Mineral mediated isotope fractionation of soil water.

ABSTRACT

RATIONALE The ability to recover the isotopic signature of water added to soil samples that have previously been oven-dried decreases with the increasing presence of silt and clay. The effects on the isotopic signature of water associated with physicochemical soil properties are not yet fully understood, for either hydration or dehydration of soil samples. METHODS: The soil sample chemistry and the crystallinity of minerals were measured by X-ray fluorescence and X-ray diffraction. The organic carbon and the cation-exchange capacity were also determined. Water of known isotopic signature was used to spike an oven-dried substrate and subsequently extracted by cryogenic vacuum extraction at a temperature of 105°C. In addition, the soils were oven-dried at 205°C and water extractions were also performed at 205°C. The isotopic signatures of the water samples were determined by cavity-ring-down spectrometry. RESULTS: The isotope effects caused by the cryogenic vacuum extraction method applied to soils with elevated clay content were reduced. First, by increasing the extraction temperature to 205°C, we improved the precision of the cryogenic vacuum extraction method and the recovery of the known isotopic signature of the spike water. Secondly, the post-correction of data based on the physicochemical soil properties and a common extraction temperature of 105°C improved the measurement trueness. CONCLUSIONS: The isotopic signature of soil water is influenced by mineral-water interaction. During the hydration of clay, different minerals deplete free water in heavy isotopes. The extracted soil water (dehydration water) gathered from clay-rich soils is generally more depleted in the heavy isotopes than the spike water, making results obtained for different soil types difficult to compare. Isotope effects observed at the mineral-water interface highlight potential explanations for eco-hydrological separation of water pools. Copyright © 2016 John Wiley & Sons, Ltd.