Scientific and Practical Considerations Related to Revising Bulletin 17B: The Case for Improved Treatment of Historical Information and Low Outliers

Since 1982, Bulletin 17B (B17B) has provided guidelines for conducting flood frequency analyses in support of federal projects in the U.S. The stability and consistency of B17B is widely recognized as a virtue, but research during the past 25 years suggests that substantially more accurate frequency estimates could be obtained if some of B17B's procedures were revised. Among other things, scientists and engineers have developed better methods for incorporating historical flood information into frequency analyses, and for addressing the problem of low outliers and zero flows, that take advantage of computational capabilities that were not available when B17B was published in 1982. Similarly, an additional 30 years of data are available to improve regional skew estimation. In this paper we consider both the technical aspects of proposed changes to B17B and some of the practical issues and implications related to altering procedures that have been in use for more than two decades.