Added seasonal forecasting skill from land surface parameterization detail

During dry spells, a large part of the Netherlands depends on water from the IJssel lake, a large surface water reservoir. Water is extracted for a number of purposes, such as irrigation, water quality, shipping and drinking water. Besides local precipitation, the main source of water flowing into the lake is the river IJssel; a distributary of the Rhine. To keep water available for extraction by the surrounding regions, lake levels cannot be allowed to fall more than about 20 cm under the regular summer maintenance level. Prior to the onset of a drought, therefore, it might be desirable to raise lake levels to maintain sufficient water availability during the dry spell. For adequate management of the reservoir, therefore, long-range forecasting of precipitation and river discharge would be extremely helpful. However, meteorological forecast skill is known to be nearly absent for lead times longer than about a month in northwestern Europe. The land surface contains a number of components that may increase forecast skill for Rhine river discharge; examples are the amount of snow in the Alps, groundwater, and soil moisture. We investigate to what extent this is the case and whether the forecast skill of Rhine river discharge forecasts increases with increasing detail in the land surface parameterization of the initial conditions. We collected streamflow reforecasts from various sources: ECMWF SEAS5, EFAS, SMHI-HYPE and a high-resolution distributed hydrological model (WFLOW), forced by ECMWF SEAS5 meteorological forecasts.