Modeling permafrost dynamics and water balance of Arctic watersheds in a changing climate

Abstract

Changes in climate across the Arctic in recent decades and especially the increase of near-surface air temperature promote signicant changes in key natural components of the Arctic including permafrost (defined as soil experiencing subzero temperature for more than two consecutive years). Recent borehole observations exhibit signicant increase in ground temperatures below the depths of seasonal variations. Modeling studies on a global scale suggest a steady decrease in area underlain by near-surface permafrost in the northern hemisphere in recent decades. Global projections for the next century predict further permafrost degradation depending on the greenhouse gas concentration trajectory. Permafrost degradation is not only associated with climate feedbacks but can also result in signicant changes in coastal and terrestrial ecosystems and increased risks of costly infrastructural damage for Arctic settlements. In addition, permafrost plays an important role in the terrestrial part of the Arctic freshwater cycle as the volumes of frozen ground are practically impermeable for subsurface moisture transport and contain excess water in the form of ground ice. Since geophysical observations bear signicant costs in the Arctic, especially in the remote areas, simulations performed with physically based numerical models allow researchers to assess the current state of permafrost in Arctic regions and make future projections of its dynamics and resulting hydrological impacts. In this dissertation we use numerical modeling in two distinct ways: 1) to estimate current and future ground temperature distribution with high resolution on a regional scale and 2) to evaluate the role permafrost degradation plays in changes in water balance of watersheds under changing climate. First, we study the permafrost evolution of the Seward Peninsula, Alaska over the 20th and 21st century using a distributed heat transfer model. Model parameters are calibrated with a variational data assimilation and are distributed across the study domain with an ecosystem type approach. Simulations suggest that the peninsula will experience a reduction in the near surface permafrost extent of up to 90% and an average increase in ground temperature across the peninsula up to 4.4°C towards the end of the 21st century under the high greenhouse gas concentration trajectory. Second, we perform an ensemble of millennia-long experiments by simulating hypothetical idealized small-scale watersheds placed in a typical Sub-Arctic setting with a physically based distributed hydrological model. In these experiments we single out the effects of temperature dependent subsurface moisture transport by applying air temperature change in our forcing scenarios only to sub-zero temperatures within a given year. Results suggest a long-term increase in annual runoff of 7-15% and a similar decrease in evapotranspiration under a prolonged (up to a millennia) air-temperature increase. The short-term (< 100 years) water balance response highly depends on soil permeability and the watersheds slope and profile curvature. The simulated changes in water balance are a direct result of the decrease in near- surface soil moisture and intensified subsurface moisture transport in the deeper soil layers due to the permafrost thaw. Additional experiments suggest that simplied models that do not include lateral subsurface moisture transport, as typically done in Earth System Models, can reproduce similar changes in equilibrium water balance to the ones predicted by more sophisticated models for the watersheds with gentle slopes. We also find that if the air temperature trend is reversed and watersheds are experiencing prolonged cooling, a high degree of hysteresis in water balance behavior can be observed, however, the long-term changes in water balance are equal in their amplitude. Additionally, we find that initial soil moisture distribution in the deeper soil which is essentially a consequence of the paleoclimate (given the same permeability and topography) determines the overall soil moisture storage deciency which in turn results in the lag between the onset of warming and the increase in total runoff. The deficit in soil moisture storage is highly dependent on the watersheds topography.