Long-term and high-frequency non-destructive monitoring of water stable isotope profiles in an evaporating soil column

Abstract

The stable isotope compositions of soil water (δ2H and δ18O) carry important information about the prevailing soil hydrological conditions and for constraining ecosystem water budgets. However, they are highly dynamic, especially during and after precipitation events. In this study, we present an application of a method based on gas-permeable tubing and isotope-specific infrared laser absorption spectroscopy for in situ determination of soil water δ2H and δ18O. We conducted a laboratory experiment where a sand column was initially saturated, exposed to evaporation for a period of 290 days, and finally rewatered. Soil water vapor δ2H and δ18O were measured daily at each of eight available depths. Soil liquid water δ2H and δ18O were inferred from those of the vapor considering thermodynamic equilibrium between liquid and vapor phases in the soil. The experimental setup allowed for following the evolution of soil water δ2H and δ18O profiles with a daily temporal resolution. As the soil dried, we could also show for the first time the increasing influence of the isotopically depleted ambient water vapor on the isotopically enriched liquid water close to the soil surface (i.e., atmospheric invasion). Rewatering at the end of the experiment led to instantaneous resetting of the stable isotope profiles, which could be closely followed with the new method. From simple soil δ2H and δ18O gradients calculations, we showed that the gathered data allowed one to determinate the depth of the evaporation front (EF) and how it receded into the soil over time. It was inferred that after 290 days under the prevailing experimental conditions, the EF had moved down to an approximate depth of −0.06 m. Finally, data were used to calculate the slopes of the evaporation lines and test the formulation for kinetic isotope effects. A very good agreement was found between measured and simulated values (Nash and Sutcliffe efficiency, NSE = 0.92) during the first half of the experiment, i.e., until the EF reached a depth of −0.04 m. From this point, calculated kinetic effects associated with the transport of isotopologues in the soil surface air layer above the EF provided slopes lower than observed. Finally, values of kinetic isotope effects that provided the best model-to-data fit (NSE > 0.9) were obtained from inverse modeling, highlighting uncertainties associated with the determinations of isotope kinetic fractionation and soil relative humidity at the EF.