Modelling radiation and energy balances with Landsat 8 images under different thermohydrological conditions in the Brazilian semi-arid region

Abstract

Four Landsat 8 images were used together with a net of seven agro-meteorological stations for modelling the large-scale radiation and energy balances in the mixed agro-ecosystems inside a semi-arid area composed by irrigated crops and natural vegetation of the Petrolina municipality, Northeast Brazil, along the year 2014. The SAFER algorithm was used to calculate the latent heat flux (lambda E), net radiation (R-n) was acquired by the Slob equation, ground heat flux (G) was considered as a fraction of R-n and the sensible flux (H) was retrieved by residue in the energy balance equation. For classifying the vegetation into irrigated crops and natural vegetation, the SUREAL algorithm was applied to determine the surface resistance (r(s)) and threshold values for r(s) were used to characterize the energy fluxes from these types of vegetated surfaces. Clearly one could see higher lambda E from irrigated crops than from natural vegetation with some situations of heat horizontal advection increasing its values until 23% times larger than R-n, with respective average lambda E ranges of 5.7 (64% of R-n) to 7.9 (79% of R-n) and 0.4 (4% of R-n) to 4.3 (37% of R-n) MJ m(-2) d(-1). The corresponding H mean values were from 1.8 (18% of R-n) to 3.2 (28% of R-n) and 5.4 (60% of R-n) to 9.2 (94% of R-n) MJ m(-2) d(-1). Average G pixel values ranged from 0.3 to 0.4 MJ m(-2) d(-1), representing 3 and 4% of R-n for natural vegetation and irrigated crops, respectively.