Summary:

A characteristic feature of the phenomena occurring in the atmospheric air is the existence of autocorrelation. With tools geostatistics is possible to identify modeling and estimation of the state of contamination of the air. The phenomenon of autocorrelation shows that the results of measurements that are adjacent to each other in time and space are more alike than more distant measurements. The existence of autocorrelation time can be the basis for estimating the value of the measurement location. To study the autocorrelation time gas concentrations and aerosols used semivariance analysis. Knowledge of semivariance parameters allows you to use the best uncompressed linear estimation method which is kriging.
The paper presents an example of application semivariance analysis of ambient air quality status for the area of the province. An evaluation of spatial and temporal autocorrelation in the Silesia province based on the basic parameters such as range of impact , and the effect of the threshold nuggets. The analysis results show the semivariance as a function of the linear relationship between distant the points and the degree of their similarity. The existence of spatial autocorrelation provides a basis for modeling and estimation using stochastic methods.