Nonparametric test for separability of spatio-temporal processes

✍ Scribed by Rosa M. Crujeiras; Rubén Fernández-Casal; Wenceslao González-Manteiga

• Publisher: John Wiley and Sons
• Year: 2009
• Tongue: English
• Weight: 226 KB
• Volume: 21
• Category: Article
• ISSN: 1180-4009
• DOI: 10.1002/env.1006

✦ Synopsis

Abstract

Spatio‐temporal correlated data are quite common in many fields, such as ecology, methereology or environmental science. In this situation, it is important to check the existence of spatio‐temporal interactions or on the contrary, if the spatial and temporal dependence components can be modelled independently. In this work, we propose a testing technique for detecting separability in the spatio‐temporal dependence structure. Our approach is based on the representation of the log‐periodogram as the response variable in a regression model. Within this context, separability can be interpreted as additivity in spatial and temporal frequency components. We also provide some simulation results comparing the performance of the asymptotic test with a marginal integration and a backfitting approach. The testing method is also applied to a real‐data case. Copyright © 2009 John Wiley & Sons, Ltd.