|
Abstract:
|
The identification of precipitation regimes is important for many purposes such as agricultural planning, water resource management, and return period estimation. Since precipitation and other related meteorological data typically exhibit spatial dependency and different characteristics at different time scales, clustering such data present unique challenges. In this paper, we develop a flexible model-based approach to cluster multiscale spatial functional data to address such problems. The underlying clustering model is a functional linear model with within-curve dependence, and the cluster memberships are assumed to be a realization from a Markov random field with geographic covariates. Based on diagnosis plots and the asymptotic distribution of posterior probabilities of cluster memberships, we provide a method for assessing the final cluster assignments of curves. The methodology is applied to a precipitation data from China to identify precipitation regions, which is shown to be better than the conventional approach. Though the focus of this study is on precipitation data, the clustering method is generally applicable to other environmental data with similar structure.
|
