Implementing an Operational Ozone Forecasting System Based on WRF/CALMET/CALGRID Models: a 5-Month Case Study over Tuscany, Italy

Abstract

An ozone forecasting system has been developed based on the CALGRID photochemical dispersion model. The system is implemented as an experimental operational web service at the LaMMA Consortium to forecast ozone pollution over Tuscany region, Italy. It takes advantage of two daily forecasts, one reproducing the meteorology over Tuscany and the other providing chemical boundary conditions. Meteorological forecasts are performed by the Weather Research and Forecasting Nonhydrostatic Mesoscale Model (WRF-NMM) mesoscale model, which is operative at LaMMA for the daily forecasting service. Calculated at a 10-km resolution, WRF-NMM outputs are downscaled to the 4-km system’s final resolution through the CALMET diagnostic model. Daily boundary conditions are forecast by the CHIMERE dispersion model of the PREV'AIR European air quality forecasting system. Since CHIMERE uses the MELCHIOR chemical mechanism, a suitable species conversion was required to the SAPRC-90 mechanism implemented in CALGRID. Emission data are provided by the Regional Inventory of the Sources of Emissions of the Tuscan Regional Authority. Each day, a 96-h simulation is run starting from 00 UTC of the preceding day, used to merely initialize the system, resulting in a 3-day forecast of ozone from today issued to the web in terms of ozone daily mean and maximum concentrations over Tuscany. The system was experimentally run over May to September 2007, resulting in a good accuracy compared to station observations.