Publicações

Simulations for the present climate (1961-1990) in the resolution of 40 km were produced on the South America with the regional model Eta forced by HADCM3 as lateral boundary conditions. The Eta Model was adapted to carry through integrations of decadal timescale for studies of climate change scenarios. This work has as objective to adjusting the simulations of the present climate for São Paulo state, using statistical correction over model grid-boxes, according to daily precipitation of 20 meteorological stations. Simulations of the present climate show that the largest precipitation amounts to the east of the State with annual average values varying between 5 and 7 mm/day. In central part of the State the daily means are below 4 mm/day. The rainy (drier) season occurs between the December and February (June and August) months with values above (below) 11 mm/day (1 mm/day) in the stations located near the coast (middle) of the State. In general, the rainy months (drier) exhibited negative (positive) bias, therefore underestimated (overestimated) the precipitation. The adjustment method consisted of determining a correction factor according to monthly precipitation bias in the period between 1961 and 1985, the corrections were applied in the period between 1986 and 1990 (test period). The method was evaluated based on the mean absolute error (MAE), the root mean square error (RMSE) and the Skill Score before and after corrections. The performance of the method has been classified subjectively such as: Excellent, when the method reduced model errors (bias, MAE e RMSE) in most of the months, Satisfactory and Unsatisfactory, when errors increased after correction. In general, it was noticed that the method satisfactorily adjusted the simulations in most of the State, mainly in the stations where largest bias occurred near the coast. In the stations where the method did not perform satisfactorily the values of the simulated precipitation were generally similar to the observations, the stations with unsatisfactory behavior were located in central north of São Paulo state. These corrections will be applied to simulations of future climate change scenarios to support vulnerability studies.