand earlier climates, scientists measure temperature-sensitive properties preserved in natural archives, though uncertainties grow larger the further back in time they look. Ice cores drilled in Antarctica and Greenland extract trapped gases and dust layers that provide climate records reaching back almost three million years.

The earth's steady temperature is set by a balance between incoming solar warming and outgoing infrared cooling, a condition known as radiative equilibrium.

According to the Stefan-Boltzmann law, the infrared radiation an object emits increases with its temperature. When the cooling by emitted infrared radiation equals the absorbed solar warming, the planet reaches a steady temperature. Without greenhouse gas insulation, the earth's average surface temperature would be a chilly -18°C or 0°F instead of the actual 15°C or 59°F. Rather than permanently trapping heat, greenhouse gases constantly intercept and impede infrared radiation that must eventually be radiated to space.

Water vapor accounts for over 90 percent of the atmosphere's heat-intercepting ability, though its concentration varies locally. Carbon dioxide has a uniform global concentration and absorbs different infrared colors, meaning even small increases have a noticeable impact. A small percentage change in heat interception produces an apparently outsized temperature change because absolute temperatures are measured on the Kelvin scale anchored to absolute zero. Physicists expect a 1 percent change in heat interception to produce a 1 percent change in temperature. When measured on the Kelvin scale where earth's average temperature is about 288 K, a 3 K rise corresponds precisely to a 1 percent increase.

2. Computing Climate Realities

Climate models are complex computer programs that require extensive tuning of subgrid parameters, leaving them with wide uncertainties and significant disagreements when simulating historical warming.

You can understand climate models as digital laboratories that simulate the atmosphere and oceans using physical laws on a three-dimensional grid. Today's largest scientific computers store 10^17 or 100 million billion numbers and process 10^18 or a billion billion operations per second.

Models divide the surface into grid squares that measure 100 km by 100 km or 60 miles by 60 miles for the atmosphere, and 10 km by 10 km or 6 miles by 6 miles for the ocean. Phenomena like clouds and thunderstorms occur on scales much smaller than these grid boxes. Modelers must therefore rely on subgrid assumptions and tuning parameters to represent these crucial processes.

Researchers at the Max Planck Institute tuned a convection parameter by a factor of ten because the original value produced twice as much warming as observed.

Model Tuning and Ensembles

Tuning climate models by adjusting subgrid parameters to match historical observations introduces subjective choices that can mask errors and predetermine outcomes. Because subgrid phenomena cannot be precisely derived from first principles, modelers adjust numerical parameters until simulation output aligns reasonably well with observed climate features. Fifteen lead climate modelers co-authored a paper highlighting the lack of transparency in model tuning.

UK researchers tuned their model by adjusting how partial snow cover changes the albedo of northern forests and altering how much dimethyl sulfide is produced by ocean microorganisms.

Climate model ensembles show wide disagreements among individual models and fail to reproduce historical warming periods, indicating that climate science is far from settled. When comparing multiple models compiled in projects like CMIP, results differ dramatically from one another and from observations. Newer model generations like CMIP5 and CMIP6 show wider spreads in uncertainty than older ones and fail to replicate the strong observed warming trend from 1910 to 1940.

An analysis of 267 simulations from 29 different CMIP6 models showed they do a very poor job of describing warming since 1950 and underestimate early twentieth-century warming.

Natural Cycles and Emissions

Natural multi-decadal cycles in oceans and the atmosphere strongly influence global temperatures, complicating efforts to attribute recent warming solely to human activities. The climate system exhibits natural repeating behaviors over decades and centuries, such as El Niño events, the Atlantic Multidecadal Oscillation, and the Pacific Decadal Oscillation. Current models struggle to reproduce the timing, strength, and duration of these slow cycles.

The Atlantic Multidecadal Oscillation displays sixty-to-eighty-year cycles inferred from sea surface temperatures in the North Atlantic.

Estimates of equilibrium climate sensitivity to doubled carbon dioxide have remained uncertain for decades, with modern models yielding a wide range of results. Equilibrium climate sensitivity measures how much average surface temperature anomaly increases if carbon dioxide concentrations double from preindustrial levels of 280 ppm. The 1979 Charney Report estimated climate sensitivity between 1.5 and 4.5°C or 2.7 and 8.1°F, with 3°C or 5.5°F as the most likely value. A 2014 IPCC report reverted to this exact same range without providing a most likely estimate, while a 20-author paper published in July 2020 found a likely range of 2.6 to 4.1°C after evaluating 40 models from the CMIP6 ensemble.

