The thermal regime of the active layer temperature (ALT) is a key variable with which to monitor permafrost changes and to improve the precision of simulations and predictions of land surface processes. The dynamics of the active layer thermal regime can differ substantially under various land surface types and climatic conditions. The proper simulation of these different processes is essential for accurately predicting the changes in water cycles and ecosystems under a warming climate scenario. In this paper, an artificial neural network (ANN) forecasting model system was developed using only two accessible parameters, air and ground surface temperatures, to predict and simulate the ALT thermal regime. The model results show that the ANN model has better real-time prediction capability than other physics-based models and performs well at simulating and forecasting variations in soil temperature with a step size of 12days in permafrost regions on the Qinghai-Tibetan Plateau. The influence of an increase in air temperature on the ALT thermal regime was more intense during the thawing process than during the freezing process, and this influence decreased with an increase in soil depth.

Recent fire activity throughout Alaska has increased the need to understand postfire impacts on soils and permafrost vulnerability. Our study utilized data and modeling from a permafrost and ecosystem gradient to develop a mechanistic understanding of the short- and long-term impacts of tundra and boreal forest fires on soil thermal dynamics. Fires influenced a variety of factors that altered the surface energy budget, soil moisture, and the organic-layer thickness with the overall effect of increasing soil temperatures and thaw depth. The postfire thickness of the soil organic layer and its impact on soil thermal conductivity was the most important factor determining postfire soil temperatures and thaw depth. Boreal and tundra ecosystems underlain by permafrost experienced smaller postfire soil temperature increases than the nonpermafrost boreal forest from the direct and indirect effects of permafrost on drainage, soil moisture, and vegetation flammability. Permafrost decreased the loss of the insulating soil organic layer, decreased soil drying, increased surface water pooling, and created a significant heat sink to buffer postfire soil temperature and thaw depth changes. Ecosystem factors also played a role in determining postfire thaw depth with boreal forests taking several decades longer to recover their soil thermal properties than tundra. These factors resulted in tundra being less sensitive to postfire soil thermal changes than the nonpermafrost boreal forest. These results suggest that permafrost and soil organic carbon will be more vulnerable to fire as climate warms.