Spatial interpolation of dry deposition using EOF models

Abstract

Random processes are monitored over space and time by a network of stations distributed across a spatial region. Auxiliary information is often gathered not only at the stations but at other points across the region. The incorporation of auxiliary information in some interpolation techniques has show improvement on the interpolation results. The Empirical Orthogonal Functions (EOF) model is a well-known eigenvector based prediction technique widely used in meteorology and oceanography for modeling the variability of the observed spatio-temporal random process. Similarity matrices are constructed using available auxiliary information and included in the EOF model to develop a spatial interpolation method. The resulting interpolation technique will be applied to real data set and the results compared to ordinary kriging.

Random processes are monitored over space and time by a network of stations distributed across a spatial region. Auxiliary information is often gathered not only at the stations but at other points across the region. The incorporation of auxiliary information in some interpolation techniques has show improvement on the interpolation results. The Empirical Orthogonal Functions (EOF) model is a well-known eigenvector based prediction technique widely used in meteorology and oceanography for modeling the variability of the observed spatio-temporal random process. Similarity matrices are constructed using available auxiliary information and included in the EOF model to develop a spatial interpolation method. The resulting interpolation technique will be applied to real data set and the results compared to ordinary kriging.