The main objective of this study is to combine theoretical and modeling work to understand the physical processes and dynamics of cold land to the atmosphere such as the effect of snow cover and soil frost with a multi-layer Soil-Snow-Vegetation Model (SSVM) coupled to an atmospheric Purdue Regional Model (PRM) over North America. It is known that the presence of snow strongly affects the surface energy budget. High albedo of snow dramatically reduces the amount of shortwave radiative energy available at the surface and its low thermal conductivity significantly restricts exchanges of heat between soil and atmosphere. The current land surface models in general have some known deficits, such as problems with snow melt, freezing snow,and no explicit implementation of wetlands in the model. Thus, improving land surface physics and new parameterizations has been developed to better understand the physical processes and dynamics of cold land on regional climate and hydrological cycle as a biosphere-atmosphere coupled system. Land Data Assimilation System (LDAS) is used to provide initial and boundary conditions for the land surface. Numerical simulations and sensitivity studies conducted show that using high-resolution land surface and soil moisture data processed from remotely sensed and ground observations can improve the location of the weather phenomena, amount of precipitation, and regional circulations. Compared with the current vegetation package, the new SSVM show significant differences in both soil moisture and ground temperature simulation in the early month. However, the early month simulation result did not show significant changes to the forecasted upper weather system, due to the fact that upper-level simulations are partly constrained by the boundary conditions from the re-analysis data. The case study being conducted is over the Northern Plains Flooding during April and May of 1997. During this period there was a severe flooding in Dakotas and Minnesota due to heavy spring snowmelt. Approximately $3.7 ~ 4.1 billion dollars of damage occurred and there were 11 casualties. Further sensitivity studies will be presesnted on the hydrological cycles of precipitation, moisture budgets, and surface energy to improve physical understandings. The coupled modeling system will help to better understand and identify the important factors influencing regional scale water and energy cycles at different spatial and time scales.