ISCO 2004 - 13th International Soil Conservation Organisation Conference – Brisbane, July 2004

Conserving Soil and Water for Society: Sharing Solutions

ANALYSIS OF THE RUSLE AND WEPP MODELS FOR A SMALL WATERSHED LOCATED
IN VIÇOSA, MINAS GERAIS STATE, BRAZIL

R.A. Cecílio, R.G. Rodriguez, L.G.N. Baena, F.G. Oliveira, F.F. Pruski, A.M. Stephan and J.M.A. Silva
Department of Agricultural Engineering / UFV, Viçosa, Brazil.

Abstract
The objective of this paper was to compare runoff and soil loss amounts observed in an experimental watershed
with those values simulated by RUSLE and WEPP models. The watershed has sandy-clay-loam soil covered with
pasture. Runoff and soil loss measured were very low, due to high infiltration rate and vegetable covering
conditions. In WEPP simulations two values were used for the hydraulic conductivity (K0): a measured in field
(scenario A) and another calculated using WEPP internal procedures (scenario B). The runoff calculated by WEPP
in scenario A was very close to measured value. In scenario B the annual runoff amount was over-predicted.
RUSLE model simulated a very low value of soil loss and, therefore, close to the field observed. WEPP, in scenario
B, also simulated a small soil loss, although greater than that one simulated by RUSLE. In scenario A, WEPP
simulated soil loss equal to zero, that is close to the observed value. Both models satisfactorily estimated soil loss
for the analyzed conditions, although RUSLE performed better. In WEPP, the water loss was simulated better using
measured K0, suggesting that WEPP internal procedures should be carefully used for Brazilian conditions.

Additional Keywords: soil erosion, hydrologic modeling, hydraulic conductivity, runoff, watershed

Introduction
Soil erosion is one of the most important agricultural problems in the world. It is a primary source of sediments that
pollutes streams and fills reservoirs. Erosion also adds to the removal of valuable plant nutrients lost with the
runoff (Schwab et al. 1981).

That soil erosion is a serious problem is widely recognized. What is difficult to assess reliably and precisely,
however, are the dimensions – the extent, magnitude, and rate – of soil erosion and its economic and environmental
consequences (Lal 1994).

Modeling soil erosion is the process of mathematically describing soil particle detachment, transport, and
deposition on land surfaces. There are at least three reasons for modeling erosion: (a) erosion models can be used as
predictive tools for assessing soil loss for conservation planning, project planning, soil erosion inventories, and for
regulation; (b) physically-based mathematical models can predict where and when erosion is occurring, thus
helping the conservation planner target efforts to reduce erosion; (c) models can be used as tools for understanding
erosion processes and their interactions and for setting research priorities (Nearing et al. 1990).

The influence of all the factors that controls soil erosion should be considered in the modeling of the erosive
process. Due to the complexity soil erosion process, several propositions have been elaborated intending to find its
appropriate modeling (Pruski 1996). The Revised Universal Soil Loss Equation (RUSLE) and the Water Erosion
Prediction Project (WEPP) are two of the widely used computer-based models in the world, but few researches has
been done to evaluate its application to tropical climate and soil conditions.

In agreement with the exposed, this paper had as objectives: (a) to evaluate the prediction of water loss by runoff
made by WEPP, comparing its value with the runoff amount measurd in a experimental watershed located in Minas
Gerais State (Brazil); and (b) to evaluate the soil loss predicted by RUSLE and WEPP models, comparing with the
amount measured in the same watershed.

Materials and Methods

Watershed description
The experimental watershed is located in the district of Palmital, near Viçosa city, Minas Gerais (Brazil) and its
area is about 4.0 ha. The watershed´s surface was all covered with pasture (about 100% of surface covering) during
the period soil and water loss data were taken. A gauging station, located on the bank of a drainage ditch, measured
surface water leaving the watershed by runoff. No subsurface water movement occurs in the experimental
watershed. A raingauge station was also located in the watershed’s area.

Paper No. 658 page 1

ISCO 2004 - 13th International Soil Conservation Organisation Conference – Brisbane, July 2004
Conserving Soil and Water for Society: Sharing Solutions

Only four rainfall events provides surface runoff during the hydrologic year of 2001/2002. (about 4 hectares). The
runoff amount measured in that period was 62 m3. The soil loss measured in the period was about 200 kg ha-1. This
soil loss is considered smaller than the average soil loss observed in the Brazilian pasture areas (400 kg ha-1)
(Bertoni and Lombardi Neto 1999).

Physical and chemical analysis accomplished to superficial layers of the watershed´s soil (0 – 0.20 m) are presented
in Tables 1 and 2. The infiltration rate after long time of wetting (Tie) was determinate by the use of the ring
infiltrometers and was equals to 60 mm h-1, a very high value.

