Drought drove forest decline and dune building in eastern upper Michigan, USA, as the upper Great Lakes became closed basins

Walter L. Loope; Henry M. Loope; Ronald J. Goble; Timothy G. Fisher; David E. Lytle; Robert J. Legg; Douglas A. Wysocki; Paul R. Hanson; Aaron R. Young

doi:10.1130/g32937.1

USGS · 70043908

Drought drove forest decline and dune building in eastern upper Michigan, USA, as the upper Great Lakes became closed basins

Walter L. Loope·Henry M. Loope·Ronald J. Goble·Timothy G. Fisher·David E. Lytle·Robert J. Legg·Douglas A. Wysocki·Paul R. Hanson·Aaron R. Young

View original source ↗Export metadata ↓

Abstract

Current models of landscape response to Holocene climate change in midcontinent North America largely reconcile Earth orbital and atmospheric climate forcing with pollen-based forest histories on the east and eolian chronologies in Great Plains grasslands on the west. However, thousands of sand dunes spread across 12,000 km 2 in eastern upper Michigan (EUM), more than 500 km east of the present forest-prairie ecotone, present a challenge to such models. We use 65 optically stimulated luminescence (OSL) ages on quartz sand deposited in silt caps (n = 8) and dunes (n = 57) to document eolian activity in EUM. Dune building was widespread ca. 10–8 ka, indicating a sharp, sustained decline in forest cover during that period. This decline was roughly coincident with hydrologic closure of the upper Great Lakes, but temporally inconsistent with most pollen-based models that imply canopy closure throughout the Holocene. Early Holocene forest openings are rarely recognized in pollen sums from EUM because faint signatures of non-arboreal pollen are largely obscured by abundant and highly mobile pine pollen. Early Holocene spikes in nonarboreal pollen are recorded in cores from small ponds, but suggest only a modest extent of forest openings. OSL dating of dune emplacement provides a direct, spatially explicit archive of greatly diminished forest cover during a very dry climate in eastern midcontinent North America ca. 10–8 ka.

Explore related subjects

Geology Article

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Walter L. Loope, Henry M. Loope, Ronald J. Goble, Timothy G. Fisher, David E. Lytle, Robert J. Legg, Douglas A. Wysocki, Paul R. Hanson, Aaron R. Young. 2012-02-28. Drought drove forest decline and dune building in eastern upper Michigan, USA, as the upper Great Lakes became closed basins. https://doi.org/10.1130/g32937.1

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related USGS reports

USGS2026

Zircon petrochronology fingerprints mantle sources of magmatic rare earth element deposits

The mantle sources of carbonatite and alkaline magmatic systems that host rare earth element (REE) deposits are difficult to constrain because whole-rock compositions are commonly altered. Alteration-resistant zircon trace-element compositions are used to fingerprint mantle sources at Mountain Pass, California, the largest REE deposit in the United States, and extend this approach to a global carbonatite−alkaline rock dataset. Autocrystic zircon samples from the Mountain Pass intrusive suite record subduction-like U-Sc-Nb-Yb compositions, oxidized fO2 (0.8−2.4 relative to fayalite-magnetite-quartz buffer [ΔFMQ]), and Ti-in-zircon temperatures up to ∼970 °C, indicating a lithospheric mantle source enriched and oxidized by Paleoproterozoic subduction and subsequently tapped during Mesoproterozoic postcollisional lithospheric thinning. This contrasts with the ocean island basalt−like enriched mantle recorded by Mesoproterozoic zircon from the Bayan Obo carbonatite complex, China, demonstrating that world-class REE deposits of comparable age can reflect fundamentally different mantle sources. New classifiers separate subduction-metasomatized from non-subduction-enriched mantle domains with ∼90% accuracy, providing a zircon-based tool for fingerprinting the mantle source regions of magmatic REE deposits.

Geology↗

USGS2026

Deformed submarine terraces in Puget Sound, Pacific Northwest, indicate only one M >~7.5 earthquake on the Seattle fault zone in the past 11,000 yr

Submerged marine terraces in Puget Sound, deformed across the Seattle fault zone (SFZ), indicate that only one earthquake as large as M~7.5 has occurred in at least the past 11 kyr. Previous paleoseismic studies document a M~7.5 earthquake between 923–4 CE, which uplifted coastal marine terraces by as much as 8 m. We demonstrate that this earthquake was the only such event since ~11 ka by mapping and quantifying deformation of older marine terraces, now submerged in Puget Sound. The submerged terraces, attributed to a late-glacial sea-level lowstand, record both glacial isostatic rebound and tectonic deformation. Vertical offset of the ~11 ka terraces within the SFZ is comparable to that of the marine terraces uplifted in 923 CE, implying no additional large (M>~7.5) earthquake on the SFZ since ~11 ka. This result implies a longer recurrence interval than current hazard estimates, which assumes recurrence of M>7.1 events every 5 kyr. Our mapping of SFZ deformation since ~11 ka also supports fault segmentation and contiguous block uplift between the Seattle and Tacoma fault zones.

Washington↗

USGS2025

Forest-floor burial in 1507 by the largest Mount Rainier lahar of the past millennium

New dating of lahar-killed trees underscores volcano hazards in the Puget Sound metropolitan area. Beginning as a landslide from the west flank of Mount Rainier, Washington, USA, the Electron Mudflow, which was the largest lahar of the last millennium, swept more than 60 km down the Puyallup River drainage into areas now densely populated. Wiggle matching of seven radiocarbon ages from buried, bark-bearing Douglas-fir ( Pseudotsuga menziesii ) trees brackets the mudflow's age between 1477 and 1522 CE with 99.7% certainty. To narrow this date, we applied dendrochronology crossdating on samples collected from 21 trees killed by the lahar, measuring 86 time series for statistical verification. The four bark-bearing trees died the same year while the final rings in all other trees had decayed, exposing rings formed in earlier years. When averaged together, the crossdated measurements form a 475 yr master chronology that was correlated against absolutely dated tree-ring chronologies in the region. The Electron chronology best matched with chronologies from low-elevation sites, especially a Douglas-fir chronology from Vancouver Island, Canada, to show that the Electron trees died in 1507 CE. Latewood in the final ring was beginning to form, indicating the mudflow likely occurred in the late-summer months. What caused the Electron Mudflow is unknown, but this precise date will help to assess possible relationships with other events, assist in interpreting Indigenous narratives about the mudflow, and increase awareness of potential lahar hazards.

Washington↗