That's tougher. Consider that the present Earth has barely been mapped, and the seafloor was unknown in detail until the 1980s. As a kid I can remember seeing maps with blank spaces in the desert of the Arabian Peninsula and in Antarctica. Only with the space age have we finally done it right, viewing the Earth from high above and filling in every detail.
Clues to the Geographic Past
To map the past, nothing can be measured directly; everything has to be inferred from the evidence of the rocks. And the farther back in time we go, the fewer rocks of that age survive. It is as if we had to map today's Earth while limited to what we could dredge up with buckets hung from balloons on a cloudy moonless night. Where do we start?
An outcrop may have fossils that indicate its rough age. And other things might indicate the environment at the time the rocks formed. Those might be current marks, or shallow-water fossils, or the red sandstones typical of volcano-filled valleys. The alignment of magnetic mineral grains would show where north was—or, just as likely, south—and how far away the pole was from that spot during that age. Also, we can learn the age of many rocks in the laboratory.
Stitching Together the Big Picture
Evidence from many outcrops must be put together to create a picture of that particular region at that particular time. We could tell if that area was land or sea, which direction the shoreline lay, and what the local latitude was.
And in turn, many regions must be stitched together to get a picture of the ancient continent, and still more to map the whole world. Such a map could be put together, but it would take many years of work.
Geologists have done that work, two centuries of fieldwork. Two other big breakthroughs have helped. One is the insight given to us by the theory of plate tectonics, and the other is a decent record of sea level over geologic time. Now that we know where the continents have been and where their shorelines lay—the Earth's paleogeography—we can assemble reliable world maps for the last several hundred million years of Earth history.
Putting the Past on the Web
Getting such maps onto the Web is another story. Most scientists work in specific regions and geologic time periods, and while they may have very good paleogeographic maps for their own region and time zone, they have no time or energy to make maps for the whole Earth. There have been a few attempts for the Web, but for the most part the databases are used on scientists' workstations, not the public's desktops.
Two people out there who have done it right are Christopher Scotese and Ronald Blakey. Unlike those who use simple geologic outlines, they have tried to recreate the real appearance of the Earth as it might look from space—oceans, mountains, and all.
Scotese is an associate professor of geology at the University of Texas at Arlington, but his Paleomap Project is an outside website where he shows his images along with DVDs, software, and publications. The University of California and others also exhibit his maps.
Blakey worked for many years at Northern Arizona University creating lean and elegant paleomaps of the whole world, with a particular focus on the Colorado Plateau. After retiring, he founded Colorado Plateau Geosystems to continue improving them. He also sells poster-size printed maps.
Their work reminds me of the great work being done by Raven Maps. Once you put a Raven map on your wall, you'll be spoiled forever.
PS: Honorable mention goes to the very cool Atlas of the Ice Age Earth, Jonathan Adams and Hugues Faure's attempt to show the kinds of vegetation that covered the world during the last 15,000 years. They're flat outline maps, not pictorial, but it takes only a little imagination to put yourself on the ancient taiga, spear in hand, hunting mammoths. Or study the present-potential vegetation maps to envision how the Earth would look today without humans.