Abstract. Thawing of ice-rich permafrost deposits can cause the formation of thermokarst terrain, thereby involving ground subsidence and feedbacks to the thermal and hydrological regimes of the subsurface. Thermokarst activity can entail manifold pathways of landscape evolution and cause rapid permafrost thaw in response to a warming climate. Numerical models that realistically capture these degradation pathways and represent the involved feedback processes at different spatial scales, are required to assess the threats and risks that thermokarst processes pose to the functioning of ecosystems and human infrastructure in the Arctic. In this study, we therefore introduce a multi-scale tiling scheme to the CryoGrid 3 permafrost model which allows to represent the spatial heterogeneities of surface and subsurface conditions, together with lateral fluxes of heat, water, snow, and sediment, at spatial scales not resolved in Earth system models (ESMs). We applied the model setup to a lowland tundra landscape in northeast Siberia characterized by ice-wedge polygons at various degradation stages. We present numerical simulations under a climate-warming scenario and investigate the sensitivity of projected permafrost thaw to different terrain heterogeneities, on both a micro-scale (ice-wedge polygons) and a meso-scale (low-gradient slopes). We found that accounting for both micro- and meso-scale heterogeneities yields the most realistic possibilities for simulating landscape evolution. Simulations that ignored one or the other of these scales of heterogeneity were unable to represent all of the possible spatio-temporal feedbacks in ice-rich terrain. For example, we show that the melting of ice wedges in one part of the landscape can result in the drainage of other parts, where surface water has been impounded a number of decades earlier as a result of ice-wedge thermokarst. We also found that including subgrid-scale heterogeneities in the simulations resulted in a more gradual response in terms of ground subsidence and permafrost thaw, compared to the more abrupt changes in simple one-dimensional simulations. Our results suggest that, under a warming climate, the investigated area is more likely to experience widespread drainage of polygonal wetlands than the formation of new thaw lakes, which is in general agreement with evidence from previous field studies. We also discuss how the presented model framework is able to capture a broad range of processes involved in the cycles of ice-wedge and thaw-lake evolution. The results of this study improve our understanding of how micro- and meso-scale processes control the evolution of ice-rich permafrost landscapes. Furthermore, the methods that we have developed allow improved representation of subgrid-scale processes such as thermokarst in ESMs.
Review: Effects of multi-scale heterogeneity on the simulated evolution of ice-rich permafrost lowlands under a warming climate.
