13.4 Results and Discussion
219
13.4.2 Sediment Processes of Tunnel Flows
(1) Within storm variations
To examine the relationship between sediment and runoff discharge, the correlation
coefficients between measured flow discharge rates and sediment concentrations for
all recorded events at Tunnel 1, 3, 4 and 6, as well as at catchment outlet are determined. Sediment concentration shows a good correlation with discharge only in 20%
of all recorded tunnel flow events, but in 60% of catchment outflow events. Among
those tunnel flow events with a significant correlation between runoff discharge and
sediment concentration, five occurred at Tunnel 3, one each in Tunnel 1 and Tunnel
4, and none in Tunnel 6. It is noted that the best correlation between discharge and
sediment concentration does not exist in small tunnels such as Tunnel 1 or Tunnel 4
but in Tunnel 3, the second-largest tunnel in the catchment. This implies that the correlation between runoff discharge and sediment concentration is not directly related
to the size of tunnel systems but the tunnel stability.
To further examine the erratic relationship between discharge and sediment concentration, timing of peak flow discharge is compared with that of peak sediment
concentration for each tunnel flow event. Overall, 65% of events show the first peak
sediment concentration precedes the first peak runoff. A wide variation also exists
between tunnels, with 63% at Tunnel 1, 22% at Tunnel 3, 88% at tunnel 4, and
88% at Tunnel 6, respectively. The early peaking of sediment concentration suggests that there are significant flushing effects in these tunnel systems. The flushing
could relate to the sediment deposition in the previous storm or collapse during the
inter-storm period. For early storms of the season, it may relate to deposition from
a range of processes (mass wasting, aeolian deposition, etc.) during the previous
fall and winter. This view is supported by observations of a substantial thickness
of fluvial deposits and wind-blown dusts in the sloping section of Tunnel 6 and the
roof collapse materials of Tunnel 1 prior to the occurrence of the first major storm
of 1990.
Tunnel flow hydrographs are characterized by a rapid recession limb, but sediment
concentration remains very high in many events. This is reflected by a counterclockwise loop in the later part of the event in hysteresis graphs (Fig. 13.6). Of a
total of 35 tunnel flow events, 19 or 56% have a counter-clockwise loop in the latter
part of the event. This pattern was found to be associated with the occurrence of
hyperconcentrated flow which maintained high sediment concentration during flow
recession (Hamilton 1990; Qian and Wan 1983; Wang et al. 1982).
Comparison of data from tunnel outflow and catchment outflow indicates that
peak sediment concentration in tunnel-flows is not distinctively higher than that in
the catchment outflow. This is different from the findings by Bryan and Harvey
(1985) in the Alberta Badlands where peak sediment concentration of pipe flow is
substantially above peak channel concentrations. It is noted that the measured peak
sediment concentrations in this area were one magnitude order higher than those
measured in the Alberta Badlands area. However, except for Tunnel 1, peak sediment
219
13.4.2 Sediment Processes of Tunnel Flows
(1) Within storm variations
To examine the relationship between sediment and runoff discharge, the correlation
coefficients between measured flow discharge rates and sediment concentrations for
all recorded events at Tunnel 1, 3, 4 and 6, as well as at catchment outlet are determined. Sediment concentration shows a good correlation with discharge only in 20%
of all recorded tunnel flow events, but in 60% of catchment outflow events. Among
those tunnel flow events with a significant correlation between runoff discharge and
sediment concentration, five occurred at Tunnel 3, one each in Tunnel 1 and Tunnel
4, and none in Tunnel 6. It is noted that the best correlation between discharge and
sediment concentration does not exist in small tunnels such as Tunnel 1 or Tunnel 4
but in Tunnel 3, the second-largest tunnel in the catchment. This implies that the correlation between runoff discharge and sediment concentration is not directly related
to the size of tunnel systems but the tunnel stability.
To further examine the erratic relationship between discharge and sediment concentration, timing of peak flow discharge is compared with that of peak sediment
concentration for each tunnel flow event. Overall, 65% of events show the first peak
sediment concentration precedes the first peak runoff. A wide variation also exists
between tunnels, with 63% at Tunnel 1, 22% at Tunnel 3, 88% at tunnel 4, and
88% at Tunnel 6, respectively. The early peaking of sediment concentration suggests that there are significant flushing effects in these tunnel systems. The flushing
could relate to the sediment deposition in the previous storm or collapse during the
inter-storm period. For early storms of the season, it may relate to deposition from
a range of processes (mass wasting, aeolian deposition, etc.) during the previous
fall and winter. This view is supported by observations of a substantial thickness
of fluvial deposits and wind-blown dusts in the sloping section of Tunnel 6 and the
roof collapse materials of Tunnel 1 prior to the occurrence of the first major storm
of 1990.
Tunnel flow hydrographs are characterized by a rapid recession limb, but sediment
concentration remains very high in many events. This is reflected by a counterclockwise loop in the later part of the event in hysteresis graphs (Fig. 13.6). Of a
total of 35 tunnel flow events, 19 or 56% have a counter-clockwise loop in the latter
part of the event. This pattern was found to be associated with the occurrence of
hyperconcentrated flow which maintained high sediment concentration during flow
recession (Hamilton 1990; Qian and Wan 1983; Wang et al. 1982).
Comparison of data from tunnel outflow and catchment outflow indicates that
peak sediment concentration in tunnel-flows is not distinctively higher than that in
the catchment outflow. This is different from the findings by Bryan and Harvey
(1985) in the Alberta Badlands where peak sediment concentration of pipe flow is
substantially above peak channel concentrations. It is noted that the measured peak
sediment concentrations in this area were one magnitude order higher than those
measured in the Alberta Badlands area. However, except for Tunnel 1, peak sediment