Quantifying the effect of a freeze-thaw cycle on soil erosion: laboratory experiments

In this paper we quantitatively test the hypothesis that soil freeze-thaw (FT) processes significantly increase the potential for upland hillslope erosion during run-off events that follow thaw. We selected a highly frost-susceptible silt to obtain an upper bound on FT effects, and completed three series of six experiments each to quantify differences in soil erosion and rill development in a bare soil following a single FT cycle. Each series represented a specific soil moisture range: 16-18 per cent, 27-30 per cent and 37-40 per cent by volume, with nominal flow rates of 0-4, 1 center dot 2 and 2 center dot 4 L/min and slopes of 8 degrees and 15 degrees. Each experiment used two identical soil bins: one a control (C) that remained unfrozen, and another that was frozen and thawed once. Standard soil characterization tests did not detect significant differences between the FT and C bins. We measured cross-sectional geometry of an imposed straight rectangular rill before each experiment, sediment load during and rill cross-sections after. Changes in cross section provided detailed measures of erosion at specific locations, while sediment load from time series run-off samples integrated the rill erosion. Several parameters, including average maximum rill width, average maximum rill depth, rill cross-section depth measures and sediment load, all followed similar trends. Each was greater in the FT than in the C, with values that generally increased with slope and How. However, soil moisture was the only parameter that affected the FT/C ratios. Average sediment load grouped by soil moisture provided FT/C ratios of 2 center dot 4, 3 center dot 0 and 5 center dot 0 for low, mid and high moisture, respectively. In contrast, a 'dry' experiment at 4-5 per cent soil moisture had FT/C of 1 center dot 02 for sediment load. These results show a dramatic increase with soil moisture in the rate and quantity of bare soil eroded due to the FT cycle. As both FT and C results were highly sensitive to initial conditions, minimum differences in soil weight, bulk density and soil moisture through each series of experiments were required to achieve consistent results, indicating that rill erosion may be chaotic. Published in 2005 by John Wiley & Sons, Ltd.