Climate, Grounded
What is measured, what is modeled, and what is solvable at what cost - so a reader can place any climate headline rather than just react to it.
(single years run hotter: 2024, the warmest on record, was 1.55°C)
(up from about 280 ppm before industrialization and 320 ppm in 1965)
(roughly six times higher per American)
A note on framing. Climate headlines often mix measurements, model projections and policy arguments. The page below walks through what is well-measured (temperature, CO2, sea level), what is modeled (regional impacts, future trajectories), and what is contested even among careful scientists (specific tipping points, exact damages, the right policy response). The aim is to make it easier to read any climate headline with a sense of where it sits in the larger picture.
What is measured
Several pieces of climate data are well-measured, well-replicated, and not seriously disputed by anyone in the relevant scientific fields. These are the foundation of everything else, and a reader who is not sure where to anchor should start here.
Atmospheric CO2 has risen sharply. Direct measurement at Mauna Loa Observatory in Hawaii has tracked atmospheric CO2 concentration since 1958. The level has risen from about 315 parts per million then to about 430 today. Ice cores and older geological records show that the current level is the highest in at least two million years. The Intergovernmental Panel on Climate Change (IPCC) attributes the rise unequivocally to human activity, chiefly fossil-fuel burning, which emitted a record 38.1 billion tonnes of CO2 in 2025 (Global Carbon Project). This is not modeled - it is measured directly.
Global average temperature has risen. Multiple independent temperature datasets (NASA, NOAA, the UK Met Office, Berkeley Earth, Japan Meteorological Agency) all show the global average has risen by roughly 1.4 degrees Celsius since the late 1800s. The agreement between datasets that use different methods and different stations is the strongest single piece of evidence for the warming trend. The IPCC's 2021 assessment found human influence on the warming "unequivocal" and responsible for essentially all of it. It is worth separating two numbers that are often merged. Individual years have already run hotter than the long-term level: the World Meteorological Organization put 2024 at 1.55°C, 2025 at 1.44°C and the 2023-25 average at 1.48°C. With a very strong El Niño developing, August 2026 tied July 2023 as the warmest month on record (Copernicus). The Paris thresholds refer to the long-term level rather than to any single year, which is why "we passed 1.5°C" headlines and "we have not passed 1.5°C" corrections can both be defensible.
Sea level has risen. More than 20 centimeters (8 inches) globally since 1900, with the rate accelerating: about 11 centimeters since 1993 alone. The causes are thermal expansion (water expands when warmed) and melting of glaciers and ice sheets. Tide gauges and satellite altimetry both show the trend; they agree closely.
Polar ice has shrunk. Arctic sea ice has lost roughly 40% of its summer extent since 1979 (when satellite measurement began). The Greenland and Antarctic ice sheets are both losing mass, Greenland faster. Mountain glaciers worldwide are mostly retreating. Each of these is directly observed, not modeled.
Some extreme events have become more frequent. Heat waves, intense rainfall events, certain types of drought, and certain wildfire conditions have all become measurably more common in the last three to five decades. Linking a specific event to climate change is harder and more disputed. A 2016 US National Academies review found confidence highest for heat and cold extremes, lower for droughts and heavy rain, and little or none for thunderstorms and mid-latitude storms; the IPCC's 2021 assessment says the evidence has strengthened since.
Who emits what
The global story is one of large differences across countries, both in current emissions and in cumulative historical contribution. The charts below show World Bank figures, drawn from the European Commission's Emissions Database for Global Atmospheric Research.
US per-capita emissions have fallen by about a third since 2000, mainly because less coal is burned for electricity - replaced first by cheap shale gas and more recently by wind and solar - and partly because of efficiency improvements. Australia, Canada, Russia and Saudi Arabia now emit about as much or more per person.
