Water Temperature: Theoretical Developments
During CROSSROADS, the model HaliBubble was created. It allows quantifying paleo-brine temperature from observed phase changes in fluid inclusions in halite crystals. The model quantitatively links the homogenization temperature of inclusions (the temperature at which vapor disappears from the inclusion) to the true water temperature at the time of trapping, accounting for surface tension, host-crystal effects, and hydrostatic pressure. This framework provides the first physically constrained halite thermometer, published in American Journal of Science (2025), with an open-access interface for community use.
To link these inclusion data to climate, a second physical model was developed that predicts lake water temperature from air temperature, wind speed, humidity, and cloud cover. It reproduces the Dead Sea’s observed warming since 1979 and predicts extreme future heating due to rising salinity and reduced evaporative cooling. Applying the model to the Messinian Salinity Crisis (~5.5 Ma) revealed similar processes, implying that shrinking hypersaline lakes undergo strong positive feedbacks enhancing heat retention. These results, available as a preprint on EarthArXiv, challenge long-held interpretations of evaporite deposits and hypersaline climate feedbacks.
Experimental Validation
To test the models, fieldwork was conducted in the Dead Sea in 2023, to sample halite that crystallized 35 m below the surface in the 1980s, when instrumental temperature data exist. Using femtosecond laser-induced vapor nucleation and homogenization temperature measurements on hundreds of inclusions at the University of Bergen, reconstructed and observed temperatures agreed within 0.1 °C. This confirmed the method’s accuracy, published in Chemical Geology (2024).
Application to the Holocene
The new paleothermometer was applied to halite from a deep sediment core from the Dead Sea spanning the last 12 ka. The resulting temperature series, with ~400-year resolution, reveals previously unresolved Holocene climate variability in the Levant. A manuscript detailing this dataset is in preparation.
Water Composition
To complement temperature reconstructions, the chemical composition of fluid inclusions was characterized. At the University of Lyon, density of biphasic inclusions was measured using Brillouin spectroscopy, while at Binghamton University elemental ratios were determined via LA-ICP-MS. Combined with chemical modeling, these data reconstruct the evolution of major ion concentrations (Mg, K, Ca,Cl, Na, SO4) in the Dead Sea throughout the Holocene, providing the most accurate Holocen etimeseries of brine composition.
Lake Level Reconstruction
Because Dead Sea level reflects regional rainfall, past levels were inferred using concentrations of “conservative” elements—those linearly related to lake volume—combined with bathymetry. The resulting Holocene lake-level record aligns with dated shorelines and fills major gaps. It shows that during the last 10 ka, the salt-saturated Dead Sea remained confined to the northern basin, flooding the southern basin only during the first 2 ka of the Holocene. This interval of higher water level coincides with the Neolithic Revolution, pointing to a wetter regional climate.
Dissemination and Open Science
The project yielded two peer-reviewed papers (AJS, Chemical Geology), one submitted preprint, and three manuscripts in preparation. Results have been presented at international conferences, and all data and models will be released under Open Science standards. The HaliBubble interface (
https://www.wolframcloud.com/env/emmanuel.guillerm/HaliBubbleDataProcessing(s’ouvre dans une nouvelle fenêtre)) allows global use of the model for halite-based climate reconstruction. Together, these outputs provide a lasting resource for reconstructing past temperature and hydrology in evaporitic systems worldwide.
