Geology Reference
In-Depth Information
growing period for soya leading to less vegetation
cover. This was accompanied by an increase in
the spatial and temporal variability of erosion.
Overall, these studies provide a good example
of a comprehensive modelling approach to evalu-
ate the impacts of climate change on soil erosion.
They are also representative of a type of early
modelling study (e.g. Favis-Mortlock & Boardman,
1995) not discussed in this section. The results
for the US cover the range of possible impacts
using hypothetical climate change scenarios
(Pruski & Nearing, 2002a) and evaluate interac-
tions between changes in rainfall, runoff and
vegetation productivity for different climatic
conditions (Pruski & Nearing, 2002b), as well as
the impacts of agricultural land use changes
(O'Neal
et al
., 2005) and the potential to imple-
ment adaptation measures (Zhang & Nearing,
2005). In China, the uncertainty inherent in cli-
mate change predictions was also explored
(Zhang, 2007). The main results point to the com-
plex interactions between different impacts of
climate change, which can lead to increases in
soil erosion even where rainfall is expected to
decrease; the importance of vegetation biomass
productivity in these regions is highlighted.
However, these studies were constrained to the
slope scale and agricultural fields; possible
impacts on rangelands, gully erosion or catch-
ment sediment yield were not studied, and should
not be inferred from the results due to the com-
plex nature of the processes involved.
reservoirs and irrigation schemes. The following
SWAT applications were focused on processes
linking hillslopes to the river network to take
advantage of these features. In contrast with the
single-slope WEPP applications detailed above,
SWAT was applied to complex catchments.
The first example is an application of SWAT to
a watershed in Finland by Bouraoui
et al
. (2004).
In contrast with the usual approach for climate
change impact assessment, the authors looked at
the impacts of recent climate change (1965 to
1998) on river flow and sediment yield, by com-
paring model results with observed climate and a
synthetic climate series where rainfall and tem-
perature increases were removed using non-
parametric methods. The model results pointed
to an increase in winter runoff and suspended
sediment caused by a combination of increasing
rainfall and decreasing snow cover. These results
indicate the likely trend for future climate change
impacts in this region. Another example is given
by Chaplot (2007), who applied SWAT to two
watersheds in the US, one with a humid climate
and agricultural land (in Iowa), and the other with
a semi-arid climate and a significant proportion
of pasture (in Texas). The author simulated the
impacts on runoff and soil erosion of two CO
2
and temperature change scenarios, combined
with rainfall changes from −40% to +40%; the
scenarios were stochastically generated using
CLIGEN. Model results point to a dominant
impact of rainfall in soil erosion rates at the
humid watershed, with the wettest scenario lead-
ing to an increase of 157%, and similar decreases
for the drier scenarios. Soil erosion in the semi-
arid watershed, however, did not show great sen-
sitivity to changes in climate, except by decreasing
for the −40% rainfall scenario; surprisingly, soil
erosion decreased in the wetter scenarios, possi-
bly due to improved conditions for vegetation
cover in the winter.
A final example is given by Nunes
et al
. (2008),
who applied SWAT to two groups of watersheds
in Portugal, also with humid and semi-arid cli-
mates. Climate change scenarios were generated
using CLIGEN with the intention of simulating
the range of climate change predictions by GCMs
15.4.2 Continuous modelling
at the catchment scale
Another model used to assess the impacts of cli-
mate change on soil erosion has been SWAT - Soil
and Water Assessment Tool (Neitsch
et al
., 2002).
Like WEPP, SWAT simulates hillslope erosion
processes, sediment transport and deposition,
vegetation growth and residue processes, and
agricultural operations. Unlike WEPP, however,
SWAT was designed to simulate mesoscale catch-
ments, trading detail at the slope scale for the
ability to represent more complex catchment
and river network structures, including large
