|MadSci Network: Earth Sciences|
Great question Jonathan.
Geologists and paleontologists are a lot like crime scene investigators. They rely on fragments of evidence to reconstruct what happened a long long long long long time ago. They use the evidence in the rocks and fossils themselves to tell a story, or in this case, to paint a picture.
Think about the world around you. The landscape is different in different places. Mountains and plains look different. The streams and rivers carrying sediment through them look different, too. As a result, the sand, gravel, and mud carried by the stream looks different in different areas and the layers that are formed by those materials look different. The beach at the edge of the ocean is different from either of the others, and rock layers formed of beach sand look different from those formed of river sand.
If I find a sand where all the grains are moderately small and extremely well-rounded, I can infer that those sand grains had spent a long time rolling around and banging into one another. The beach is the most likely place for a grain to do this as they wash back and forth all day long. In contrast, if I find a shale, a rock made of clay, I know the water had to be very quiet in order to allow the tiny clay particles to settle. That only happens on a floodplain, a lake, or offshore in the ocean. If the shale contains fossils of marine life or if the shale layer goes across mountain after mountain without much change (showing it was laid down over a very large area) then I can decide that it was probably formed in an ocean.
Then, geologists use fossils and cross-cutting igneous rocks to determine the ages of sedimentary layers.
Once you know the age and you can figure out the depositional environment, you have a very good handle on the paleogeography, what the geography of the land was in the past.
Sedimentary rocks give the best evidence for paleogeography, but if rocks of a certain age are not present that tells us something, too. In those cases, we can often use the patterns of nearby rocks to fill in the paleogeography. If I start in some place where the rocks are shales and move west and find that as I go west the rocks change from shales to river sandstones to conglomerates (gravel) from mountain streams, then it's not too hard to infer that just to the west of the conglomerates, where no sediments are found of that age, would likely have been the high mountain areas that were eroding to form all that gravel, sand, and clay.
The final pieces to the puzzle are the big picture of plate tectonics and the global climate. We know from a variety of lines of evidence, some of it the very sorts I describe above, where the continents were over most of the history of multi-celled life on earth. See http://scotese.com/ for paleogeographic maps that show the distribution of land on earth over time. Earth's climate has also gone through changes in the past. In Mississippian time, there is a profound episode of warming associated with the formation of coals all over the earth. These coals themselves are the evidence for the warming, containing abundant fossils of trop[ical plants, evenin parts of the word that were at the usually temperate mid-latitudes. The same thing happened in the Early Jurassic, when cryolophosaurus was running around. Much of the evidence is in plant fossils, not just leaves and trunks, but also pollen. If you find deciduous or tropical plant fossils in the same general sequence of rocks as Cryolophosaurus fossils, then it is reasonable to picture them running through tropical forests.
All of that said, there is still a significant amount of any illustration of the past that is left up to the artist. Artists may stick close to the facts or they may get more fanciful. This is especially true of things that do not get fossilized, like the color of skin on animals (see "weird-cryolophosaurus_ellioti") or foliage on plants. Not all artists will agree, either. Here: http://prehistoricsillustrated.com/pg_gbo_02.html is a cryolophosaurus in a pine forest, perhaps a more typical high latitude landscape, even during a warm spell, but I don't know what plant fossil evidence indicates this landscape.
Bottom line, we use the rock record and plant fossils to sleuth out a lot of information about what the landscape was like, but it's never a complete picture. Then the artist uses her own ideas and imagination to fill in the missing parts.
David Smith, Da Vince Science Center, Allentown, PA
Try the links in the MadSci Library for more information on Earth Sciences.