RESUMO:Os dados meteorológicos são de grande importância para os estudos científicos, pois auxiliam na tomada de decisões em diferentes áreas do conhecimento. As estações automáticas realizam o trabalho de captar esses dados, porém, problemas podem ocorrer nos instrumentos causando falhas nas séries de dados e inutilizando um período ou até mesmo toda a série. Visto que a análise desses dados é prejudicada com esse problema, as falhas devem ser tratadas para garantir uma maior qualidade na obtenção das informações. Este trabalho visa comparar métodos estatísticos de preenchimentos de falhas e verificar qual método possui melhores resultados no preenchimento de falhas em séries de dados meteorológicas. Falhas foram simuladas em séries de dados reais e o desempenho de quatro métodos foram comparados: média simples, média móvel, regressão linear simples e regressão linear múltipla. Para verificar os resultados obtidos, foram usados o erro médio absoluto e o coeficiente de correlação. Os resultados mostraram ótimo desempenho do método de regressão linear múltipla para as variáveis de temperatura, umidade e ponto de orvalho, enquanto que a média simples teve o melhor resultado para a variável de pressão atmosférica. Nenhum dos quatro métodos obteve bons resultados para a variável de radiação solar. PALAVRAS-CHAVE: tratamento, processamento, séries temporais, regressão linear.
ANALYSIS METHODS OF APPLICATION FOR STATISTICAL DATA IN METEOROLOGY ABSTRACT:The meteorological data are important for scientific studies, to assist in decision-making in different areas of knowledge. The automatic stations make it possible to obtain meteorological data. However, problems may occur in equipment causing failures in data series and making it useless a period or even the entire series. Analysis of these data is impaired by this problem, so gaps should be treated to ensure a higher quality in obtaining the information. This work has the objective of comparing statistical gap filling methods and check which method has better results in gap filling in meteorological datasets. Gaps were simulated in real datasets and the performance of four methods was compared: simple average, moving average, simple linear regression and multiple linear regression. To verify the results obtained were used the mean absolute error and the correlation coefficient. The results showed good performance of the multiple linear regression method for temperature, humidity and dew point, while the simple average had the best result for the atmospheric pressure variable. None of the four methods achieved good results for the solar radiation variable.
This paper presents Mannga (Multiple variables with Artificial Neural Network and Genetic Algorithm), a method designed for gap filling meteorological data. The main approach is to estimate the missing data based on values of other meteorological variables measured at the same time in the same local, since the meteorological variables are strongly related. Experimental tests showed the performance of Mannga compared with other two methods typically used by researches in this area. Good results were achieved, with high accuracy even for sequential failures, which is a big challenge for researchers. The core advantages of Mannga are the flexibility of handling different types of meteorological data, the ability of select the best variables to assist the gap filling and the capacity to deal with sequential failures. Moreover, the method is available to public use with the Java programming language.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.