3. Weather Extremes and Data Reality

Official observational records show that most extreme weather events, hurricanes, tornadoes, and droughts have not increased significantly, contradicting media reports driven by attribution studies and running-record metrics.

While media reports frequently implicate human-induced climate change for extreme weather disasters, IPCC reports express low confidence regarding global trends in floods, droughts, severe thunderstorms, and extratropical cyclones. Event attribution studies attempt to link individual weather events to climate change after the fact, a methodology that lacks observational testing. The World Meteorological Organization states that any single event, such as a severe tropical cyclone, cannot be attributed to human-induced climate change based on current scientific understanding.

Government climate reports have utilized misleading running-record metrics to create the false impression that US daily high temperature records are increasing rapidly. The 2017 Climate Science Special Report featured a prominent graph showing an increasing ratio of record high to record low daily temperatures. However, this running-records method naturally guarantees a structural decline in the number of all records over time because each new record raises the bar, masking the reality that absolute record highs have not increased.

Professor John Christy reproduced the CSSR running records figure and then conducted an absolute records analysis using 725 US stations, revealing no significant trend in record highs since 1895. The 2016 running records analysis cited by the CSSR used 1,408 stations.

Storms and Tornadoes

Official government climate assessments often feature alarming graphs about hurricane intensity while burying statements from their own text that admit no significant long-term trends exist beyond natural variability. Koonin analyzed the 2014 National Climate Assessment and found that its Key Message highlighted an upward trend in the Power Dissipation Index for North Atlantic hurricanes since 1981. However, buried deep within the text and supported by IPCC reports, the authors acknowledged low confidence in any long-term increase, noting that pre-1949 periods were just as active.

Koonin discovered a direct contradiction when comparing the alarming figures in the main text of the 2014 US National Climate Assessment to Appendix 3, page 769, which explicitly stated there was no significant trend in US land-falling hurricanes.

The 2012 PDI reconstruction reached back to 1880, showing pre-1949 activity.

The apparent doubling of US tornado numbers since 1950 is an artifact of improved weather radar and observation practices rather than an actual increase in storm frequency. Historically, weak tornadoes went unrecorded, especially in sparsely populated areas. Modern weather radar and a shift toward logging weak storms have skewed the data. When focusing only on strong storms (EF1 or greater), there is no upward trend over the past sixty years, and the most destructive storms (EF3 or greater) have actually decreased in frequency.

NOAA records show that 60 percent of recorded US tornadoes are now category EF0 compared to only 20 percent in 1950, proving that the rise in total count comes from counting weaker storms. In the sixty years following 1954, the number of strongest tornadoes (EF3 or above) decreased by 40 percent. NOAA's contiguous US tornado records started in 1950.

Precipitation and Wildfires

Extreme single winter events like Washington, DC's Snowmageddon are often misattributed to climate change, whereas long-term records show declining snowfall totals. While the 2009-10 winter produced a record 28 inches of snow in Washington, DC, historical records indicate a 40 percent drop in annual cold-season snowfall over the 130-year period from 1889 to 2018, aligning with warming temperatures that reduce total snow cover.

During the Snowmageddon winter of 2009-10, Koonin and his wife were trapped in their Maryland home for several days without electricity or heat.

Global precipitation over land shows little detectable or significant positive trend at a global scale, contradicting the simple assumption that a warming globe accelerates the hydrological cycle uniformly. While regional increases have occurred in mid-latitude land areas of the Northern Hemisphere, global satellite and ground observations demonstrate high variability and low confidence in long-term global precipitation trends. Precipitation has, however, become lumpy with heavier single-day downpours in specific regions like the US Northeast.

Satellite observations covering global latitudes between 60 degrees S and 60 degrees N found no detectable and significant positive trends in total global precipitation despite rising global temperatures. The average annual precipitation over the globe is 980 mm, while over land it is 818 mm. Since 1901, global land precipitation anomalies have increased at a rate of 0.2 percent per decade.

Recent multi-year droughts in the American Southwest are mild when placed in historical context alongside centuries-long megadroughts that occurred long before human industrial influences. Tree-ring reconstructions of the Colorado River flow and Palmer Drought Severity Index data reveal that droughts lasting for decades occurred between 900 and 1300 AD due to natural variability. Current human influences appear weak compared to these historical precedents.

