The Intergovernmental Panel on Climate Change has identified abrupt thaw caused by thermokarst formation as a permafrost disturbance process that could lead to large-scale environmental change in a warming Arctic. However, there are significant uncertainties associated with that assessment. To address those uncertainties, NGEE Arctic scientists simulated a small catchment comprising 465 ice wedge polygons using the Advanced Terrestrial Simulator (ATS) configured as an intermediate-scale permafrost thermal hydrology model. The spatially distributed model included, for the first time, subsidence, and evolution of ice-wedge polygon microtopography caused by melting of excess ice. The model uses data from the NGEE Arctic study site near Utqiaġvik, AK and builds on our previous model evaluations against multiple types of observation data, thus providing a rare degree of confidence in the projections. In the RCP8.5 scenario, the simulated catchment is projected to experience about 0.5 m of average subsidence between 2006 and 2100, with locally larger values that cause transition from low-centered polygons to high-centered polygons. That thermokarst formation causes only a modest acceleration of thaw; projected end-of-century active layer thickness is 1.8 m when subsidence is included compared to 1.45 m without subsidence. These results suggest that abrupt thaw caused by thermokarst formation is unlikely to occur over large areas. However, the simulations do show a trend to drier conditions throughout the century because landscape drainage is improved by the transition from low-centered polygons to high-centered polygons, leaving less water stored on the landscape. That is, subsidence may help maintain landscape runoff and thus Arctic streamflow in the face of increased evapotranspiration but may also lead to drier tundra.