Statistical Aspects of Ice Gouging on the Alaskan Shelf of the Beaufort Sea.

摘要

The probability density function of the gouge depths into the sediment is represented by a simple negative exponential over four decades of gouge frequency. The exceedance probability function is, therefore, e to the -lambda d, where d is the gouge depth in meters and lambda is a constant. The value of lambda shows a general decrease with increasing water depth, from 9/m in shallow water to less than 3/m in water 30 to 35 m deep. The deepest gouge observed was 3.6 m, from a sample of 20,354 gouges that have depths greater than or equal to 0.2 m. The dominant gouge orientations are usually unimodal and reasonably clustered, with the most frequent alignments roughly parallel to the general trend to the coastline. The value of N(bar) sub 1, the mean number of gouges (deeper than 0.2 m) per kilometer measured normal to the trend of the gouges, varies from 0.2 for protected lagoons to 80 in water between 20 and 38 m deep in unprotected offshore regions. The distribution of the spacings between gouges as measured along a sampling track is a negative exponential. The form of the frequency distribution of N sub 1 varies with water depth and is exponential for lagoons and shallow offshore areas, previously skewed for 10 to 20 m depths off the barrier islands, and near-normal for deeper water. As a Poisson distribution gives a reasonable fit to the N sub 1 distributions for all water depths, it is suggested that gouging can be taken as approximating a Poisson process in both space and time. The distributions of the largest values per kilometer of gouge depths, gouge widths, and the heights of the lateral embankment of sediments plowed from the gouges are also investigated.