Study On Thermodynamic Watershed Hydrological Model For Cold Regions

Representative Elementary Watershed (REW) approach first proposed by Reggiani et al. (1998) presents a different perspetive for hydrological modeling via developing scale adaptable equations. Thermodynamic watershed Hydrological Model (THModel) developed by Tian et al. (2006) further extends the REW theory by explicitly incorpotating energy balance equations. In this extensible framework, the energy sensitive hydrological processes such as evaporation, transpiration, snow and glacier melting, soil freezing and thawing can be modeled in a physically reasonable way, which provides a flexbile modeling framework for cold regions.Catchment is a complex system with a certain degree of self-organization. The interaction of different processes plays an important role in evolution of this particular system. For cold regions, water movement is intensively coupled with energy processes which should be coped with in a simultaneously manner. In this thesis, a thermodynamic hydrological model for cold regions is developed based on the THModel framework. The energy balance equations for the glacier zone, snow zone and unsaturated zone are closed aided by energy balance method, degree-day approach and the maximum unfrozen-water content model. Also, the characteristics of constitutive relationships at the REW scale are analyzed by virtual experimental method, and the promising method to develop the general closure relations is proposed.The Thermodynamic Hydrological Model for Cold regions is applied to two catchments, the headwaters of the Urumqi River basin and Reynolds Creek Experiment Watershed with different climates. In both basins the model can simulate the streamflow reasonable well. In the application of the headwaters of the Urumqi River basin, the significant influences of seasonality of energy input on the runoff generation which presents the tight coupling of water and energy. In the Reynolds Creek Experiment Watershed, the model can not reproduce the annual daily discharge below 0.2m3/s very well, but the simulated snowcover area ratio, snow temperature and snow density follow the physical law of snow accumulation and depletion.The sensitivity analysis of model parameter is carried out by both"one-at-a-time"approach and Generalized Likelihood Uncertainty Estimation (GLUE) method. The results show the heterogeneity coefficient of radiation has significant influences on the watershed hydrological responses, and additional data can reduce the uncertainty to a certain extent. The prediction uncertainty relating to the form of snow delption curve is also estimated for Reynolds Creek Experimental Watershed.