USING GIS TO INVESTIGATE BIAS IN THE FOSSIL RECORD: A CASE STUDY OF THE LATE CRETACEOUS WESTERN INTERIOR SEAWAY OF NORTH AMERICA

There are many phenomena in the geologic record that can conspire to cloud paleobiologic analyses. Most often cited of these are taphonomic issues of preservation, bias in sampling, and bias in outcrop availability. Discussion has focused on identifying the influence of the rock record on our understanding of biodiversity in the Phanerozoic. Paleobiogeography is another important tool for uncovering patterns in the fossil record that directly reflect macroevolutionary processes. In particular, paleobiogeographic studies shed light on the controls of species distributions and range size through time, reflecting processes related to species origination, migration, and extinction. Consequently, it is important to ask how geologic completeness may be influencing observed paleobiogeographic patterns.

We focus on whether geologic incompleteness masks paleobiogeographic patterns in the Late Cretaceous Western Interior Seaway (WIS), specifically pertaining to patterns of competitive replacement. We provide quantitative tests using GIS-calculated ranges and outcrop area, in conjunction with resampling techniques and other approaches, to assess the effects of outcrop and sampling bias on paleobiogeographic analyses of key vertebrate taxa. The WIS is ideal for this type of study; it has benefitted from over 60 years of comprehensive research by Bill Cobban and others, and has been extensively mapped and sampled. Bill’s work has provided critical insight into the relationship between the geologic history and biological evolution of the WIS. The results presented here suggest that many paleobiogeographical questions can be effectively investigated in the WIS. We argue that much of the data gathered over the years by specialists like Bill is still critically relevant today, and further that at times too much focus has been on the incompleteness of this record when all scientific data, including modern biological data, are incomplete.