Joseph Fourier’s 1824 memoir on the temperatures of the Earth and planetary space marks the first rigorous scientific articulation of what is now called the greenhouse effect, even though neither the term nor a fully correct quantitative theory yet existed. In this work, Fourier framed planetary temperature as an energy-balance problem, calculated that the Earth ought to be far colder if heated only by incoming solar radiation, and therefore inferred that the atmosphere must act as a warming envelope that changes how heat is exchanged between the surface and space. Drawing on Horace-Bénédict de Saussure’s solar oven experiments, Fourier recognized that air and glass can be transparent to sunlight yet relatively opaque to the “dark heat” (infrared radiation) emitted by warmed surfaces, and he generalized this insight to the Earth’s atmosphere. His key conceptual advance lay in distinguishing heat in the state of light from non-luminous heat, arguing that the atmosphere admits the former more readily than the latter and thereby elevates surface temperature. Although he misattributed some of the unexplained warmth to hypothetical interstellar radiation and did not identify specific greenhouse gases, Fourier established the basic physical picture later refined by Pouillet, Tyndall, and Arrhenius. While Postma’s 2015 note confirms Fourier’s discovery is analogous to the modern concept, it also finds this analogy paradoxical when set against the modern definition of heat.
John Tyndall’s mid-nineteenth-century experiments established the fundamental principle that different gases absorb and emit radiant heat to vastly different degrees. Using a long tube fitted with rock-salt windows, a thermopile, and a galvanometer, Tyndall compared a reference beam of heat against a beam transmitted through a gas sample, measuring the relative reduction in radiant heat caused by each gas. He reported in 1859 that carbon dioxide and water vapor were exceptionally strong absorbers of infrared radiation, while oxygen, nitrogen, and hydrogen were nearly transparent. Water vapor proved the most powerful absorber among all atmospheric constituents tested. These were laboratory measurements of bulk absorption properties, not direct molecular absorption coefficients in the modern spectroscopic sense. A century and a half later, precise spectroscopic studies have confirmed and expanded Tyndall’s insights. For instance, laboratory measurements of CO₂+H₂O mixtures between 2400 and 2600 cm⁻¹ show that the H₂O-induced continuum absorption of CO₂ is significantly larger than pure CO₂ absorption, confirming that this mixture must be accounted for in radiative transfer calculations. Astrophysical observations toward young stellar objects in the Large Magellanic Cloud have detected H₂O and CO₂ ices spectroscopically, deriving a CO₂/H₂O ice ratio of 0.45, higher than in Galactic sources, suggesting environmental influences on ice composition. Laboratory work on mixed H₂O:CO₂ ices demonstrates that CO₂ alters H₂O band profiles and strengths, including the appearance of a free-OH stretching band near 2.73 μm, making it crucial to constrain ice environments for correct spectral assignments.
Svante Arrhenius, in his landmark 1896 paper, produced the first quantitative estimate of climate sensitivity by developing a hand-calculated radiative balance model. His method was not a modern general circulation model but a single-layer, energy-balance style calculation that treated the atmosphere as a simplified absorber of infrared radiation. Arrhenius examined six discrete scenarios for atmospheric CO₂ levels—0.67, 1.0, 1.5, 2.0, 2.5, and 3.0 times the concentration of his day—and computed the resulting temperature change at different latitudes, averaging the outcome globally. He concluded that doubling or halving CO₂ would change Earth's temperature by roughly 5–6°C, which stands as the first widely cited climate sensitivity estimate. A crucial element of his calculation was the inclusion of a water-vapor feedback; Arrhenius recognized that warmer air would hold more moisture, amplifying the initial warming. Later analysis based on multiple physical constraints, including the fractal climate response function introduced by van Hateren, yields a millennial-scale equilibrium climate sensitivity for doubled CO₂ of 2.0 ± 0.3 K, with a transient climate response of 1.5 ± 0.2 K. Paleoclimate evidence from glacials and interglacials further constrains the fast-feedback sensitivity to 3 ± 1°C per 4 W/m² of CO₂ forcing, with the slower, full Earth-system sensitivity being even larger due to feedbacks involving ice sheets and atmospheric CO₂ itself.
