Geoscience Reference
In-Depth Information
scaled precipitation data from six global climate mod-
els for the twenty-first century, De Wit and Stankiewicz
(2006) calculate a decrease in perennial drainage area
across 25 % of Africa by the end of the century. In a series
of sensitivity tests, they show that for regions receiving
500 mm/yr of precipitation, a 10 % decrease in precip-
itation results in a drop of 50 % of surface drainage. A
significant dimension of this study is that its focus was
to understand implications for population access to water,
not landscape/geomorphological change per se. Its geo-
morphological significance is, however, considerable.
A similarly empirical approach was adopted by Ellis
et al.
(2008) for semi-arid central Arizona. Combining a
statistical downscaling routine of climate change simula-
tions from six global climate models with a climatic water
budget model based on evapotranspiration, precipitation
and soil moisture capacity, they show that runoff varies
from 50 to 127 % of observed levels.
Where hydrological models have been applied to assess
the hydrological impacts of climate change on runoff, it
is similarly the case that small differences in the inputs
to the hydrological model, such as temperature from dif-
ferent climate models, result in an amplification of the
uncertainty by the hydrological model (e.g. Seguı
et al.
,
2010).
The studies of changing arid zone hydrology under cli-
mate change discussed thus far all make use of either
annual or monthly precipitation. For some hydrological
regimes in arid regions, flow results from one or two pre-
cipitation events during the course of the year. The ca-
pacity to simulate these systems accurately and precisely
hinges on projections of precipitation that capture numer-
ous characteristics, including amount, intensity, duration,
type and timing (Goudie, 2006). Very few of these param-
eters have ever been extensively investigated in climate
models and, for many arid regions, observed data with
which to confront the models is in any case extremely
scarce. The potential to take projections further to actu-
ally assess impacts on specific fluvial processes within
river basins or channels is extremely limited. This is not
simply due to issues associated with the outputs that cli-
mate models can generate, but because of the problems
of systematically interpreting the behaviour and conse-
quences for processes and forms of dryland rivers today.
70
60
50
40
30
20
10
0
1991- 2000
2091- 2100
123456
7 8 9 10 11 12 13 14 15 16
Windspeed (m/s)
Figure 24.5
Histogram (frequency of occurrence on vertical
axis versus wind speed on horizontal axis) of January to March
Bodele 925 hPa winds in m/s for 1991-2000 and 2091-2100
from the MRI model.
refinement effort is aided substantially by a number
of field campaigns in desert and dryland regions such
as BoDEx (Bodele Dust Experiment), AMMA (African
Multidisciplinary Monsoon Analysis) and the Fennec (the
Saharan Climate System) project.
24.6
Climate change and fluvial systems
Changes to hydrology, including runoff, have already
been briefly discussed at the global scale earlier in this
chapter. A clear result from many models forced with
increased greenhouse gas concentrations is an intensifica-
tion of the hydrological cycle (Meehl
et al.
, 2007), with
decreases in precipitation in many parts of the already
arid subtropics. Since precipitation is zero bounded, pre-
cipitation decreases in model simulations tends be ac-
companied by decreases in precipitation variance (e.g.
Wetherald, 2009).
The dynamics of runoff and hydrology cannot be prop-
erly simulated at the resolution of the global climate mod-
els. As a result, some of the more significant research
results have emerged from efforts to represent the large-
scale forcing of climate change at higher resolution and to
simplify complex, often nonlinear hydrological relation-
ships. One such study, representative of the intermedi-
ate approach to assessing the impact of climate change
on arid zone geomorphology, approaches the problem
of assessing changes in streamflow in Africa by relat-
ing high-resolution drainage patterns to precipitation (De
Wit and Stankiewicz, 2006). Using a digitised database
of 2 million km
2
of river networks, cross-checked against
a digital elevation model from the SRTM (Shuttle Radar
24.7
Conclusions
Numerous components of arid zone geomorphology and
the potential impact of climate change upon them have
not been covered in this chapter. These include the in-
