HOW IS STRATIGRAPHIC QUANTIFICATION INFLUENCED BY PRESERVATION BIAS?

In most depositional environments, even casual analysis of sequences of depositional events (e.g. floods, storms) indicates that only a vanishingly small fraction of them can be preserved in strata. This has led many to suggest that only the most extreme events make it into the record – an example of preservation bias that, if generally true, would likewise bias our ability to reconstruct past conditions. We will review a spectrum of theoretical and experimental results that bear on the problem of preservation bias. In some cases, the results indicate very strong biases, such as important signals that are missing from the record altogether and so could never be reconstructed. But there are also examples in which natural variability of sedimentation processes works in our favor, such that ‘normal’ transport conditions are reasonably represented in preserved deposits. All in all, although quantitative reconstruction must always be done with awareness of preservation bias, in my view neither theory nor experiments nor field observations support sweeping generalizations to the effect that physical strata preserve only extreme conditions. Rather than fretting excessively about what we may be missing, it would be more useful to devote our energy to developing frameworks that allow us to test quantitative reconstruction by using them to make testable predictions about unobserved aspects of the stratigraphic record.