ABSTRACT
Geomorphological data are of two types, serially independent and serially dependent. In serially independent data the values at one time are not affected by those occurring before or after that time for a given time scale and the observations may be regarded as discrete events. There are three basic types of serial dependence, trend, periodicity and auto-regressive moving average dependencies or persistence. Where trend exists it is often easily recognized by a simple plot of values of the series against time. In the temporal sense the occurrence or non-occurrence of rapid failure of a clay slope per unit of time on a section of cliffed coast and the distribution of probabilities of a glacier surge ought both to be described by the Poisson. Data which records if a particular threshold level in a system is or is not exceeded can be characterized easily by modified Poisson models.
