SOIL PROPERTIES, CONDITION AND SOIL LOSSES FOR SOUTH AND EAST BRAZILIAN FOREST AREAS

The slope-length exponent m is related to the ratio (β) of rill to interrill
erosion by the following equation (Foster et al., 1977):
m β 1 β ^{Eq. 5}
Values for the ratio β of rill to interrill erosion were computed according

to the equation proposed by McCool et al. (1989):
_0.8 Eq. 6

β sinθ 0.0896 3.0 sinθ 0.56
where θ is the slope angle (degree).
The slope-steepness factor (S) was evaluated according to equations

presented in McCool et al. (1987) for slopes greater than 4 m, which were
applied in several studies like Bhattarai & Dutta (2007), Pandey et al. (2007),
Dabral et al. (2008) and Beskow et al. (2009):
S 10.8 sinθ 0.03 for slope < 9% Eq. 7
S 16.8 sinθ - 0.50 for slope ≥ 9% Eq. 8
where S is the slope-steepness factor, and θ corresponds to the slope angle
(degree). The combined LS factor was calculated by multiplying L and S factor
maps and the final map was generated through GIS software.

d) Cover-management factor (C)
The cover-management factor (C) reflects the effect of cropping and

management practices on erosion rates, indicating how the conservation
measures used can affect the average annual soil loss. It varies with activities,
such as crop rotations, or other management practices (Renard et al., 1997). The
concept of the C factor is the ratio of soil loss from an area with specific cover
and management to soil loss from an identical area in a clean-tilled continuous
fallow condition (Wischmeier & Smith, 1978). The soil loss ratio (SLR) is then
estimated by the ratio between soil loss under actual conditions and soil loss
experienced under the reference conditions:

57

SLR SL SL ^{Eq. 9}
c r
where SL_c is the soil loss from a specific cropping/management, and SL_r is the
soil loss from reference conditions (bare-soil).
The C factor was calculated by multiplying the SLR for each time
interval by its corresponding EI value (Wischmeier & Smith, 1978), using the
following equation:
C ... ^{Eq. 10}
SLR₁ EI₁ SLR₂ EI₂ SLR_n EI_n EI_t
where C is the cover-management factor, SLR_i is the soil loss ratio for the time
step i, EI_i is the EI occurring during that time step, n is the number of time steps
used, and EI_t is the sum of EIs for the entire period of time.

e) Support practice factor (P)
The P factor is the ratio of soil loss with a specific support practice, such
as terracing, strip cropping, or contouring, to the corresponding soil loss with up
and downslope tillage (Renard et al., 1997). For this study, no support practices
were considered, so the P factor was set equal to 1.0 for the entire area. A
similar assumption was also adopted by Gaffer et al. (2008), Ozcan et al. (2008),
Beskow et al. (2009), and Kouli et al. (2009).

4.3 Soil Loss Tolerance (T)
The soil loss tolerance (T) is closely related to soil formation. In other
words, the tolerable value can be considered equal to soil formation rate. Thus,
the term soil loss tolerance denotes the maximum rate of soil erosion that can
occur and still allows crop productivity to be economically sustainable (Renard
et al., 1997). This way, agricultural soils can ‘‘tolerate’’ a certain amount of
erosion without adversely impacting on long-term productivity because new soil
is constantly being formed to compensate losses (Bazzoffi, 2008). Within
tolerance context, the erosion value can be considered as the weathering-limited

58

process (Stallard, 1995), where the transport processes remove weathered
material from an area more rapidly than the weathering can generate soil
material (Bazzoffi, 2008). The tolerable value depends on soil depth, on soil
organic matter amount, on soil permeability, on clay ratio between A and B
horizon, and on clay amount (Galindo & Margolis, 1989). For such soils, the
clay amount and soil permeability might be the main properties to consider
because they showed high clay amount and low permeability, mainly in the B
horizon.
The tolerance values used in this study were generated by Martins
(2005) who applied the methodologies suggested by Smith & Stamey (1964),
Lombardi Neto & Bertoni (1975), and Galindo & Margolis (1989). Martins
(2005) presented a T value for each soil class as the average value obtained
through the aforementioned methods. Thus, the T values were 10, 13 and 11 t
ha^-1 yr^-1 for the PA1, FX and PA2, respectively. These values were compared to
USLE predictions using GIS tools to identify areas where land-use and
management were appropriate, and to identify areas that need more attention in
order to reduce and to prevent long-term soil degradation.

4.4 Geographic Information System (GIS)
GIS is a tool for making and using spatial information. It can be defined
as a computer-based system to aid in the collection, maintenance, storage,
analysis, output, and distribution of spatial data and information (Bolstad, 2005).
Within the GIS environment a raster data model describe the area as a regular set
of cells in a grid pattern. Thus, for EAW the cell dimension was defined as 10
meters on each side. The integration of USLE model and GIS framework can be
established by converting all parameters of USLE into a raster-based format and
by evaluating these digital parameter layers. Thus, a map illustrating the water
erosion potential for the watershed was created. To do this, the USLE model

59

(Equation 1) was applied by multiplying different layers with GIS software: K,
R, LS, C and P factors.

