The clay mineralogy rather than the clay content determines radiocaesium adsorption in soils on a global scale

Abstract. The transfer of radiocaesium (137Cs) from soil to crops is the main long-term radiation risk after nuclear accidents. The prevailing concept is that 137Cs sorption in soil – and, hence, its bioavailability – is controlled by soil clay content (0–2 µm). This study tested this assumption using 24 soils collected worldwide. The radiocaesium interception potential (RIP), i.e., 137Cs adsorption, was measured for the bulk soils and for their clay and silt fractions. The RIP varied by a factor of 438 among soils and was unrelated to the clay content (p > 0.05). The RIP in the clay fractions was lowest for young volcanic soils with allophane and mica and for highly weathered tropical soils with kaolinite. In contrast, RIP values about 2 orders of magnitude higher were found in intermediate-weathered temperate soils dominated by illite. Soil RIP was, hence, related to soil illite content (R2= 0.50; p < 0.001). A significant fraction of soil RIP originated from clay minerals embedded in the silt fraction. The sum of RIP in clay and silt fractions overestimated the soil RIP by, on average, a factor of 2, indicating that the isolation of clay opens selective 137Cs sorption sites inaccessible in intact soils. Soil mineralogy, not just clay content, governs soil RIP. In terms of validity, existing 137Cs bioavailability models require recalibration for use on a global scale.