CONTROLS ON DRAINAGE DENSITY IN THE RECENTLY GLACIATED CENTRAL LOWLANDS OF THE USA

Portions of the Central Lowlands (CL) physiographic province of the United States have been repeatedly glaciated during the Quaternary, leading to the disruption, re-routing, and infilling of pre-glacial river channels. Glacial advances have had variable extents and deposited many recessional moraines, leading to a patchwork of land surfaces of different ages. Recently compiled and calibrated radiocarbon ages for North American glacial deposits allow us to confidently assign ages to the young fraction of these surfaces. Surface ages indicate that some river systems have been developing for ~0.5+ million years (e.g. the Driftless region) while others have been developing for only tens of thousands of years (e.g. the bed of glacial lake Agassiz). The range in surface ages allows us to ask: how well does drainage density correlate with surface age in the CL? Are there other factors that explain drainage density patterns? Regional datasets including LiDAR topography, soil texture, mean annual temperature, climatological precipitation, and detailed surface hydrology, are available for much of the CL region. To assess the relationships between drainage density, surface age, and other factors we reconstructed pre-agricultural stream networks using the National Hydrography Dataset. Flowlines from the NHD were modified to remove ditches and other engineered channels to best approximate stream networks prior to intensive agriculture. We subdivided the CL into a grid with 10 km spacing and calculated partial drainage densities (PDD) for each grid box by summing the lengths of natural streams. For a given age surface, PDD typically has a unimodal distribution. The mean PDD increases with surface age and reaches a saturation value by ~30 ka, but with a wide distribution. After 30 ka PDD distributions increasingly tighten around this saturation value. We present relationships between PDD and a number of physical parameters, including topographic position index, soil texture (sand/silt/clay content), landform type (e.g. ground moraine, glacial lake, etc.), distance to major river, and percent non-contributing area. Preliminary results suggest that, in addition to time, soil texture and non-contributing drainage area most strongly correlate with drainage density in the recently glaciated (<30ka) Central Lowlands.