Early in the 20th century, chemists attempted to measure atmospheric CO₂ using wet-chemistry methods, but the results were too imprecise and variable to establish a global baseline or reliably detect any long-term increase. The foundational measurements came from Brown and Escombe in 1905, who used a method involving caustic sodium hydroxide and double titration; their data were later used by G. S. Callendar to estimate a turn-of-the-century level of 274 ± 5 ppmv. Recognizing the limitations of these early data, Callendar refined his estimate after incorporating additional datasets, arriving at a late-19th-century background of 288–290 ± 3 ppmv. Despite these efforts, the inherent flaws in the methods—poor precision, vulnerability to contamination, and a tendency to reflect local or regional rather than global background air—meant that surveys of the first half of the century showed measurements ranging widely, from the high 200s to 400–500 ppm. Consequently, scientists of that era could not confidently detect an upward trend; it was thought that a clear increase might take 100 years or more to discern. These early, imperfect data nonetheless provided the historical anchor points that modern scientists use to compare with the precise Mauna Loa record, establishing that the industrial-era rise in CO₂ has been rapid and unequivocal.
The foundational science of anthropogenic climate change was established through the coordinated work of Roger Revelle, Hans Suess, and Charles David Keeling. In 1957, Revelle and Suess published a landmark paper arguing that the ocean would absorb fossil-fuel carbon dioxide much more slowly than previously assumed, because vertical mixing into deeper waters is limited. This meant that human CO₂ emissions should accumulate in the atmosphere faster than natural carbon sinks could remove them. Recognizing the need for direct observation, Revelle recruited Keeling to the Scripps Institution of Oceanography to measure atmospheric CO₂ with high precision. Keeling had developed an instrument capable of very accurate readings, and Revelle selected the remote Mauna Loa Observatory in Hawaii to avoid local pollution, ensuring the data would reflect background atmospheric change. Keeling began continuous measurements in 1958. His record, now known as the Keeling Curve, showed that atmospheric CO₂ was rising steadily year after year, superimposed on a regular seasonal cycle. This provided the first clear observational evidence that natural carbon cycling was not keeping pace with human emissions, directly confirming the Revelle–Suess hypothesis. The causal chain is thus: Revelle and Suess framed the ocean uptake problem, Revelle sponsored the measurement program, and Keeling produced the observed record that proved CO₂ was accumulating. The Keeling Curve remains the most iconic empirical demonstration that human activities are altering the global atmosphere.
Syukuro Manabe developed the first general circulation model (GCM) for climate by progressively adding physical processes to an existing atmospheric model at the U.S. Weather Bureau’s General Circulation Research Section. Between 1959 and 1963, director Joseph Smagorinsky recruited Manabe to code the GCM, and by 1963 the team had completed a nine-level, hemispheric primitive-equation model. Manabe first constructed a one-dimensional radiative-convective equilibrium model in the early 1960s as a stepping stone. By 1965, he had produced a three-dimensional global model with nine vertical levels, incorporating radiative transfer, convection, moisture, and hydrologic-cycle physics to study how changing CO₂ affected surface temperature, including water-vapor feedback. In the late 1960s, Manabe extended this into a coupled atmosphere–ocean system with Kirk Bryan, creating the first coupled GCM capable of simulating ocean dynamics. Manabe thus transformed a weather-prediction model into a full climate model by systematically integrating radiation, convection, moisture, and later ocean coupling. These pioneering simulations established the framework for modern climate modeling, demonstrating that increasing CO₂ warms the surface through greenhouse trapping, amplified by water-vapor feedback. Manabe’s work remains foundational for understanding climate sensitivity and the physical basis of anthropogenic global warming.
The Charney Report, published in 1979 by the U.S. National Academy of Sciences, produced the first formal, quantitative assessment of how Earth’s climate would respond to a doubling of atmospheric carbon dioxide. Its central finding was that such a doubling would most probably warm the planet by about 3°C, with a likely error range of 1.5°C to 4.5°C. This estimate defined the canonical range for equilibrium climate sensitivity. The report concluded that known imperfections in the climate models of the era were not serious enough to overturn the expectation of appreciable, long-term warming. It identified feedbacks involving clouds, relative humidity, and lapse rates as the primary sources of uncertainty driving the wide sensitivity range. The report also noted that the ocean’s thermal inertia would slow the rate at which that warming became observable. The 1.5°C to 4.5°C range proved remarkably durable, remaining close to the values used by later Intergovernmental Panel on Climate Change assessments for decades. By framing CO₂ increases as a driver of meaningful, non-negligible warming, the Charney Report shaped both the scientific framing of global warming and the subsequent policy conversation around it.
