Estimativa da temperatura de superfície na região metropolitana de Goiânia por meio de imagens Landsat e previsão de temperaturas para períodos posteriores

• Autor(a) principal: Siqueira, Rubens Villar
• Data de Publicação: 2015
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional da UFG
• dARK ID: ark:/38995/00130000014z9
• Texto Completo: http://repositorio.bc.ufg.br/tede/handle/tede/5406
• topic: Temperatura de superfície; Ilha de calor urbano (ICU); Imagens de satélites; Previsão de temperaturas; Surface temperature; Urban heat island (UHI); Satellite images; Temperature forecast; ENGENHARIA CIVIL::ENGENHARIA HIDRAULICA

Climate analysis, whether at global, regional or local level, it has been the subject of research in various fields of earth sciences. Among the climatic parameters, temperature and precipitation have gained importance in recent decades because of significant changes in their magnitudes. Thus, this work performs a detailed analysis of the temperature for the Greater Goiânia, using satellite images to generate surface temperature for the study area, at first, through an analysis between the years 1997 and 2008 and after in about twenty years, periodically every four years, for the years 1997, 2001, 2005, 2009 and 2014. The elaborate maps, besides showing the spatial variation of urban heat islands, show that there was significant changes to the minimum temperature, maximum and average. Between the period 1997 and 2008, the minimum decrease about 1.4°C and maximum jump of 31.2°C to 36.0°C. Test results for the five periods between 1997 and 2014, show that the year 2014 is presented as the hottest in the years studied. Through the resulting maps of this analysis, it can see that the range of temperatures, the difference between the maximum and minimum, grow with the years. An estimated temperature of satellite validation model was performed by direct comparison between the surface temperature and the data of GOIÂNIA weather station belonging to INMET, with differences of 0.7°C to 1.9°C between the temperatures demonstrating the applicability of satellite images to estimate temperatures in areas that do not have a dense meteorological network. The last analysis performed is forecast monthly temperatures for the period between the years 2040-2047, using the method of Holt-Winters. The model used for predicting allowed the computation of the seasonality of the minimum monthly temperatures, average and maximum for the historical period between the years 1970 to 2015. The predicted temperatures renew the expectation of increased minimum temperatures, average and maximum presented by the analysis of Historic data. As shown, in addition to the monthly increases in temperature, the occurrence of these will be situated in the highest classes of about 1.0° C warmer. We can see that, too, after 2000, all temperatures rise significantly, where their amplitudes between the minimum and maximum are located at a higher level than in previous years.