Paleogeography

Paleogeography is the study of ancient geography: the reconstruction of how continents, ocean basins, mountain ranges, shallow seas, and coastlines were arranged at successive moments in Earth’s past. It synthesises evidence from paleomagnetism, stratigraphy, paleontology, paleoclimatology, and geochemistry to produce maps of former worlds — maps that reveal not merely where the landmasses stood, but how their arrangement controlled ocean circulation, climate, the distribution of life, and the formation of economically important mineral and energy resources. The discipline lies at the intersection of nearly every branch of Earth science and provides the spatial framework without which the history of the planet would be a chronology of events with no stage on which to set them.^{1, 18}

The intellectual roots of paleogeography extend to Alfred Wegener, whose 1915 reconstruction of a single ancestral landmass he called Pangaea drew on coastline fits, matching geological provinces, fossil distributions, and paleoclimatic indicators.¹⁴ Wegener’s maps were qualitative, hand-drawn, and widely rejected, but the basic methodology — assembling converging lines of evidence to constrain the positions of continents at a given time — remains the foundation of the field today. What has changed is the precision. Quantitative paleomagnetic data now fix paleolatitudes and continental orientations to within a few degrees for much of the Phanerozoic, ocean-floor magnetic anomalies record the spreading history of post-Jurassic ocean basins in continuous detail, and open-source software allows researchers to animate the entire history of tectonic plate motion in four dimensions.^{1, 6}

Methods of paleogeographic reconstruction

No single data source is sufficient to reconstruct the geography of a vanished world. Paleogeographic maps are built from the convergence of independent lines of evidence, each with its own strengths and characteristic uncertainties. The reliability of a reconstruction is judged by how many of these lines agree and by how well the resulting map predicts observations not used in its construction.⁷

Paleomagnetic data and apparent polar wander. When igneous rocks cool through the Curie temperature of their magnetic minerals, they lock in the direction of the ambient geomagnetic field. Sediments acquire a weaker but analogous record as magnetically susceptible grains settle and align. Because the time-averaged geomagnetic field closely approximates a geocentric axial dipole, the inclination of a paleomagnetic measurement determines the paleolatitude at which the rock formed, and the declination constrains the orientation of the continent relative to the pole.¹⁵ A compilation of paleomagnetic poles from successively older rocks on a given continent traces an apparent polar wander (APW) path — the apparent migration of the magnetic pole through time as seen from that continent. When APW paths from different continents diverge, the continents must have moved relative to one another; when the paths converge upon reassembly into a former supercontinent, the reconstruction is independently confirmed. Torsvik and colleagues compiled high-quality APW paths for the major Phanerozoic continents, providing the quantitative backbone of modern paleogeographic models from 540 Ma to the present.¹

Geological matching. Continents that were once joined share geological features that terminate abruptly at their modern coastlines and resume on the opposite side of the ocean. The Appalachian–Caledonian mountain belt continues from eastern North America through Britain and Scandinavia; Precambrian cratons in South America align with those in West Africa; Grenville-age orogenic belts (~1.0 Ga) are traceable from eastern North America through Scandinavia, East Antarctica, and southern Australia when Rodinia is reassembled.^{3, 12} Matching rock types, structural trends, and radiometric ages across ocean basins provides geometric constraints that are independent of paleomagnetism.

Fossil distributions. Wegener drew heavily on biogeographic evidence: the Permian seed fern Glossopteris occurred across South America, Africa, India, Australia, and Antarctica; the freshwater reptile Mesosaurus was found only in southern Brazil and South Africa. These distributions are inexplicable if the continents had always occupied their present positions but follow logically from a Gondwanan assembly.¹⁴ More broadly, the degree of faunal similarity between two regions at a given time reflects their physical connectivity: shared taxa indicate land bridges or narrow seaways, while endemic radiations signal isolation. Paleobiogeographic data remain essential for constraining the longitudinal positions of continents, which paleomagnetism alone cannot determine.^{13, 18}

Paleoclimatic indicators. Certain rock types form under restricted climatic conditions. Evaporites precipitate in arid belts; coals accumulate in humid equatorial or temperate zones; tillites and dropstones record glaciation. Mapping these lithofacies through time and comparing their distributions to known climate zones constrains the paleolatitude of the host continent. When glacial deposits of identical age appear in regions that now straddle the equator — as Carboniferous–Permian tillites do in India, Australia, southern Africa, and South America — the implication is that those regions were once clustered near the South Pole, exactly as Gondwanan reconstructions predict.^{5, 18}

Ocean-floor magnetic anomalies. For the past ~200 million years, the seafloor itself preserves a continuous tape recording of plate motions. Symmetric stripes of normally and reversely magnetised crust flanking mid-ocean ridges record the history of seafloor spreading, and identification of these magnetic anomalies against the geomagnetic polarity timescale yields both the age and the rate of ocean-floor creation at each ridge segment. Fitting the anomaly patterns on conjugate margins back together reconstructs the opening history of an ocean basin in fine detail.^{2, 16} This is the most precise constraint available for post-Jurassic paleogeography, but it cannot extend further back in time because all older ocean floor has been destroyed by subduction.

