Estudo para identificação de tendências do regime pluvial na região metropolitana de Belo Horizonte a partir de métodos estatisticos e modelos climáticos

• Ano de defesa: 2009
• Autor(a) principal: Gladstone Rodrigues Alexandre
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; UFMG
• Palavras-chave em Português: Modelos climáticos; Regime pluvial; Recursos hídricos Desenvolvimento; Precipitação (Meteorologia); Engenharia sanitária; Chuvas
• Link de acesso: http://hdl.handle.net/1843/REPA-82TJGL

This research aimed to detect eventual trends in the rainfall regime of the Metropolitan Region of Belo Horizonte (RMBH). Have been used for this purpose, a collection of parametric and non-parametric statistical tests and simulations of regional climate models. Four groups of rainfall series have been considered in statistical analysis: the annual rainfallseries, the quarterly rainfall series, the half-yearly series and maximum total monthly rainfall series. The four groups of precipitation series passed for the process of date visual analysis with construction of time series plots of 10-period moving average and moving median, LOWESS smoothing (Locally Weighted Scatterplot Smoothing), linear regression and RAPS (Rescaled Adjusted Partial Sums) graphs, and, in the sequence, for the statistical tests analysis for detecting trends, through of software TREND. In the assessment of simulations of regional climate models, have been analysed future scenarios of annual precipitation of ETA and PRECIS models. Have been constructed maps and time series plots of scenarios simulated for the two models and, subsequently, the scenarios baseline of models have been compared with rainfall data observed in RMBH. The analysis of the results obtained through the statistical methods have not been evidenced trends in the annual rainfall and maximum total monthly rainfall series. However, the results indicated for the quarterly rainfall and the half-yearly series, positive trends for the less rainy semester to the detriment of a decrease of rainfalls in the rainiest period. Those results were beacon and attested by the graphic analysis of the data. Already the results obtained in the evaluation of the climate models, have been evidenced a great discrepancy among the simulated data for the same ones and the observed data, suggesting that the analysis of future scenaries generated by those models should be surrounded of a lot of caution. In a preliminary way, have been analyzed in that study trends in the annual average temperature in Belo Horizonte, using the same methodology suggested for the study of the rainfall regime. The result of the statistical analysis for the temperaturehave been indicated strong positive trends. However, it result might has been caused for intrinsic characteristics of urban areas, for example, heat island effect. Already the evaluation of the simulated data for the climate models when compared with the measured data, havebeen indicated similar results to the observed in the analysis of the rainfall regime.