Intending optimize the models simulations, the experimental watershed´s area was divided in six different areas
with different representative hillslopes (Figure 1). The simulations were done to each one of the different areas.

Table 1. Textural classification of the watershed’s soil

Depth Coarse Very fine Silt Clay Textural classification
sand sand
0 – 10 cm 40 % 14 % 21 % 25 % Sandy-clay-loam
10 – 20 cm 48 % 15 % 15 % 22 % Sandy-clay-loam

Table 2. Chemical analysis of the watershed’s soil

Depth pH P K CEC Organic matter
(mg/dm3) (mg/dm3) (cmol/dm3) (dag/kg)
0 – 10 cm 5.43 1.4 101 3.74 3.23
10 – 20 cm 6.01 0.7 100 4.42 3.21

Figure 1. Division of the experimental watershed in hillslopes.

Paper No. 658 page 2

ISCO 2004 - 13th International Soil Conservation Organisation Conference – Brisbane, July 2004
Conserving Soil and Water for Society: Sharing Solutions

RUSLE model
The revised universal soil loss equation (RUSLE) expresses the soil loss per unit of area. The equation is
A=R K LSC P (1)

where
A = soil loss per unit of area, t ha-1;
R = rainfall-runoff erosivity factor, MJ ha-1 mm h-1;
K = soil erodibility factor (t ha-1)/(MJ ha-1 mm h-1);
L = slope length factor, unit-plot conditions;
S = slope factor, unit-plot conditions;
C = cover-management factor unit-plot conditions; and
P = conservation practice factor, unit-plot conditions.

The R factor, calculated using precipitation data of Viçosa´s pluviometric station, was equal to 6769.5 MJ mm ha-1
h-1. The value of soil erodibility (K) was taken equal to 0.017 (t ha-1)/(MJ ha-1 mm h-1) (Bertoni and Lombardi Neto
1999). The factors L and S were calculated directly by the RUSLE software, being used the slope lengths and
slopes. The C factor was calcuated considering the original covering of the soil as "bluegrass" (RUSLE database),
with 100% of surface covering during all the year. As no conservation practices were adopted in the area of the
experimental watershed, the value associated to the factor P was equal to one. The factors R, K, C and P didn't vary
in all the six hillslopes of the experimental watershed.

WEPP model
The WEPP watershed computer model requires a minimum of four input data files, including climate, slope,
plant/management, and soil.

The climate input data required by the WEPP model was generated by CLIGEN computer model. Based on long-
term statistics from historical climate data of Viçosa city, the CLIGEN model generated daily values of
precipitation amount, duration, maximum intensity, time to peak intensity, maximum and minimum temperature,
solar radiation, dew point temperature, wind speed and direction for the experimental watershed area.

The slope input file includes physical features such as slope length, slope steepness, and profile aspect. As in the
application of RUSLE model, six different representative hillslopes were considered in the watershed area. The
watershed configuration and watershed channels are presented in Figure 1.

The plant/management file requires land use (agriculture, range, or forest) and information about the specific plants
present and management practices used. In the simulations were used pasture covering (“bluegrass” variety) for all
the watershed’s area.

Soils input files include soil parameters presented in Tables 1 and 2 and Green-Ampt effective hydraulic
conductivity (K0). Two different simulations were conduced considering different K0 values, being kept constant all
the other soil, slope and climate input data. In the first simulation (scenario A), the K0 value was used as the
infiltration rate after long time of wetting (Tie), in agreement with recommendations of Cecílio et al. (2003) for
Brazilian soils. In the second simulation (scenario B), the K0 value was calculated by the software WEPP taking
account to the physical and chemical characteristics of the soil.

Results and Discussion

Table 3 presents the annual soil loss and runoff values obtained by the simulation made with WEPP an RUSLE.

Table 3. Annual soil loss and runoff measured and calculated by simulation made with RUSLE and WEPP
Field- RUSLE WEPP WEPP
measured (Scenario A) (Scenario B)
Runoff 62.0 m3 - 63.4 m3 275.4 m3
Soil loss 200 kg ha-1 140 kg ha-1 0 kg ha-1 300 kg ha-1

In WEPP simulations, scenarios A and B predicted soil loss values close to the field-measured value. Scenario A
simulation predicted soil loss lower than the real value due to the high hydraulic conductivity considered. The soil
Paper No. 658 page 3
ISCO 2004 - 13th International Soil Conservation Organisation Conference – Brisbane, July 2004
Conserving Soil and Water for Society: Sharing Solutions

loss values calculated by this scenario under-predicted the real value, as found by Zeleke (1999) in Ethiopian
Highlands. Scenario B simulation over-predicted soil loss due to a very low mean hydraulic conductivity calculated
by WEPP model (about 12 mm h-1). Scenario B simulation result is in agreement with the trend showed by Zhang
et al. (1996) that found that small annual soil erosion values tended to be over-predicted by WEPP model.

