Despite advances in permafrost thermal hydrology models, strong coupling among surface and subsurface processes, and heterogeneities in the subsurface spatial structure and in the surface elevation (microtopography) lead to significant uncertainties in model projections. Careful model evaluation against field observations is important to identify critical processes but remains a major challenge. Ahmad Jan (ORNL) and his colleagues evaluate the Advanced Terrestrial Simulator (ATS) against field observations from polygonal tundra at the Barrow Environmental Observatory to resolve certain model uncertainties. ATS couples a multiphase, three-dimensional representation of integrated surface and subsurface thermal hydrology models including overland nonisothermal flows, snow processes, and surface energy balance. The 3-D simulations were forced by meteorological data and observed water table elevations in ice-wedge polygon troughs. The NGEE Arctic researchers show that the three-year simulations agree reasonably well with multiple field observations (e.g., snow depth, water table, soil temperature measurements, and evapotranspiration). Observations show that troughs remain inundated most of the summer but not polygon centers. The water table in polygon centers is strongly controlled by the flow from troughs to centers, illustrating the importance of lateral conductivity (trough-to-center) in keeping ice-wedge polygon centers wet/dry. The role of lateral trough-to-center subsurface flow is more critical in late summer than early summer. These results explore the existence of a null space in the parameter space. The evaporation model parameters and the saturated hydraulic conductivity can be varied simultaneously in a way that leaves the water level in the polygon center approximately unchanged. Thus, independent measurements are needed to provide additional constraints.