LAST GLACIAL MAXIMUM PALEOWIND DIRECTIONS FROM SAND DUNES IN NORTHEAST ARKANSAS

Last Glacial Maximum (LGM) paleoclimate simulations of North America are based primarily on temperature and precipitation estimates derived from plant community reconstructions of fossil pollen. In addition to temperature and precipitation trends, these simulations predict major atmospheric pressure cell locations, jet stream configurations, and thus the directions of strong winds. There are a few high-quality paleowind proxy data In the mid-continent to constrain these LGM paleoclimate models where most inferences of wind directions come from observed thinning and fining of eolian silt (loess) deposits along east-west transects. These studies conclude WNW and W winds during LGM. In the Lower Mississippi Valley, Late Pleistocene eolian sand dunes contain abundant high quality paleowind proxy data. We investigated the internal architecture, morphology, and depositional setting of these dunes at three localities in northeastern Arkansas (Augusta 35.31°, -91.35°, Diaz 35.67°, -91.28°, and Success 36.48°, -90.68°) to interpret LGM paleowind directions. These dunes are 4 to 9 m high and are typically stratified with cross-bedded eolian fine sand horizons, glacial outwash units, and paleosols. Crescent ridges, and arcuate and linear dunes are elements of large parabolic dunes that may be several kilometers long, locally reworked into fields of transverse dunes. The inclination direction of cross beds measured in sand pits and on ground penetrating radar profiles and the azimuth of the arms and vertex of parabolic dunes record a dominant paleowind direction from the SW and SSW. OSL ages from eolian sand of these dunes range from 19 ka to 39 ka, suggesting dune mobilization was coeval with LGM loess deposition. LGM loess and sand are commonly assumed to be entrained seasonally during December-January-February. Current numerical General Circulation Models (GCM) show a north-to-south transition from NW to SW surface winds in the North American mid-continent during LGM. This transition is sensitive to paleowind proxy data, and our results agree with June-July-August model outputs and least for December-January-February outputs. Thus, incorporating our paleowind proxy data may improve GCMs with respect to their December-January-February strength and locations of major LGM atmospheric pressure cells.