Stable isotope ratios of hydrogen (δ2H values) in local meteoric water correlate with those of nonexchangeable H (δ2Hn) in bulk soil and organic matter, allowing for probabilistic spatial assignments to soil samples. We hypothesized that there is a similar correlation between the δ2H values of local meteoric water and the δ2Hn value of soil clay fractions. The H pool of topsoil clay fractions is usually dominated by pedogenic clay minerals, which contain structural hydroxyl-H, preserving the δ2H values of ambient water at the time mineral formation. We applied a steam equilibration method with water vapors of known δ2H values to determine δ2Hn values in soil clay fractions by eliminating the influence of easily exchangeable H. We collected topsoil samples from 24 locations on five continents with a wide variation in δ2H values of meteoric water. To calculate δ2Hn values via steam equilibration, the equilibrium fractionation factor between exchangeable H in the clay fractions and equilibration waters (αex–w) is needed, which is, however, unknown. We therefore assessed the effect of αex-w values from 1 (no fractionation) to αex-w = 1.08, which has been used in steam equilibration approaches of organic matter and bulk soil, on the δ2Hn values of soil clay fractions. The δ2Hn values of the clay fractions ranged from –167 ± SD1‰ for a northern Siberian sample to –44 ± 4‰ in western Kenya for αex-w = 1 and from −191 ± 5 to –81 ± 5‰ for αex-w = 1.08. The δ2H values of local meteoric water correlated significantly with the δ2Hn values of the clay fractions (r = 0.85 for αex-w = 1 and 0.65 for αex-w = 1.08, p < 0.001). A multiple regression model including hillslope as a possible measure of the accumulation of clay minerals with a slightly different δ2Hn value from nearby locations by eolian or aqueous transport in addition to the mean seasonal δ2H values of local meteoric water explained up to 89% of the observed variation in δ2H values of the clay fractions. Our results demonstrate that the δ2Hn values of soil clay fractions are driven by the δ2H values of local precipitation on a global scale, irrespective of the used αex–w value and influenced by the mineral type. Overcoming previous methodological limitations opens up opportunities for probabilistic spatial assignments of unknown organic matter-poor topsoil samples, applications to paleosoil problems or to experimentally track clay mineral neoformation in soils by tracer experiments.
Read full abstract