A Statistical Approach for Estimating the Radiocesium Interception Potential of Soils

Abstract The solid‐liquid distribution of radiocesium (137Cs) varies extensively among soil types. A statistical approach was used to relate 137Cs sorption with soil properties. Eighty‐eight pasture soils were sampled in Belgium. All soil samples were characterized for texture, cation exchange capacity, exchangeable K, Ca, and Mg content, pH and organic C content. Soil samples were equilibrated in a 10−1 M CaCl2 — 5.10−4 M KCl solution. Adsorption of 137Cs was measured after 16 h in this medium. The solid‐liquid concentration ratio, KD, of 137Cs in this specific ionic scenario allows for the calculation of the Radiocesium Interception Potential (RIP), which in turn can be used for KD predictions for a wide range of different ionic scenarios. The RIP varied from 54 to 11 200 mmolc kg−1. Stepwise multiple regression between RIP and soil characteristics yielded statistically significant models (P < 0.001) for samples from each northern and southern Belgium. For northern Belgium the regression model only included texture fractions (R = 0.86, n = 62) whereas for southern Belgium the regressors were texture fractions and soil pH (R = 0.88, n = 26). The correlations between the RIP and soil characteristics depend strongly on the geological origin of the soil.