Estudo das relações preditivas entre o número de dias de chuva e a Temperatura da Superfície do Mar (TSM) para o Rio Grande do Sul

Abstract

A pluvial precipitation measured in meteorologic stations, is not always a good parameter to evaluate the hydric availability in a determined period. The total pluvial precipitation corresponds the whole period considered, not taking into account if they were distributed well or badly. With the Number of Rainfall Days, there is an idea of the intensity of the pluvial precipitation, as analysing the same total of rain in intervals of distinct time, obtaining the intensity as well as the quantitative variability of the same. To know and be able to predict grandeur meteorologics has been the purpose of researchers in the whole world. The forecast must contemplate the temporal scale as much as the spatial. Starting from the theory that the scales of model prediction at long term, now existing, does not contemplate the diverse climatical regions State of Rio Grande do Sul and the regional studies can improve the information demanded by society, this study had as main objective to determine the predicted relations between the Number of Rainfall Days (NRD) of some meteorologic stations of Rio Grande do Sul and the Sea Surface Temperature (SST). In this research, were used two sets of data; the first formed by monthly datas of Number of Rainfall Days in 17 meteorologic stations in the State, from the period of 1982 to 2005. The second, composed of datas of the Sea Surface Temperature, from period 1982 to 2005. The series were divided into two periods, the dependent, comprehending the gap from 1982 to 2002, for determination of predicted equations as well as the factor of regression, and the independent period, which gap was from 2003 to 2005, for validation of the model. The datas of SST were used to, through the equations of regression, establish the relations between the variables. After establishing the equations, the values predicted of NRD were calculated, and them compared with the values observed, in order to verify the efficiency of the model. For all the regions analysed, were obtained good results in the prediction of Number of Rainfall Days for all the months analysed. The series of observed data, proceeds the same standard of distribution of this variable, although there are some foreseen values that present differences in observed values, but are not significant. In the independent period, the foreseen series show the biggest differences in relation to the observed values. The region in which the model presents the best dexterity is the echoclimatic region of Campanha (R9) and the month of best prediction is July.