Summary and Conclusions

through this range. Conversely, it may be anticipated that FT effects on Hanover silt or any other soil

would diminish and eventually disappear as soil moistures approach zero. To test this hypothesis we

repeated an experiment from the low moisture series at greatly reduced soil moisture. The soil in this

experiment was visibly much drier than that of the low moisture series. Both the conditions and results of

this low moisture 2dry experiment are presented in Table 12. The conditions include low slope, mid

flow, and soil moistures in the 45% by volume range. The FT/C ratio of each concentration measure is

slightly greater than 1, indicating marginal increases in soil erosion as a result of FT. The absolute and

percentage increases in FT relative to C concentrations are also relatively small for this "dry" experiment.

For comparison, the FT/C ratios of each concentration measure for the corresponding low moisture

experiment all exceed 2. The dry experiment sediment load and total transport also increased by just a

few percent in the FT relative to the C, while increases in the corresponding low moisture experiment

were more than double. Dry experiment cross-sectional measures also provide very similar comparisons

that indicate only slightly greater erosion of the FT. Differences between the dry and low moisture

experiments show convergence of the various FT and C erosion measures as soil moisture approaches

zero, and greatly reduced effects of FT on rill erosion. These results confirm the fundamental importance

of soil moisture to FT-induced changes in soils that affect subsequent flow-induced erosion.

Summary and Conclusions

In this paper we quantitatively tested the hypothesis that soil FT processes can significantly

increase the potential for upland hillslope erosion during runoff events that follow thaw. Our conceptual

understanding is that ice formed in soil voids during freezing reduces particle cohesion and soil strength,

and makes the soil more erodible. Frost-susceptible silt was used to obtain an upper bound on increases

in bare soil erodibility and rill development due to a single FT cycle over ranges of soil moisture, applied

flow, and slope. For each experiment identical soil bins, an unfrozen control (C), the other frozen and

thawed, were tested in parallel to quantify FT effects. Standard soil characterization tests were unable to

detect significant differences between FT and C soils, especially at high (saturated) soil moisture where

increases in erosion were the greatest. Important changes caused by FT to enhance soil erodibility must

occur at smaller scales than those of the tests, probably approaching that of an ice lens, soil grain, or pore.

Measurements at these scales have a greater potential to be definitive. Without small-scale measurements

we cannot advance understanding of soil FT mechanics, and instead focus on quantifying the effects.

Three replicated experiments showed a convergence of results as differences in bin weight, bulk

density, and soil moisture decreased. To support quantitative comparisons of results, tightly controlled

bin preparation minimized these differences over all soil bins of each prescribed soil moisture series of

experiments. Sediment load obtained from runoff samples gave an integrated measure of erosion through

time for each bin, while cross-sectional data provided measures of erosion at specific locations. Both

types of data were obtained to provide independent evidence, allowing us to assess the concurrence of

trends, and yield a stronger basis for conclusions. The maximum eroded rill width and depth averaged for

nine locations, root-mean-square and maximum rill cross-sectional depth change for two locations, and

average sediment loads and total mass transport during each experiment were all greatly increased by the