Drag the time scale to run plate motion backwards with two published plate models. Go to 250 Ma for the supercontinent, or keep going to 400 Ma to watch it assemble.
| Place | Latitude then | Moved |
|---|
Click any land on the map to pin it and follow it through time.
The continents now move with two published, peer-reviewed plate models, each with its own coastlines cut into hundreds of rigid pieces. They agree on the big picture: a C-shaped Pangaea centred on Africa, with the Tethys Ocean cutting into its east and the China blocks still islands in that ocean. They differ in the details, and switching between them is the honest way to see how uncertain those details are.
Some things are expected, not errors. Coastlines are today's shapes, so they overlap or leave gaps where crust has since stretched, eroded or been squashed; the sunken-crust layer shows the true continental edges. Pieces that did not exist yet (Iceland, Hawaii, young island arcs) or were not yet attached simply aren't drawn at that time. Plate boundaries are shown in 5-million-year steps. Global positions before about 200 Ma are uncertain by several degrees of latitude and much more in longitude.
Ice, seas and seafloor. Ice sheets and inland seas are drawn as simplified outlines from the geological record, not modelled, so treat their edges as approximate. The sea-level curve is a long-term global average after Haq and colleagues, uncertain by around ±100 m. Seafloor age comes from the EarthByte isochrons, the magnetic-stripe record of each piece of ocean floor; most ocean crust older than about 180 Ma has been subducted, so the oldest oceans are blank.
About the future. "Show the future" takes each plate's motion over the last 10 million years and keeps it going unchanged. That is a fair guide for a few tens of millions of years: the Atlantic widens, Africa squeezes the Mediterranean shut, Australia heads for Southeast Asia. Real plates change course when they collide or when new subduction zones form, which this cannot predict, so beyond about 50 Myr treat it as a thought experiment. Where continents overlap, the extrapolation has run into a collision it cannot model.
Sources: Müller et al. 2019, Tectonics; Scotese PALEOMAP PaleoAtlas; continental crust extents from the EarthByte COB terranes; oceanic LIPs from Johansson et al. 2018; data as distributed with GPlates.