A Hierarchical Bayes Model for Combining Precipitation Measurements from Different Sources

• Activity Number: 70
• Type: Contributed
• Date/Time: Monday, August 12, 2002 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics & the Environment*
• Author(s): Tamre Cardoso*+ and Peter Guttorp and Paul Sampson and Sandra Yuter
• Affiliation(s): University of Washington and University of Washington and University of Washington and University of Washington
• Keywords: Rain rates ; Precipitation ; MCMC

Abstract

A hierarchical Bayes model provides the framework for estimating rain rates through time, conditional on rain gauge and ground radar observations. The key feature of this model is the characterization of a drop-size distribution (DSD). The unobserved DSD is modeled as two independent components: 1) an AR(1) time-varying mean with GARCH errors for the total number of drops, and 2) a time-varying lognormal distribution for the size of drops. From the modeled DSD, precipitation parameters of interest, including rain rates, can be calculated along with associated uncertainty. The use of this model deviates from the common notion of rain gauges as "ground truth"; rather, information from the various precipitation measurements is incorporated into the parameter estimates. The model is implemented using Markov chain Monte Carlo methods and is applied to a dataset from Eureka, California.