Groundwater changes in evaporating basins using gypsum crystals’ isotopic compositions

Abstract:

While the dynamics of groundwater evaporation are well known, it is still challenging to reconstruct the water patterns in areas where water is not available anymore. We selected a specific location in White Sands National Monument (WSNM), New Mexico, to validate a method to extract information from hydrated minerals regarding past groundwater evaporation patterns in evaporitic basins. WSNM has gypsum (CaSO₄.2H₂O) dunes and crystals precipitated from the evaporation of an ancient lake. Our approach aims to extract the water of crystallization of gypsum and measure its oxygen and hydrogen isotopic compositions, in order to reconstruct the groundwater history of the area. The idea is that as the mother brine evaporates its isotopic composition changes continuously, recorded as water of crystallization in successive growth zones of gypsum. To check if the isotopic composition of the salt could effectively differentiate between distinctive humidity conditions, the methodology was tested first on synthetic gypsum grown under controlled humidity and temperature conditions. T and RH% were maintained constant in a glove box and precipitated gypsum was harvested every 24 hours. d²H and d¹⁸O of water of crystallization from the synthetic gypsum was extracted using a specially developed technique on a TC/EA. The brine was measured using a Gas Bench II for d¹⁸O and an H-Device for d²H on a Thermo Finnigan MAT 253 mass spectrometer. With the method tested, we measured natural gypsum. In order to identify the growth zones we mapped the surface of the crystals using an experimental space flight XRF instrument. Crystals were then sampled for isotopic analyses. Preliminary results suggest that site-specific groundwater changes can be described by the isotopic variations. We will show that the methodology is a reliable and fast method to quantify hydrological changes in a targeted environment. The study is currently ongoing but the full dataset will be presented at the conference.