Statistical tests for common bankfull frequency in rivers

Abstract

Two statistical tests are described for investigating the claim that rivers exceed bankfull level with a common frequency. The first test is necessary to demonstrate that vertically adjacent bench level categories at a site are separate in terms of expected recurrence intervals. The second test examines the homogeneity of bankfull frequencies from site to site for each bench level category. The tests were used to examine data from nine stream gaging sites in New South Wales, Australia. It was shown that three distinct bench levels could be identified in the field. The bankfull frequencies associated with the present flood‐ plain level were shown to form a homogeneous population of bankfull frequencies.