A specific characteristic of permafrost landscapes is the occurrence of patterned ground, such as polygons, circles, stripes, and thermokarst lakes. The formation of these patterns through periglacial processes is primarily determined by the climate as one factor, but soil characteristics and availability of water, time (duration of formation) and topography are others. The goals of this project are to (i) determine which processes in the surface structures and between the structures and the atmosphere are significant and (ii) investigate their relative importance on scales from centimetres to hundreds of kilometres. Three contrasting arctic permafrost sites are studied in detail: Siberia (Lena Delta), Spitsbergen (Ny-Ålesund) and Alaska (North Slope). These sites differ in their history (glaciation record), past and current climate, topography, surface and subsurface characteristics. Patterned ground is dominant at all three sites at varying scales from meters to kilometres: non sorted circles (Spitsbergen), polygons and thermokarst lakes (Siberia) and water tracks (Alaska). The hydrologic, thermal and solute dynamic processes are analysed using experimental data of these fluxes as well as one and two dimensional modelling. It is found that differences between the sites in radiative, atmospheric fluxes and ground heat fluxes are relatively small during the summer. The summer water budget, however, varies significantly between the sites and depends largely on the precipitation (snow/rain) input and timing. At all sites, patterned ground exerts a major influence on the heat and water budget. At the Spitsbergen site, small scale differences in surface and subsurface heat budget induce and maintain small scale differences such as greater thaw depths underneath the centre of a mud boil. On a larger unit, such as for polygonal ponds in Siberia, the heat release during fall freeze-back affects the surface temperature. At the watershed scale (Alaska, Siberia) the water storage and runoff depends on topography and patterned ground features, such as water tracks and polygonal ponds and lakes. Thus, predicting changes in surface energy and water budget requires landscapes models where key processes, mainly thawing and routing of water, are accurately predicted.