| This dissertation presents development, testing and applications of an algorithm for the simulation of soil freezing and thawing fronts (FTFs) in land surface models. A Two-direction Stefans Algorithm (TDSA), derived from the solution of the Stefan Problem, was developed to predict the position of FTFs in soil. TDSA was tested using observed soil temperature and moisture data from several sites across a north-south transect in North America. TDSA performed well for both permafrost and seasonal frost regions. TDSA was then implemented in the Community Land Model 3 (CLM3), used in a global climate model, to predict FTFs. Several modifications, including incorporation of peat layer, canopy heat capacity, and unfrozen soil water, were also made to make CLM3 compatible to northern cold regions.; Coupled CLM3-TDSA Model testing in a boreal forest site in Saskatchewan for multiple years indicated significant improvements in the thermal response of the model, and TDSA provided better simulations of FTFs than the zero-isothermal method. Defining the start and end of the growing season using FTFs significantly improved the simulation of photosynthetic uptake. Projected warming may cause an earlier start of growing season and a higher photosynthesis during spring. Warmer temperatures would increase ecosystem respiration, causing annual net ecosystem productivity to decline.; The coupled CLM3-TDSA model was also applied to study the development of ALD (active layer depth) in permafrost regions. The roles of climate warming, soil texture and vegetation were investigated. Results showed that warming caused deeper and earlier thaw of permafrost, soil texture had a significant effect on the development of ALD, and vegetation cover affected the ALD development through its effect on ground surface energy, and the start and length of thawing season. These results showed that TDSA can be applied to study the long term changes in ALD in permafrost regions under projected climate change. |