Chinese per-capita emissions have risen sharply since 2000 and now exceed the European Union average. China is by far the largest emitter in absolute terms but per person remains below the United States. Whether Chinese emissions have peaked is one of the most-watched questions in climate policy. The Global Carbon Project estimated a rise of about 0.4% in 2025; analysts at the Centre for Research on Energy and Clean Air see a plateau since a peak in March 2024.
Indian per-capita emissions are still less than half the world average and roughly a sixth of US per-capita emissions. India argues that the developed countries' historical emissions have already used most of the available "carbon budget," and that asking poor countries to cap their emissions before they have developed is unfair. Others reply that the warming still to come depends on emissions from here on, and India is now the third-largest emitter.
How countries compare
The per-capita differences are enormous, and the political conversation about who should do what depends partly on which framings of fairness one accepts. The numbers below are 2024 per-person CO2 emissions in tonnes (World Bank), plus rough shares of cumulative emissions since 1750 (Global Carbon Project).
The takeaway: the highest per-person emitters today are petrostates and resource-rich economies such as Saudi Arabia, Australia, Canada and Russia, with the United States close behind. The largest absolute emitter is China. Sub-Saharan Africa combines some of the greatest exposure to climate damage with the smallest historical responsibility. The political fight about who should pay for what flows directly from this asymmetry, and the asymmetry is not easily resolved by any clean framing.
What is modeled (and what depends on the model)
Beyond the directly measured data, much of what is said about climate depends on models - simulations of how the climate system will respond to different levels of CO2, how regional impacts will play out, and how much damage at what cost. Models are not arbitrary; they are tested against historical data and have improved over decades. But they are not direct measurements, and the further into the future they project, the wider the uncertainty bands become.
Some things that climate models predict with reasonable confidence:
- Temperatures will continue to rise as long as net emissions stay positive. The amount of warming per amount of cumulative CO2 emissions is reasonably well-constrained.
- Some regions will warm faster than others. The Arctic is warming roughly three to four times the global average, depending on the dataset and time window. Land warms faster than ocean.
- Heavy precipitation events will become more frequent in many regions, even where total annual rainfall does not change. Warmer air holds more water vapor.
- Sea level will continue to rise. The pace depends on how Greenland and Antarctica respond, which is one of the major uncertainties.
Some things models predict with much less confidence:
- The exact timing and magnitude of any tipping points (Amazon dieback, Greenland collapse, ocean circulation changes). These are real risks; the probability of crossing them is not well pinned down.
- Regional impacts at fine scale. A model may correctly predict that Mediterranean droughts intensify on average without correctly predicting which year, which province, or which crop is hit.
- Total economic damages at any given temperature. Estimates range from a few percent of GDP to severe disruption, with serious researchers across the range.
- The pace of social and technological adaptation, which would partly offset the physical impacts.
A reader confronting a climate headline can usefully ask: is this from a measurement (high confidence), a near-term model (moderate confidence), or a distant-future model (significant uncertainty)? Each is useful information; treating all three as equally certain produces both panic and dismissal.
The paths from here
The world is not on the high-emissions path that some 2010s scenarios assumed. It is also nowhere near the path needed to limit warming to 1.5 degrees; at current emissions, the remaining carbon budget for that limit runs out in about three years. The realistic trajectories sit in a middle band where the question is not whether warming continues but how much, how fast, and with what response.
Current trajectory: about 2.5 to 3°C of warming by 2100
If the world keeps the policies it has actually implemented (not its more ambitious pledges), most modeling suggests it is heading for 2.5 to 3 degrees of warming by 2100. The UN Environment Programme (UNEP) put it at 2.8 degrees in 2025. This is well above the international goal of 1.5-2 degrees but well below the worst-case 4-5 degree scenarios that some 2010s analyses assumed.
Will it happen? This is the base case. It assumes continued steady but insufficient policy action, continued cost declines for renewable energy, and no major surprises (positive or negative). The United States withdrew from the IPCC and left the Paris Agreement in January 2026, and leaves the UN climate convention in February 2027. The 2026 Hormuz oil shock may cut global fossil emissions by around 0.5%, through prices rather than policy (Carbon Brief, from International Energy Agency forecasts). It is neither a disaster scenario nor a success scenario - it is what the data currently extrapolates to.
