Extreme events such as droughts and heat waves have serious and damaging impacts on terrestrial processes. Under climate change, these extreme weather events are likely to shift in both magnitude and frequency at regional and local scales. The resulting interactions and feedbacks between the terrestrial and atmosphere systems could lead to non-linear and/or threshold responses in the eco-climate system, and raise a concern as to the resiliency of natural as well as managed ecosystems under extreme changes. This study investigates the response of ecosystem to droughts at different time scales and magnitudes. Four land surface models with different bio-geophysical parameterizations and representations are used to simulate soil-canopy processes, such as evapotranspiration, during these extreme events. The Terrestrial Ecosystem Model (TEM) is a process-based ecosystem model that uses spatially referenced information on climate, elevation, soils, vegetation and water availability to make monthly estimates of vegetation and soil carbon and nitrogen fluxes and pool sizes. There are two versions of TEM model, the TEM-Hydro daily model and the TEM monthly model. The Advanced Canopy-Atmosphere-Soil Algorithm (ACASA) is a multi-layered land surface model based on eddy-covariance theory to calculate the biosphere-atmosphere exchanges of carbon dioxide, water, and momentums. The Community Land Model (CLM) is a community-based model consists of biogeophysics, hydrological cycle, biogeochemistry and dynamic vegetation. Model simulations are evaluated using the biogeophysical and micrometeorological field observations from the AmeriFlux sites across the US. Preliminary results indicate that during a severe drought the link between evapotranspiration and Net Ecosystem Productivity (NEP) in the models is weaker than what observations indicate. This and other interpretations are presented and discussed in the context of planned experimental simulations with fully coupled regional ecosystem-climate model(s) driven by an integrated earth-system model.