During the campaign in August 2010, five controlled meteorological (CMET) balloons where launched from Ny-Ålesund. Paul B. Voss developed the CMET balloons, and they are adapted to operate in the Arctic environment. They are known for their light weigh (approximately ½ kg) and the ability to perform repeated vertical soundings. Data from four of the balloons were used. The flight duration ranged between approximately 24 hours to over 3 days, and 9 soundings were performed. The balloon data were compared with the Lagrangian particle dispersion and trajectory model, FLEXPART. FLEXPART is developed by Andreas Stohl. The study shows that the wind speed used by FLEXPART is generally too high and that the relative humidity spanned by the model is generally too narrow. These errors might be originating from the ECMWF data. The potential temperature, on the other hand, is mostly really good. How close resemblance the calculated trajectories from FLEXPART have with the balloons depends on the stability to the modeled air. More sinking air leads to bigger trajectory errors. The calculated average relative horizontal transport deviation (RHTD) for the first 16 hours is constant around 28 %, which are a few percentage points greater than in earlier studies (e.g. Riddle et al. (2006)). This would probably have been more reliable if the ensemble of the trajectories were greater.