Faster transition reduces warming to about 2°C
Renewable energy scales faster than expected, electric vehicles displace gasoline cars more quickly, building electrification proceeds, and a few major policies (carbon pricing, methane regulation, deforestation controls) deliver more than expected. Warming peaks around 2 degrees and declines slowly through net-zero policies.
Will it happen? Possible. The pace of clean-energy cost decline has consistently been faster than even optimistic forecasts. Whether that pace continues, and whether grid investment, materials supply, and political will keep up, are the open questions. At the UN climate conference in Belém (COP30) in November 2025, 82 countries backed a roadmap away from fossil fuels; the final text omitted it, and the COP30 presidency is drafting one outside the talks for COP31 in Antalya, Turkey, in November 2026. This path is consistent with the current direction if it accelerates.
Adaptation becomes the central frame
As it becomes clear that the world will not hold warming to 1.5 degrees, attention shifts from emission cuts (still important) to adaptation (more directly survival-focused). Investment in heat-resistant infrastructure, flood defenses, drought-tolerant agriculture, climate-resilient health systems, and migration management grows substantially.
Will it happen? Already partly happening. UNEP now expects long-term warming to exceed 1.5 degrees at least temporarily, and COP30 called for tripling adaptation finance by 2035. But UNEP puts developing countries' needs by then at over $310 billion a year, about 12 times the international public finance they received in 2023. The political conversation is starting to shift but lags the physical reality.
Carbon removal scales up
Direct air capture, enhanced rock weathering, ocean alkalinity, biochar, and other carbon-removal technologies move from research and pilot stage to commercial deployment. By the 2040s, removing CO2 from the air becomes a meaningful supplement to the slower task of emission reduction. Costs fall enough that gigaton-scale removal becomes affordable.
Will it happen? Uncertain. The technology is real but still extremely expensive, and novel methods such as direct air capture remove only about 2 million tonnes of CO2 a year (State of Carbon Dioxide Removal report, 2026). Whether the cost-decline curves of solar and wind can be repeated for direct air capture is the central question. Large-scale deployment is at least a decade away.
Tipping points are crossed
One or more of the climate-system tipping points (Amazon transition, Atlantic ocean circulation slowdown, large-scale ice loss in Antarctica or Greenland) is crossed in the next thirty to fifty years. The consequences are large, regionally specific, and difficult or impossible to reverse on human time scales.
Will it happen? The probability is not well-constrained. The Global Tipping Points Report 2025 found warm-water coral reefs already passing their tipping point and the Atlantic circulation at risk of collapse below 2 degrees; the IPCC's 2021 assessment judged, with medium confidence, that the circulation will not collapse abruptly before 2100. Some serious researchers think the risks are exaggerated. It is a real possibility worth taking seriously without being sure exactly how likely it is.
The most likely warming range by 2100 is now roughly 2 to 3 degrees - a serious outcome but not a civilization-ending one. Adaptation will rise as a parallel agenda alongside emission cuts. Carbon removal will scale slowly. Tipping-point risks remain uncertain. Climate will become more politically central, not less, regardless of whether one finds that welcome.
Where serious analysts disagree
Climate is one of the topics where the loud public debate is often quite different from the careful debate among researchers. The named voices below represent different positions held by serious people whose work is worth engaging directly.
The mainstream IPCC framing is roughly right
The Intergovernmental Panel on Climate Change is the largest scientific consensus exercise in modern science. It is conservative by design and its summaries reflect a careful averaging of many lines of evidence. The headline findings - that human emissions are causing warming, that warming is producing measurable impacts, and that limiting warming requires substantial emission cuts - are well-established and worth treating as the baseline for serious policy discussion.
Held by: the IPCC's working groups, the major climate science institutions worldwide, and the substantial majority of working climate scientists.
