Spatio-temporal Cokriging Interpolation For Air Pollution Index Analysis

Meteorological data,such as observations of air pollutant,temperature and precipitation,are typical spatio-temporal data.They not only have regular values in temporal dimension,but also have obvious spatial dimensions.With reviewing existing literatures,some problems about data quality and its remedies can be identified: On the one hand,because the meteorological data are from the monitoring stations whose distribution is uneven and the number is limited,along with the lack of monitoring data caused by various mechanical and human factors,The uneven and absent of data greatly increase the difficulty of analyzing and modeling.On the other hand,some existing methods for filling the missing data not only simply ignore the spatiotemporal correlation of the variables,but also fail to consider the influence of other related variables,and lose a lot of valuable information.Therefore,it is necessary to analyze and study the spatio-temporal characteristics of meteorological data,and design appropriate and reasonable spatio-temporal data model to meet the different application needs of the real world.In recent years,with the wide application of spatio-temporal data,the various spatio-temporal data models are presented in related research communities,and become the research hot spots with important theoretical and application values.Based on the relevant theories and methods of geostatistics,this paper proposes a spatio-temporal cokriging interpolation method in considering the reduction of data from the perspective of application,and verifies the effectiveness,practicability and generality of the model through the analysis of concentration of fine particulate matter in Beijing air pollutants.The main research work is as follows:(1)The characteristics of spatio-temporal variability and spatio-temporal autocorrelation are studied and analyzed,and data inspection and processing methods are given,and two time and space expansion methods are introduced by an example.At the same time,the spatio-temporal kriging interpolation method is derived in detail,and the calculation formulas of different spatiotemporal variation functions are given.The characteristics of different spatiotemporal variation function models are discussed in depth.(2)With observation that the influence factors of the research object are numerous,the principal component analysis method is used to solve the problem that there are too many related variables which are unable to be fully considered in the spatio-temporal data interpolation process.The detailed data processing flow is given.The reduction of data provides a new idea of high efficiency.(3)The analysis and modeling method of spatio-temporal variables is deeply discussed.Based on the comprehensive consideration of the research objects and other influential factors,a spatiotemporal cokriging interpolation model is constructed,and the solution and evaluation of the parameters of the interpolation model are given,which have some reference significance for the multivariate spatiotemporal data analysis.(4)Taking the PM2.5 concentration of air pollutants in Beijing as source data,the case study and result analysis of the proposed spatio-temporal cokriging interpolation method are carried out.The interpolation experiments are discussed from three aspects,such as the test and preprocessing of spatiotemporal variables,the selection of cooperative variables and the construction of interpolation models.The different selection methods and the effects of different variation functions on the results are discussed,and some suggestions are given to provide some references for future research.