ABSTRACT

Do sedimentary rocks record mainly average, continuous, day-to-day processes or relatively rare, large-magnitude ones separated by long nondepositional intervals? Subtle legacies from Lyellian uniformitarianism may still impose a subconscious abhorrence of unique events, discontinuities, and large deviations from "average" magnitudes. Where repeated sharp changes of sedimentation are inescapable, periodic cycles are commonly invoked to preserve uniform, orderly variations from some supposed norm. The sedimentary record rarely reflects such uniformity, however, as sedimentologists have gradually realized.

Magnitude versus frequency of processes has long been debated in geomorphology but has received less attention in sedimentology. Recurrence interval, recovery time, and preservation potential are critical factors for evaluating their significance for the sedimentary record. Large-magnitude processes, which represent positive deviations from the norm and are rare on the human time-scale, must be significant over geologic time. But how significant? Could not everyday processes have obliterated much of the evidence? Flood deposits have relatively low preservation potential because they lie above base level. Marine gravel layers dispersed by abnormal waves have greater preservation potential because most ordinary processes are not competent to modify them. Sandy or shelly de osits formed by large waves and displaying either hummocky cross-stratification or graded bedding have a moderate preservation potential, especially if too thick for burrowing animals to homogenize them. Turbidites, which provide exceptional records of episodicity, have excellent preservation potential because they lie well below base level. Many bedding planes are important records of episodicity, too; some are surfaces of erosion, others of nondeposition. These represent negative deviations from average process magnitudes.

The sedimentological importance of rare events is difficult to assess because the record of such events may be very subtle. This is especially true if a deposit has been thoroughly bioturbated or if the record of an event is simply an erosional surface. It has now become possible to evaluate quantitatively ancient episodic sedimentation using modern-process rates as well as refined biostratigraphic and isotopic dating. I predict that such evaluation will necessitate revision of our favorite depositional models, which have become so important for exploration as well as for research.