The clean-energy transition is going faster than official forecasts assumed
Solar, wind, batteries, and electric vehicles have all consistently scaled faster than even the optimistic scenarios projected. The rate of cost decline has been remarkable. The world is moving toward a decarbonized energy system faster than the official climate models incorporate, and the trajectory of warming is consequently better than the pessimistic scenarios suggest.
Held by: Hannah Ritchie (Our World in Data), Kingsmill Bond (RMI, an energy think tank), and other analysts who track deployment data. Their data on cost declines and scaling is solid; the open question is whether the necessary supporting infrastructure and political will keep up with the technology.
Adaptation matters more now than further mitigation
Given that the world is locked into substantial warming over the next several decades regardless of any plausible emission cuts, the marginal benefit of more aggressive mitigation is smaller than the marginal benefit of investing heavily in adaptation - flood defenses, drought-tolerant agriculture, heat-resilient cities. The current allocation between mitigation and adaptation is unbalanced toward mitigation.
Held by: Roger Pielke Jr. (American Enterprise Institute, previously the University of Colorado), parts of the development-economics community, and a strand of climate-policy thinking. Critics reply that every increment of warming adds damage and that adaptation has limits, which the IPCC says some regions and ecosystems have already reached.
The catastrophic-risk case is underweighted
The strong version of the alarmed reading argues that the mainstream IPCC framing is conservative by construction and systematically underweights tail risks. These include tipping elements that could cascade (Amazon dieback, ocean circulation change, large-scale ice loss, permafrost feedback) and warming trajectories above 3 degrees that produce disruption beyond current policy planning. They also include second-order effects on food systems, migration, and political stability that climate models do not fully capture. On this account, current policy is calibrated to a central estimate and is leaving the long tail of bad outcomes inadequately hedged.
Held by: James Hansen (formerly NASA, whose 1988 testimony first put climate on the public agenda), Tim Lenton (Exeter) and the broader Earth-system tipping-point research community, Kevin Anderson (Tyndall Centre), and journalists like David Wallace-Wells. His 2017 essay "The Uninhabitable Earth" was faulted by climate scientists for dwelling on worst cases. The position's strength is the literature on tipping elements and tail risk; the IPCC itself says low-likelihood outcomes such as ice-sheet collapse cannot be ruled out. Critics, including the climate scientists Zeke Hausfather and Glen Peters, reply that the worst-case emissions scenarios have become implausible and that presenting them as business as usual misleads policy.
Impacts are likely smaller than the headlines suggest, and policy costs higher
The "lukewarm" reading accepts that human emissions are warming the planet and that the warming has measurable effects. But it argues that the most-cited damage estimates are sensitive to assumptions that may be too pessimistic, and that aggressive near-term mitigation policies often have costs that are themselves substantial. It adds that humanity has historically adapted to climate variability quite effectively given enough time and capital. The implication is not "do nothing" but "do less than the maximalist policy proposals, focus on adaptation, and let cheap clean technology do most of the work as it gets cheaper."
Held by: Bjørn Lomborg (Copenhagen Consensus), Steven Koonin (NYU, formerly Department of Energy), parts of the energy-economics community, and a strand of policy thinking centered around think tanks like the Breakthrough Institute. The position is contested by mainstream climate scientists. It is sometimes used by less-careful commentators to dismiss the underlying science, which the named figures themselves explicitly do not. The empirical case for "lukewarm" is weaker on the physical science (where the IPCC consensus is durable) and stronger on the economic and policy questions, where reasonable estimates do diverge.
Climate framing has become entangled with political agendas on both sides
Climate is a real physical problem; the conversation about climate has become entangled with political agendas about energy, lifestyle, growth, and identity that are partly separate from the science. Both alarmist and dismissive coverage exaggerate in ways that make it harder for serious citizens to engage with the actual choices. A more honest conversation would separate the physical science (well-established), the economic estimates (uncertain), and the political preferences (contested) more clearly.
