NALPS19: Sub-orbital-scale climate variability recorded in northern Alpine speleothems during the last glacial period

Abstract

Sub-orbital-scale climate variability of the last glacial period provides important insights into the rates at which the climate can change state, the mechanisms that drive such changes, and the leads, lags, and synchronicity occurring across different climate zones. Such short-term climate variability has previously been investigated using δ18O from speleothems (δ18Ocalc) that grew along the northern rim of the Alps (NALPS), enabling direct chronological comparisons with δ18O records from Greenland ice cores (δ18Oice). In this study, we present NALPS19, which includes a revision of the last glacial NALPS δ18Ocalc chronology over the interval 118.3 to 63.7 ka using 11, newly available, clean, precisely dated stalagmites from five caves. Using only the most reliable and precisely dated records, this period is now 90 % complete and is comprised of 16 stalagmites from seven caves. Where speleothems grew synchronously, the timing of major transitional events in δ18Ocalc between stadials and interstadials (and vice versa) are all in agreement on multi-decadal timescales. Ramp-fitting analysis further reveals that, except for one abrupt change, the timing of δ18O transitions occurred synchronously within centennial-scale dating uncertainties between the NALPS19 δ18Ocalc record and the Asian monsoon composite speleothem δ18Ocalc record. Due to the millennial-scale uncertainties in the ice core chronologies, a comprehensive comparison with the NALPS19 chronology is difficult. Generally, however, we find that the absolute timing of transitions in the Greenland Ice Core Chronology (GICC) 05modelext and Antarctic Ice Core Chronology (AICC) 2012 are in agreement on centennial scales. The exception to this is during the interval of 100 to 115 ka, where transitions in the AICC2012 chronology occurred up to 3000 years later than in NALPS19. In such instances, the transitions in the revised AICC2012 chronology of Extier et al. (2018) are in agreement with NALPS19 on centennial scales, supporting the hypothesis that AICC2012 appears to be considerably too young between 100 and 115 ka. Using a ramp-fitting function to objectively identify the onset and the end of abrupt transitions, we show that δ18O shifts took place on multi-decadal to multi-centennial timescales in the North Atlantic-sourced regions (northern Alps and Greenland) as well as the Asian monsoon. Given the near-complete record of δ18Ocalc variability during the last glacial period in the northern Alps, we also offer preliminary considerations regarding the controls on mean δ18Ocalc for given stadials and interstadials. We find that, as expected, δ18Ocalc values became increasingly lighter with distance from the oceanic source regions, and increasingly lighter with increasing altitude. Exceptions were found for some high-elevation sites that locally display δ18Ocalc values that are heavier than expected in comparison to lower-elevation sites, possibly caused by a summer bias in the recorded signal of the high-elevation site, or a winter bias in the low-elevation site. Finally, we propose a new mechanism for the centennial-scale stadial-level deple tions in δ18O such as the Greenland Stadial (GS)-16.2, GS-17.2, GS-21.2, and GS-23.2 "precursor" events, as well as the "within-interstadial" GS-24.2 cooling event. Our new highprecision chronology shows that each of these δ18O depletions occurred in the decades and centuries following rapid rises in sea level associated with increased ice-rafted debris and southward shifts of the Intertropical Convergence Zone, suggesting that influxes of meltwater from moderately sized ice sheets may have been responsible for the cold reversals causing the Atlantic Meridional Overturning Circulation to slow down similar to the Preboreal Oscillation and Older Dryas deglacial events.