Overview of Estimation using L-moments
Loading...
Authors
Kikuchi, Tokiko
Issue Date
2023
Type
Thesis
Language
Keywords
Alternative Title
Abstract
Traditional estimation methods such as maximum likelihood or method of moments do not work very well for heavy tailed distributions often present in water resources research. In addition, they usually require relatively large samples to provide reason- ably accurate estimates. To remedy these shortcomings, Hosking (1990) introduced the method of L-moments for estimation of parameters. The motivation was the need to estimate high quantiles of precipitation and stream flow data using relatively small samples of observations. This work explores the benefits of L-moments based esti- mation for a variety of distributions. In addition, we define a new model for weather (or climate) episodes based on dependent sequence of observations and explore the L-moments in that setting.