Modeling weather patterns with Markov fields

Storm cell identification and tracking has always been considered to be a vital endeavor in severe weather operations. However, development of high fidelity algorithms to address this task has proven to be a challenge. This is largely due to the fact that cell identification and tracking are an ill-conceived problem. For convenience, storm cells have always been treated as individual and independent entities. Unfortunately, this is a poor model for the true dynamics of complex weather patterns. As a result, the potential performance of any algorithm based on such assumptions is severely limited. For this purpose, a simple and flexible model of weather patterns that utilizes Markov random fields is suggested. This paper describes an initial framework for modeling the dynamic fields with complex interactive neighboring entities. The basic idea presented here has potential in a broad range of applications. One such example is the study of the dynamic flow of storm systems and how they influence the development of neighboring systems and cells.