5 RESULTS AND DISCUSSION

5.1 USLE Parameters
Annual rainfall-runoff erosivity ranged from 4,536 MJ mm ha^-1 h^-1 yr^-1
to 17,056 MJ mm ha^-1 h^-1 yr^-1 (Figure 4). The large variation shows that the

assessment of the rainfall-runoff erosivity for this particular watershed was very
important to achieve a reliable estimate of the R factor. The R factor calculation
based on rainfall amount and/or geographic location can be useful for regions of
Brazil where intensity values have not been recorded. However, when such
information is available, the rainfall-runoff erosivity factor can be estimated
with more accuracy.

60

FIGURE 4 Rainfall-runoff erosivity index for the studied watershed. For 2004 it
was showed a partial value from events between January and July
(modified from Martins, 2005).

The greatest R value (17,056 MJ mm ha^-1 h^-1 yr^-1) was observed in 2004.
It is important to mention that the R value for March/2004 corresponded to
12,540 MJ mm ha
^-1 h^-1. This monthly value was greater than the annual values
for previous six years (Figure 4). A gap in the data did not allow the
determination of the total erosivity factor in 2004; nevertheless, it can be
inferred that it would be high since higher erosivity values have been observed
during December months. In this study, the spatial distribution of the R factor
was assumed to be constant in the entire watershed. The EAW area is about only
286 ha, where small rainfall spatial variability is expected, consequently a small

61

variation in rainfall-runoff erosivity. This procedure was also adopted by Dabral
et al. (2008) for a much larger watershed with 127,878 ha in Northeastern India.
The soil erodibility map (Figure 5) was generated using results from
field plots which were measured by Martins (2005). It is worthwhile to point out
that the greater the erodibility value, the lesser resistance of the soil to water
erosion. Analyzing the K factor map, it can be noticed that the major part of the
watershed (92%) can be considered as having low soil erodibility, with values
_{less than 0.010 t h MJ}^-1 mm^-1, while the remaining part has moderate soil
erodibility (Foster et al., 1981).

FIGURE 5 Erodibility factor (K) for the experimental watershed.
62

The LS factor was calculated by Equations 4, and 7 or 8 taking into
account its spatial variation which ranged from 0.03 to 6.88 (Figure 6). The
spatial analysis of the LS factor indicated that 65% of the EAW had a
topographic factor less than 1.0. It means that in only a small part of the
watershed the LS factor resulted in a water erosion risk. In addition, through
slope categories (Figure 3), it is also found that a considerable area (52%) had
gradients less than 5%, indicating a low risk for soil losses. On the other hand,
the steeper slopes may cause more runoff and result in greater soil erosion.

FIGURE 6 Topographic factor (LS) of the studied area.
63

The watershed is composed of only three types of land-use, namely
native forest (30%), planted eucalyptus forest (59%) and forest roads (11%).
Figure 7 shows the C factor map which was generated according to Equation 10.
From such equation, the C factors calculated were 0.297 and 0.017 for the
eucalyptus plantation and the Atlantic Forest (native forest), respectively. The C
value for the forest roads was assumed to be equal to 1.0. To our knowledge,
these cover-management factor values for eucalyptus and Atlantic Forest are the
first ones obtained directly from field experiments in Brazil or even in South
America. Researchers have come up with a certain variation of the covermanagement
factor for forest areas. Roose (1977) recommended a value of
0.001 for dense tropical forest in Africa. The C factor estimated for broad-leaved
forest in Greece was 0.130 (Kouli et al., 2009) based on satellite image analysis.
Among the USLE factors the cover-management is the easiest one that can be
modified in order to substantially reduce the erosion risks. In this context,
keeping native forest may reduce soil erosion risk by hundreds of times
compared to forest roads.

64

FIGURE 7 Cover-management factor (C) for the studied area.

5.2 Spatial Distribution of Soil Loss
The map of long-term annual average soil loss (Figure 8) was generated
using Equation 1. This calculation was done multiplying the USLE factors using
GIS. The soil loss values calculated ranged from 2x10^-3 to 983 t ha^-1 yr^-1, with a
weighted average value equal to 6.2 t ha^-1 yr^-1. The annual soil loss values were
reclassified (Table 1) according to the classes suggested by Bahadur (2009). In
Table 1 we can observe that 72.7% of the watershed area had an erosion rate
under “very slight” category, with annual soil loss less than 6 t ha^-1 yr^-1. This

65

behavior can be explained due to the predominant low slope gradient (Figure 3)
and the soil mineralogy in the watershed, which has a high kaolinite content with
very low gibbsite and iron-oxide (Duarte et al., 2000), favoring blocky structure
and increasing cohesion and, consequently, decreasing soil erosion. In addition,
the greatest part of the watershed (65%) has LS factor less than 1.0 (Figure 6)
and an adequate cover management (Figure 7). We can conclude that such
factors contributed to the low erosion rate, although the erosivity factor was
considered high according to Foster et al. (1981). The results indicated that
“severe” to “very extremely severe” erosion risk occurs in 11.5% of the area
(Table 1). The greatest soil erosion values were found in sites occupied by forest
roads (Figure 7) and with high LS values (Figure 6). This kind of information is
extremely valuable since it can be used to plan a conservation management with
focusing on sites with a high potential for water erosion. Sediment delivery
control practices like sediment basins, barriers, containment structures, and
vegetable drainage ditches should be constructed in order to keep the sediments
from moving, thus reducing the process of sediment transport. According to
Antonangelo & Fenner (2005), forest roads have been one of the main reasons
for soil erosion and siltation of rivers in forest areas. The construction of forest
roads removes natural protection and makes soil movement easier, thus making
these roads more vulnerable to the effect of rainfall-runoff erosivity.

66