AN EQUATION FOR ESTIMATING THE (R) USLE SLOPE

STEEPNESS FACTORS
Dr. Matthew K. Mulengera
Department of Agricultural Engineering and Land Planning,
Sokoine University of Agriculture, P. O. 3003, Chuo Kikuu, Morogoro, Tanzania
email: [email protected]

Abstract
Currently three equations are used for calculating slope steepness factor for the Revised Universal Soil Loss
Equation (RUSLE); one for slopes less than nine percent, the second for slopes equal or greater than nine
percent, and the third equation that has been proposed for use in China for slopes equal or greater than 17
%. The three equations have linear relationships between slope steepness factor, S and the sine of the slope
angle, θ. The three equations under predict the factor, S when used on slope steepness beyond the data sets
used to develop them. In addressing this problem several alternative forms of equations (linear, power, and
polynomial) were tested using field plot soil loss data gathered by several researchers in different countries
for slopes ranging from three percent to 55 %. A single power function relating the sine of the slope angle, θ
to the slope steepness factor, S has been identified to be more suitable and accurate for estimating the RUSLE
slope steepness factor, S.

Keywords: Slope steepness factor; Soil loss prediction; Statistical regression

Introduction s n
S = ( ) (1 )
9
The slope steepness factor, S for the (Revised) Where, s is the tangent of the slope angle and n
Universal Soil Loss Equation is the ratio of soil is 1.4 or 1.35 respectively.
loss from a field with given slope gradient to
that from a field with slope gradient of nine Smith and Wischmeier, (1957) using data from
percent under otherwise identical conditions runoff plot with gradients varying from three
(slope length, soils, soil moisture regimes, percent to 18% developed the slope steepness
rainfall, and management) (Wischmeier and factor equation of the form:
Smith, 1978). The slope steepness (gradient)
affects soil erosion because it affects soil S = 0.0065s 2 + 0.0453s + 0.065 (2)
detachment and transport by rainfall splash
(Brian, 1979; Savat, 1981) and runoff
Wischmeier and Smith (1978), realising that
(Wischmeir and Smith, 1978). The relationship
shear stress of the surface flow is related to the
of slope steepness to soil loss is influenced by
sine of the slope angle (Chow, 1959), used
its interactions with soil properties, vegetation
same data for equation 2 to develop the slope
types, surface roughness, and residue
steepness factor of the form:
management (Wischmeier and Smith, 1978;
Liu et al,1994), and rainfall characteristics (Liu
et al, 1994). S = 65.4 sin 2 θ + 0.0453sin θ + 0.0654 (3)

Different researchers have found varying McCool et al., (1987) using data from
relationships between slope steepness and soil simulated rainfall for slopes of 0.1% to three
loss. Zingg (1940), using simulated rainfall on percent and data from natural rainfall for slopes
slopes up to 20%, and Musgrave (1947), using of three percent to 18 % and Liu et al. (1994)
composite data from slopes up to 16% found using data from natural rainfall for slopes of
the slope steepness factor, S equation of the nine percent to 55 %, developed equations of
form: the form:
 M. K. Mulengera

S = A sin θ + B (4) of slope angles and sine values of the slope

angles were used to test the three different
Where, A equals to 10.8, 16.8, or 21.91 and B functional forms, (i.e. linear, polynomial, and
equals to 0.03, -0.5, and –0.96 respectively. power) to identify the most consistently
accurate equation(s) for estimating the slope
Although for the slope steepness ranges used steepness factor, S.
the linear relationship between soil loss to the
sine of the slope angle was statistically more Results and Discussion
accurate the increasing gradients of the
equations with increasing slope steepness Figures one to three and table 1 show some of
suggest that using the equations for land slopes the statistical regression analysis results for the
beyond that used to develop them results in slope steepness and soil loss relationships. The
under prediction of soil loss. This poses three expressions shown are the best results
problems in planning soil conservation in from the regression analysis. Table one shows
mountainous regions where cultivation of that the equations are almost equally accurate.
steeper slopes is common. In this study an All are statistically significant at more than
equation which can be used for all land slopes, 99.95 %. The slope steepness factor equation of
giving accurate predictions is presented. the polynomial form (Fig. 2) predicts zero soil
loss when gradient equals 1.3 % and below this
Materials and Methods value the equation predicts negative soil loss
(deposition). Equation 3 by Wischmeier and
Data on slope steepness and soil loss for slopes Smith (1978) developed using data for slopes
ranging from three percent to 55 % were ranging from 3 % to 18 % does not predict zero
assembled from different literature sources soil loss even when the slope is zero. Because
(Wendelaar, 1978; Liu et al., 1994; Vogel, its development used more data from low
1994; Mulengera; 1996). All the data came slopes it should be considered more accurate
from natural runoff plot research. All soil loss within its data range. Thus, the polynomial
data not normalized to nine percent slope as equation developed in this study is not as good
required for use in the RUSLE were normalized as that by Wischmeier and Smith (1978) for
(Murphree and Mutchler, 1981). land slopes below the lowest value of runoff
plots data used for its development (i.e. three
The Microsoft Excel software was used for percent).
statistical regression analysis. Tangent values

