Arctic regions have been determined to be particularly sensitive to a warming global climate both on the basis of climatic modeling and observation of dramatic changes in arctic landscapes and sea ice. As early as the 1990s, air temperatures in interior northwest Alaska were warming at a rate of 0.75 °C per decade. The resulting thawing of permafrost causes significant damage to buildings, roadways, and can lead to increased mass wasting (e.g., active layer detachments and thaw slumps) by melting the soil ice that " cements?? the grains together to resist soil movement, as well as ground subsidence. These climate-induced ground movements can threaten infrastructure, such as road, bridges, and pipelines, either by direct physical damage or indirectly, such as through changes in drainage patterns, increased risk of flooding and forest fires.
Because these geohazards often occur in remote locations with harsh weather conditions and limited access, and the precursor conditions for initiating them can occur gradually, methods for remotely monitoring changes in ground conditions and estimating ground failure risks have significant engineering and economic value. The research described in this paper addresses this need by developing techniques for detecting changes in permafrost and seasonally frozen soil terrains using satellite and airborne remote sensing data, and combining these data with mathematical models to estimate the risk of ground failures due to soil thawing. The methodology consists of combining multiple sources of satellite-acquired synthetic aperture radar (SAR) data with high resolution optical-band data and aerial photography to map frozen ground and associated changes in soil moisture, and to detect vertical and lateral ground movements.
The remote sensing data interpretations along with traditional soil and vegetation mapping are used to inform mathematical models of permafrost and frozen soil stability. These models are used to develop maps of the probability of ground movements associated with permafrost degradation and seasonally frozen soil under current and future climate conditions. The models and slope stability risk algorithm were applied to a portion of Kobuk Valley National Park, Alaska for which soil and vegetation land cover maps were available. The cryosphere model results suggest that the same relative change in active-layer thickness occurs across the landscape, but warmer locations experience a larger absolute change in active-layer thickness and may experience permafrost loss as a consequence. The slope failure risk algorithm indicates that the upland areas are most susceptible to slope failure, particularly south-facing slopes, but low-cohesion low-land soils and steep river banks are also susceptible to failure.