Loading field map
Climate science is the study of the Earth's climate system: what determines its average state, how it varies naturally on timescales from seasons to ice ages, and how it responds to human and non-human perturbations. It is a physical science rooted in meteorology, oceanography, and physics, but it draws heavily on chemistry, biology, geology, and mathematics. Its central object is not weather—the short-term, chaotic behavior of the atmosphere—but the statistical description of that behavior over decades and longer, together with the slower components of the Earth system that set the boundary conditions for weather.
The field's defining questions are deceptively simple: Why is the climate the way it is? How much can it change, and how fast? What are the mechanisms that drive change? And, increasingly, what will happen to the climate in the future under different scenarios of greenhouse gas emissions? The stakes are high because climate is not a background condition but an active force shaping ecosystems, agriculture, water availability, sea level, and human settlement. Understanding climate is therefore both a fundamental scientific pursuit and a practical necessity for anticipating risks and planning adaptation.
At its core, climate is governed by the Earth's energy budget. The planet receives shortwave radiation from the Sun, reflects a fraction of it back to space (the albedo), and absorbs the rest. To maintain a stable temperature, the Earth must re-emit an equal amount of energy as longwave infrared radiation. The temperature at which it does so depends on how easily that infrared radiation escapes to space. Greenhouse gases—water vapor, carbon dioxide, methane, nitrous oxide, and others—absorb and re-emit infrared radiation in the atmosphere, effectively raising the altitude from which the Earth's thermal radiation escapes. Because the atmosphere is colder at higher altitudes, this raises the surface temperature needed to balance the budget. This is the greenhouse effect, a mechanism first recognized in the nineteenth century by scientists including Joseph Fourier, John Tyndall, and Svante Arrhenius. Without it, the Earth's average surface temperature would be roughly −18°C rather than the observed +15°C.
The greenhouse effect is not a hypothesis but a well-measured physical phenomenon. The scientific challenge lies in quantifying how much the system warms when the concentration of a greenhouse gas changes, and how that warming is distributed across regions, seasons, and components of the climate system. This is complicated by feedbacks. For example, warming melts snow and ice, reducing albedo and causing more absorption of sunlight—a positive feedback. Warming also increases atmospheric water vapor, itself a greenhouse gas—another positive feedback. But warming may also increase cloud cover, which can either cool (by reflecting sunlight) or warm (by trapping infrared) depending on cloud type and altitude. The net effect of clouds remains the largest source of uncertainty in estimating climate sensitivity, the equilibrium warming for a doubling of atmospheric carbon dioxide.
Climate science emerged from two distinct traditions. The first was the empirical study of past climates—paleoclimatology—which grew out of geology and natural history. In the nineteenth century, geologists recognized evidence of past ice ages in glacial deposits and carved valleys, and they debated whether such changes could be explained by variations in the Earth's orbit. The Serbian mathematician Milutin Milankovitch later formalized this idea, showing that cyclical changes in the Earth's orbital eccentricity, axial tilt, and precession alter the seasonal and latitudinal distribution of sunlight. These Milankovitch cycles are now accepted as the pacemaker of glacial-interglacial cycles, though they require amplification by feedbacks involving carbon dioxide, ice, and albedo to explain the full magnitude of temperature change.
The second tradition was the physics of the atmosphere. In the late nineteenth and early twentieth centuries, scientists began to treat the atmosphere as a fluid obeying the laws of thermodynamics and fluid dynamics. The Norwegian school of meteorology, led by Vilhelm Bjerknes, developed the concept of air masses and fronts, while Lewis Fry Richardson made the first attempt to compute weather numerically by hand, laying the conceptual groundwork for computer modeling. The connection between these traditions was made explicit in the mid-twentieth century, when researchers such as Gilbert Plass and Roger Revelle began to calculate the warming effect of rising carbon dioxide from fossil fuel combustion. In 1958, Charles David Keeling began continuous measurements of atmospheric CO₂ at Mauna Loa, providing the first unambiguous record of its rapid increase.
