Notice and Invitation
Oral Defense of Doctoral Dissertation
The Volgenau School of Engineering, George Mason University 


Michael Wright
Bachelor of Arts, University of Virginia, 2008
Master of Science, George Mason University, 2011
 
Evaluation of heterogeneity statistics for regional frequency analysis

 

Monday, April 28th, 2014, 1:00 PM to 3:00 PM
The Nguyen Engineering Building, Room 1602

All are invited to attend.

Committee
Dr. Mark Houck, Chair
Dr. Celso Ferreira, Co-Chair
Dr. Shanjiang Zhu

Dr. Stephen Nash

Dr. Jason Giovannettone

 

 

Abstract

Precipitation gauge records are often pooled to reduce the error of quantile estimates by increasing sample size. This introduces a new error component proportional to heterogeneity, the degree to which the constituent gauges diverge from the regional average. Precipitation analysts, especially those working in data-sparse regions of the world, often use heterogeneity statistics to determine whether a candidate regionalization of gauges would increase or decrease the error of quantile estimates. This dissertation assesses the relationship between error in quantile estimates and five heterogeneity statistics proposed in the literature. Analysts can use this research to defend the choice of certain heterogeneity statistics over others and to establish thresholds below which candidate regionalizations can be declared acceptably homogeneous. All five-or-more-site regionalizations of a twelve-gauge Minnesota dataset are enumerated and Monte Carlo simulation is used to estimate quantile error and the heterogeneity statistics. Linear relationships found between heterogeneity estimators and quantile error across these real data are compared to relationships found in simulation experiments isolating the heterogeneity-related component of quantile error. Two statistics have highly linear relationships to error in both the simulation and enumeration studies. The less linear statistic is more robust to deviation from the hypothesis that regional coefficient of variation and skewness ranges increase in tandem as heterogeneity rises. Novel heterogeneity thresholds are defined for this statistic.

A copy of this doctoral dissertation is on reserve at the Johnson Center Library.