Strontium isotopes in tap water from the coterminous USA

Lesley A Chesson,James R Ehleringer,Scott A Hynek,Glen N Mackey,Diego P Fernandez,Brett J Tipple

doi:10.1890/es12-00122.1

Abstract

Strontium isotope analysis has proven useful in geo‐location investigations of organic and inorganic materials and may complement the region‐of‐origin information provided by hydrogen and oxygen stable isotope analysis. In this study, we analyzed 99 drinking (tap) water samples collected from 95 municipal water systems across the USA to investigate the potential that 87Sr/86Sr can be used to provenance samples from managed hydrological systems. Results from a leaching and exchange experiment demonstrated that non‐ideal storage conditions did not prohibit Sr isotope analysis of previously archived water samples stored in glass. Tap water samples were analyzed via multi‐collector inductively coupled plasma mass spectrometry, which was preceded by a novel, automated, in‐line Sr purification method. Measured tap water 87Sr/86Sr was compared to expected 87Sr/86Sr for collection location, which was predicted using four published isotope landscape (isoscape) models: age of bedrock (bedrock model), age plus major and minor lithology of bedrock (major bedrock model), weathering of Sr from rock (local water model), and surface fluxes within watershed (catchment model). The geologic history of the geographic regions represented by collected tap waters was diverse and we therefore expected significant covariation in measured and modeled 87Sr/86Sr values. Tap water exhibited large ranges in both Sr concentration (0–1.9 mg/L) and 87Sr/86Sr (0.7037–0.7320). Measured tap water 87Sr/86Sr ratios were significantly and positively correlated with predictions based on bedrock and catchment models. However, these bedrock and catchment models explained relatively little of the tap water Sr isotopic variation (∼10% and 17%, respectively), suggesting that the factors affecting drinking water 87Sr/86Sr are complex and more numerous than the variables included in current water models. This could be due to the reliance of some municipal water systems on groundwater, rather than surficial water sources; the natural movement of water across distinct geologic gradients; and/or the managed transport of water from source to point‐of‐use. Although published isoscape models for predicting Sr isotopic variation within the continental USA are reasonable approaches for estimating surface water 87Sr/86Sr, additional efforts are needed to develop a prediction model specifically for tap water 87Sr/86Sr.

Highlights

There is growing curiosity in stable isotope applications to understand the origins of goods, materials, and organisms; this interest is demonstrated by the recent publication of several reviews of the analytical technique for source investigation (e.g., Hobson 1999, Kelly et al 2005, West et al 2006, Ehleringer et al 2010, Oulhote et al 2011)
As expected based on the diversity of geologic history in different regions of the country, we observed large variations in both Sr concentration and 87Sr/86Sr for 99 tap water samples collected from municipalities across the coterminous USA
Combined with the results of a storage container leaching and exchange experiment, the large variations in measured 87Sr/86Sr of USA tap water suggested that it might be possible for investigators to analyze previously collected water samples and generate useful data, despite less-than-ideal storage conditions

Summary

Introduction

There is growing curiosity in stable isotope applications to understand the origins of goods, materials, and organisms; this interest is demonstrated by the recent publication of several reviews of the analytical technique for source investigation (e.g., Hobson 1999, Kelly et al 2005, West et al 2006, Ehleringer et al 2010, Oulhote et al 2011). The analysis of other elements, such as strontium (Sr), may prove useful for geo-location (Aggarwal et al.2008) and could complement the region-of-origin information provided by hydrogen and oxygen isotopes. This is because the relative abundances of Sr isotopes in the environment vary spatially and are largely dependent upon local geology (Aberg 1995, Bentley 2006). There are four natural isotopes of Sr, 3 nonradiogenic (84Sr, 86Sr, and 88Sr) and 1 radiogenic, the product of radioactive decay. The radiogenic isotope, 87Sr, is formed through the b-decay of 87Rb, which has a half-life of 48.8 3 109 years. By convention, measured Sr isotope ratios are reported as 87Sr/86Sr and calculated as

Methods

Results

Discussion

Conclusion