The modern field took shape in the 1960s and 1970s with the development of general circulation models (GCMs)—computer programs that solve the equations of fluid motion and thermodynamics on a global grid, coupled to representations of radiation, clouds, and land surface processes. These models were initially developed for weather forecasting but were soon adapted for climate. By the 1980s, they had become the primary tool for projecting future climate change, and they have grown steadily more comprehensive, now including interactive oceans, sea ice, vegetation, and biogeochemical cycles. The Intergovernmental Panel on Climate Change (IPCC), established in 1988, has since provided periodic assessments that synthesize the state of climate science for policymakers, and its reports have become the authoritative reference for the field's consensus findings.
Climate science is not organized into rival schools in the way that, say, economics or psychology might be. It is a mature physical science with a shared theoretical foundation. However, it does contain distinct research traditions that address different aspects of the climate problem, and understanding their relationship is essential for a working map of the field.
One major tradition is the empirical reconstruction of climate from observations. This includes the instrumental record—thermometers, rain gauges, satellites, and ocean buoys—which extends back roughly 150 years for global coverage and somewhat longer for specific regions. It also includes the much longer paleoclimate record, reconstructed from natural archives such as ice cores, tree rings, corals, lake sediments, and ocean sediments. Ice cores from Antarctica and Greenland preserve trapped air bubbles that allow direct measurement of past atmospheric CO₂ and methane concentrations, as well as isotopic ratios that indicate temperature. These records show that atmospheric CO₂ over the past 800,000 years has oscillated between roughly 180 and 300 parts per million, closely tracking temperature through glacial cycles, and that the current concentration, over 420 parts per million, is unprecedented in that record.
The role of this tradition is to establish what the climate has actually done, to identify the mechanisms that drove past changes, and to provide targets and constraints for models. Its limitation is that the past is an imperfect analogue for the future: boundary conditions such as the configuration of continents, the presence of ice sheets, and the rate of change differ from today. Nevertheless, paleoclimate evidence is crucial for testing whether models can reproduce large, slow changes that are not visible in the short instrumental record.
A second tradition focuses on understanding the physical, chemical, and biological processes that govern the climate system through theory and targeted observation. This includes radiative transfer—how radiation interacts with gases, aerosols, and clouds—as well as the dynamics of atmospheric and oceanic circulation, the thermodynamics of sea ice, and the biogeochemical cycles of carbon, nitrogen, and sulfur. Researchers in this tradition often work with simplified models, laboratory experiments, or field campaigns designed to isolate a particular process. For example, the study of cloud microphysics asks how aerosol particles serve as cloud condensation nuclei, how droplets form and grow, and how this affects cloud reflectivity and precipitation. This process-level understanding is the foundation on which comprehensive models are built, and it is also where many of the field's persistent uncertainties reside, particularly around clouds and aerosols.
The third major tradition is numerical modeling, which integrates process understanding into a coherent, quantitative representation of the whole climate system. Modern Earth system models couple the atmosphere, ocean, land surface, cryosphere, and carbon cycle, and they are run on supercomputers at horizontal resolutions of tens to hundreds of kilometers. They are used for several distinct purposes: to simulate the historical climate and test whether models can reproduce observed trends; to attribute observed changes to specific causes, such as greenhouse gases versus solar variability or volcanic eruptions; and to project future climate under different emission scenarios.
Modeling is not a single approach but a hierarchy. At one end are simple energy balance models that treat the Earth as a single box or a few boxes and compute its temperature from radiative forcing and a prescribed sensitivity. At the other end are comprehensive Earth system models. In between are intermediate-complexity models that resolve some processes but not others, and regional models that downscale global simulations to finer detail. The hierarchy is useful because simpler models are transparent and fast, allowing exploration of ideas that would be computationally prohibitive in full models, while complex models are more realistic and can represent interactions that simple models cannot.
The relationship between these traditions is complementary rather than competitive. Observational and paleoclimate research provides the data that models must reproduce; process studies provide the parameterizations—simplified representations of sub-grid-scale processes—that models need; and models, in turn, generate predictions that motivate new observations and experiments. Disagreements within the field are typically about the magnitude of specific effects or the best way to represent an uncertain process, not about the reality of the greenhouse effect or the basic physics of the system.
