Estimation of sensible heat flux using the Surface Energy Balance System (SEBS) and ATSR measurements

This paper describes a modified version of the Surface Energy Balance System (SEBS) as regards the use of radiometric data from space and presents the results of a large area validation study on estimated sensible heat flux, extended over several months. The improvements were made possible by the characteristics of the Along Track Scanning Radiometer (ATSR-2) on board the European Remote Sensing satellite (ERS-2) and relate to: (a) the use of bi-angular radiometric data in two thermal infrared channels to estimate column atmospheric water vapor: (b) the use of bi-angular radiometric data in four spectral channels in the 550-1600 nm spectral regions to estimate aerosols optical depth: (c) determination of bottom of atmosphere (BOA) spectral reflectance using column water vapor, aerosols optical depth and a two-stream radiative transfer scheme to relate BOA spectral reflectance to top of atmosphere spectral radiance (d) direct and inverse modeling of radiative transfer in a vegetation canopy to relate BOA spectral reflectance to canopy properties, such as spectrally integrated hemispherical reflectance (albedo). A parameterization of the aerodynamic resistance for heat transfer (in term of kB(-1)) was applied for the first time at large spatial scales. For such large area analyses SEBS requires wind speed, potential temperature and humidity of air at an appropriate reference height. The latter was taken as being the height of the planetary boundary layer (PBL) and the data used were fields generated by an advanced numerical weather prediction model, i.e. regional atmospheric climate model (RACMO), integrated over the PBL. Validation of estimated sensible heat flux H obtained with the ATSR radiometric data was done using long-range, line-averaged measurements of H done with large aperture scintillometers (LAS) located at three sites in Spain and operated continuously between April and September 1999. The root mean square deviation of SEBS H estimates from LAS H measurements was 25.5 W m(-2). (C) 2003 Elsevier Science Ltd. All rights reserved