The triple oxygen isotope composition of leaf waters in Mpala, central Kenya

The triple oxygen isotopic composition of water is an emerging tool for investigating the hydrological environment. The δ18O-δ17O relationship differs during kinetic and equilibrium isotope fractionation, such that the 17O depletion can be sensitive to relative humidity (Rh) during kinetic fractionation, mixing among different pools, and to the specific mode of kinetic fractionation. It has been proposed that the δ18O-δ17O relationship during evapotranspiration, as characterized by the slope λ(stem-leaf) on a ln(δ17O+1) vs. ln(δ18O+1) plot, is mainly controlled by Rh but not affected by other environmental conditions or by plant species. In order to understand the sensitivity of λ(stem-leaf) to Rh and the utility of 17O-excess (the deviation of δ17O from a reference slope) in the terrestrial biosphere as a tracer of Rh conditions today and in the past, this study expands the triple oxygen isotope measurements of leaf waters to additional species in a semiarid environment. Paired stem and leaf waters of Acacia and grasses were collected in the Mpala Research Center in central Kenya and analyzed for their triple oxygen isotope composition. Leaf waters that were sampled diurnally (8 sampling intervals between 6 am and 5 pm) exhibit a range in δ18O and 17O-excess values of 11.2‰ and 107 per meg respectively for Acacia brevispica, and 14.4‰ and 147 per meg for the grass Panicum maximum. Except for one sample collected at 7am, the λ(stem-leaf) values for grasses are systematically lower (0.0012 to 0.0110) than Acacia λ(stem-leaf) values at the corresponding time of day; this might be explained by the progressive evaporative isotopic enrichment and mixing processes of leaf water along parallel veins of grass leaves. Most of the triple oxygen isotope composition of the Acacia leaf waters can be predicted using Craig-Gordon model. We built a mass balance model of an evolving leaf water system from nonsteady-state to steady-state conditions during evapotranspiration to explain the diurnal 17O-excess variation of the Acacia leaf waters. The results from this model indicate that 17O-excess of the leaf waters may be affected by variations in stomatal conductance, leaf water volume, and water residence time, in addition to variations in Rh. Our study emphasizes that the triple oxygen isotope composition of leaf waters is controlled both physically by Rh conditions but also by the physiological responses of a plant to its environment. The triple oxygen isotope composition of grass leaf water is a result of both evapotranspiration and mixing processes, and thus species effects should be considered in the discussion of λ(stem-leaf)-Rh relationships. A better understanding of the λ(stem-leaf)-Rh relationships could provide a framework for interpreting variations of triple oxygen isotopes of tree cellulose or fossil tissues as proxies for paleo-hydrological change.