TerrSysMP-DART Interface: An Integrated Data Assimilation Platform for Coupled Atmosphere, Land Surface and Groundwater Model.

The Terrestrial Systems Modeling Platform (TerrSysMP) has evolved as a highly modular interface, which uses an external coupler (OASIS3 coupler interfaced with MCT; OASIS3-MCT) and three component models for the atmosphere (Consortium for Small-Scale Modeling; COSMO), land surface (the NCAR Community Land Model; CLM) and groundwater (ParFlow) to improve our understanding of time-space patterns of energy, water and carbon fluxes and feedbacks in the terrestrial system. For predictions of the terrestrial system, uncertainty estimates of states and fluxes of the atmosphere, land surface and groundwater, and the connection with commensurate observations via ensemble data assimilation with corresponding observations is also crucial. The Data Assimilation Research Testbed (DART) provides a community-based and flexible data assimilation (DA) capability across all components, which is easy to implement and to customize according to model and data needs. In this study, we present the development of a TerrSysMP-DART interface for COSMO, CLM and ParFlow and its experimental application for weakly coupled DA for each model component in fully coupled runs. We use the synthetic observations of boundary layer temperature profile, soil temperature and soil moisture for the atmosphere, land surface and groundwater model respectively. The DA experiment examines the impact of the above observations on the behavior of the individual component models and their influence on the propagation of patterns of energy, moisture and carbon fluxes along the assimilation cycle.