MEASUREMENT OF RUNOFF AND SOIL-EROSION ON REGULARLY CULTIVATED FIELDS IN SWITZERLAND - SOME CRITICAL CONSIDERATIONS

In field measurements of soil erosion spatial variability is usually neglected. Due to natural site conditions, however, considerable variations of runoff and soil loss can occur even within apparently homogeneous fields. This was observed on large (75 m(2)) erosion test plots in triple replication in a forage-maize cropping experiment. The variability, with mean coefficients of variation ranging from 30% to 50%, partially obscures the expected differences between the compared treatments. Soil loss computations based on unit area (e.g. single plots) thus hardly allow comparisons among dissimilar sites and treatments or an extrapolation to larger areas. Thus for the presented experimental design we suggest as many replications as possible. At a similiar level of measurement accuracy, this is easier attainable with a flexible set-up and open plots than with permanently installed large experimental stations.