|
Activity Number:
|
235
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Tuesday, August 8, 2006 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Statistics and the Environment
|
| Abstract - #306298 |
|
Title:
|
Modeling Precipitation Network Data When Station Reporting Times Are Misaligned
|
|
Author(s):
|
Jarrett Barber*+ and Alan E. Gelfand and Douglas W. Nychka
|
|
Companies:
|
Montana State University and Duke University and National Center for Atmospheric Research
|
|
Address:
|
Department of Mathematical Sciences, Bozeman, MT, 59717-2400,
|
|
Keywords:
|
dynamic model ; Gaussian spatial process ; precipitation ; temporally misaligned data
|
|
Abstract:
|
We present a two-stage stochastic model for precipitation events and amounts using data from a network of monitoring stations recording hourly or daily precipitation. Daily stations report accumulations for nominal 24-hour periods, but these periods differ from station to station. That is, the reporting periods are misaligned. In addition to temporal correlation, we expect spatial correlation among events and/or amounts at nearby stations, but temporal misalignment masks spatial correlation to some degree. Our model uses a pair of conditionally independent latent Gaussian spatial processes within a dynamic framework to align data to common reporting times across stations effectively and to recover information masked by misalignment. We use the model to reconstruct an aligned daily precipitation network data product.
|
