Defining monitoring strategies for runoff and erosion studies in agricultural catchments: A simulation approach

The number of runoff and erosion studies in cultivated catchments in north west Europe is increasing. Measurements are usually costly and time-consuming, in particular of soil hydrological properties because agriculture causes considerable spatial and temporal variation. A subdivision of the catchment is necessary to optimize the choice of sampling sites and to define extrapolation procedures. Four sampling strategies were examined, consisting of subdividing the catchment according to the crops (1), or according to the surface structure (2), followed by increasing the number of sampled fields by adding fields that are presumed to have maximum runoff activity (3) or intermediate activity (4). These strategies were compared by modelling the runoff and erosion in a virtual catchment, containing 15 arable fields with various crop sequences, at three times during a winter season. Strategy 2 gave better results than 1, because of the large variability of surface structure of the bare fields, which was not correlated with the crop type. Runoff prediction benefited from extra sampling only when groups of fields with a large heterogeneity in hydrological behaviour were included. The quality of erosion prediction increased when the steeper fields were included in the sampling scheme. A monitoring strategy is proposed that consists of: (i) repeated, fast and exhaustive descriptions of the soil surface using surface structure classes, (ii) subdivision of the catchment into groups of fields with identical surface characteristics before measurements are made, and (iii) sampling and extrapolation according to these groups.