FT cycle, reflected by average FT/C ratios for the full series of experiments between 1.2 and 4.9. These
results leave little doubt that soil FT is of primary importance to overland flow-induced erosion following
thaw. These measures all followed similar trends when grouped and averaged for equal soil moisture,
slope, and flow. Average FT and C trends both increased with slope and flow when those data groupings
were considered. Approximate proportionality of sediment load to slope and flow is consistent with
current understanding of rill erosion. However, grouping by slope and flow yielded a consistent
relationship between FT and C parameters, represented by the corresponding FT/C ratio. All these FT/C
ratios were nearly equal to the overall ratio for the complete set of experiments, indicating that the effects
of FT were independent of both slope and flow. The only experimental variable that affected the
relationship between FT and C for each of the measured parameters was soil moisture. Sediment load and
rill measurement FT/C ratios increased dramatically with soil moisture content. As an example, averaged
sediment load ratios increased from 2.4 to 5.0 over the soil moisture range considered. In addition, an
experiment at very low soil moisture, 4.4% by volume, yielded an FT/C ratio of 1.02, indicating minimal
FT effects in very dry soils. These results confirm that soil moisture is the primary parameter controlling
the enhancement of frost-susceptible soil erosion following FT. Because the soil used in our experiments
was frost-susceptible, the effect of FT on erosion was probably near its maximum.
The results of this paper demonstrate that FT is a primary process contributing to upland soil
erosion in cold climates, and that FT effects can increase up to several hundred percent with soil moisture.
Watersheds or sub-watersheds with high soil moisture and silty soils that freeze have greatly enhanced
erodibility during runoff events that follow thaw. These results also suggest that models of flow-induced
rill erosion must properly account for the effects of FT in order to successfully predict hillslope erosion
and sediment yield for watersheds that experience ground freezing. Algorithms that accurately describe
near-surface soil moisture and moisture redistribution during freezing and thawing are required elements
of such models. Analysis of these experimental data will continue with goals that include
parameterization of FT effects on soil erosion, and quantitative comparison with existing models of FT
effects. Further investigation of fundamental FT effects on flow-induced soil erosion for other soils and
at larger scales would expand on the results presented here.
The authors thank Dennis Lambert for his help in designing and fabricating the CRREL soil-erosion
simulator, and Lauren Raymond and Reed Harrison for assistance in conducting the experiments and
analyzing the samples. Funds for this research were provided by the Engineer Research and Development
Center Regional Sediment Management project "FreezeThaw Effects on Soil and Bank Erosion and
Stability," and BT25 project "Soil Erodibility and Runoff Erosivity Due to Soil Freezing and Thawing."