Archean cratons and the earliest supercontinents

The deep Precambrian record is the most challenging arena for paleogeographic reconstruction. The geological evidence is fragmentary, much of it metamorphosed or structurally overprinted, and the paleomagnetic database thins dramatically with increasing age. Nevertheless, the outlines of a pre-Pangaean tectonic history have emerged with increasing clarity over the past several decades.¹²

The oldest relatively stable continental fragments are the Archean cratons: the Kaapvaal and Zimbabwe cratons of southern Africa, the Pilbara and Yilgarn cratons of Western Australia, the Superior and Slave cratons of Canada, and their counterparts on other continents. Whether these cratons were ever assembled into a true supercontinent during the Archean remains debated. The concept of Vaalbara — a postulated ~3.1 Ga grouping of the Kaapvaal and Pilbara cratons — is supported by similar stratigraphic successions and broadly compatible paleomagnetic poles, but the data are sparse and the reconstruction is considered speculative.¹²

The first supercontinent for which a reasonably detailed paleogeographic reconstruction exists is Columbia (also called Nuna), which assembled between approximately 2.0 and 1.8 Ga and persisted until about 1.3 Ga. Columbia brought together the core cratons of what would later become Laurentia, Baltica, Siberia, West Africa, and Amazonia, along with portions of India and Australia. The reconstruction is constrained primarily by the correlation of Paleoproterozoic orogenic belts — such as the Trans-Hudson orogen in North America and the Svecofennian orogen in Scandinavia — and by a growing body of paleomagnetic data from ~1.8 Ga rocks.^{12, 18}

After a prolonged period of relative stability and gradual fragmentation, the continental fragments reassembled into Rodinia, whose formation spanned roughly 1.3 to 0.9 Ga through a series of global orogenic events, the most prominent of which is the Grenville orogeny. Li and colleagues synthesised the paleomagnetic, geological, and geochronological evidence into a widely cited reconstruction showing Laurentia at the core of Rodinia, flanked by Baltica, Amazonia, West Africa, and the combined East Antarctic–Australian block.³ The so-called SWEAT hypothesis (Southwest US–East Antarctic connection) placed East Antarctica against the western margin of Laurentia, an arrangement supported by the continuation of Grenville-age rocks between the two. Rodinia’s breakup, beginning around 750 Ma and driven by mantle plume activity, opened the Pacific and Iapetus Oceans and set the stage for the dramatic environmental upheavals of the Neoproterozoic, including the Snowball Earth glaciations.^{3, 4}

Gondwana, Laurasia, and the assembly of Pangaea

The fragmentation of Rodinia produced a suite of continental blocks that, over the following 400 million years, underwent a complex choreography of collision and dispersal. By the late Neoproterozoic to early Cambrian (~550–520 Ma), most of the southern continents — South America, Africa, India, East Antarctica, and Australia — had coalesced through the Pan-African and Brasiliano orogenies into the vast southern supercontinent of Gondwana.¹³ Gondwana was immense, covering nearly 100 million square kilometres and encompassing what is today about 64 percent of all land area. Its paleogeographic history is reconstructed from dense networks of paleomagnetic data, the matching of orogenic belts across its internal sutures, and the distribution of distinctive Gondwanan fossil assemblages, especially the Glossopteris flora of the Permian.^{13, 14}

Meanwhile, the northern continents followed their own trajectory. Laurentia (the core of modern North America), Baltica (Scandinavia and Eastern Europe), and Siberia assembled through the Caledonian and Uralian orogenies into the northern landmass of Laurussia (also called Euramerica) by the Late Silurian to Devonian. The final collision of Gondwana and Laurussia during the Carboniferous, approximately 335–300 Ma, produced the supercontinent Pangaea and raised the Variscan–Hercynian mountain belt across what is now central Europe.^{1, 12}

