Predictive Mapping of Air Pollutants: a Spatial Approach

Integrated transport, land-use and air quality monitoring are prioritized especially in metropolitan areas. The complexity is straight forward, since data, indicators, variables, methods and approaches vary and isolated. The Spatial Information Sciences (SIS) provides mature solutions for data and policy integration, since the nature of problem is specific to geo-spatial distribution. The interaction between transport and land-use has been studied in several international, national studies; however air quality policies have not fully integrated. For integration, limited number of sample points and sparse spatial observations should be interpolated in order to obtain areal coverage. Geostatistics provides the most probable solution depending upon measurements and other relevant information, where the accuracy of the prediction is known. Within this study, air quality parameters of the European side of the Istanbul Metropolitan area were evaluated by means of geostatistics. For assessing the uncertainty and accuracy; time series plots and comparison of predicted and observed concentrations of parameters for each monitoring station were produced. For mapping the air pollution levels, kriging method was used. Air quality emission parameters, sulfurdioxide (SO2), carbonmonoxide (CO) and particulates (PM), were collected from six sampling stations between years 2003-2006. Different point combinations were tested for producing SO2, CO and PM maps, in order to investigate the effect of point distribution on kriging. Research results indicated that the integrated usage of geostatistical methods, remote sensing and spatial analysis can introduce valuable information to identify, visualize and explore the complex relationships between transport, land-use and air quality.