The main problem with nitrogen (N) at dairy farms on sandy soils in the Netherlands is the leaching of nitrate to the groundwater. Experimental farm De Marke was set up in 1991 on poor sandy soils with the objective to develop a prototype of an economically feasible farming system with acceptable nutrient losses. For the quantification of the inputs and outputs of the De Marke farming system, measurements have been carried out on subsystems of the farm. This thesis uses the data collected for assessing the performance of the crop-soil-nitrogen subsystem.Monitoring of soil moisture conditions and nitrate concentrations was carried out during the years 1991-1995 at six experimental sites. With these data, studies were carried out on nitrate leaching affected by land use in relation to the occurring soils at the farm and the groundwater regimes, using simulation models. The use of simulation models can only be valid when the right input data are used. Soil physical characteristics are important input parameters and for the monitoring sites a comparison was made using soil physical characteristics from either laboratory measurements or from the Staring series as input. It was found that simulation results were not significantly different, implying that the Staring series could be used in studies like these for simulating the unsaturated water flow regime in sandy soils.Cattle grazing at the experimental farm was reduced to eight hours per day, but urine-affected areas had great influence on the nitrate measurements. At one 'wet' site the probability of exceeding the EC-directive for drinking water (11.3 mg/l nitrate-N) under a urination deposited in either July or September was respectively 10 and 25%. At the dry site the directive will be exceeded under any urine patch in almost 100% of the years, affecting the field average concentration. In this field careful grazing management would result in less nitrate leaching, but the environmental goals would not be reached.A precision agriculture technique which allows ...