Uncertainties in estimating water fluxes and residence times using environmental tracers in an arid unsaturated zone

Environmental tracers are used widely to evaluate flow processes and estimate fluxes and ages of pore water in arid regions. The purpose of this study was to evaluate uncertainties in water flux and age on the basis of data from environmental tracers, including meteoric Cl, 36 Cl, 3 H, d 2 H, and d 18 O in porous media. Representative profiles of environmental tracers from drainage and interdrainage areas at a site in the Chihuahuan Desert of Texas were evaluated. The chloride mass balance approach (CMB) was used to evaluate water fluxes and ages. The long residence times indicated by the Cl data in interdrainage areas (55,000 to 105,000 years to 25 m depth) were generally corroborated by residence times estimated from radioactive decay of 36 Cl (39,000 6 13,000 to 59,000 6 14,400 years). Uncertainties in the CMB approach include uncertainties in transport processes, Cl input, and Cl output. Although the CMB approach assumes one- dimensional, downward piston flow, water potential and stable isotope data in interdrainage areas suggest net upward water movement. Cl data indicate that drying of the profiles may have persisted throughout the Holocene (;10,000 years). Therefore the downward flow assumption may only be applicable in the older, deeper sections of the profiles. Cl diffusion is significant near the surface where Cl concentration gradients are steep. Anion exclusion may affect calculated water fluxes based on Cl in clay-rich zones. Although it is difficult to quantify uncertainties in diffusion and anion exclusion processes, they act in concert and result in overestimation of water flux and underestimation of age by the CMB approach. Therefore, in interdrainage areas the CMB approach provides an upper bound on actual water fluxes and a lower bound on actual ages. Error bars on these bounding estimates were evaluated on the basis of uncertainties in Cl input (;635%) and in Cl output (63%) that result in 638% uncertainty in water flux and 224 to 56% uncertainty in water age in interdrainage areas. In drainage areas it is much more difficult to apply the CMB approach because of preferential flow, large uncertainties in Cl input as a result of run-on, reduced sensitivity of Cl to water flux, and analytical uncertainties in Cl measurements. Although preferential flow was shown by 3 H data, mixing calculations suggest that 36 Cl/Cl ratios cannot be used to evaluate preferential flow when Cl concentrations in the matrix exceed 10 to 100 g m 23 , as is found in the playa and the fissure. Neglecting Cl input from run-on results in underestimation of water flux by about an order of magnitude. Therefore the apparent CMB water flux, which ignores preferential flow and run-on, represents a lower bound on the actual water flux in contrast to an upper bound for interdrainage areas. These results have important implications for waste disposal in arid regions because they suggest that water fluxes estimated using the CMB approach are conservatively high in interdrainage areas characterized by porous media.