At maximum assembly, Pangaea stretched from pole to pole. Its northern component, Laurasia (comprising Laurentia, Baltica, Siberia, and parts of Asia), was separated from Gondwana by the east-facing embayment of the Tethys Ocean. The vast superocean surrounding Pangaea, called Panthalassa, occupied roughly 70 percent of Earth’s surface. This configuration profoundly influenced global climate: general circulation models show that Pangaea’s enormous size generated an extreme monsoon system with deep continental interiors that were hyper-arid, while coastal regions experienced seasonal torrential rains.^{5, 18}

The breakup of Pangaea and the modern world

Pangaea did not fragment all at once. Its disintegration proceeded in a well-documented sequence spanning over 150 million years, and the details of this breakup are recorded with high precision in ocean-floor magnetic anomalies and the stratigraphy of rifted margins.²

The first stage began in the Late Triassic to Early Jurassic (~200–180 Ma), when rifting between Laurasia and Gondwana opened the Central Atlantic and established the initial separation of North America from Africa. This event was accompanied by the emplacement of the Central Atlantic Magmatic Province, one of the largest igneous events in Earth history. By the Middle Jurassic, seafloor spreading was underway in the Central Atlantic, and the Western Tethys had begun to narrow as Africa rotated relative to Europe.^{2, 16}

The second phase, during the Late Jurassic to Early Cretaceous (~160–130 Ma), saw Gondwana itself begin to fragment. East Gondwana (India, Australia, Antarctica) separated from West Gondwana (Africa, South America) as the South Atlantic and Indian Oceans began to open. India broke free as a discrete continental block and began its rapid northward journey toward Asia, a voyage that would culminate in the Himalayan collision beginning around 50 Ma.^{2, 13}

The third phase, in the Late Cretaceous to early Cenozoic (~90–50 Ma), completed the fragmentation. Australia separated from Antarctica and began drifting northward. South America and Africa achieved full separation as the South Atlantic widened. The North Atlantic propagated northward, eventually splitting Greenland from Europe. By the Eocene (~50 Ma), the modern ocean basins were recognisably established, though their widths and the positions of the continents continued to evolve to their present configuration.^{2, 5}

Major stages of Pangaea breakup^{2, 16}

Ocean gateways and their climatic consequences

Among the most consequential legacies of paleogeographic change is the opening and closing of ocean gateways — narrow seaways whose existence or absence controls the routing of major ocean currents and, through them, the distribution of heat across the planet. Three gateways have received particular attention for their profound climatic effects: the Tethys Seaway, the Drake Passage, and the Central American Seaway.^{8, 9}

Tethys closure. The Tethys Ocean was the great equatorial seaway that separated Gondwana from Laurasia during the Mesozoic. As the African and Arabian plates converged with Eurasia during the Cenozoic, the Tethys progressively narrowed. A major pulse of terrigenous sedimentation between 24 and 21 Ma reflects the collision of Arabia with Eurasia and the effective closure of the Indian Ocean–Mediterranean seaway.¹⁷ The ongoing convergence fragmented the western Tethys into the Paratethys and the proto-Mediterranean, drastically reducing equatorial ocean circulation. Modelling studies indicate that the closure of the eastern Tethys seaway during the Middle Miocene contributed to the global cooling trend of the time by redirecting warm equatorial currents and reducing heat transport to high latitudes.¹⁹ The final vestiges of this process are the modern Mediterranean Sea, which nearly dried out entirely during the Messinian salinity crisis (~5.96–5.33 Ma), and the enclosed Black, Caspian, and Aral seas, remnants of the Paratethys.

Drake Passage. The separation of South America from Antarctica opened the Drake Passage, creating an unobstructed circum-Antarctic ocean pathway. Kennett proposed in 1977 that this gateway enabled the formation of the Antarctic Circumpolar Current (ACC), thermally isolating Antarctica from warmer waters to the north and triggering the onset of major Antarctic glaciation near the Eocene–Oligocene boundary (~34 Ma).⁸ Subsequent plate tectonic modelling by Livermore and colleagues refined the timing, suggesting that a deep-water gateway may have developed during the Middle Eocene, consistent with geochemical proxy evidence for changes in Southern Ocean circulation and biological productivity around that time.⁹ The ACC remains the strongest ocean current on Earth today, transporting approximately 100–150 Sverdrups, and models demonstrate that the global thermohaline conveyor belt functions only in the presence of an open Drake Passage subject to westerly wind forcing.

