Advanced four-dimensional air traffic management tools necessitate very accurate prediction of wind and temperature in the very short term. A method based on optimal interpolation of observational data was developed and tested within the framework of Eurocontrol's PHARE programme and the 4MIDaBLE programme under the auspices of the European Commission. It takes advantage of the growing density of automated aircraft reports to update background forecasts and produce nowcasts valid twenty minutes to a couple of hours ahead with an improved accuracy. The emergence of surveillance techniques such as the Automatic Dependent Surveillance based on the use of aircraft-to-aircraft and aircraft-to-ground data links and the Aeronautical Telecommunications Network is generating much interest in nowcasting techniques because they may facilitate the acquisition of observational data on selected areas with a given observation density. The basis of the nowcasting method is presented along with different evaluations, one of which lasted several months, using routine aircraft data over France. The evaluation shows that operational forecasts can be improved by taking into account the latest aircraft observations, but it is necessary to select aircraft data to obtain a positive response.