A fourth tradition, which has grown in importance since the 1990s, is detection and attribution. This is the statistical and physical discipline of determining whether a change in climate has occurred (detection) and whether that change can be ascribed to a particular cause, such as human greenhouse gas emissions, as opposed to natural variability (attribution). The approach typically compares observed changes in temperature, precipitation, or other variables with the patterns simulated by models under different forcing combinations. Because greenhouse gases, aerosols, solar variability, and volcanic eruptions each produce distinct spatial and temporal fingerprints, it is possible to separate their contributions with high confidence for some variables—most notably global mean temperature—and with lower confidence for others, such as regional precipitation. This tradition is the bridge between the physical science of climate and the societal questions of responsibility and adaptation.
A proper understanding of climate science requires a clear-eyed view of uncertainty, because the field is often misrepresented as being more or less certain than it is. There are several distinct types of uncertainty, and they behave differently.
The first is uncertainty about the current state: how much has the climate already warmed, and how much of that is due to human activity? This is well constrained. The instrumental record shows a global mean warming of about 1.1°C since the late nineteenth century, and attribution studies consistently find that most of this warming is due to greenhouse gas increases, with natural factors playing a minor role. The basic finding that the planet is warming and that humans are the dominant cause is not seriously disputed within the scientific community.
The second is uncertainty about the future: how much will the climate warm for a given emission pathway? This depends on climate sensitivity, which is not known exactly. The IPCC's best estimate for the equilibrium warming from a doubling of CO₂ is about 3°C, with a likely range of 2.5 to 4°C. The uncertainty arises primarily from cloud feedbacks and from the behavior of the carbon cycle—how much of the emitted CO₂ will be absorbed by the oceans and land, and how much will remain in the atmosphere. This uncertainty is irreducible in the sense that it reflects genuine gaps in understanding, but it is bounded: even the low end of the range represents substantial warming, and the high end represents severe risk.
The third is uncertainty about regional and local impacts. Global mean temperature is a useful summary statistic, but it does not tell a farmer in the Sahel or a city planner in Jakarta what will happen to rainfall, heat waves, or sea level in their specific location. Regional projections are less certain than global ones, partly because the models are coarser than the scales at which impacts are felt, and partly because regional changes are influenced by atmospheric circulation patterns that are harder to predict. This uncertainty is not a reason for inaction but a reason for robust planning that performs well across a range of possible futures.
The contemporary field is characterized by several converging developments. Earth system models are becoming more comprehensive, incorporating ice sheet dynamics, permafrost carbon, and dynamic vegetation, which allows them to address questions about tipping points—thresholds beyond which changes become self-sustaining and potentially irreversible. Examples include the collapse of the West Antarctic Ice Sheet, the dieback of the Amazon rainforest, and the release of methane from thawing permafrost. Whether such thresholds exist, and where they lie, is an active area of research with high stakes.
Another major development is the use of machine learning and statistical emulation to accelerate modeling, to improve the representation of sub-grid-scale processes, and to analyze the vast output of model ensembles. These tools are not replacing physical understanding but are being integrated with it, often by learning the behavior of expensive models and then using the learned approximation to explore parameter space more thoroughly.
The field is also increasingly engaged with the social and policy dimensions of climate change. This includes the study of climate impacts on agriculture, health, and infrastructure; the development of adaptation strategies; and the analysis of mitigation pathways—the combinations of emission reductions, carbon removal, and land-use changes that would limit warming to specific targets such as 1.5°C or 2°C above pre-industrial levels. This work is not a separate discipline but an extension of climate science into the human systems that interact with the physical climate, and it requires collaboration with economists, engineers, and social scientists.
Finally, the field is marked by an unusual degree of public visibility and contestation. Because the findings of climate science have profound economic and political implications, they have been subjected to organized campaigns of doubt and misrepresentation. Scientists have responded by emphasizing transparency, reproducibility, and the explicit communication of uncertainty. The result is a field that is, by the standards of science, unusually careful about stating what is known, what is uncertain, and what is simply unknown—an epistemic posture that is both a strength and, at times, a source of public confusion.
Climate science is not a finished edifice but a working enterprise. Its foundations—the physics of the greenhouse effect, the reality of past climate change, the influence of human activity on the present climate—are secure. Its frontiers—the precise magnitude of future warming, the behavior of clouds and ice sheets, the regional expression of global change—remain open. The field's enduring contribution is to have transformed the climate from a backdrop of human affairs into an object of rigorous, quantitative study, and in doing so to have made the future a subject of scientific inquiry rather than mere speculation.