Air pollution from traffic is a major health problem in urban areas, and there is a great interest in modeling and measuring the dispersion of pollutants from near- ground sources. Due to the complex nature of the motions in the boundary layer, it is not straightforward to obtain the necessary meteorological data to perform dispersion calculations. This thesis present a study on the use of modeled meteorological data and locally observed data in dispersion calculations. The thesis will also investigate the impor- tance of different micrometeorological variables on the calculations. The Research LINE model (RLINE) and the dispersion model EPISODE have been run with mete- orological input from both locally measured observations, and modeled meteorology from the Weather Research and Forecasting model (WRF). To evaluate the model results, concentrations of NO 2 have been measured with passive samplers near a road in close proximity to the meteorological observation site. The measurements were conducted in winter and spring conditions. The results indicate that the effects of thermal instability have a larger impact on dispersion calculations than mechanically generated turbulence. The differences in wind speed between measured and modeled meteorology were generally large. A sensitivity test gave concentration reductions in the range of 38%- 57% when increasing the lowest wind speed to the same magnitude as the highest wind speeds. The study reveals several challenges connected to the use of both modeled and measured meteorology for dispersion purposes, and it has not been possible to say that one is more advantageous than the other.