Isthmus of Panama. The gradual emergence of the Central American land bridge, culminating in the closure of the Central American Seaway around 2.8 Ma, produced consequences for both ocean circulation and terrestrial biogeography.¹⁰ By blocking the direct exchange of surface water between the Atlantic and Pacific, the isthmus strengthened the Atlantic thermohaline circulation, increased the salinity of Atlantic surface waters, and is widely considered to have influenced the intensification of Northern Hemisphere glaciation during the Pliocene–Pleistocene transition. The closure simultaneously divided previously continuous marine populations into Caribbean and Pacific lineages, driving speciation on both sides of the barrier — an event termed the Great American Schism.¹⁰

Biogeographic implications

The arrangement of continents and ocean basins determines the boundaries of the biosphere. Paleogeography explains why biogeographic patterns that puzzled nineteenth-century naturalists follow logically from the history of plate motions.²¹

Wallace’s Line. In 1876, Alfred Russel Wallace published The Geographical Distribution of Animals, the foundational work of zoogeography, in which he documented a sharp boundary running through the Indonesian archipelago between the islands of Bali and Lombok and between Borneo and Sulawesi.²¹ To the west of this line, the fauna is characteristically Asian (tigers, rhinoceroses, primates); to the east, it is Australian (marsupials, cockatoos, monotremes). Wallace attributed this discontinuity to a deep-water barrier separating the two faunal realms, but the ultimate cause is tectonic. The islands west of Wallace’s Line sit on the Sunda Shelf, the submerged extension of the Asian continental plate, which was repeatedly exposed as dry land during Pleistocene glacial lowstands, permitting overland dispersal of Asian mammals. The islands to the east sit on the Sahul Shelf, the extension of the Australian plate, which was similarly joined to Australia during low sea levels. Between the two shelves lies a deep-water channel — the product of the collision zone between the Indo-Australian and Eurasian plates — that has persisted as a barrier to terrestrial dispersal for tens of millions of years. Wallace’s Line is, at root, a plate boundary.¹⁸

The Great American Interchange. The emergence of the Isthmus of Panama initiated one of the best-documented episodes of inter-continental biological migration. For most of the Cenozoic, South America was an island continent with a highly endemic mammalian fauna that had evolved in isolation for over 60 million years — including notoungulates, litopterns, and the marsupial carnivores known as sparassodonts. The formation of a continuous land bridge in the Pliocene permitted reciprocal dispersal: North American placental carnivores, deer, horses, tapirs, and proboscideans moved south, while South American armadillos, ground sloths, opossums, and porcupines moved north.¹¹ The exchange was ultimately asymmetric: North American invaders diversified extensively in South America and contributed to the extinction of many endemic South American lineages, whereas relatively few South American taxa established lasting populations in the north. Woodburne demonstrated that the timing and intensity of these migratory pulses correlated with Pliocene and Pleistocene glacial cycles, with the development of non-tropical ecologies across Central America facilitating trans-isthmian passage during cooler intervals.¹¹

The broader lesson is that paleogeography sets the boundary conditions for evolution. Continental connections permit gene flow and faunal homogenisation; ocean barriers enforce isolation and drive endemism. The explosive diversification of animal life during the Cambrian coincided with the dispersal of Rodinia’s fragments into isolated blocks separated by widening oceans, and the great mammalian radiations of the Cenozoic unfolded against the backdrop of progressive continental separation following Pangaea’s breakup.^{12, 18}

Paleogeography and the distribution of resources

The economic geology of the modern world is, in large part, a consequence of ancient geography. The location and character of mineral deposits, petroleum source rocks, and coal measures reflect the climatic, oceanographic, and tectonic conditions that prevailed at the time and place of their formation, and those conditions are controlled by paleogeography.^{5, 20}

Coal, the product of accumulated and buried plant material, formed preferentially in equatorial and temperate humid zones. The great Carboniferous coal measures of Europe and eastern North America accumulated when those regions lay in the equatorial belt of Pangaea, a paleolatitude confirmed by paleomagnetic data. Today those coal seams sit at temperate latitudes, their tropical origin evident only from the paleogeographic record. Conversely, the Permian coal deposits of India, Australia, and southern Africa formed in high southern latitudes within Gondwana, under cool-temperate conditions quite different from those of the Carboniferous tropics.^{5, 18}

