| In the general context of global warming, the influence of permafrost in QinghaiTibet Plateau(QTP) is becoming increasingly prominent. Understanding of the current distribution status and pattern of permafrost in QTP has a scientific and practical significance to global climate change, water resource management, and engineering planning in cold region. Statistical-empirical models have been traditionally adopted in modelling permafrost distribution in the QTP. Some recent researches have employed land surface model coupled with permafrost modelling component to simulate waterheat exchange processes in permafrost at point scale. However, traditional empirical models take insufficient account of water-heat physics. Land surface models not only need a great many of driving data, but mainly focus on heat simulation on the surface layer and are difficult to accurately simulate deep soil temperature. Numerical models have been widely applied to engineering in cold regions but rarely to the evolution of the permafrost. This study aims to develop a physical water-heat coupling model based on the heat conduction and water balance equations, to simulate temperature and moisture in both shallow and deep soil layers, which is a further development and improvement of the works of pioneer contributors. By establishing heat transfer and moisture migration equation in layered soil, the model considers the impacts of phase transformation in the freezing / melting front, dynamic changes of soil water, and unfrozen water. The distribution of permafrost in the QTP was obtained by driven the established permafrost model with field data and remote sensing data.A one-dimensional water-thermal coupling numerical model of frozen soil was built based on the pan with partial differential equations(PDE model) in MULTIPHYSICS 3.5a COMSOL. First, the universal partial differential equations were selected to build heat conduction and moisture transfer equations. Second, the geometric modeling was built. Then, water and thermal parameters were set. Finally, the temperature and moisture content of each soil layer is computed by solving the heat conduction and moisture transfer equations. The model was verified using observed data at the Tanggula station. To apply the model to the whole plateau surface, the 0cm surface temperature derived from MODIS LST product was used as the upper boundary condition for temperature, and the 0cm soil moisture, estimated by the NOAH model in the Global Land Data Assimilation System(GLDAS), was used as the upper boundary condition for soil moisture. The lower boundary condition for temperature was geothermal, which was estimated from the heat conduction coefficient combined with the temperature gradient. The volumetric water content of the lower boundary was determined according to the SHAW model. Based on the definition of permafrost and the estimated temperature and volumetric water content of the soil profile, the distribution of permafrost was judged. The result was compared with the survey maps of two typical permafrost regions, the simulation result from the Mean Average Ground Temperature Model.The main conclusions of this paper are as follows:1. The model performs well in soil temperature simulation at the Tanggula station, with an R2 above 0.88 and an RMSE less than 1 ?. The performance of soil moisture simulation is acceptable, with an R2 above 0.7 and an RMSE less than 7.65(m3·m-3). The estimated active layer thickness is about 3.6 m, and the depth of zero amplitude of soil temperature is approximately 15 m and consistent with the measured depth. The developed model is concluded to be applicable to study the changes of soil water flow and heat transport in permafrost underlain areas.2. The spatial distribution of permafrost estimated from the proposed numerical model is in good agreement with the referenced maps previously mentioned, with an permafrost area of 1,158,000 km~2. |
PDF Full Text Request