Our conclusion is that this standard suite of bulk soil characterization measurements does not
detect significant FT-induced changes in conditions between a soil that has been frozen and thawed, and
the same soil that had not been frozen. This conclusion holds particularly well for saturated soil where
the most pronounced effects of FT on flow-induced erosion are expected. The implication is that
important changes caused by FT to enhance runoff-induced erosion following thaw occur at smaller
scales approaching that of a pore, and measurements must sample these scales to be definitive.
Results
Overview
Rill development occurred with a repeatable progression through each experiment (Fig. 3),
starting as a shallow, uniform sheet flow across the entire width of the imposed rill. In the early stages
channels gradually developed near the edges of the imposed rectangular rill. The time required for an
eroded channel to be discernible in the FT bin averaged over 10 minutes in the low-moisture series, 6
minutes in the mid-moisture series, and less than 3 minutes in the high-moisture series. Not all
experiments developed a channel in the C bin, but for those that did, the overall average time was 21
minutes, a time that did not vary with soil moisture. These initially small channels typically grew in size
and persisted for several minutes until one became the primary channel. This channel continued to
enlarge and progressively captured the entire flow. The average time for a primary channel to develop
along the entire length of the bin was 17.5 minutes in the FT, and when it developed, 34 minutes in the C.
By experimental series average times in the FT decreased significantly as soil moisture increased, while
the C average times also showed some decrease with increasing moisture. These data suggest that the FT
soil was generally more erodible than that of the C, and this difference increased with soil moisture.
On average threefour knick points formed along the FT channel in all the moisture series, while
three formed in the C during the low-moisture series and one or no knick points formed in the C channel
during the mid- and high-moisture series. More significant channel development and deepening
followed, primarily via knickpoint growth and upstream migration. While erosion progressed, the rills
developed large-scale roughness and exhibited undercutting and sidewall sloughing, processes commonly
observed in larger hillslope-scale rills. Experiments were terminated when erosion in the FT bin resulted
in a deep rill over the entire channel length that reduced the final overall slope relative to the original
applied slope. Erosion in the C bin frequently did not complete this rill development progression prior to
termination of the experiment.
Duration (min), pre-experiment soil moisture content, and post-experiment water surface slope in
each bin are given for all experiments in Figure 4. Test duration is grouped by soil moisture series, with
8 slope experiments 1 through 3 and 15 slope experiments 4 through 6 corresponding to the same
progressive flow increase (low → mid → high). Increases in applied flow caused a general decrease in
experiment duration. High slope experiments were somewhat shorter in duration than low slope
experiments at the same flow rate. Pre-test soil moisture and post-test water surface slope are not related
to the flow condition, but results are presented in this format for uniformity. The C and FT soil moistures
were kept very close by preparing and homogenizing a single volume of soil that was divided and placed
in each bin. The low, mid, and high soil moisture series are clearly distinguished in Figure 4, but by
6
Previous Page
