
CASE STUDY: HEISONGLIN RESERVOIR, CHINA 25.11
flood, 1.4 Mm
3
, or 23 percent of the inflow flood volume, seeped from the muddy lake
and into the overlying clear water, as computed from the rate of lowering of the water
level, the turbidity interface, and the outflow discharge. The muddy water interface fell by
0.03 to 0.05 m/h, which was faster than the water level (Table 25.3).
TABLE 25.3 Settling of Turbidity Interface Measured August 21-24, 1971, Heisonglin
Reservoir
W
ater level in
reservoir, m
Depth of
interface, m
Interface
settling rate, m/h
Amount of clear
water seeped, 10
3
m
3
T
ime of observation
Aug. 21, 11:30 1761.98 0.6 - 324
Aug. 22, 11:00 1760.62 1.41 0.034 620
Au
g. 23, 11:00 1758.80 2.56 0.047 1050
Aug. 24, 10:34 1756.69 4.05 0.063 1460
Source: Northwest Institute of Hydraulic Research (1972b).
Previous flume experiments and field data (Fan, 1980) showed that the interface settling
velocity is inversely proportional to the suspended-solids concentration of the muddy
suspension. For loessal silts, the settling velocity of the interface (V
z
, cm/s) may be
expressed as a function of suspended solids concentration (C, g/L) by the following
formula:
V
z
0.02C
2/3
(25.1)
For the mean concentration of 110 g/L on August 23, the interface settling velocity may
be computed as V
z
= 0.00087 cm/s = 0.031 m/h, which is of the same order of magnitude
as the average interface settling rate of 0.048 m/h observed in the field (Table 25.3). As
the sediment thickness in the muddy lake decreased, clear water was aspirated into the
bottom outlet and diluted the outflow concentration (Fig. 25.8).
The coarsest materials in a muddy lake settle first, thereby sorting the deposits from
individual flood events. Floodplain sediments deposited by the August 20, 1971, event
were sampled as a function of depth, showing that grain size in the deposit increased
with depth because coarse particles settled faster and became concentrated toward the
bottom of the deposit (Fig. 25.9). Deposit depth was relatively uniform from range 1 to
8 and decreased further upstream.
25.7 TURBIDITY CURRENT VENTING
In dry years characterized by low runoff during the early flood season, the reservoir was
not emptied but sediment was released by venting turbidity currents. Hydrographs for
two turbidity current venting events are shown in Figs. 25.10 and 25.11. Both events
involved relatively small inflowing floods in which the inflowing sediment-laden water
plunged beneath the already-impounded water and accumulated as a submerged muddy
lake behind the dam, where it was gradually released through the bottom outlet. There is
a 1.5-hour hydraulic travel time over the 7.5 km from the inflow gage station to the dam.
Of this distance, the first 5.8 km is above the plunge point and flow is nonstratified.
During the larger 1964 event (Fig. 25.10) the peak outflow sediment concentration slightly
exceeded peak inflow concentration. This may be a result of either scour and
entrainment of sediment downstream of the inflow gage station or increased sediment
concentration due to settling within the muddy lake. In the smaller event (Fig. 25A1), the