CRITICAL EVALUATION OF ISOTOPE END-MEMBERS FOR GROUNDWATER SOURCE PARTITIONING

Implementations of isotope end-member mixing analysis in hydrologic investigations is deserving of additional critical scrutiny similar to analyses conducted in other fields, such as trophic ecology. In many cases, the spatiotemporal variability of isotopic ratios in precipitation is subject to multiple dependent variables making definition of end-members, for use in mixing models, a major challenge. In such cases where the system cannot be fully constrained with stable isotopes (δ²H and δ¹⁸O) alone, the hydrologic model must be carefully conceptualized, and ancillary information taken into account. Additionally, spatiotemporal coverage of precipitation data is often insufficient to easily determine end-members from empirical data alone. It has too often been the case in hydrologic investigations that end-members are poorly conceptualized or even statistically indistinguishable from one another, bringing into question the utility of the subsequent mixing analysis. A case study for the Grand Canyon region of northern Arizona is presented, where a strong seasonal and elevational difference of δ²H and δ¹⁸O in precipitation is empirically observed. Groundwater δ²H and δ¹⁸O which fall outside of the range observed in precipitation suggest empirical precipitation data is insufficient to fully capture the system and modeled values of δ²H and δ¹⁸O in precipitation are used to fill the gap. Because the influence of seasonal and elevational changes on isotope end-members in precipitation are similar, statistical tests are used to determine the significance of the respective drivers and whether a statistically meaningful end-member can be determined. Using an iterative process of model conceptualization and quantitative analysis, a seasonal end-member mixing model was developed to improve understanding of recharge sources on the South Rim of the Grand Canyon and suggests a previously unquantified recharge source from the eastern Coconino Plateau.