Geology Reference
In-Depth Information
for this area, with temperature increases from
2°C to 6°C and rainfall decreases from 2.5% to
40%. The authors simulated changes to each
parameter separately; they also simulated two
sets of scenarios of combined rainfall and tem-
perature change with wetter and drier conditions
(rainfall decreasing by 1.6% and 6.2% respec-
tively per 1°C increase in temperature). The
results point to rainfall changes as the main driv-
ing forces for soil erosion in all landcovers except
in wheat croplands, where temperature increases
were more important due to the negative impact
on biomass production and soil cover. For the
combined changes, model results varied signifi-
cantly between vegetation cover types; range-
lands and managed forests showed a decrease in
soil erosion in all scenarios, while agricultural
lands (wheat croplands and vineyards) responded
differently according to the combination of rain-
fall and temperature changes. For the drier scen-
arios, soil erosion decreased in both agricultural
landcovers; for the wetter scenarios, soil erosion
decreased slightly in vineyards (−25%) and
increased in wheat croplands (up to 149%). These
results are important since these landcover types
represent the most important sediment sources
in the study area. The authors also found greater
responses in the humid watershed, as in the pre-
vious study, and note that in one of the test sites
the shallow soils (c. 10 mm) were responsible for
a relatively low sensitivity of surface runoff to
rainfall decreases.
In short, these studies are not as comprehen-
sive as those presented in the previous section,
but they do provide additional information at the
large catchment scale. The simulated areas are
quite large (up to 1000 km
2
in all studies), and
include different vegetation cover and soil types,
which impact differently upon similar climate
change scenarios; these different impacts com-
bine to determine changes to watershed sediment
yield. In particular, these results confirm one of
the conclusions of the previous case studies,
namely that the impact of decreasing rainfall
rates on soil erosion is complex and depends upon
the impact on vegetation biomass growth; how-
ever, these results also indicate that croplands in
drying climates are particularly vulnerable to
increases in soil erosion rates.
A final note should be made on the robustness
of both this and the previous modelling
approaches. While the SWAT approach has a
larger spatial domain, there was in all cases a lack
of data for validating the erosion simulations; the
model was assessed using sediment yield meas-
urements in the channel network, which is not
sufficient to ensure that sediment sources and
sinks are being correctly simulated (Jetten
et al
.,
2003). The WEPP approach was more robust,
since it was applied to heavily monitored slopes
with data available for calibration and validation.
This difference illustrates how the lack of meas-
ured erosion data may hamper climate change
impact studies using watershed-scale models.
15.4.3
Grid-based continuous modelling
A different example of erosion modelling at the
slope scale was performed using the PESERA -
Pan-European Soil Erosion Risk Assessment
model (Kirkby
et al
., 2004). PESERA simulates
erosion at the slope scale in a similar way to
SWAT and WEPP, including a vegetation growth
component; however, the model has been applied
using a 1 × 1 km grid for western and central
Europe, taking into account the spatial variability
of rainfall, relief, soil and vegetation properties.
This grid-based approach allows an estimation of
soil erosion for large areas while taking into
account some degree of spatial variability,
although the processes represented are still at the
slope scale (e.g. gully erosion and channel proc-
esses are not taken into account by PESERA). This
approach could also be used to generate global
predictions for the impacts of climate change on
soil erosion, similar to the work done for surface
runoff in recent years (e.g. Manabe
et al
., 2004).
This approach was performed by Mantel
et al
.
(2003) using the A2b climate scenario based on the
HadRM3 RCM. The study was performed for
two windows in northwestern and southwestern
Europe, with contrasting climates (humid and dry,
respectively) and land uses. Land-use change scen-
arios (switching other arable crops to maize) were
