12.4 Results and Discussion
201
was at the highest value of 9.3 for clay fraction, 1.7 for silt fraction and at the lowest
value 0.6 for sand fraction. R ed s of the clay, silt and sand fractions in channel flows
were 9.3, 1.4 and 0.9, respectively, when water erosion occurred only. It can be seen
that the change of R ed s of the suspended sediment particle sizes caused by gravity
erosion and water erosion in the channel flow followed the similar trends to that in
the gully flows. However, the difference of R ed s in the channel flows influenced by
gravity erosion and water erosion became smaller than in the gully flows.
12.4.4 Causes of the Changes of PSDSS
The changes of PSDSS before and after mass failures reflected the combined effects of
soil sources, erosion types, as well as sediment transport and depositional processes.
The soil type and characteristics were the dominant factors controlling the particle
size (Wendling et al. 2016). Soil textures at each location determine the range and
availability of particle sizes for possible erosion (Grismer et al. 2008). In addition,
disaggregation of unstable aggregates due to the impact of raindrops and runoff
turbulence can also affect the particle size distribution of the eroded sediment (Meyer
et al. 1992). Moreover runoff velocity might also have an effect on the sediment
particle size distribution to a certain extent (He et al. 2017).
Prior to the mass failures, the four suspended sediments samples taken from gully
flows showed a graduate decrease in sand-sized fraction over time, in the order of
83, 77, 63 and 61% (Table 12.2). The initial high contents of sand-sized particles
might be the loose and large particles on the surfaces which were washed away by the
overland flow. With the depletion of those large and loose particles, the detachment
of cohesive soil particles required more energy and therefor might be size-selective,
which led to the decrease of sand-sized fraction and the enrichment of fine particles.
These results are consistent with other studies that show the dynamic changes of
particle size selectivity in water erosion processes (Koiter et al. 2017; Shi et al. 2013;
Issa et al. 2006).
The smaller fraction of sand-sized particles in channel flows than that in gully
flows, with a mean value of 64 versus 71%, indicates the selective deposition of
coarse sediments on channel beds due to the abrupt decrease of slope gradient from
gully slope to channel bed. It is expected no or very little deposition occurred in
the steep gullies. After mass failures, the fraction of sand-sized particles in gully
flows further decreased, with a mean value of 51% ranging from 40 to 59%. Such
a decrease was likely to be caused by the source materials from mass movements
and the selective detachment by overland flows. Mass failures occurred within a
thicker soil layer with lower content of sand-sized particles than that on surface
soils. As shown in Table 12.3, the average proportion of sand-sized particle was 70%
on surface soils (0–5 cm), in comparison of 60% on the depth of 40–45 cm and 59%
on the depth of 120–125 cm. The median particle size also decreased with the depth
(Fig. 12.5), which may be ascribed to the different climatic conditions during the
deposition of wind-blown loess particles in Quaternary and Holocene. The fraction
201
was at the highest value of 9.3 for clay fraction, 1.7 for silt fraction and at the lowest
value 0.6 for sand fraction. R ed s of the clay, silt and sand fractions in channel flows
were 9.3, 1.4 and 0.9, respectively, when water erosion occurred only. It can be seen
that the change of R ed s of the suspended sediment particle sizes caused by gravity
erosion and water erosion in the channel flow followed the similar trends to that in
the gully flows. However, the difference of R ed s in the channel flows influenced by
gravity erosion and water erosion became smaller than in the gully flows.
12.4.4 Causes of the Changes of PSDSS
The changes of PSDSS before and after mass failures reflected the combined effects of
soil sources, erosion types, as well as sediment transport and depositional processes.
The soil type and characteristics were the dominant factors controlling the particle
size (Wendling et al. 2016). Soil textures at each location determine the range and
availability of particle sizes for possible erosion (Grismer et al. 2008). In addition,
disaggregation of unstable aggregates due to the impact of raindrops and runoff
turbulence can also affect the particle size distribution of the eroded sediment (Meyer
et al. 1992). Moreover runoff velocity might also have an effect on the sediment
particle size distribution to a certain extent (He et al. 2017).
Prior to the mass failures, the four suspended sediments samples taken from gully
flows showed a graduate decrease in sand-sized fraction over time, in the order of
83, 77, 63 and 61% (Table 12.2). The initial high contents of sand-sized particles
might be the loose and large particles on the surfaces which were washed away by the
overland flow. With the depletion of those large and loose particles, the detachment
of cohesive soil particles required more energy and therefor might be size-selective,
which led to the decrease of sand-sized fraction and the enrichment of fine particles.
These results are consistent with other studies that show the dynamic changes of
particle size selectivity in water erosion processes (Koiter et al. 2017; Shi et al. 2013;
Issa et al. 2006).
The smaller fraction of sand-sized particles in channel flows than that in gully
flows, with a mean value of 64 versus 71%, indicates the selective deposition of
coarse sediments on channel beds due to the abrupt decrease of slope gradient from
gully slope to channel bed. It is expected no or very little deposition occurred in
the steep gullies. After mass failures, the fraction of sand-sized particles in gully
flows further decreased, with a mean value of 51% ranging from 40 to 59%. Such
a decrease was likely to be caused by the source materials from mass movements
and the selective detachment by overland flows. Mass failures occurred within a
thicker soil layer with lower content of sand-sized particles than that on surface
soils. As shown in Table 12.3, the average proportion of sand-sized particle was 70%
on surface soils (0–5 cm), in comparison of 60% on the depth of 40–45 cm and 59%
on the depth of 120–125 cm. The median particle size also decreased with the depth
(Fig. 12.5), which may be ascribed to the different climatic conditions during the
deposition of wind-blown loess particles in Quaternary and Holocene. The fraction