This study analyzes the changes of NO 2 vertical tropospheric columns (VTCs) over Europe during the period 2004–2009 using a statistical model, based on a homogeneous high-quality data set of observations of the Ozone Monitoring Instrument OMI. At each point of a regular grid, a Generalized Additive regression Model (GAM) with non-parametric model terms was fitted to the observed columns to describe the most relevant factors contributing to the observed variability in NO 2 VTCs. These factors include annual cycle, day of week, wind, precipitation, retrieved cloud radiance fraction, and trend. Significant negative changes are found in areas with large anthropogenic sources over Western Europe (mostly from −4 to −8% year −1 ). The overall negative changes are consistent with EMEP/CEIP (European Monitoring and Evaluation Programme/Center on Emission Inventories and Projections) emission estimations and previous trend studies. However, we found remarkably large spatial variations in NO 2 column changes within individual regions. Our analysis shows that in particular the NO x emissions from Spanish power plants (from −10 to approx. −20% year −1 ) and over the center of England (up to approx. −12% year −1 ) have been strongly reduced in the past few years, at a rate exceeding the reported emission changes averaged over the individual country. A number of other features of the temporal behavior of the time series of tropospheric NO 2 distributions over Europe were quantified, including clear annual and weekly cycles. Modeling the influence of wind considering both wind direction and wind speed not only improves the accuracy of the trend results, but can be particularly interesting for identifying the sources of the NO 2 VTCs and the transport pathways of air pollutants. The effects of precipitation are observed to vary obviously during warm and cold months, due to the strong seasonal dependence of soil NO x emissions.