thumbnail

Methods for Estimating Regional Skewness of Annual Peak Flows in Parts of Eastern New York and Pennsylvania, Based on Data Through Water Year 2013

Scientific Investigations Report 2021-5015
By:  and 
https://doi.org/10.3133/sir20215015

Links

Abstract

Bulletin 17C (B17C) recommends fitting the log-Pearson Type III (LP−III) distribution to a series of annual peak flows at a streamgage by using the method of moments. The third moment, the skewness coefficient (or skew), is important because the magnitudes of annual exceedance probability (AEP) flows estimated by using the LP–III distribution are affected by the skew; interest is focused on the right-hand tail of the distribution, which represents the larger annual peak flows that correspond to small AEPs. For streamgages having modest record lengths, the skew is sensitive to extreme events like large floods, which cause a sample to be highly asymmetrical or “skewed.” For this reason, B17C recommends using a weighted-average skew computed from the skew of the annual peak flows for a given streamgage and a regional skew. This report presents an estimate of regional skew for a study area encompassing parts of eastern New York and Pennsylvania. A total of 232 candidate U.S. Geological Survey streamgages that were unaffected by extensive regulation, diversion, urbanization, or channelization were considered for use in the skew analysis; after screening for redundancy and pseudo record length (PRL) of at least 36 years, 183 streamgages were selected for use in the study.

Flood frequencies for candidate streamgages were analyzed by employing the expected moments algorithm, which extends the method of moments so that it can accommodate interval, censored, and historical/paleo flow data, as well as the multiple Grubbs-Beck test to identify potentially influential low floods in the data series. Bayesian weighted least squares/Bayesian generalized least squares regression was used to develop a regional skew model for the study area that would incorporate possible variables (basin characteristics) to explain the variation in skew in the study area. Ten basin characteristics were considered as possible explanatory variables; however, none produced a pseudo coefficient of determination greater than 1 percent; as a result, these characteristics did not help to explain the variation in skew in the study area. Therefore, a constant model that had a regional skew coefficient of 0.32 and an average variance of prediction at a new streamgage (AVPnew, which corresponds to the mean square error [MSE] of 0.11) was selected. The AVPnew corresponds to an effective record length of 68 years, a marked improvement over the Bulletin 17B national skew map, whose reported MSE of 0.302 indicated a corresponding effective record length of only 17 years.

Suggested Citation

Veilleux, A.G., and Wagner, D.M., 2021, Methods for estimating regional skewness of annual peak flows in parts of eastern New York and Pennsylvania, based on data through water year 2013: U.S. Geological Survey Scientific Investigations Report 2021–5015, 38 p., https://doi.org/10.3133/sir20215015.

ISSN: 2328-0328 (online)

Study Area

Table of Contents

  • Abstract
  • Introduction
  • Methods
  • Results and Discussion
  • Summary
  • Acknowledgments
  • References Cited
  • Appendix 1. Assessment of a Regional Skew Model for Parts of Eastern New York and Pennsylvania by Using Monte Carlo Simulations
Publication type Report
Publication Subtype USGS Numbered Series
Title Methods for estimating regional skewness of annual peak flows in parts of eastern New York and Pennsylvania, based on data through water year 2013
Series title Scientific Investigations Report
Series number 2021-5015
DOI 10.3133/sir20215015
Year Published 2021
Language English
Publisher U.S. Geological Survey
Publisher location Reston, VA
Contributing office(s) WMA - Integrated Modeling and Prediction Division
Description Report: vi, 38 p.; Data Release
Country United States
State New York, Pennsylvania
Online Only (Y/N) Y
Additional Online Files (Y/N) N
Google Analytic Metrics Metrics page
Additional publication details