On the accuracy of noble gas recharge temperatures as a paleoclimate proxy

Abstract

Dissolved noble gases in groundwater are an important terrestrial temperature proxy for the last glacial maximum (LGM). Noble gas temperatures (NGT) provide a record of long‐term mean water table temperature (WTT) during groundwater recharge. For NGT to accurately represent surface air temperatures (SAT), the difference between mean annual air temperature (MAAT) and WTT must be known through time. Many paleoclimate studies reference NGT without articulating the potential difference between NGT and air temperature. Recognizing the array of climatic changes that have occurred since the LGM, it is possible some of these changes have altered the relationship between WTT and MAAT in groundwater recharge zones. The coupling of WTT and MAAT was evaluated in numerical modeling experiments that examined WTT sensitivity to changes in (1) precipitation amount, (2) water table depth, and (3) air temperature. Moderate changes in precipitation amount (±20%) and water table depth (1–2 m) caused WTT‐MAAT decoupling of ∼0.2°C. Varying air temperature, either MAAT or annual amplitude, changed the duration of snow cover which caused seasonal decoupling of WTT from SAT. Assuming SAT was actually 5–7°C cooler at the LGM than at present, these modeling experiments suggest that errors associated with WTT‐MAAT decoupling at snow‐free sites are insignificant given the precision of NGT. However, results indicate that WTT‐MAAT decoupling could cause an underestimation of the actual SAT change by ∼1.4°C at sites having seasonal snow cover.