Modelling Study for Assessment and Forecasting Variation in Urban Air Pollution

This work proposes the development of an airpollution model based on a joint application of Kalman filter and Kriging technique. The use of modelling techniques in data environmental analysis allows engineers to characterize in a better way the behaviour of air pollutants, in order: (i) to validate the measured data; and (ii) to predict the values of contaminant substances emissions. These techniques are, therefore, a very useful analysis tool, especially when the analyzed time series are characterized by numerous missing or erroneous data. The joint application of both Kalman filter and Kriging algorithms allows users to benefit from the main advantages of two independent methods, in order both to improve the performance of the developed model and to reduce its uncertainty. The agreement of estimated and measured values highlights how well the developed model reproduces the reality and so confirms its efficiency.