Modeling Loess Sedimentation

Quantitative models of suspended particle transport can be used to interpret spatial and temporal patterns of loess deposition rate and grain size, potentially providing information on dust source locations and the direction and speed of loess-transporting winds. Several data sets from central North America record the spatial patterns of grain size and/or thickness on upland summits for Peoria Loess, deposited at relatively high rates during the Late Pleistocene. Thickness on upland summits provides an approximation of the long-term average loess deposition rate, assuming low rates of post-depositional erosion on such stable landscape positions. Well-defined regional trends of both grain size and thickness have been documented in previous work, and have been interpreted as evidence of Peoria Loess transport direction. In this study I compared observed thickness and grain size trends to predictions based on one of the simplest process-based modeling approaches, applying steady-state solutions of diffusion-deposition equations proposed by Huang (1999, Journal of Applied Meteorology, v. 38, p. 250-254). Using these solutions, spatial patterns of deposition from line sources or arrays of point sources (representing dust production over extensive areas) were calculated for multiple particle size classes, over distances up to 200 km downwind. For each particle size class, deposition velocity is set equal to the Stoke’s law settling velocity, resulting in distinct downwind fining through more rapid settling of coarser particles. Observed trends of Peoria Loess grain size and thickness can be reproduced with this approach, using plausible assumptions about the wind speed profile and the vertical and lateral eddy diffusivity. For example, observed data from central Nebraska can be approximated by assuming source areas northwest of the loess region, with transport by northwesterly winds that would have had speeds of 8-10 m s -1 at 10 m height, well within the range of modern observations. I also considered patterns of deposition rate and grain size integrated over typical wind speed distributions. The results are dependent on the assumed relationship between wind conditions and dust production at the source. Assuming that dust production rate is proportional to the horizontal saltating particle flux and increases with friction velocity cubed, then the pattern of downwind dispersion is largely controlled by the high-velocity tail of the wind speed distribution, mainly associated with strong cyclones and especially cold fronts in the midlatitude settings of most major loess deposits. Ongoing work is directed toward effects of varying wind direction and irregular or shifting sink-source boundaries. Beyond the inherent limitations of a steady-state modeling approach, one major difficulty is that the dust grain size distribution at the source is unknown, although large changes in that distribution result in relatively modest change in downwind sedimentation patterns. Other important sources of uncertainty include the use of typical fully dispersed particle size analyses, despite evidence that some loess was transported as aggregates, and the effects of spatially variable post-depositional erosion and compaction on observed patterns of loess thickness.