The culmination of the glaciers in the European Alps during the Last Glacial Maximum (LGM) is one of the most intensively studied paleoglaciological events, but its trigger and forcing remain incompletely understood. Here, we provide evidence that the timing of this glacier maximum coincided within age uncertainties with a 3100 yr-long interval of subsurface warming (26.6 to 23.5 ka BP) as recorded by an archive preserved in caves, cryogenic carbonates. This interval of sustained permafrost degradation during one of the coldest intervals of the last glacial period calls for a fundamental change in the dry Arctic-style precipitation regime. Instead, heavy snowfall during autumn and early winter led to the accumulation of a seasonal snowpack insulating the ground from the winter chill. Combined with thermal modelling, the data provide compelling evidence that the LGM glacier advance in the Alps was fueled by intensive snowfall late in the year, likely sourced from the Mediterranean Sea.

A flowstone from the central European Abaliget Cave (Mecsek Mts, Hungary) provides a record of uninterrupted calcite deposition between ∼160 and ∼124 ka, covering most of Marine Isotope Stage (MIS) 6 and part of 5e. δ18O values of three lateral drill cores show synchronous high‐frequency (millennial‐scale) variability during MIS 6, interpreted as stadials and interstadials, and a 3.4‰ rise at the MIS 6/5e boundary. The interstadials are mostly symmetrical in shape and show consistently lower δ18O values than calcite formed during MIS 5e. The rises (decreases) in δ18O are followed by drops (increases) in δ13C with a delay of 1–2 ka, implying enhanced (reduced) soil bioproductivity. This period of highly variable climate is bracketed by two broad δ18O minima. The first minimum between ∼160 and ∼148 ka coincided with a maximum in ice‐rafted detritus in the eastern North Atlantic. The second one from ∼134 to ∼129 ka occurred during Heinrich 11, before the rapid and large δ18O increase at ∼128 ka.

The relationship between the atmospheric concentration of cosmogenic isotopes, the change of solar activity and hence secondary neutron flux has already been proven. The temporal atmospheric variation of the most studied cosmogenic isotopes shows a significant anti-correlation with solar cycles. However, since artificial tritium input to the atmosphere due to nuclear-weapon tests masked the expected variations of tritium production rate by three orders of magnitude, the natural variation of tritium in meteoric precipitation has not previously been detected. For the first time, we provide clear evidence of the positive correlation between the tritium concentration of meteoric precipitation and neutron flux modulated by solar magnetic activity. We found trends in tritium time series for numerous locations worldwide which are similar to the variation of secondary neutron flux and sun spot numbers. This variability appears to have similar periodicities to that of solar cycle. Frequency analysis, cross correlation analysis, continuous and cross wavelet analysis provide mathematical evidence that the correlation between solar cycle and meteoric tritium does exist. Our results demonstrate that the response of tritium variation in precipitation to the solar cycle can be used to help us understand its role in the water cycle.