96 Tanzania Journal of Engineering and Technology, (TJET) Vol. 31 (No.2), Dec, 2008
 A. A. Shariff, S. A. Tamba & A. A. Mdimi

12
y = 24.748x1.2656
10 R2 = 0.9449

topographic factor, S 8

0
0 0.1 0.2 0.3 0.4 0.5 0.6
sine of land slope angle

Fig. 1: Power function relating topographic factor, S to sine of slope angle,

θ

12
2
y = 11.252x + 15.418x - 0.202
10 2
R = 0.9491
Topographic factor, S

0
0 0.1 0.2 0.3 0.4 0.5 0.6
Sine of land slope angle

Fig.2: Polynomial function relating topographic factor, S to sine of slope

angle, θ

Tanzania Journal of Engineering and Technology, (TJET) Vol. 31 (No.2), Dec, 2008 97
 M. K. Mulengera

Fig. 3: Linear function relating topographic factor, S to sine of slope angle, θ

Table 1: Some statistical analysis output for the regression equations

Equation Coefficient of Residual standard p-value

determination, R2 deviation, S
S = 24.748(sinθ ) 1.2656 0.945 0.76 << 0.0005
S = 11.254 sin θ + 15.41sin θ − 0.202
2
0.950 0.73 << 0.0005
S = 21.03 sin θ + 0.684 0.950 0.73 <<0.0005

The regression line relating soil loss to slope equations for calculating slope steepness factor
gradient in this study (Fig. 3) is almost similar of the RUSLE. The equation proposed by
to that developed by Liu et al., (1994). This is McCool et al., (1987) for slopes less than 9 %
due to the fact that out of the 24 data set used, does not predict deposition as is the equation 3
only 7 data sets were those not collected by developed by Wischmeier and Smith (1978).
them and were below the slope range they used Therefore, the linear equation developed in this
to develop their equation, (i.e. 9 %). Although study is also not as good as that by McCool et
the coefficient of determination for the linear al., (1987) for use on low slope gradients.
function relationship is the highest (Fig. 3), the
regression line passes below all the observed The expression for slope steepness factor in
soil losses for slopes less or equal to 4.5 %. The Figure 1 predicts zero erosion when slope
equation predicts zero soil erosion when land gradient is also zero and soil erosion prediction
slope is 3.25 % and the equation predicts increases exponentially with slope. All the
deposition for lower slope gradients. McCool et equations developed by researchers for low
al., (1987) using more data for slopes ranging slopes predict soil erosion almost equal to zero
from 0.1 % to 18 %found that there was a break when land gradient is zero (Wischmeier and
in regression line slope at 8 % land slope. As a Smith, 1978, McCool et al., 1987, Liu et al.,
result they proposed the two currently used 1994) and do no predict any deposition.

