Geoscience Reference
In-Depth Information
improve water use efficiency would be demanding techni-
cally, economically and politically. For example, Micklin
(2007) calculates that to overhaul 6 million hectares of
irrigation systems, and so provide an additional 12 km
3
of
inflow to the Aral, would cost at least $16 billion. The like-
lihood of such large-scale restoration schemes is unlikely
in the foreseeable future. However, partial rehabilitation
schemes involving hydrological engineering in conjunc-
tion with increased irrigation efficiency, ecological con-
servation and improved crop management may offer a
future in the short to medium term (Micklin, 2007).
mum mean daily flow of 2-8 m
3
/s. The flood inundated
the town of Calama, destroyed several bridges and caused
major channel erosion, overbank flooding and channel re-
organisation throughout the catchment. Incision of new
channels reached 2 m in depth and 5 m in width, caus-
ing widespread road disruption, and silt deposition to a
thickness of
1 m was evident over hundreds of square
kilometres. Such floods can also come with little warning.
The lag time between the centroid of rainfall and flood
peak for the Atacama flood was a mere 5 hours, even
though the drainage basin covers a large area of almost
33 000 km
2
(Houston, 2006). Smaller ephemeral catch-
ments can reach peak flood in less than an hour (Ben-Zvi,
Massoth and Schick, 1991).
Such floods are difficult to analyse and predict in dry-
land areas where rainfall is often of high intensity and,
because of its convective nature, highly localised in space
and time. Rainfall may therefore cover only a portion
(
∼
23.6
Fluvial hazards
Flash floods in deserts are a common but poorly under-
stood hazard. Indeed, the complexity of arid fluvial sys-
tems, the lack of data on rainfall-runoff relationships and
the rarity of observations of flood events in channels that
are dry for the majority of the time make assessments and
predictions of flood hazards extremely difficult. However,
the extremes of rainfall that may occur in drylands, in com-
bination with rapid runoff from sparsely vegetated slopes,
can result in major flooding events that have the capacity
to bring about lasting change to both the geomorpholog-
ical and human environment. For example, the 12 largest
floods ever recorded in the USA in basins
10 km
2
) of any individual catchment (Sharon, 1972).
The resulting partial contribution of any catchment to
a flood wave makes the system extremely difficult to
measure and model with complexity in both rainfall/
runoff (Srikanthan and McMahon, 1980; Wainwright and
Parsons, 2002) and runoff/sediment relationships (Parsons
et al.
, 2006). However, the recognition and quantification
of the hazard from flooding is becoming more urgent
as growing population centres begin to encroach upon
mountain piedmonts and drainage zones (Rhoads, 1986;
Grodek, Lekach and Schick, 2000). In such areas flash
floods can be a major hazard to both human life and in-
frastructure (Figure 23.8), and yet our lack of knowledge
concerning such floods can compound the problem as
<
1000 km
2
have all occurred in arid or semi-arid areas (Costa, 1987).
Houston (2006) reports on the Atacama flood of Febru-
ary 2001, where estimated maximum daily river flows
(the gauging station was destroyed) for the perennial Rıo
Loa River reached 136 m
3
/s in comparison to the maxi-
<
