关键词: 时空, 空间, 插值, 变异函数, 克里金

Abstract: The spatial-only Kriging interpolation is usually used to solve the problem which is short of sampled data. When sampled and unsampled locations are relative with time and space, it will bring on the loss of worthy data in time dimension if apply directly the spatial-only Kriging interpolation to spatiotemporal domain, which has led to the study on spatiotemporal Kriging interpolation. The goal of this paper is to extend spatial-only Kriging interpolation to spatiotemporal domain and implement the spatiotemporal variograms, spatiotemporal interpolations and spatiotemporal cross validation. Firstly the maximum likely variograms model and effective spatiotemporal sill, nugget, range are derived, secondly the spatiotemporal kriging interpolation is implemented. Lastly cross validation is done to test the validity of spatiotemporal interpolation. The experiment shows that the spatiotemporal extending of spatial-only Kriging interpolation can provide enough information on random domains at some certain accuracy. It provides an effective approach to spatiotemporal estimation and interpolation for various spatiotemporal phenomena.

Key words: Spatiotemporal, Spatial-only, Interpolation, Variograms, Kriging