Petroleum source rocks are concentrated in six stratigraphic intervals that together account for more than 90 percent of discovered global reserves: the Silurian, Upper Devonian–Tournaisian, Pennsylvanian–Lower Permian, Upper Jurassic, Middle Cretaceous, and Oligocene–Miocene. Each of these intervals corresponds to paleogeographic configurations that promoted high biological productivity (upwelling zones, restricted basins, warm shallow seas) and anoxic bottom-water conditions that preserved organic matter from oxidation. The prolific source rocks of the Middle East, for example, formed in restricted basins along the margins of the Tethys Ocean, where warm equatorial waters and limited circulation created ideal conditions for organic-matter preservation.^{5, 20}

Evaporite deposits — beds of halite, gypsum, and potash — are another paleogeographic product. They precipitate when restricted marine basins in arid subtropical latitudes undergo intense evaporation. The Permian evaporites of the Zechstein Basin in northern Europe, the Silurian Salina Group of the Michigan Basin, and the Messinian salt deposits of the Mediterranean all formed under specific paleogeographic conditions: basins with limited connection to the open ocean, situated within the arid subtropics.⁵ Their modern distribution across a range of latitudes reflects subsequent continental drift rather than any climatic anomaly.

The practical implication is that paleogeographic reconstruction has become a standard tool in mineral and petroleum exploration. Companies and geological surveys use paleogeographic maps to identify regions that occupied favourable positions for resource formation at specific times in the past, guiding the allocation of exploration effort to areas with the highest geological probability of hosting deposits.^{5, 20}

Software and modern reconstruction tools

The creation of paleogeographic maps was, until recently, an artisanal exercise requiring deep geological expertise and considerable manual effort. Christopher Scotese’s PALEOMAP Project, initiated in the 1970s, produced some of the most widely reproduced paleogeographic maps in textbooks and the scientific literature. Scotese’s atlas of 114 Phanerozoic maps, spanning the past 750 million years, remains a standard reference, illustrating the evolving distribution of deep oceans, shallow seas, lowlands, and mountains at intervals of roughly five million years.^{5, 20}

A transformative development came with GPlates, an open-source, cross-platform plate tectonic geographic information system developed primarily by the EarthByte Group at the University of Sydney in collaboration with the California Institute of Technology and other institutions. GPlates allows users to load plate reconstruction models, visualise the motion of tectonic plates and associated data through geological time, and interactively modify reconstructions using their own datasets. The software has made quantitative plate reconstruction accessible to researchers outside the traditional paleogeographic community — including climate modellers, paleobiologists, and resource geologists — and has catalysed a proliferation of new applications.⁶

The plate models implemented in GPlates draw on the work of several research groups. Müller and colleagues produced a detailed global model for the past 200 million years, incorporating lithospheric deformation along major rifts and orogens and calibrated against ocean-floor magnetic anomalies, fracture zone geometries, and continental geological data.¹⁶ Merdith and colleagues extended the full-plate tectonic framework back to 1 billion years ago, linking Neoproterozoic reconstructions to the Phanerozoic record and providing, for the first time, a continuous global plate model spanning the entire known history of the supercontinent cycle.⁴ Seton and colleagues published a foundational 2023 review in Nature Reviews Earth & Environment that systematically documented the data types, assumptions, and uncertainties underlying modern plate reconstructions, providing non-specialists with a guide to using (and not misusing) the software and its outputs.⁷

The PALEOMAP PaleoAtlas for GPlates combines Scotese’s paleogeographic maps with GPlates’ plate reconstruction framework, allowing users to visualise not just plate positions but the distribution of land, sea, mountains, and ice at each time step. These integrated tools have become standard in Earth system modelling, where paleogeographic boundary conditions are required to run general circulation models of ancient climates.^{5, 6}

Uncertainties and the loss of ancient ocean floor

Every paleogeographic reconstruction carries uncertainties, and these grow substantially with increasing geological age. Understanding the sources and magnitudes of these uncertainties is essential for interpreting paleogeographic maps critically rather than treating them as established fact.⁷

The most fundamental limitation is the destruction of pre-Jurassic ocean floor. Oceanic lithosphere is continuously created at mid-ocean ridges and destroyed at subduction zones, and the average age of the ocean floor is only about 80 million years. No ocean crust older than approximately 200 Ma survives on the modern seafloor. This means that the most precise constraint available for paleogeographic reconstruction — the identification and fitting of ocean-floor magnetic anomalies — extends back only to the Early Jurassic. Before that point, the entire record of ocean-basin geometry has been consumed, and reconstructions must rely entirely on continental geological and paleomagnetic data.^{2, 7}

