Modelagem para estimativa de dados faltantes em série de dados meteorológicos

Abstract

In this work is presented the implementation of an autoregressive-predictive method for the estimation of missing data in meteorological data series, using just the information from the same serie. The Burg method was used for the calculation of the autoregressive coefficients, which reduces the error generated in the calculation of these coefficients. For the implementation of the method was used the multiplatform framework Qt through the programming language C++. With Qt we were able to obtain a friendly graphical interface that facilitates the use of the model. The validation of the model was done with data of the variables: air temperature, wind speed and direction, atmospheric pressure, relative humidity and precipitation of the Meteorological Station located at the IPRJ/UERJ. For precipitation no favorable results were obtained with the use of the model, nonetheless the results for the other variables are in agreement with the reported tolerance, particularly with orders 1 and 2 of the model and the probability of success in the estimates does not depend on the amount of data of the series. The estimated data retains the statistical characteristics of the original series.