Soil Erosion and Conservation - Environmental Modelling: Finding Simplicity in Complexity - page 356

N 2 O, CFC-11, CFC-12, and HCFC-22 (Edwards and
Slingo, 1996), a parameterization of simple background
aerosol climatology (Cusack et al ., 1998), and several
other improvements over the previous Hadley Centre
model, HadCM2. Results from the model are reported on
a2.5 ◦ latitude by 3.75 ◦ longitude grid.
The Canadian Global Coupled Model (CGCM1) (Boer
et al ., 2000), is composed of an atmospheric component
based on the model GCMII (McFarlane et al ., 1992)
coupled with an ocean component based on the model
GFDL MOM1.1 (Boer et al ., 2000). For the current study
we used results from the simulation GHG

Centre model results also indicated a reduction in ero-
sivity across much of the southern plains, again from
Texas to Nebraska, but extending somewhat west of the
corresponding area shown in the Hadley Centre results.
The Canadian Centre model did not show consistent
results for the south-eastern United States. Results of
the computations using the annual precipitation (see
maps in Nearing, 2001) indicate changes in parts of the
southeast United States tending toward lower erosivity,
corresponding to a tendency toward a decrease in the
annual precipitation in that region. Results of the erosiv-
ity computations using the Fournier coefficient indicate
the possibility of little change or increases over part of the
region for the 80-year comparison (see maps in Nearing,
2001). Calculated increases in erosivity using the Fournier
coefficient suggest a change in the distribution of rainfall
patterns through the year.
Erosivity results calculated from the Canadian Centre
for Climate Modelling and Analysis and the Hadley Cen-
tre models show major differences in the south-western
United States, including California, Arizona, Nevada, and
Utah. Whereas the Hadley Centre model results suggest
a definite trend towards lower erosivity in this area, the
Canadian Centre for Climate Modelling and Analysis
model results suggest a definite, strong trend toward
greater erosivity through the twenty-first century.
The amount of inconsistency in the calculations from
the twomethods of calculating erosivity trendswas, for the
most part, similar between the two models (Table 22.1).
Overall, between 16 and 20% of the calculations resulted
in negative values of the R-factor calculated from total
annual rainfall, RP, when the R-factor calculated from the
Modified Fournier coefficient, RF, was positive, or vice
versa . For the cases where both RP and RF were large, i.e.,
greater than 10%, those percentages were much smaller,
although 7.6% of the pairs were inconsistent in this case
for the Canadian model results for the 80-year time
interval (2000-2019 to 2080-2099). It is not out of the
question to expect inconsistencies between results of RP
andRF, since RP is based on total annual precipitation and
RF is based on the monthly distributions of precipitation.
Both relationships are statistically based, and we have no
reason to favour one over the other.
One might expect a consistent trend for the change of
erosivity as a function of time, and in general this was true
(Table 22.2). In this case, the Canadian model exhibited
more inconsistency as function of time when using the
monthly precipitation values to calculate erosivity, though
it was consistent temporally in terms of the erosivity
calculated using the annual precipitation.

A1, which
incorporated an increase of atmospheric concentration
of greenhouse gases (GHG) corresponding to an increase
of 1% per year for the time period studied, as well as
the direct forcing effect of sulphate aerosols (Reader and
Boer, 1998). The data from this model were presented on
a Gaussian 3.75 ◦ by 3.75 ◦ grid.
Changes in rainfall erosivity for the two models were
computed for two time intervals, 40 and 80 years. In
the first case the values of erosivity from the 20-year
period from 2040 to 2059 were compared to the period
2000-2019, and in the second case the values of erosivity
from the 20-year period from 2080 to 2099 were com-
pared to the period 2000-2019. Erosivity changes were
computed in two ways: (a) as a function of change in
average annual precipitation for the twenty-year periods
using equations 11 and 12 from Renard and Freimund
(1994), and (b) as a function of the Fournier coefficient
for the twenty year periods using equations 13 and 14
from Renard and Freimund (1994).
The erosivity results calculated from the Hadley Centre
model analyses indicated a general increase in rainfall
erosivity over large parts of the eastern United States,
including most of New England and the mid-Atlantic
states as far south as Georgia, as well as a general
increase across the northern states of the United States
and southern Canada (see maps in Nearing, 2000). The
Hadley Centre results also indicated a tendency for ero-
sivity increases over parts of Arizona and New Mexico.
Decreases in erosivity were indicated in other parts of the
south-western United States, including parts of Califor-
nia, Nevada, Utah, and western Arizona. Decreases were
also shown over eastern Texas and a large portion of the
southern central plains from Texas to Nebraska.
The erosivity results calculated from the Canadian
Centre for Climate Modelling and Analysis model also
showed an increase in erosivity across the northern states
of the United States, including New England, and south-
ern Canada (see maps in Nearing, 2001). The Canadian

+