Geoscience Reference
In-Depth Information
(a)
TLBS
LBS
BS
25
20
15
15
10
10
BS
r = 0.979
y = 0.80+0.868x
TLBS
r = 0.990
y = -1.07+1.16x
5
5
0
0
0
5
10
15
0
5
10
15
20
25
RBS
25
25
20
20
(b)
15
15
10
10
5
LBS
r = 0.977
y = -0.32+1.03x
RBS
r = 0.858
y = -0.04+1.09x
5
5
0
0
0
5
10
15
20
25
0
5
10
15
20
25
0
0
Number of upstream tributaries
1
2
Time, h
Figure 13.30
(a) Drainage nets and catchments of four unnamed contiguous sand-bed ephemeral streams, northern Kenya,
showing bed material sample locations, and the relations between the number of coarse-fine couplets of horizontal laminae and
the number of influent tributaries upstream of any sampling trench in the channel-fill of the most recent flash flood (after Frostick
and Reid, 1979). (b) Flood hydrograph of the Il Kimere, northern Kenya, showing the discharge pulses that arise from tributary
contributions to the trunk stream (after Frostick, Reid and Layman, 1983).
erosion on the poorly vegetated slopes of a water catch-
ment, as well as from the channel bed. Sediment transport
is not supply-limited. As a result, suspended sediment
concentrations reach record levels, as do bedload trans-
port rates. However, the consequences of such efficiency
are far from beneficial. For instance, the life-expectancy
of water impoundment structures is shortened consider-
ably as deltas prograde into dam lakes and fine-grained
material settles right up to dam walls. Even where the
dammed trunk river is perennial, it may be being supplied
with sediment by its highly efficient ephemeral tributaries
as it passes through a desert region. As a result, the half-
life of the reservoirs - so essential to human existence in
such regions - will be much reduced. This is the case with
the Elephant Butte Reservoir on the Rio Grande (Vanoni,
1975) and with the Tarbella Reservoir on the Indus, where
6 % of the capacity had been lost within five years of its
construction (Ackers and Thompson, 1987). In fact, loss
of capacity can be so rapid as to call into question the
sustainability of this means of eking out water supply in
drylands. Thus, for example, the capacity of the Sefid Rud
reservoir in northern Iran was reduced from 1.8 to 1.2 bil-
lion m
3
in only 16 years after closure of the dam, a loss
of 2 % per year (Tolouie, West and Billam, 1993).
However, the spottiness of rainfall in desert regions,
together with the highly variable effects of transmis-
sion losses, make ephemeral rivers highly unpredictable
in comparison with perennial equivalents. As a result,
there is less chance of achieving a stochastic model of
flows and, therefore, less chance of predicting their nui-
sance - though recent research in semi-arid environments
gives encouragement that predictive tools might even-
tually emerge. There is a great need for an increase in
information about channel form, flood hydrology and
sediment transport through an increase in fixed gauging
stations, however unrewarding it may seem to document
floods that often occur as infrequently as less than once
per year.
