Observed and simulated (RegCM4) climatology of the cold fronts over the cities of São Paulo (1981-2002) and Rio Grande (1991-2008)

The objective of this work was to investigate as RegCM4 model forced in the boundary with as Era-Interim dataset simulates the climatology of frontal systems over the states of Rio Grande do Sul and São Paulo. The cold fronts obtained from RegCM4 were compared with Água Funda (IAG-USP, period of 1981-2002) and FURG-RS weather stations for the period 1991-2008. The cold fronts were identified using an algorithm that checks the tuming of the wind at same time the there is temperature decreases. This algorithm was applied to the four seasons: summer, auturnn, winter and spring. The results show that RegCM4 identifies smaller number of fronts over the city of Rio Grande, underestimating about 15% of total fronts and this underestimation persists in other seasons. In the São Paulo city, RegCM4 overestimates (positive bias) the number of fronts in only 1%, but this overestimates does not persist in all seasons summer and autumn presenting underestimation of about -20% and -2%, respectively. As in São Paulo, in Rio Grande the largest underestimation of fronts by RegCM4 occurs in the summer (~-19%). The climatology of the some meteorological variables was obtained from two day before until two days after the passage of the fronts. In this climatology it was noted large agreement between RegCM4 and observations. Over the city of São Paulo (Rio Grande), from one day before the front (prefrontal) until the day of the front the variab,les . me.ridional wind, relative humidity (except in the autumn and air temperature (zonal wind and air temperature) simulated by RegCM4 have similar values of the observation. Already to zorial wind and precipitation the errors are large for São Paulo city. Therefore, these results show that RegCM4 simulates according with observations the main meteorological features associated with cold fronts in the two cities. The simulation of zonal wind has the greatest errors in São Paulo city, indicating that the location of the extratropical cyclones associated with the cold fronts would not be well represented by RegCM4. Maps of compositions of the simulated and observed fields will be used to verify this hypothesis