Analysis Sensitivity of the Physics Parameterization WRF Model for the Rio Grande do Sul State on the winter of 2014

Abstract

The objective of this study is to observe the sensitivity of parameterizations of the WRF model to quantify the variables in surface: atmospheric pressure, air temperature, relative humidity and precipitation during the winter of 2014 in the State of Rio Grande do Sul (RS). The results were demonstrated from analysis of statistical indices, bias and Mean Squared Error root (RMSE) calculated for comparisons between the data extracted from 6 experiments of the WRF model simulations with data from the National Institute of Meteorology monitoring stations (INMET) in RS. The experiments were configuring with different physical parameterization, so that it may examine what combination performs better in the representation of the RS winter conditions. From the recognition of different physical interpretations that each set of parameterization can represent, a case study was made in order to diagnose the precipitations that occurred in the State, mainly in the municipality of Irai. The analysis came from a monitoring rain event occurred between 25 and 30 June 2014, using meteorological fields of 850hPa stream lines and rainfall. However, realizes that both temperature as pressure, the bias and the RMSE obtained no significant differences between experiments. UR, in the calculation of bias showed a big difference between the experiments, because of the manner of calculation only considers the systematic errors, which may cause cancellation of errors between underestimation and overestimation. The RMSE for the same variable showed no differences in significant amounts in the experiments, only in experiments 3 and 5, smallest error value when compared to the other experiments (~ 2%). To develop some considerations on the precipitation, the bias diagnosed underestimates the experiments for the rains during the winter of 2014; however, in the calculation of RMSE the experiments had not consent to each other, except 4 and 6, where the values of total errors were lower to 2mm. For the case study, which was accompanied rainfall occurred during the passage of an extratropical cyclone, in all experiments showed the characterization of the precipitation event. Thus, to diagnose the amount of precipitation during the event occurring on the Irai weather station with model data, combined with statistical analysis, the experiment 6 from the parameterization of combinations shown in this study had the best performance to characterize the atmospheric state during the winter period in the RS.