Establishing reference inventories of 137Cs for soil erosion studies: Methodological aspects

•Statistically sound procedures for identifying eroding sites are presented.•Effects of sample sizes on minimum detectable erosion rates are quantified.•Sensitivities of commonly used sampling schemes are systematically studied.•Diagrams are provided that assist in planning radiocaesium soil erosion studies.

The radiocaesium technique has become an attractive methodology for quantifying long-term average soil erosion and accumulation rates. It is based on comparing the 137Cs inventory in soils of a site which is suspected to be affected by soil redistribution to that of an adjacent undisturbed reference site. A major complication arises from the spatial variability of 137Cs generally present even at small areas of < 1 ha. In this paper we present statistically based procedures that efficiently allow to discriminate between soil erosion and pure spatial variation of 137Cs. We systematically explore the effect of study design and spatial variability of 137Cs on the sensitivity of the radiocaesium technique for soil erosion studies. The procedures and simulation results presented assist (i) in checking whether the sensitivity of the 137Cs technique is adequate for an intended study, and (ii) in developing an optimal study design.In addition it is shown that some of the statistical criteria proposed in the literature for discriminating soil redistribution from 137Cs variability show high error risks and should be avoided.