Land surface schemes in atmospheric and hydrological models describe the exchange of momentum, heat and moisture between the soil and the atmosphere. These schemes are usually calibrated and optimized with respect to the turbulent fluxes with the forcing data (mean atmospheric variables) and validation data (fluxes) being measured at a tower respresentig the local scale (e.g. PILPS 2a). On the other hand, runoff is being considered as an integral quantity representing the outflow of a river system. With respect to runoff land surface schemes are usually calibrated on regional or catchment scale (e.g. PILPS 2c).
In order to overcome this scale distinction between fluxes and runoff the land surface scheme SEWAB was validated with respect to runoff on a local scale. A six month time series - 20 min averages - of forcing data(atmospheric variables) and validation measurements (surface and subsurface runoff, soil moisture, water level) is used from a small research area of 14.5 hectare size in the south of Ireland. The site is agricultural grassland homogeniously distributed over the small catchment area.
The rainfall regime is characterised by long duration events of variable intesity and total depth which occur at any time of the year. Short duration, high intensity events occur mainly in summer. The rainfall - runoff relationship reflects a fast response of runoff to rainfall events with some reduction in the amount of water due to ponding.
Among different hydrological components of SEWAB (VIC approach, water table change, distinction of fast and slow baseflow components, depth dependend saturation hydraulic conductivity) we included a description of the infiltration (micro- and macropores) and ponding process which allowed a reasonable simulation of the runoff at this small scale.
The hydrological processes, the tuning of SEWAB and the comparison between simulated and observed runoff is explained.