Geology Reference
In-Depth Information
location and was complicated by changes to sea-
sonal climate patterns. However, one trend
emerged from this work: in the US, erosion can
be expected to increase where rainfall increases
significantly, but where rainfall decreases the
impacts are more complex and erosion can either
increase or decrease, depending upon the interac-
tions between the impacts of plant biomass and
rainfall on erosion.
Finally, this work also focused on the impacts
of adaptations to climate change on soil erosion.
O'Neal
et al
. (2005) studied the combined impact
of changes to climate and crop management on
soil erosion in the Midwestern US using WEPP,
with a similar climate change scenario to the one
described above, and downscaling results using a
stochastic weather generator, CLIGEN (using a
method similar to the one described by Zhang
et al
., 2004). Management practices were adapted
to fit the climate scenarios by adjusting planting,
tillage and harvesting dates, and changing crop
rotations; the scenario used a future shift from
maize and wheat to soybeans. Model results point
to an increase in soil erosion between 33% and
274% by the 2050s in most of the study areas; the
increase in erosion can be attributed to higher rain-
fall, later planting dates leaving the soil exposed
for longer, and shifts towards greater cultivation of
soybeans. Vegetation changes led to more erosion
even in regions with lower rainfall. Zhang and
Nearing (2005) used WEPP to study the impacts of
three climate change scenarios (A2a, B2a and
GGa1) on soil erosion in central Oklahoma. The
climate scenarios were downscaled from HadCM3
predictions for the 2070s, also with CLIGEN, and
predicted less rainfall and higher temperatures.
However, WEPP predicted an increase in soil ero-
sion of between 18% and 82% due to the combined
impacts of higher rainfall variability (resulting in
increased frequency of large storms) and, in some
scenarios, a decrease in wheat yield. The authors
also studied the impacts of adopting conservation
tillage and no tillage to counteract soil erosion
increase, with model results indicating their effec-
tiveness as adaptation measures.
A similar approach was subsequently applied
to the Yellow River basin in China, focusing on
the Loess plateau drylands, a region which already
experiences high levels of soil erosion (Zhang &
Liu, 2005) and where climate change is expected
to increase rainfall and, in particular, rainfall ero-
sivity (Zhang
et al
., 2005). High soil erodibility
and different climate and cropping systems pre-
sented different challenges to this work. Zhang
and Liu (2005) applied the WEPP model to two
slopes in this region, using a stochastic weather
generator (CLIGEN) to downscale three climate
change scenarios (A2a, B2a, and GGa1) from the
HadCM3 GCM for the 2080s. The results point
to an increase in soil erosion of between 2% and
81% despite a significant increase in crop yield;
in this region the rise in rainfall was the domi-
nant driving force for soil erosion changes. The
authors also concluded that the adoption of con-
servation tillage could be sufficient to adapt to
climate change and reduce the negative impacts
on soil erosion. In a subsequent work, Zhang
(2007) tested the impacts of different downscal-
ing methods on WEPP predictions. The downs-
caling approach described above was refined by
introducing an intermediate step, where GCM
results were first downscaled spatially using cur-
rent climate data for local stations, using a trans-
fer function; the spatially downscaled results
were then used to drive the stochastic weather
generator. Using this approach, the author reached
soil erosion predictions of 4 to 10 times higher
than previously. These results point to the impor-
tance of correctly downscaling GCM predictions
when studying the impacts of climate change on
soil erosion.
Finally, the WEPP model was applied to
hillslopes cultivated with soya on a tropical
hillslope in Brazil (Favis-Mortlock and Guerra,
1999). Future climate scenarios were taken from
three GCMs (HADCM2, CSIRO9 Mk2 and
ECHAM3TR) for 2050; two of the models predict
a large increase in summer rainfall, while the
third points to a slight decrease. These scenarios
were also downscaled using a statistical approach
based on CLIGEN. WEPP predicted changes to
soil erosion from −9% to +55%, following rela-
tively modest changes in rainfall (−2 to +10%), a
result also of increased water stress during the