Paleomagnetic ambiguities compound the problem. Paleomagnetism constrains paleolatitude and continental orientation but not paleolongitude: a continent at 30° south latitude could have been at any longitude around the globe, and paleomagnetic data alone cannot discriminate among those options. This longitudinal indeterminacy is particularly acute for Precambrian reconstructions, where geological tie-points between continents are fewer and the magnetic database is sparser. Different research groups have produced mutually incompatible reconstructions of the same supercontinent, differing not in the latitudinal positions of the cratons (which are reasonably constrained) but in their east–west arrangement.^{3, 7}

Remagnetisation is a persistent concern. Rocks can acquire secondary magnetic overprints through burial, metamorphism, or chemical alteration, partially or wholly replacing the original paleomagnetic signal. Progressive demagnetisation techniques can often isolate the primary component, but in some cases the original signal is irrecoverably lost. For Precambrian rocks that have experienced one or more orogenic cycles, the reliability of any individual paleomagnetic pole requires careful assessment of the demagnetisation behaviour and field tests (fold tests, conglomerate tests, baked-contact tests) before it can be used in a reconstruction.¹⁵

True polar wander introduces an additional complication. True polar wander (TPW) is the reorientation of the entire solid Earth relative to its spin axis, driven by changes in the planet’s moment of inertia as mass is redistributed by mantle convection and surface processes. TPW shifts the geographic positions of all continents simultaneously without changing their relative positions, and its signal in the paleomagnetic record is indistinguishable from individual plate motion unless additional constraints (such as hotspot tracks or seismic tomographic images of mantle structure) are available. Torsvik and colleagues identified four episodes of Mesozoic TPW with rates of 0.45–0.8 degrees per million years, noting that cumulative TPW has been nearly zero since the Late Carboniferous.¹

Despite these challenges, the convergence of independent lines of evidence provides robust constraints on paleogeography for most of the Phanerozoic and increasingly reliable outlines for the Proterozoic. The quality of a reconstruction is ultimately determined not by any single dataset but by the consistency of the picture that emerges when paleomagnetic, geological, paleontological, and paleoclimatic evidence are all brought to bear simultaneously.^{7, 18}

A paleogeographic narrative through deep time

Drawing together the evidence described in previous sections, the history of Earth’s surface geography can be summarised as a series of transformations, each with profound consequences for the global environment.

During the Archean and early Proterozoic (4.0–2.0 Ga), small protocontinental cratons grew through the accretion of volcanic arcs and the internal differentiation of the mantle. Whether these fragments ever coalesced into a single large landmass is uncertain, but by ~2.0 Ga they had begun to assemble into Columbia/Nuna, Earth’s earliest well-documented supercontinent.¹²

The Mesoproterozoic (1.6–1.0 Ga) saw Columbia’s gradual fragmentation and the subsequent re-assembly of its constituents into Rodinia through the Grenville and equivalent orogenies. Rodinia’s breakup around 750 Ma coincided with the Neoproterozoic Snowball Earth glaciations, possibly because the disruption of continental weathering patterns and the opening of new ocean basins altered the global carbon cycle.^{3, 4}

The late Neoproterozoic to early Paleozoic (650–420 Ma) was dominated by the assembly of Gondwana from the fragments of Rodinia’s southern components. Gondwana stretched from the South Pole into the tropics and persisted, in progressively modified form, for over 300 million years. The northern continents remained separate, drifting through the Iapetus, Rheic, and other now-vanished ocean basins before converging to form Laurussia.¹³

The late Paleozoic (420–250 Ma) witnessed the collision of Gondwana and Laurussia into Pangaea, the last true supercontinent. Pangaea’s assembly coincided with the Carboniferous–Permian glaciation of the Southern Hemisphere, the formation of extensive tropical coal swamps, and the largest mass extinction in Earth history at the end of the Permian.^{12, 18}

The Mesozoic (250–66 Ma) saw Pangaea fragment in stages, opening the Atlantic and Indian Oceans, closing the Tethys, and launching India on its collision course with Asia. The Cenozoic (66 Ma to present) completed the dispersal, opening the Drake Passage, closing the Central American Seaway, and establishing the modern configuration of continents and ocean basins that shapes today’s climate.^{2, 5}

Throughout this narrative, the recurring theme is interdependence: the geography of the surface controlled the circulation of the oceans and atmosphere, which controlled the climate, which shaped the biosphere, which in turn altered the chemistry of the oceans and atmosphere. Paleogeography provides the physical stage on which all of these interactions played out, and reconstructing that stage — however imperfectly — is essential to understanding why the Earth system behaved as it did at any given moment in its history.¹⁸