|
Activity Number:
|
386
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 1, 2007 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Statistics and the Environment
|
| Abstract - #309386 |
|
Title:
|
A Spatial-Temporal Point Process Model for Nowcasting Radar Reflectivities
|
|
Author(s):
|
Yong Song*+ and Christopher K. Wikle
|
|
Companies:
|
University of Missouri-Columbia and University of Missouri-Columbia
|
|
Address:
|
701 S Providence Rd, Department of Statistics, Columbia, MO, 65203,
|
|
Keywords:
|
Spatial-temporal ; Point process ; Movement ; Nonlinear ; Bayesian ; Nowcasting
|
|
Abstract:
|
Short-term forecasts of heavy rainfall involve forecasting the movement of the radar-based estimates of precipitation over fairly short time-scales. The tradition methods of nowcasting weather radar reflectivities are based on linear extrapolation approaches and physical (deterministic) approaches. These methods have limitations. The deterministic approach is difficult to apply in near real time, and does not account for uncertainties in dynamical parameterizations. Extrapolation methods often fail due to the nonlinear nature of the process evolution. In this project, the authors employ a spatial-temporal point process model for short-term forecasts of heavy rainfall events. A nonlinear structure and the dependence between the rainfall cells have been considered. Bayesian approaches have been applied in developing this model.
|
