AMS RADIOCARBON DATING DOCUMENTS CLIMATE EVENTS IN THE UPPER MIDWEST COEVAL WITH THE BØLLING/ALLERØD AND YOUNGER DRYAS EPISODES

Radiocarbon dating of late-glacial sediment cores from the Upper Midwest has been confounded by variable sedimentation rates, too few dates, hardwater error, and pre-Quaternary organic matter. Detailed AMS radiocarbon dating of cores from Nelson Lake and Island Lake in northeastern Illinois and Devils Lake in south-central Wisconsin reveal climate and vegetation changes coeval with the Bølling/Allerød and Younger Dryas (YD) episodes in North Atlantic region. The Illinois sites show an early tundra phase immediately after deglaciation (before 17,000 cal yr B.P.) and spruce forest (>60% Picea pollen) 17,000 to 14,700 cal yr B.P. Spruce pollen declined and black ash (Fraxinus nigra) increased precipitously at the beginning of the Bølling (14,700 cal yr B.P.). At Devils Lake, sedimentation began at the beginning of the Bølling. At all three sites, black ash declined and spruce increased at the beginning of the YD (12,900 cal yr B.P.). At the end of the YD, spruce disappeared from the region, and black ash increased again. The decrease in black ash and the increase in spruce especially characterize the YD. Other taxa indicate some regional differences between the Illinois and Wisconsin sites. In climate terms, the pollen data indicate a dramatic warming at the beginning of the Bølling, cooling during the YD, and rapid warming at the end of the YD.

At Nelson Lake, sediment accumulated very slowly during the YD. If plotted against depth or with a time scale determined with a small number of radiocarbon dates, the YD pollen zone appears to be a brief fluctuation; whereas detailed dating shows that the episode is 1200 years long. The chronology for Devils Lake was originally based on a suite of decay-count radiocarbon dates, which are internally consistent; and the site has been considered one of the better dated sites in the Midwest. However, the new AMS dates show that the decay-count dates near the base are about 1000 years too old. According to the decay-count chronology, the pollen data indicate warming; whereas the AMS chronology indicates cooling. We strongly suspect that detailed dating of other sites in the Midwest will demonstrate a pattern of climate changes consistent with the North Atlantic climate episodes.