Vegetation cover, land cover and surface wetness are few of the many factors exerting control on the partitioning of energy to latent, sensible and ground heat flux. Spatial estimates of these factors can be inferred from remote sensing data. A process-based understanding of the surface energy balance, however, needs to consider both the temporal and the spatial variations of the surface.The fractionated polygonal tundra landscape of Samoilov Island of wet and dry surfaces induces strong spatial variations of resistance to evapotransipration. The development of low-centered ice-wedge polygons results in a prominent microrelief that is the most important factor for small-scale differences in vegetation type and near surface soil moisture. Depressed polygon centers alternate with elevated polygon rims with elevation differences of up to 0.5 m over a few meters distance. In the depressed polygon centers, drainage is strongly impeded due to the underlying permafrost resulting in water-saturated soils or small ponds. In the course of the summer season, however, the surface wetness changes significantly since the water table falls about 5 cm below the surface. This change in surface wetness is likely to be associated with changing evapotranspiration rates. We consider the effect of seasonal changes in land cover, vegetation cover and surface wetness on latent heat flux by investigating a time-series of high-resolution aerial and hyperspectral satellite imagery and comparing them to ground-based measurements of near-surface soil moisture and latent heat flux.Two sets of aerial images from August 15 and September 11, 2008 in the VNIR provide detailed information of the polygonal landscape with a resolution of 0.3m. CHRIS Proba imagery provides hyperspectral data with 18 spectral bands in the VNIR range (400 - 1050 nm) and a resolution of 17 m. Acquisition dates are June 21, July 23 and September 10, 2008.