One outstanding challenge in predicting weather and simulating regional climate is to accurately represent the land-atmosphere interactions cross various temporal and spatial scales, given the tremendous heterogeneity in land surfaces. I will report recent enhancements to the Noah and Noah-MP (Multiple Parameterization) land-surface models in the Weather Research and Forecast (WRF) model and their impacts on numerical weather prediction and regional climate simulations. Recent results of evaluating the ability to simulate the evolution of seasonal evolution of snowpack in complex terrains by six land surface models will be discussed.