Coupled atmospheric-hydrological modeling for feedback investigations in the Poyang lake catchment, China

Coupling terrestrial hydrological and atmospheric models allows investigations of the role of water in the earth system in a more integrative way. In particular, for the analysis of possible changes of the hydrological cycle due to human-induced climate change, land cover conversions, and water resources management feedback mechanisms among the earth surface, subsurface and atmosphere are crucial. Investigations of such feedback, which is primarily caused by water and energy fluxes, require a cross-compartment two-way coupled atmospheric-hydrological modeling system. For this purpose, we coupled the regional atmospheric model WRF-ARW and the distributed hydrological model HMS. Both models share the same land surface model: the Noah-LSM. This model system enables long term simulations for investigations of land use and/or climate changes on the hydrological cycle. The model system is applied for the Poyang lake basin in China with a catchment size of approximately 160,000 km2 using a spatial resolution of 10x10 km2. In addition to the projected climate change signals, human-induced rapid changes of land use occurred in the past and will likely continue for this region. Prior to the application of the coupled model system, offline simulations of the advanced weather research and forecast model (WRF-ARW) are performed to identify a suited setup for the study area. For this purpose, several configurations of WRF-ARW, using ECMWF's ERA-INTERIM reanalysis data as driving data, are compared and validated with publicly available observational data sets for the period 2003-2005. For the coupled system, the integration of HMS into WRF-ARW required in addition the implementation of an interface between saturated zone (groundwater model) and the LSM which enables the feedback between the different compartments. With this coupled model system, the potential of the integrated WRF-HMS simulations is evaluated by analyzing and comparing first simulation results with and without feedback.