Reconstructed Precipitation Tritium Leads to Overestimated Groundwater Recharge

Abstract

Time series of precipitation tritium contents are often used for groundwater recharge estimation based on the tritium mass ratio of the (un) saturated zones to precipitation. But records of the precipitation tritium measurements are sparse and often need to be reconstructed. Uncertainties associated with reconstructed records can affect the recharge estimates, but this has received little attention. Here we leverage four >15 m deep, exceptionally well preserved unsaturated zone tritium profiles to quantify the uncertainty in Tritium mass balance from precipitation tritium reconstruction. The tritium profiles have bomb tritium peaks at 7.2–10.5 m below the surface and peaks ranging from 46 to 235 tritium unit. We first estimate the diffuse recharge by the tritium peak method. Since this method is independent of any precipitation tritium reconstruction, the estimated recharge was used as truth to evaluate four precipitation tritium reconstruction methods: two interpolation methods and two reference curve methods. Direct comparison between the observed and simulated precipitation tritium showed that the reference curve methods performed more poorly than the interpolation methods. Indirect evaluation by comparing the recharge rates estimated by the tritium storage method with our truth measurements showed that tritium storage method overestimated recharge by 100% to 200%. Our results suggest that the atmospheric tritium flux was greatly underestimated, especially by the reference curve methods. As such, groundwater recharge at our sites would be overestimated. This has implications for using these standard approaches elsewhere.

Global Isotope Hydrogeology―Review

“…Some of the thermonuclear‐test‐derived tritium was transferred into reservoirs such as seas and aquifers by rain and recharge (Michel, ; Vogel et al, ); widespread sampling of thousands of groundwater springs and well waters confirms its presence (Figure ). Subsequently, tritium has been measured widely to assess downward soil water fluxes (e.g., Allison et al, ; Allison & Hughes, , ; Cook et al, ; Daniels et al, ; Gvirtzman & Magaritz, ; Li et al, ; Li & Si, ; Lin & Wei, ; Phillips, ; Phillips et al, ; Si & de Jong, ; Verhagen et al, ), groundwater velocities and the presence of recently recharged waters in the aquifer (e.g., Allison, ; Allison & Holmes, ; Cook et al, ; Currell et al, ; Ette et al, ; Hendry, ; Libby, ; Martinec, ; McMahon et al, ; Price et al, ; Qin et al, ; Robertson & Cherry, ; Scanlon et al, ; Schlosser et al, ; Solomon et al, ; Stuart et al, ; Su et al, ; Theis, ; Visser et al, , ; Vogel et al, ; Wang et al, ), groundwater discharges to rivers and river water “ages” (e.g., Brown, ; Cartwright & Morgenstern, , , , ; Dinçer et al, ; Duvert et al, ; Hofmann et al, ; Howcroft et al, ; Koeniger et al, ; Małoszewski et al, ; Martinec, ; Michel, ; Michel et al, ; Rank et al, ; Stewart et al, , ; Stewart & Morgenstern, ; Visser et al, ), timespans elapsed between rainfall and root water uptake (Zhang et al, ), and well water vulnerabilities to pollut...…”

Reconstruction of the record of tritium in precipitation in the temperate zone of South America

“…Despite their usefulness, all these approaches have different advantages, disadvantages and degrees of uncertainties that must be analysed when choosing the most suitable for a particular area. Examples of application of these methods, mostly restricted to specific regions, can be found in Ba and Xu (2010) in the North Pacific, Zhai et al (2013) and Li and Si (2018) in China, Michel et al (2018) in the United States, and Jerbi et al (2019) in Tunisia. In South America, 3 H input functions were calculated for Cochabamba, Bolivia, by Stimson et al (1996) applying an extrapolation method and for an area of the state of Mina Gerais, Brazil, by Silva and Cota (2021) using correlation and triangulation methods.…”

Groundwater recharge in typical geomorphic landscapes and different land use types on the loess plateau, China

“…The model runs passed the Gelman-Rubin and Geweke diagnostics tests. with an age of approximately 50 years (Chen et al, 2012;Li & Si, 2018). Likewise, 14 C dating results from several studies show groundwater ages between 136 and 23 412 years, indicating paleo precipitation recharge (Li et al, 2019).…”