Assessing the strength of the monsoon during the late Pleistocene in southwestern United States

Improved predictions of drought require an understanding of natural and human-induced climate variability. Long-term records across glacial–interglacial cycles provide the natural component of variability, however few such records exist for the southwestern United States (US) and quantitative or semi-quantitative records of precipitation are absent. Here we use the hydrogen isotope (δD) value of C28n-alkanoic acid in lacustrine sediments of Pleistocene age to reconstruct δD values of precipitation in northern New Mexico over two glacial–interglacial cycles (∼550,000–360,000 years before present) and obtain a record of monsoon strength. Overall, reconstructed δD values range from −53.8‰ to −94.4‰, with a mean value of −77.5 ± 8‰. Remarkably, this variation falls within the measured present-day summer monsoonal and winter weighted means (−50.3 ± 3‰ and −106.4 ± 20‰ respectively), suggesting that processes similar to those of present time also controlled precipitation during Marine Isotope Stage (MIS) 13 to 10. Using the δD summer monsoonal and winter mean values as end-members, we interpret our reconstructed δD record of precipitation as a direct, and semi-quantitative, indicator of monsoon strength during MIS 13 to 10. Interglacial periods were characterized by greater monsoon strength but also greater variability compared to glacial periods. Pronounced cycles in the strength of the monsoon occurred during interglacial periods and in general were positively correlated with maximum mean annual temperatures. Our estimates of monsoon strength are supported by independent proxies of ecosystem productivity, namely, TOC, δ13C of TOC and Si/Ti ratio and warm pollen taxa Juniperus and Quercus. Interglacial variability in the strength of the monsoon resembles a response to the land-sea surface temperature contrast (LSTC) except for the early part of MIS 11. During this period, LSTC would have remained relatively strong while monsoonal strength decreased to a minimum. This minimum occurred following the warmest interval of MIS 11, suggesting a more complex driving of monsoon strength during warm periods. In addition, this period of monsoon minimum coincided with a core section of mud-cracked sediments that suggest low monsoonal precipitation was an important factor in the onset of drought. Our estimates of monsoon strength represent a record of natural variability in the region that is relevant to present time, in particular the variability during interglacial MIS 11, which is considered an analog for the current interglacial. Our results suggest that natural variability can cause significant reductions in monsoonal precipitation with the implication of a potentially adverse effect from sustained warming