17O as a Tracer of Evaporation Losses for Waters in Arid Regions

Isotope signals δD and δ18O are widely used to study hydrological processes. The signals of evaporating water are modified by equilibrium evaporation and by diffusion. While equilibrium fractionation depends on temperature, the magnitude of diffusion fractionation depends mostly on humidity in the evaporation zone. δD of waters suffered evaporation deviates from the Global Meteoritic Water Line; this deviation is expressed as D-excess. D-excess can be used to estimate humidity conditions. However due to larger temperature dependence for δD than that for δ18O, D-excess values are temperature dependent. Recently, a new tracer of hydrological processes, namely 17O-excess was proposed (Barkan & Luz, 2007). 17O-excess is of the same nature as D-excess and is defined in a similar way. The advantage is that 17O-excess is temperature independent (Barkan & Luz, 2007). In order to analyze 17O in water samples we have set up the CoF3 fluorination method after (Barkan & Luz, 2005). The data of waters from two arid regions (Sistan, Iran and Atacama, Chile) demonstrate consistent decrease in 17O-excess with increasing δ18O (Surma et al., AGU-2013). The trends are consistent with the data of residual waters in evaporation experiments where water vapors underwent diffusion (Barkan & Luz 2007; Luz & Barkan 2009). We have also performed water vapor diffusion experiment followed by isotope analysis of vapors collected (Surma et al., AGU-2013). The advantage is that vapors demonstrate an effective isotope fractionation of evaporation (equilibrium fractionation and diffusion). Water vapors collected in our experiment demonstrate 17O increase and δ18O decrease, in agreement with marine water vapors measured (Uemura et al., 2010). Thus we conclude that 17O deviations due to vapor diffusion are common in nature. We speculate that based on some assumptions, 17O-excess values can be used to evaluate the magnitude of δ18O shifts in natural waters resulted from diffusion fractionation and hence to evaluate the magnitude of evaporation fractionation. Based on assumptions of boundary conditions, these values can be used to estimate the magnitude of evaporative losses for waters in arid regions. Discussion on the use of 17O-excess and possible limits will be presented at the meeting.