Nitric oxides (NOx) play a very important role among the anthropogenic trace gases. They affect human health and have an impact on ozone chemistry and climatic change. Here we describe a new method for the quantification of the global NOx budget from image sequences of the Global Ozone Monitoring Experiment (GOME) spectrometer on board the ERS 2 satellite. In contrast to measurements using ground-based or balloon-or aircraft-borne sensors, this instrument provides, for the first time, the possibility of observing global maps of NO2 column densities. As part of this work, algorithms were developed to analyze GOME spectra numerically and to extract physically relevant parameters from the resulting maps using image-processing techniques. Column densities of NOx were determined using differential optical absorption spectroscopy (DOAS) <Platt, 1994>. By the combined use of an efficient B-spline interpolation and an inversion algorithm based on Householder transformations, the numerical algorithms accelerate the retrievals by a factor of 26 with respect to previous methods. Moreover, techniques are presented for separating tropospheric and stratospheric NO2 columns and estimating the lifetime of NO2 in the troposphere. This allows determination of regional NOx source strengths. Independent of traditional methods, a global source strength of (43¿20) Tg N yr-1 is estimated. The accuracy of this method is comparable to that of established statistical approaches. ¿ 2001 American Geophysical Union