Footprint Modelling Applied to ASTER Derived Heat Fluxes

Land surface heat fluxes were estimated using the LUMPS scheme with ASTER data from Cairo, Egypt. The LUMPS scheme is built on the premise that heat fluxes can be modelled using net radiation, the ground heat flux, air temperature and two empirical parameters that describe the surface cover (availability of water) and morphometry (roughness elements, and density respectively). These two empirical parameters were deduced from eddy covariance measurements at three typical sites from a measurement campaign in Cairo in 2007/2008 and applied to the ASTER image according to the land use. The energy balance of the in situ measurements was not closed; hence a Bowen-ratio based closure forcing was applied. Spatial net radiation was calculated from the short- and longwave radiance data of the ASTER images and associated radiative transfer modelling runs using MODTRAN. The ground heat flux was assessed using different methods based on net radiation and vegetation indices. Three different flux footprint models were used to assess land surface heat fluxes. The used footprint models are the model from Kormann & Meixner, the model from Hsieh and the model of Horst & Weil. The output of these models was fit to the spatial resolution of the ASTER data and overlaid over latter according to the wind direction. Finally, the results were compared to the in situ data. A first constraint in the methodology is given by the limitation of control points: The tested stations are the same as the input stations for the empirical constants of the LUMPS scheme. A second constraint is given by the presence of clouds in the footprint area of some of the ASTER data, limiting the number of useful pixels in the footprint analysis. Even though the models produced different flux prints, the resulting turbulent heat fluxes at the three stations did not differ significantly. The agreement between the in situ measured and the ASTER retrieved sensible heat fluxes was improved in some cases.