The 2009 National Climate Assessment included a 1,200-year reconstruction of the Colorado River flow demonstrating that past droughts were far more severe and longer-lasting than any experienced in the last century.

Global satellite data shows a decline in the area burned by wildfires over recent decades, driven primarily by human agricultural expansion rather than climate change. Although western US forest fires have increased since the 1980s due to fuel aridity and temperature, global satellite monitoring from 1998 to 2015 revealed a 25 percent decrease in annual burned area. This was caused by the conversion of African and South American savannas into farmland, replacing traditional grassland burning with machinery.

Despite severe West Coast fires in 2020, that year was among the least active globally for fires since satellite tracking began in 2003. The share of US wildland fires that have a human cause is 85 percent.

Climate model predictions used by financial and political leaders to project extreme rainfall increases often fail when tested against long-term observational data. Mark Carney, former Bank of England governor and UN climate envoy, warned in a 2015 speech that winter rainfall in England would increase by at least 10 percent based on Met Office model projections. However, rainfall during the subsequent six winters averaged 39 percent less than the 2014 record, remaining entirely consistent with historical norms and driven by natural North Atlantic Oscillation variability.

Carney cited the wet winter of 2014, noted as England's wettest since King George III's reign, to support model-based warnings of worsening floods. The England and Wales rainfall amount during the record winter of 2014 was 455.5 mm, while the average rainfall during the six winters following 2014 was 278 mm.

4. Societal Impacts and Responses

Sea level rise follows historical decadal patterns, weather-related mortality has plummeted, and projected economic impacts from future warming remain minimal, while energy system inertia and demographic realities constrain zero-carbon transitions.

Global Mean Sea Level is rising at about 3 mm per year. Historical tide gauge records demonstrate similar high rates of rise between 1925 and 1940, indicating strong decadal variability driven by natural cycles. Reports like the Climate Science Special Report emphasize total cumulative rise over a century while comparing it to short recent intervals. They ignore the fact that past twenty-five-year periods experienced nearly identical rates of increase.

The Battery tide gauge at the tip of Manhattan has recorded sea level anomalies since 1856, showing a long-term rate of rise of 2.87 mm per year with distinct sixty-year cycles. Global sea level rose by 120 meters since the Last Glacial Maximum 22,000 years ago. The average rate of Global Mean Sea Level rise from tide gauges since 1880 is 1.8 mm per year. Satellites measured an average rate of 3.0 mm per year over twenty-seven years. Sea level rose by 7 to 8 inches since 1900, with about 3 inches occurring since 1993, and 6 cm during the twenty-five-year period between 1935 and 1960.

Despite projections of future temperature-induced mortality spikes, actual historical data shows that weather-related death rates have plummeted over the past century. Data from the Emergency Events database indicates that weather-related death rates fell dramatically as societies developed better storm tracking, flood control, and medical care, despite a global temperature increase. Weather-related death rates are 80 times less frequent today compared to a century ago. Extreme temperatures caused 0.16 deaths per 100,000 each year in the most recent decade. Michael Greenstone testified to Congress predicting an additional 85 deaths per 100,000 by 2100 under high-emissions scenarios, relying on models that omit historical hindcasting checks.

Public health leadership has conflated poverty-driven indoor air pollution deaths with climate change impacts, misleading the public. Tedros Ghebreyesus published an article entitled "Climate Change Is Already Killing Us" that attributed millions of indoor air pollution deaths to human-induced climate change. Ambient and household air pollution cause 100 per 100,000 premature deaths each year, affecting up to three billion people in poor countries.

Global food supplies and crop yields have surged over the past century, driven by agricultural improvements and carbon dioxide fertilization. The New York Times ran a warning headline citing a UN report that climate change threatens the world's food supply, ignoring that per capita food supplies and crop yields have drastically increased since 1960. The global food supply of calories per capita increased by 38 percent since 1961. Global yields of wheat, rice, and corn more than doubled between 1961 and 2011, and 25 to 50 percent of vegetated areas of the globe showed marked greening over four decades.

