RECONSTRUCTING MIDWESTERN CLIMATE VARIABILITY DURING THE LATE PLEISTOCENE

Oceanic and terrestrial records indicate that rapid climate changes occurred globally throughout the last glacial period. These changes affected precipitation patterns in the United States, as illustrated in continental speleothem records. While oxygen isotopic variability in southwestern records has been shown to be a result of changes in Pacific sourced winter-precipitation, controls on oxygen isotope variation in the mid-continent during the last glacial period remain poorly understood.

This research provides a record of paleoclimate variation in SE Minnesota throughout the late Pleistocene. Preliminary data includes a high resolution study of two speleothem samples from Spring Valley Caverns. Speleothem sample SVC05 grew without interruption from 104-93 ka BP, then experienced six multi-decadal to centennial growth intervals. Growth intervals correlate with Greenland interstadials - specifically GI 4, 9, 10, and 14 - indicating that the mean annual temperature in SE Minnesota was above freezing and that there was enough regional moisture during these periods to spur stalagmite growth. From 104-93 ka BP, the stable isotope record of sample SVC05 suggests a transition from C4-dominant prairie to C3-dominant forest and shows an enrichment in δ¹⁸O by almost 4‰, changes that would be consistent with cooler temperatures and a drier trajectory of Gulf of Mexico moisture. Speleothem sample SVC06 exhibits continuous growth from 63-44 ka BP. SVC06 appears largely C3-dominant, with variability in both δ¹³C and δ¹⁸O most likely driven by changes in regional temperature and moisture availability.

Holocene δ¹⁸O values of Spring Valley Caverns speleothems range from -6‰ to -4‰. Even with a correction for glacial seawater composition, the δ¹⁸O values of Pleistocene samples SVC06 and of SVC05 growth intervals remain above modern values. Due to our site’s proximity to the Laurentide during the last glaciation and its position along the Gulf of Mexico moisture trajectory, this may indicate that previous estimates of southern Laurentide δ¹⁸O values are too low. Future work includes expanding the late Pleistocene record in SE Minnesota to verify patterns seen in samples SVC05 and SVC06, and correlating our research to other mid-continental records in order to track Gulf of Mexico moisture variability across the region.