98 Tanzania Journal of Engineering and Technology, (TJET) Vol. 31 (No.2), Dec, 2008
 A. A. Shariff, S. A. Tamba & A. A. Mdimi

Further comparison of the equation with the almost same soil loss (Fig. 4). Only the
three currently used RUSLE slope steepness equation by McCool et al., (1987) predicts less
factor equations (McCool et al., 1987, Liu et soil erosion for slopes greater than about 17 %
al., 1994) for slope ranges proposed by Liu et as was found by Liu et al., (1994).
al. (1994) shows that all the models predict

12

10

8
S te e p n e s s fa c to r S

0
0 20 40 60 80
S lo p e g r a d ie n t ( % )

M e a s u r e d S F a c to r
D e v e lo p e d e q u a t io n , F ig . 1
E q . b y M c C o o l e t a l. , 1 9 8 7
E q , b y M c C o o l e t a l, 1 9 8 7
E q , b y L iu e t a l. , 1 9 9 4

Fig. 4: Measured and predicted slope steepness factor values

Summary and Conclusions plot soil loss data gathered by several

researchers in different countries for slopes
Several equations have been proposed for ranging from three percent to 55 %. A single
calculating slope steepness factor of the revised power function relating the sine of the slope
USLE. Currently three equations are used, one angle, θ to the slope steepness factor, S has
for slopes less than nine percent, the second for been identified to be more suitable and accurate
slopes equal or greater than nine percent, and for estimating the RUSLE slope steepness
the third equation that has been proposed for factor, S for all the slope ranges. As it has
use in China for slopes equal or greater than 17 shown to be consistently accurate for all the
%. The three equations have linear relationships slope steepness range of zero to 55%, it is
between slope steepness factor, S and the sine intuitively judged to be more accurate beyond
of the slope angle, θ. The three equations under the slope gradient range used to develop it.
predict the factor, S when used on slope
steepness beyond the data sets used to develop References
them.
Brian, R. B., (1979). The influence of slope
Several alternative forms of equations (linear, angle on soil entrainment by sheetwash
power, and polynomial) were tested using field
Tanzania Journal of Engineering and Technology, (TJET) Vol. 31 (No.2), Dec, 2008 99
 M. K. Mulengera

and rainsplash. Earth Surface Processes Savat, J., (1981). Workdone by splash:
4:43 – 58. Laboratory Experiments. Earth Surface
Processes, 6: 275 – 283.
Chow, V. T., (1959). Open Channel
Hydraulics. McGraw Hill ,New York. Smith, D. D., and Wischmeier, W. H., (1957).
Liu, B. Y., Nearing, M. A., and Risse, L. M., Factors affecting sheet and rill erosion.
(1994). Slope gradient effects on soil loss Trans. Am. Geophys. Union, 36 (6): 889
for steep slopes. Trans. of the ASAE. – 896.
Vol. 37(6):1835 – 1840.
Vogel H., (1994). Conservation tillage in
McCool, D. K., Brown, L. C., Foster, G. R., Zimbabwe: Evaluation of several
Mutchler, C. K., and Meyer, L. D., techniques for development of
(1987). Revised steepness factor for the sustainable crop production systems in
Universal Soil Loss Equation. Trans. Of small holder farming, PhD thesis, Univ.
the ASAE 30(5):1387 – 1396. of Berne, Switzerland.Geographica,
Bernensia, African Studies Series, No.
Mulengera, M. K., (1996). Soil loss prediction A11, 150 p.
in the semi-arid tropical savanna zone: A
tool for soil conservation planning in Wendelaar, F. E., (1978). Applying the
Tanzania, PhD Thesis, Univ. of Universal Soil Loss Equation in
Newcastle upon Tyne, U. K., 232 p. Rhodesia, Unrefenced Report, Inst.
Agric. Eng., Harare, Zimbabwe.
Murphree, C. E., and Mutchler, C. K., (1981).
Verification of the slope factor in the Wischmeier, W. H., and Smith. D. D., (1978).
Universal Soil Loss Equation for low Predicting rainfall erosion losses –
slopes. J. of Soil and Water Conserv. Vol. Aguide to conservation planning. U. S.
36(5): 300 – 302. Dept. of Agric., Agric. Handb., No. 537,
58 p.
Musgrave, G. G., (1947). The quantitative
evaluation of factors in water erosion – Zingg, A. W., (1940). Degree and length of
Afirst approximation. J. Soil and Water land slope as it affects soil loss in runoff.
Conserv., 2(3): 133 – 138, 170. Agric. Eng. 21(2):59 – 64.

100 Tanzania Journal of Engineering and Technology, (TJET) Vol. 31 (No.2), Dec, 2008