Spatial Temporal Analysis And Modeling Of Meteorological Data

The spatiotemporal analysis and modeling is fundamental for climate data applications. As an important input parameter of some kind of hydrology model, the spatial and temporal information of meteorological data will have important influence on the research of hydrology model. This paper attempts to model the spatiotemporal data of the monthly average minimum temperatures at 66 rain gauges in Iowa State, US, from January 1951 to December 2000. The temperatures is a collection of spatial-correlated time series. The model parameters of time series varies along with the spatial location. The analysis of parameters will give an answer to the relationship between time domain and space domain. The spatiotemporal data is decomposed into two parts, i.e. spatiotemporal trend and spatiotemporal residuals. Based on spatial and temporal analysis, a mixed model with linear and periodical parameters is optimized to estimate the first spatiotemporal part. The spatial and temporal characteristics of the trend parameters and residuals are also explored. Case study shows that a consistent temporal trend exists in all climate stations, which enables a uniform model to simulate the spatiotemporal trend. These spatial and temporal autocorrelation exists in the residuals. Based on the findings in spatiotemporal analysis, the authors argue to develop a new method for spatiotemporal data interpolation.This paper is composed of four parts. The first part summarizes the background of the spatio-temporal analysis of meteorological data and introduces contents, significance, objects of the research and so on. The second part reviews research development of spatio-temporal analysis, summarizes the research methods, discusses relative theory and basic technology of spatio-temporal analysis, details the route taken by the research and development tools. The third part describes the case study of Iowa State. This paper attempts to model the spatiotemporal data of the monthly average minimum temperatures at 66 rain gauges in Iowa State, US, from January 1951 to December 2000. Based on the analysis of time domain and space domain respectively, the paper build a a mixed model with linear and periodical parameters to simulate the trend of meteorological data. The last part summarizes the result, major innovation and research prospect.