Published estimates reviewed by the IPCC show that a substantial global temperature rise of up to 3 degrees Celsius by 2100 would impact the global economy by 3 percent or less, representing a minor fractional change in projected long-term growth. Following the release of NCA2018, headlines warned of grim economic consequences, yet the report's underlying figures show that a massive 5 degrees Celsius warming would reduce the US economy by only 4 percent by the end of the century. A 7.2 degrees Celsius warming would cause a 10 percent hit, equating to a five-year growth delay.

Energy Realities and Adaptation

Global efforts to eliminate carbon dioxide emissions face insurmountable demographic, economic, and infrastructure realities. Developing economies like China and India are building numerous coal-fired power stations that will substantially increase their emissions despite Western reduction pledges. The current atmospheric carbon dioxide concentration is 415 ppm, with annual emissions reaching 37 billion tons and a 2 ppm increase in concentration from annual emissions. Fossil fuels supply 80 percent of the world's energy today, while global energy demand is expected to grow by 50 percent through midcentury alongside population growth from 8 billion to over 9 billion.

The Paris Agreement targets are voluntary, nonbinding, and entirely inadequate to achieve the emission cuts required to stabilize temperatures. The UN Emissions Gap Report of 2019 issued a bleak assessment confirming that countries collectively failed to stop the growth in global greenhouse gas emissions. Stabilizing warming below strict temperature thresholds requires total global emissions to vanish within decades. Pledges made under the Paris accord will reduce global emissions by less than 10 percent by 2030, falling drastically short of required zero-emission trajectories.

Energy systems change very slowly because they require immense capital investments, strict reliability, and compatibility across infrastructure. The amount of energy provided by wood today is the same as it was at the time of the Civil War, even though total energy sources have grown enormously. Comparing the energy transition to a Manhattan Project or Moon Shot is misleading because those historical efforts had single customers, secret environments, and unconstrained budgets.

Effective greenhouse gas policies require consistency, significance, focus, systems thinking, technical practicality, and an emphasis on conservation over efficiency. Processing all two hundred million metric tons of global vegetable oil into biodiesel would satisfy just one day of global diesel demand. California recorded 7,000 hydrogen vehicles out of 35 million after 16 years, alongside 44 hydrogen filling stations. You must evaluate climate policies against specific criteria:

• Maintain consistency over decades to give investors long-term confidence.
• Ensure policies are significant enough to compel real changes in emissions.
• Focus policies strictly on reducing emissions rather than diluting them with other goals.
• Apply systems thinking so that changing one part of the energy puzzle does not break another.
• Ground policies in technical practicality and physical constraints.
• Promote conservation over efficiency to ensure total energy use actually declines.

Adaptation will inevitably be humanity's primary response to a changing climate because it is local, proportional, autonomous, and agnostic to the causes of climate change. The author and his family lived in Pasadena, California, where they bolted their house to its foundation, bought earthquake insurance, and drilled their children in earthquake safety amidst 10,000 earthquakes occurring in the southern area each year. You can prepare for environmental hazards with practical steps:

• Bolt houses to foundations and secure shelves to walls.
• Carry hazard insurance.
• Store days of food and water.
• Drill family members in safety procedures.
• Formulate emergency contact and travel plans.

10 Key Ideas

1. Conflating short-term weather fluctuations with long-term climate trends distorts public understanding of environmental data.
2. Four independent analyses show that the global average surface temperature anomaly has risen by about 1 degree Celsius since 1850.
3. Oceans hold more than 90 percent of the climate system's heat and change over decades to centuries.
4. Without greenhouse gas insulation, the earth's average surface temperature would be a chilly -18°C or 0°F instead of the actual 15°C or 59°F.
5. Tuning climate models by adjusting subgrid parameters to match historical observations introduces subjective choices that can mask errors and predetermine outcomes.
6. Official observational records show that most extreme weather events, hurricanes, tornadoes, and droughts have not increased significantly, contradicting media reports driven by attribution studies and running-record metrics.
7. Historical tide gauge records demonstrate similar high rates of sea level rise between 1925 and 1940, indicating strong decadal variability driven by natural cycles.
8. Published estimates reviewed by the IPCC show that a substantial global temperature rise of up to 3 degrees Celsius by 2100 would impact the global economy by 3 percent or less, representing a minor fractional change in projected long-term growth.
9. Public health leadership has conflated poverty-driven indoor air pollution deaths with climate change impacts, misleading the public.
10. Adaptation will inevitably be humanity's primary response to a changing climate because it is local, proportional, autonomous, and agnostic to the causes of climate change.