Estimation of multiyear changes in nitrogen oxide emissions in megalopolises from satellite measurements

The influence that megalopolises have on the atmospheric composition on regional and global scales is the subject of intense investigations; however, data on the emissions of pollutants used for such investigations are often insufficiently reliable. In this work the possibilities for diagnosing long-term changes in nitrogen oxide emissions in megalopolises are investigated based on a combined use of data from satellite measurements and modeling of the tropospheric nitrogen dioxide content. Primary emphasis is placed on analyzing possible situations when emission changes are of a nonlinear character. The proposed methodology includes an original method for the nonlinear approximation of changes in a physical quantity from a noised time series of its measurements. Changes in NO _x emissions are investigated in 12 megalopolises of Europe and the Middle East in the period from 1996 to 2008. Statistically significant changes in NO _x emissions are detected in five megalopolises (Baghdad, Madrid, Milan, Moscow, and Paris). By using three megalopolises (Madrid, Milan, and Paris) as an example, it is shown that a nonlinear approximation of NO _x emission changes agrees better with independent ground-based measurements than an analogous linear approximation.