UC Berkeley

Both biogeophysical and biogeochemical feedbacks to climate change from high-latitude terrestrial ecosystems may be mediated by permafrost thaw. For instance, lakes may expand as relatively intact permafrost begins to thaw but experience drainage with continued permafrost degradation; thawing permafrost also increases the vulnerability of soil carbon to decomposition. In order to quantify these feedbacks, Earth System Models (ESMs) need to adequately represent a number of sub-surface processes.

I improved the lake model in the land-surface component (CLM4) of an ESM by including the physics of snow, ice, and underlying sediment, allowing shallow high-latitude lakes to be adequately simulated. After evaluating the model, I investigated the sensitivity of regional surface fluxes to included lake model processes. The inclusion of snow insulation and lake water phase change each cause 15-30 W m^-2 changes in seasonal surface fluxes, altering the forcing of lakes on the atmosphere. With realistic inclusion of these processes, I predict that the presence of high-latitude lakes causes little net change in mean annual air temperature. However, due to increased sub-surface thermal inertia, lakes moderate the seasonal and diurnal cycle in regions of large lake area like the Canadian Shield. Consequently, the loss of lake area in areas of disappearing permafrost could exacerbate increases in summer daily maximum temperatures due to climate warming by up to 1 °C.

Snow insulation has long been recognized as a key control on permafrost soil thermal regime. I show that the snow thermal rectifier can interact with hydrology to cause significant changes in soil temperature associated with non-thermal anthropogenic forcings. Elevated CO₂ or increased summer rainfall could increase soil water-filled pore space by 0.1-0.2 in some permafrost areas, associated with 1-2 °C increases in mean soil temperature. This warming results from both the increased thermal conductivity of the soil and the increased latent heat of fusion. Because these mechanisms saturate once soils are relatively saturated, the initial soil moisture state is crucial in determining susceptibility to these warming mechanisms. In this model, these warming mechanisms were not effective in substantially increasing permafrost vulnerability to thaw in a severe 21st century transient warming scenario.

In summary, I have shown that explicit representation of snow insulation and of freezing and thawing in lakes and soils can alter the predicted responses of soil and air temperatures to changes in climate.