Espacialização de variáveis meteorológicas combinando informações de imagens de satélite com técnicas de interpolação para o estado de Minas Gerais

• Autor(a) principal: Amorim, Ranieri Carlos Ferreira de
• Data de Publicação: 2009
• Tipo de documento: Tese
• Idioma: por
• Título da fonte: LOCUS Repositório Institucional da UFV
• Texto Completo: http://locus.ufv.br/handle/123456789/1489
• topic: Interpolação; Satélites geoestacionários; Interpolation; Geostationary satellites; CNPQ::CIENCIAS AGRARIAS::AGRONOMIA::AGROMETEOROLOGIA

The meteorological variables required for the application in the management of hydric resources and agrometeorological and climatic studies, among others, are usually measured in meteorological stations. The reach of these data are limited to the areas near the points of measurement. To estimate meteorological variables in places where there are no stations, it is possible to use the spatial interpolation (Borga & Vizzaccaro 1997, Caruso & Quarta 1998, Campling et al. 2001). So, the main objective of this work is to evaluate the performance of the IDW interpolation method (Inverse Distance Weighted) in the spatialization of meteorological variables for the state of Minas Gerais. The more specific objectives were: (i) evaluate, by means of the statistical analyses achieved through the cross validation, the performance of the IDW interpolator with a power of 2 in the spatialization of hourly data of air temperature, air relative humidity, solar radiation, wind speed and precipitation; (ii) evaluate the performance of the IDW interpolator with the distance weighting exponent varying from 0,5 to 4.0 for the spatialization of hourly data of temperature and air relative humidity for the state of Minas Gerais; (iii) evaluate the performance of the IDW interpolator, without the assistance of the altitude, to spatialize all the hourly data of air temperature in all the region of the state of Minas Gerais and (iv) propose a spatialization method for air temperature, precipitation and solar radiation, combining the information of the GOES 10 satellite and the data of the automatic meteorological stations. The IDW interpolation method with a power of 2 proved to be adequate for the interpolation of the meteorological variables analyzed in this study (temperature, air relative humidity, solar radiation, wind speed and precipitation) because it presented errors relatively small in comparison to the errors found in other studies that used daily and annual data. When the interpolation was carried out varying the distance weighting exponent, it was not possible to determine which exponent was the best to interpolate hourly data of temperature and air relative humidity in the state of Minas Gerais as a whole. However, when the areas of study are analyzed separately, the IDW interpolation method presents different answers. The area of study 1, which comprises the mesoregions of the Vale do Jequitinhonha, Vale do Rio Doce and Vale do Mucuri, presented the best results for the coefficient 4,0 in all the months analyzed. For the area of study 2, which is formed by the following mesoregions: Belo Horizonte Metropolitan Area, South and South West of Minas, Campo das Vertentes and Zona da Mata, the IDW with a power of 1,5 produced better results and for the area of study 3 (Triângulo Mineiro, Central Mineiro, North West, North and West of Minas), the IDW, either with a power of 0,5 or 1,0 provided better results. For the state of Minas Gerais, in spite of the wavy relief, the traditional interpolation (without altitude assistance) produced smaller errors, in comparison to the interpolation that considers altitude. The statistics of the errors of interpolation clearly indicate that the traditional interpolator is more accurate than the interpolator assisted by altitude. The performance of the traditional interpolator was, in average, about 1,73°C better than the interpolator with the assistance of altitude. According to the correlation between the data observed in the stations and those estimated by both interpolators, the interpolator with the help of the DEM presented, in average, a better correlation (79%), compared to that of the traditional interpolator (71%). The proposal of the creation of a methodology of meteorological variables spatialization using data of automatic stations in surface and the information contained in the satellite images as a complement to the series of data of these stations did not provide good results. The curve adjusted between the data of air temperature, precipitation and solar radiation, observed in the stations, depending on the color of the image of the GOES 10 satellite in the visible band, presented a maximum correlation of 23% for temperature, 61% for precipitation and 17% for radiation.