Simulations of energy and water balances with WRF and WRF-Hydro models: the role of model coupling and parameterizations

We tested the widely used atmospheric WRF (Weather Research and Forecasting) model in a coupled mode to the hydrological model WRF-Hydro. The coupled WRF/WRF-Hydro model adds the simulation of horizontal surface and subsurface flow of water relative to the standalone WRF. We conducted simulations for the Mediterranean island of Cyprus and 31 small mountainous river basins for the hydrological year 2011-2012. We found higher soil moisture (20%), more evapotranspiration (33%) and a small increase in rainfall (3%) for the coupled WRF/WRF-Hydro model, compared to the WRF model without horizontal flows. We also forced WRF-Hydro with observed rainfall and five different set-ups of WRF and examined the modelled streamflow. The WRF set-ups were adapted from combinations of different microphysics, cumulus cloud, planetary boundary layer and surface layer schemes. We found that WRF-Hydro with observed rain underestimated the average streamflow by 6%, during a two-year simulation (2011-2013). The best of the five WRF set-ups showed a 19% underestimation of the average streamflow, thus, an optimized ensemble of WRF set-ups is needed to model the streamflow. Our study suggests that the coupling of WRF with the WRF-Hydro model can improve land-atmosphere simulations. We will also present the calibration of parameters of the land surface component of the coupled model with observations of soil moisture and transpiration that could further enhance the ability of the model to represent the different parts of the combined terrestrial-atmospheric water cycle.