RUSLE predicted annual soil loss equals to 140 kg ha-1, a little lower than the field-measured value. This result is
in agreement with those described by Spaeth et al. (2003) that found a consistent trend of RUSLE to under-predict
soil loss. However, the predicted soil loss was very close to the field observed value, what reveals that RUSLE
predicted well the soil loss in the analysed soil and surface covering conditions. The analysis of these values
showed that RUSLE predicted soil loss better then WEPP model, what was also observed by Tiwari et al. (2000).

In the simulation made to scenario A of WEPP, it was observed that annual runoff volume was well predicted. The
percentile difference between the annual field-observed runoff and the one simulate by WEPP scenario A was only
2.3%, a very small difference. In this simulation, it was observed that the hillsides didn't contribute with runoff. All
the lost water comes from the channels of preferential flow. These channels are only supplied in rain times by water
that infiltrate in the soil. Finally, scenario A simulation over-predicted runoff, as also found by Zeleke (1999).

In the simulation of WEPP´s scenario B the runoff amount was 275.4 m3, what differs of that measured in 344%.
The trend of WEPP model to over-predict runoff on the small events was observed by Risse et al. (1994) and is due
to the low value of hydraulic conductivity calculated by the model.

The annual soil loss and runoff values predicted in scenario B simulation shows that WEPP must be used with
extreme precaution in soil conditions that differs from those where the model equations were developed, what was
also observed by Yu et al. (2000) in a subtropical Australian area and by Zeleke (1999) in Ethiopian Highlands.
The physical and chemical differences between Brazilian and United States´ soils are considerable (for a same soil
textural class), being necessary the determination of the Brazilian soils characteristics for the use of WEPP model.

Conclusions
The obtained results allowed us to conclude that:
a) both RUSLE and WEPP models satisfactorily predicted soil loss for the analyzed conditions, and RUSLE
performance better than WEPP;
b) the WEPP model had better performance in the prediction of annual runoff using the Tie value of the Green-
Ampt hydraulic conductivity;
c) the use of WEPP outside its United States database requires calibration with locally obtained data.

References
Bertoni, L. and Lombardi Neto, F. (1999) Conservação do solo. Editora Ícone, São Paulo.
Cecílio, R.A.; Silva, D.D.; Pruski, F.F. and Martinez, M.A. (2003) Modelagem da infiltração de água no solo sob condições de estratificação
utilizando-se a equação de Green-Ampt. Revista Brasileira de Engenharia Agrícola e Ambiental 7, 415-422.
Lal, R. (1994) Soil erosion: research methods. Ankeny: Soil and Water Conservation Society.
Nearing, M.A.; Lane, L.J.; Alberts, E.E.; and Laflen, J.M. (1990) Prediction technology for soil erosion by water: status and research needs.
Soil Science Society of America Journal 54, 1702-1711.
Pruski, F.F. (1996) Conservação de água e solos. ABEAS, Viçosa, MG.
Risse, L.M.; Nearing, M.A. and Savabi, M.R. (1994) Determining the Green-Ampt effective hydraulic conductivity from rainfall-runoff data
for the WEPP model. Transactions of ASAE 37, 411-418.
Schwab, G.O.; Frevert, R.K.; Edminster, T.W. and Barnes, K.K. (1981) Soil and water conservation engineering. John Wiley & Sons, Inc.
Spaeth, K.E.; Pierson, F.B., Weltz, M.A. and Blackburn, W.H. (2003) Evaluation of USLE and RUSLE estimated soil loss on rangeland.
Journal of Range Management 56: 234-246.
Tiwari, A.K.; Risse, L.M. and Nearing, M.A. (2000) Evaluation of WEPP and its comparison with USLE and RUSLE. Transactions of the
ASAE 43, 1129-1135.
Yu, B.; Ciesiolka, C.A.A.; Rose, C.W. and Coughlan, K.J. (2000) A validation test of WEPP to predict runoff and soil loss from a pineapple
farm on a sandy soil in subtropical Queensland, Australia. Australian Journal of Soil Research 38, 537-554
Zeleke, G. (1999) Application and adaptation of WEPP to the traditional farming systems of the Ethiopian Higlands. In: 10th International
Soil Conservation Organization Meeting, Proceedings…p903-912. International Soil Conservation Organization, Purdue University.
Zhang, X.C.; Nearing, M.A.; Risse, L.M. and McGregor, K.C. (1996) Evaluation of run-off and soil loss predictions using natural run-off
plot data. Transactions of the ASAE 39, 855-863.

Paper No. 658 page 4