The state-of-the-art meteorological model WRF will be made operative in the target area/areas of the project to be able to produce weather forecasts at a resolution of 1 km horizontally. The forecast horizon will be of 72 hours starting, every day, from the GFS analysis of the 00UTC. Frontier AI strategies will be exploited to calibrate the model forecasts for the wind speed in the target area/areas. Knowing in advance meteorological conditions is essential for air-quality monitoring in at least two respects: i) it is a well-known fact that the concentration of pollutants, given the same emissions, crucially depends, e.g., on the atmospheric stability, with stable atmospheres being much more efficient to trap pollutants close to the ground than unstable atmospheres; ii) output of weather forecasts can be passed as input to air-quality models to obtain a detailed, accurate, prediction of pollutant concentration several (here 3) days in advance. The present product aims at making operative the full modelling chain, from weather forecast to air-quality forecast. The modelling strategy will work synergistically with the aerial monitoring service, serving as a warning step to identify situations where the aerial monitoring activity must be particularly intense.