Geoscience Reference
In-Depth Information
fatal for people caught unawares (Hjalmarson, 1984). This
is undoubtedly because the time of rise of the flood hy-
drograph is short. Reid
et al
. (1994) have measured rates
of rise in water-stage as high as 0.25 m/min and average
water velocities
(a)
10
3
10
3
3 m/s within a few minutes of the pas-
sage of the bore. A person caught mid-channel in these
circumstances can easily lose footing and be swept away.
Reports of the interval between the arrival of the bore and
peak discharge for catchments of small or moderate size
range from 18 to 23 minutes (Renard and Keppel, 1966),
10 minutes (Schick, 1970) and 14 to 16 minutes (Reid and
Frostick, 1987).
The second characteristic of ephemeral stream hydro-
graphs is the steepness of the flood recession (Figure 13.5).
This may reflect the overriding importance of Hortonian
overland flow in arid environments. Observations of the
limited penetration of the wetting front in the soil (e.g.
≤
>
10
2
10
2
10
1
10
1
10
0
10
0
10
-1
10
-1
10
-2
10
-2
runoff
event sediment yield
210 mm, Reid and Frostick, 1986) suggest no substan-
tial subsurface routing of water that might sustain stream
discharge (Pilgrim, Chapman and Doran, 1988). Besides
this, the extreme dryness of the soil always means a high
gradient of matric potential, which would draw water
downward into the soil profile and reduce the chance of
achieving pressures greater than atmospheric that would
permit soil interflow or pipeflow.
This leads to the third characteristic that is peculiar
to ephemeral systems in modest-sized catchments. The
flood is often extremely short-lived. The whole event from
initial to final dry bed might have taken no more than a
few hours (Figures 13.5 and 13.6) and rarely more than a
day. Because the number of floods that might be expected
ranges from around six (Reid and Frostick, 1987) to much
less than one per year (Schick and Lekach, 1987), as one
moves from semi-arid to arid environments, this means
that the river system is active for much less than 1 or
2 % of the time. The probability of being on-site to make
observations and to take measurements during an event is
extremely small. This and the low rate of data acquisition
are the chief reasons for the paucity of information that
we have regarding desert flash floods.
0.01
0.10
0.50
0.90
0.99
0.05
0.30
probability
0.70
0.95
(b)
10
3
10
2
10
1
10
0
10
-1
winter
autumn-spring
0.01
0.10
0.50
0.90
0.99
0.05
0.30
probability
0.70
0.95
Figure 13.4
(a) Probability (
p
=
r
/(
n
−
1), where
r
is rank
order and
n
is the total number of events) of exceedance for
runoff volume and total (suspended sediment load plus bed-
load) sediment yield of flash floods in the Nahal Eshtemoa,
1991-2006. (b) Probability of exceedance for total sediment
yield of flood events arising from convective (autumn-spring)
and frontal (winter) storms (after Alexandrov
et al
., 2009).
13.2.3
Transmission losses
Another factor that sets ephemeral systems aside from
those that flow perennially or even intermittently is the
loss of water to the stream bed. The process has been
quantified by Renard and Keppel (1966), Burkham (1970)
and Butcher and Thornes (1978) for streams in Ari-
zona and Spain, and noted in many other instances
(such as Murphey, Lane and Diskin, 1972; Schick and
distinguishing between those travelling over antecedently
dry or wet channel beds, and indicates remarkably mod-
est velocities
≤
1 m/s. These are corroborated by other
observations made in Israel by Reid, Laronne and Powell
