Modeling rainfall data using a Bayesian Kriged-Kalman model

Jona Lasinio, Giovanna, Sahu, Sujit K. and Mardia, Kant V. (2007) Modeling rainfall data using a Bayesian Kriged-Kalman model. In, Upadhy, S.K., Singh, Umesh and Dey, Dipak K. (eds.) Bayesian Statistics and its Applications. Tunbridge Wells, UK, Anshan, 61-86.


Download (694Kb)


A suitable model for analyzing rainfall data needs to take into account variation in both space and time. The method of kriging is a popular approach in spatial statistics which makes predictions for spatial data. Kalman filtering using dynamic models is often used to analyze temporal data. These approaches have been combined in a classical framework termed kriged Kalman filter (KKF) model. In the combined model, the kriging predictions dictate the optimal regression surface for incorporating spatial structure and the dynamic linear model framework is used to learn about temporal factors such as trends, autoregressive components and cyclical variations. In this article we consider a full Bayesian KKF (BKKF) model for rainfall data and its MCMC implementation. The MCMC techniques provide unified estimation of spatio-temporal effects and allow optimal predictions in time and space. The methods are illustrated with two real data examples. Using many well known validation methods we highlight the advantages of the BKKF model.

Item Type: Book Section
ISBNs: 190574000 (hardback)
Related URLs:
Keywords: gibbs sampler, kalman filter, kriging, markov chain monte carlo, spatial temporal modeling, state-space model
Subjects: G Geography. Anthropology. Recreation > GE Environmental Sciences
Q Science > QA Mathematics
H Social Sciences > HA Statistics
Divisions : University Structure - Pre August 2011 > School of Mathematics > Statistics
ePrint ID: 30052
Accepted Date and Publication Date:
Date Deposited: 10 May 2007
Last Modified: 31 Mar 2016 11:56

Actions (login required)

View Item View Item

Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.

View more statistics