The Vostok ice cores, by extracting ancient air trapped in Antarctic ice and pairing it with temperature reconstructed from the ice’s isotopic composition, directly demonstrated that CO₂ and temperature closely covaried across four glacial–interglacial cycles spanning 420,000 years. Warm periods consistently held higher CO₂; cold periods held lower CO₂. A 2001 Nature analysis, applying a deuterium-excess correction, raised the correlation coefficient between CO₂ and reconstructed Antarctic temperature to r² = 0.89 for the past 150,000 years and r² = 0.84 for the interval 350,000 to 150,000 years ago. The rise of CO₂ at deglaciation was either nearly in phase with Antarctic warming or lagged it by less than roughly 1,000 years. This tight coupling—repeated across multiple cycles—provided strong evidence that CO₂ acted as both a climate forcing and a feedback within glacial-interglacial oscillations, not that the ice core alone proved causation. Hansen and Sato (2011) drew on the full paleoclimate record to argue that Earth in warmest interglacials of the past million years was less than 1°C warmer than the Holocene, implying that polar amplifying feedbacks can now be triggered by moderate warming. They conclude that goals limiting human-made warming to 2°C are insufficient, as ice-sheet disintegration is nonlinear and satellite gravity data, though too brief for certainty, are consistent with a doubling time for mass loss rate of 10 years or less. The evidence suggests that preserving a planet resembling the one on which civilization developed requires reducing atmospheric CO₂ from 385 ppm to at most 350 ppm.
The landmark 1976 analysis of deep-sea sediment cores by Hays, Imbrie, and Shackleton provided the first continuous, time-ordered evidence validating Milankovitch cycles, showing that climate signals preserved in microfossils and geochemistry over 450,000 years matched the orbital periodicities of precession (~19–23 kyr), obliquity (~41 kyr), and eccentricity (~100 kyr). The result was persuasive because the deep-ocean cores accumulated layer by layer, offering an undisturbed record; independent dating allowed reliable measurement of cycle lengths; and spectral analysis confirmed that dominant periodicities clustered at orbital frequencies rather than occurring randomly. As noted by NASA, this established that ice ages coincided with particular stages of orbital variation. Subsequent modeling has explored the limits of this pacing: conceptual models analyzed by Crucifix (arXiv 1302.1492v1) demonstrate that glacial cycles could theoretically remain synchronized with astronomical forcing across many parameter ranges, though random climatic fluctuations might cause a skipped obliquity cycle. Ashwin et al. (arXiv 1804.08331v2) show that quasiperiodic Milankovitch forcing can produce chaotic or strange non-chaotic responses in low-order ice-age models, limiting predictability. Raghuraman (arXiv 1511.06008v1) provides further evidence that obliquity and, to some extent, eccentricity trigger deglaciation and glaciation in a conceptual model coupling energy balance and ice-sheet dynamics. Finally, Quarles et al. (arXiv 2108.12650v3) extend this framework to exoplanets, showing that large obliquity variations from spin-orbit resonances in binary star systems can induce snowball states or runaway glaciation, illustrating how orbital forcing governs habitability beyond Earth.
Climate models evolved from simple energy balance models (EBMs) that aggregated Earth’s radiation budget, to general circulation models (GCMs) simulating three-dimensional atmosphere and ocean dynamics, and finally to Earth system models (ESMs) that integrate interactive biogeochemistry including the carbon cycle and vegetation feedbacks. EBMs in the 1950s provided early insight into climate sensitivity and ice-albedo feedback, while radiative-convective and early GCM approaches in the 1960s–1970s helped represent atmospheric structure. By the 1990s and 2000s, models increasingly incorporated terrestrial vegetation, ecosystems, and carbon cycle processes because these feedback on atmospheric composition, leading to the term Earth system models. The cGENIE intermediate-complexity ESM exemplifies this modern approach, including parameterized atmospheric photochemistry and schemes for microbial methanogenesis, aerobic methanotrophy, and anaerobic oxidation of methane to simulate ocean-atmosphere methane cycling under both modern and reducing conditions characteristic of most of Earth’s history. This evolution reflects both better computing power and a growing scientific recognition that climate change depends not only on physics but on coupled feedbacks among atmosphere, ocean, land, ice, biosphere, and biogeochemical cycles.
Install this pack and your MIND begins smart — then every answer is grounded in your own knowledge graph.
Try MIND free →