Publicações

Improved forecasts of the timing and location of thunderstorms and associated lightning are of great interest to all persons concerned with protecting life and property. Particularly during the warm season, it is essential the use of highresolution forecast tools to provide a detailed description of the mesoscale weather features, which are frequently responsible by the primary forcing convection. Comparing to the past, the skill of numerical weather prediction has improved enormously, mainly because of substantial increases in computational power and more efficient numerical techniques. Moreover, due to the chaotic behavior of the atmosphere (Lorenz, 1965), considerable efforts are continuously employed for a more comprehensive and accurate representation of the physical processes within the models. However, in terms of the regional models, one complex limitation is the uncertainty with respect the initial and lateral boundary conditions provided by the global models (Warner et al., 1997). The uncertainties associated with the model initialization can decrease the reliability of the deterministic forecasts. Since it is not easy to exactly separate the errors due to the initial conditions from those due to model deficiencies, there has been considerable interest in the investigation of the sensitivity of forecast errors to initial conditions. The present study examines the impact of initializing the Weather Research and Forecasting (WRF) model (Skamarock et al., 2008) with Global Forecast System (GFS) analyses and forecasts fields of 1 and 0.5- degree grid increment resolutions. Short-range simulations of ten thunderstorm cases were performed with high space WRF model resolution to develop a methodology to qualitatively forecast lightning occurrence with few hours in advance during the summer season in southeastern Brazil. The influence of the same initialization dataset but with different resolutions over the WRF model will be evaluated comparing the lightning forecasting maps for two additional thunderstorm cases, one without lightning occurrence, and calculating statistical scores to help in the forecasting verification.