|
Activity Number:
|
276
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Tuesday, July 31, 2007 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
IMS
|
| Abstract - #310349 |
|
Title:
|
Mixture of Thresholded, Transformed Gaussian Processes Modeling with Application to Rainfall Extremes
|
|
Author(s):
|
Jie Zhou*+ and Richard L. Smith
|
|
Companies:
|
The University of North Carolina at Chapel Hill and The University of North Carolina at Chapel Hill
|
|
Address:
|
210 Smith Building, Chapel Hill, NC, 27599,
|
|
Keywords:
|
|
|
Abstract:
|
Precipitation data are collected at point-source locations (rain gauges), or averaged over grid cells via a weather forecasting model or reanalysis. In previous work, a transformed Gaussian process model has been used to represent the spatial distribution of rainfall on any given day, to characterize the relationship between point-source and grid-averaged measurements. The current work extends this to a mixture of Gaussian processes, to allow for different weather patterns with different spatial characteristics of precipitation. A Monte Carlo procedure is proposed for estimating the model, that is then applied to US data both from rain gauges and from weather model reanalyses. The two data sources are compared by computing 50-year return values and provide an evaluation of the quality of reanalysis data for estimating extremes.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2007 program |