Held by: Mike Hulme (Cambridge), Ted Nordhaus (Breakthrough Institute), and a broader tradition of climate-science-and-society scholars. Critics such as the historians Naomi Oreskes and Erik Conway reply that the distortion has not been symmetrical: their book "Merchants of Doubt" (2010) traces organized campaigns to cast doubt on well-established science.
Reasonable people who have followed the science closely sit anywhere from "this is among the most consequential threats humanity faces" to "real but manageable, and policy ambition has overshot the underlying economics." That range is not a sign that one side is uninformed. It reflects real uncertainty about tipping-point thresholds, economic damage estimates, technology trajectories, and how to weigh long-tail risks against present-day costs of action. On the available evidence, the physical science is well-established and worth taking seriously. The economic and political response is contested even among careful analysts. The technology trajectory is more positive than the headlines suggested a decade ago but still insufficient on current emission paths, and adaptation is rising in importance and remains under-funded relative to need. The tails of the distribution - the catastrophic outcomes some serious researchers warn about, and the more manageable scenarios others find more likely - both deserve to remain inside the conversation rather than being dismissed as fringe.
What this means for you
Climate sits in the background of decisions about housing, retirement, career, business, and politics. The trajectory is slow enough that no single year requires panic and steady enough that no single year is irrelevant either. A few practical observations:
If you are buying a home or property
Climate exposure is now a real input to property value over a 30-year mortgage. Coastal flooding, wildfire risk, prolonged heat waves, and water-stress trends should be checked alongside the standard variables before signing a long mortgage. Insured catastrophe losses topped $100 billion again in 2025 (Swiss Re: $107 billion). Reinsurers differ on the main driver: Swiss Re stresses growing exposure, Munich Re the warming climate. Several US insurers have already pulled out of high-risk markets; that pattern is likely to spread. A property in a climate-resilient location is a different kind of investment from one in an exposed area, and the difference is widening.
If you are choosing where to live for the long term
Some regions are clearly better positioned for the next 30-50 years than others. Areas with reliable freshwater, moderate heat, low flood exposure, and good infrastructure for the warming climate are likely to do better: Northern Europe, parts of the Pacific Northwest, parts of the Great Lakes region in the US, parts of Canada. Coastal Southern Europe, much of the American Southwest, parts of South Asia, and low-lying coastal areas worldwide are likely to do worse. Long-term residence and investment decisions can factor this in.
If you invest
Climate-related opportunities and risks are now embedded in most major asset classes. Companies in the energy transition (renewable hardware makers, grid operators, electric-vehicle suppliers, certain utilities) have long-term tailwinds. Companies heavily exposed to fossil-fuel demand or to climate-vulnerable infrastructure face long-term headwinds. Climate-related insurance, real estate, and agricultural exposures are increasingly being repriced. None of this is investment advice; it is observing where the underlying physical and political realities point.
If you are thinking about your own emissions
Personal emission reductions matter symbolically and at the margin. The largest individual moves are usually where you live, how often you fly (the highest-impact discretionary choice for most people who fly), and what you eat (less so than is sometimes claimed but real). In one large US study, households in city cores averaged about 40 tonnes of greenhouse gases a year, those in outer suburbs about 50. Most other personal decisions are smaller in impact than the mainstream conversation suggests. If you want to do something useful, focus on the few large levers and on supporting policy at the political level.
If you are voting and engaging politically
The most consequential climate choices are made by governments, not by individuals. Voting on climate policy with attention to the specific proposed actions (carbon pricing, transmission permitting reform, building electrification, vehicle standards, climate finance for developing countries) is more useful than voting on rhetoric about climate generally. The politicians who will deliver on this are not always the ones who talk about it the most.
The mechanics behind this
The climate story sits on top of three deeper mechanisms covered elsewhere on this site. If the analysis above depends on ideas you want to understand first, these fundamentals make the conversation more legible:
