Constraining the evolution of past continental climates is key for evaluating modern estimates of climate sensitivity and, in turn, improving predictions of future climate change. The Last Glacial Maximum (LGM; ~26–18 kyr-ago (ka)) represents a benchmark period for evaluating the relationship between global temperature and radiative forcings (e.g., a weaker LGM greenhouse effect due to lower atmospheric CO₂ concentrations). However, reliable proxies of continental paleotemperatures are scarce and often associated with large uncertainties, which has limited our understanding of the dynamics of past climate changes on land. One of the most promising avenues of investigation to  document the evolution of continental climates since the LGM is the use of noble gas abundances in ancient groundwater as a direct proxy of past mean annual surface temperatures at the time and location of recharge.

In this presentation, I will expose the working principle of groundwater noble gas paleothermometry and present a new, quantitative noble gas temperature (NGT) record of the last ~40 kyr from the Albian aquifer, Eastern Paris Basin (latitude ~48°N, France). These data indicate that mean annual temperatures were ~ 5°C between 42–30 ka, before cooling to ~2°C between 28–25 ka. Then, post glacial warming between 25 and 10 ka led to Holocene temperatures of ~11°C, indistinguishable from the average modern ground surface temperature in Eastern France. Such LGM cooling (∆NGT_LGM = 9.1±0.9°C) is consistent with previous studies of noble gas paleothermometry in Europe, but markedly larger than estimates of low-to-mid latitude LGM cooling, hence supporting a stark amplification of LGM cooling with latitude. Comparing Eastern France LGM cooling estimates from state-of-the-art climate models with our ∆NGT_LGM of the Albian aquifer indicates that most of the models actually predict lower ∆T_LGM than observed in our study. Interestingly, the strong correlations between NGTs and water stable isotopes open the door to using water stable isotope variations in precipitation as an empirical paleothermometer, although the exact temperature sensitivity of water stable isotopes in precipitation remains poorly constrained. At last, I present a compilation of European ∆NGT_LGM that, once combined with other continental proxies of paleo-temperatures at high and low altitude, will allow providing a comprehensive assessment of spatial gradients (lapse rate, latitudinal and continental amplifications) of equilibrium warming across Europe.