Uso de técnica de sensoriamento remoto e imagens Landsat-TM e NOAA-AVHRR na estimativa de balanço de energia à superfície.

Abstract

The main objective of the present work was determinate of energy balance components at the surface, based on the spectral data collected through the seven bands of Landsat 5 and five bands of NOAA-16 satellites and some data of a meteorological station within the studied area. Four images of Landsat and two images of NOAA, obtained on December 04, 2000, on October 04, 2001, on September 24, 2003 and on October 12, 2004, involving irrigated areas of Project Nilo Coelho, part of Sobradinho Lake and San Francisco river, and areas of native vegetation were used. The Surface Energy Balance Algorithm for Land (SEBAL) procedures, proposed by Bastiaanssen (1995) and improved by Allen et al. (2002), was applied. The albedo of surface and vegetation indexes, based on reflective spectral bands of the Landsat 5 (1, 2, 3, 4, 5 and 7) and NOAA (1,2), were obtained. The land surface temperature for each pixel, corrected by its emissivity, was obtained. The incident solar radiation was assessed using the image header data and the atmospheric radiation was estimated available weather station data. With these informations the net radiation (Rn) was obtained. Rn that Landsat presented average values of 570.7 Wm"2 in 2000 and 551.5 Wm"2 in 2001, and 468.9 Wm"2 to NOAA in 2001. In 2003 average values to Landsat and NOAA were the 597.2 2 2 Wm and 540.9 Wm " respectively. The soil heat flux (G) was calculated as a function of the net radiation, land surface temperature, albedo and NDVI for each pixel. It average values of Landsat were 94.5 Wm"2 in 2000 and 112.3 Wm"2 in 2001, while a NOAA satellite was 109.5 Wm"2 in 2001. In 2003 average values were 124 Wm"2 and 120.3 Wm"2 to Landsat and NOAA respectively. The sensible heat flux (H) was estimated according to an iterative process that takes into account the atmospheric stability and vertical wind speed profile With two reference pixels (hot and cold) is possible to determine the vertical temperature difference between levels. Then it is possible to make a first determination of H. In the next step the atmospheric stability is obtained and then H is corrected. This processed continued till the stability of dT and H is reached. The average values of H to Landsat in 2000 was equal 131.4 Wm"2 in 2000 and 158.3 Wm"2 in 2001, while the NOAA satellite in 2001 was 218.7 Wm"2. In 2003 average values H were 331.3 Wm"2 and 274.7 Wm"2 to Landsat and NOAA respectively. The latent heat flux (LE) was obtained as residue of the energy balance equation, and they were equal to 366.4 Wm"2 in 2000 and 279.4 Wm"2 in 2001 to Landsat and while NOAA image in 2001 was 161.7 Wm"2. In 2003 average values of LE were 270 Wm"2 and 161.9 Wm"2 to Landsat and NOAA respectively. The average value of H in 2000 was less than the 2001 value and average value LE in 2000 was greater than 2001 one. It might have been a consequence of occurrence of some rain event in the studied area in December of 2000, while affected the albedo and temperature of each pixel. The evapotranspiration (ET) average daily obtained by Landsat in 2000 was 2.5 mm/day, while 2001 the ET was 2.2 mm/day, with maximum value the 4.6 mm/day in 2000 and 4.9 mm/day in 2001. The NOAA image was 1.5 mm/day in 2001, with maximum value 4.7 mm/day. In 2003 the ET to Landsat was 2.1 mm/day, while the NOAA image was 1.6 mm/day, with values maximum of 5.1 mm/day and 4.6mm/day to Landsat and NOAA respectively. The validation obtained that all flux estimated with SEBAL algorithm sub estimated a measure, and the ET daily SEBAL reach 5.0 mm/day, with average the 2.3 mm/day, while that the ET measured was 4.6 mm/day, with of the 2.0 mm/day.