HOLOCENE DUNE STABILIZATION IN INTERIOR EASTERN UPPER MICHIGAN: SIGNALS OF CLIMATE-SENSITIVE BIOGEOMORPHIC CHANGE OR INTERBASIN MELTWATER PASSAGE (OR BOTH)?

The origin of hundreds of parabolic dunes scattered across the low relief plains of interior eastern Upper Michigan (EUM) has long been attributed to demise of glacial Lake Algonquin, to wit, dunes formed by deflation of sandy shoals as lake level dropped abruptly (Bergquist, 1936). Recent optical (OSL) dating of quartz sand from 21 parabolic dunes SW of Whitefish Point supports this model. Reported dune stabilization peaked at 8.8 ka, just after the demise of Lake Minong, an Algonquin remnant.

Inconsistent with Bergquist's model, however, are six OSL ages obtained from similar terrain in EUM, interpreted to reflect widespread dune stabilization after mid-Holocene climate amelioration and rebound-driven water table rise ~6.3 ka. This latter scenario draws from models of climate-sensitive biogeomorphic change developed on the high plains and forest prairie boundary to the south and west of EUM.

To further explore the chronology of dune stabilization in EUM and both geomorphic models, we obtained 54 additional optical ages from dunes in a variety of geomorphic settings >100 km SW of Whitefish Point, spanning the Lake Superior-Lake Michigan watershed divide. Our composite OSL chronology (n = 75) shows peaks of dune stabilization at 9.8, 8.6 and 7.8 ka. GIS reconstruction of the last stages of Lake Algonquin suggests the possibility that, prior to ~10.3 ka, Lake Minong could have drained SW to Lake Chippewa via EUM. Dune building and stabilization may have occurred along this path after meltwater was diverted elsewhere. This scenario and mechanisms of sand nourishment across EUM are far from clear.

The most recent peak of dune stabilization occurs after levels of the upper Great Lakes had fallen to low levels and lies closer to the ~6.3 ka OSL sequence. A detailed pollen/charcoal study documented that at least three "state shifts" (pine forest to sparse, degraded pine savannah) occurred in EUM during the early Holocene. The latter paleosavannahs are described as small (~300 ha) and short-lived (200-300 years) and no dune forms are evident in the study area. That many dunes stabilized ~7.8, even as aridity peaked, suggests a flora well-adapted to drought and fire. We therefore suggest that the application of biogeomorphological models to dunes of EUM is questionable and that the observed pattern at ~6.3 ka has an alternate explanation.