Available online at www.sciencedirect.

com

ScienceDirect
Aquatic Procedia 4 (2015) 1291 – 1298

INTERNATIONAL CONFERENCE ON WATER RESOURCES, COASTAL AND OCEAN

ENGINEERING (ICWRCOE 2015)

Simulation of Sediment Yield over Un-gauged Stations Using

MUSLE and Fuzzy Model

P.Sundara Kumara*, T.V.Praveenb, M. Anjaneya Prasadc

a
Department of Civil Engineering, KL University, Vaddeswaram, Guntur Dist., Andhra Pradesh. India – 522 502
b
Department of Civil Engineering, Andhra University, Visakhapatnam, A.P. India – 530003
c
Department of Civil Engineering, K L University, Vaddeswaram, Guntur, A.P India – 522 502

Abstract

Land and water are two most vital natural resources of the world and hence these resources must be conserved carefully to
protect environment to maintain ecological balance. Estimation of runoff and sediment yield is one of the prerequisites for
conservation and management of water resources and also for many hydrological applications. The present study has been taken
up to predict runoff and sediment yield for a reservoir basin viz. Megadrigedda Reservoir situated in Visakhapatnam District,
Andhra Pradesh, India which is having a drainage area of 363 Sq.km. Modified Universal Soil Loss Equation (MUSLE) has been
used to estimate the sediment yield. The runoff factors of MUSLE were computed by the measured values of runoff and peak
rate of runoff at outlet of the reservoir. Topographic factor (LS) and crop management factor(C) were determined using
geographic information system (GIS) and field-based survey of land use/land cover. The conservation practice factor (P) is
obtained from the literature. Sediment yield at the outlet of the study reservoir has been simulated for fifteen storm events spread
over the period of 2010-2013 and is validated with that of measured values. The resulted coefficient of determination value has
been obtained as 0.84 for the study area which indicates that MUSLE model is working satisfactory for the selected basin. Fuzzy
logic based model has also been developed to estimate sediment yield. The resulted correlation obtained between Fuzzy
logic model and MUSLE model is 0.97 and with that of measured value it is 0.937.
© 2015The
© 2015 TheAuthors.
Authors.Published
Publishedby by Elsevier
Elsevier B.V.B.V.
This is an open access article under the CC BY-NC-ND license
Peer-review under responsibility of organizing committee of ICWRCOE 2015.
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of organizing committee of ICWRCOE 2015
Keywords: SCS-CN, Land use/land cover, MUSLE, Fuzzy model, Soil erosion, Sediment yield, Storm event.

*
Corresponding author. Tel.: +91-995933984; fax: 08645-247249.
E-mail address: pskumarphd@gmail.com

2214-241X © 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of organizing committee of ICWRCOE 2015
doi:10.1016/j.aqpro.2015.02.168
1292 P. Sundara Kumar et al. / Aquatic Procedia 4 (2015) 1291 – 1298

1. Introduction

Soil erosion is an important item for consideration in the planning of watershed development works. It reduces
not only the storage capacity of the downstream reservoirs but also deteriorates the productivity of the watershed.
Erosion involves the detachment, transport and deposition of soil particles and aggregates. Sediment yield is
defined as the total amount of eroded material to be delivered from its source to a downstream control point
(Gottschalk, 1964). Thus, sediment yield rates are directly dependent upon both soil loss rates and the transport
phenomenon of surface runoff and channel flow. Accurate estimation of sediment-transport rates, in general,
depends on an accurate prior estimation of overland flows. Thus, any errors in the estimation of overland flows
would be magnified through grossly inaccurate erosion estimations. Sediment yield is a complex phenomenon and
the variables involved in erosion modelling makes it difficult to measure and also to predict the sediment yield in a
precise manner. Among available soil erosion and sediment yield models, the universal soil loss equation (USLE),
the revised version, Revised universal soil loss equation (RUSLE), and its modified version (MUSLE) are widely
used in hydrology and environmental engineering for computing the potential soil erosion and sediment yield. The
USLE (Wischmeier and Smith 1978) was developed for estimation of the annual soil loss from small plots with an
average length of 22 m and its application for individual storm events and large areas leads to large errors. It is
reported that its accuracy increases if it is coupled with a hydrological rainfall excess model. In the USLE model,
there is no direct consideration of runoff, although erosion depends on sediment that is being discharged with flow
and varies with runoff and sediment concentration (Kinnell 2005). It has been observed that delivery ratios to
determine sediment yield from soil loss equation can be predicted accurately with considerable variation. The reason
for variation may be due to the change in rainfall distribution over time from year to year. Williams and Berndt
(1977) proposed MUSLE with the replacement of the rainfall factor with a runoff factor to consider variability of
delivery ratio.
The proposed model was intended to estimate the sediment yield on a single storm basis for the outlet of the
watershed based on runoff characteristics. It is reported that it is the best indicator for sediment yield prediction
(ASCE 1970; Williams 1975a, b; Hrissanthou 2005). MUSLE increases sediment yield prediction accuracy and as
well as it eliminates the need for delivery ratios. The MUSLE equation has been used previously by many
researchers (Tripathi et al. 2001;) and, in some cases, the equation was subjected to different modifications. The
sediment yield model like MUSLE is easier to apply because the output data for this model can be determined at the
watershed outlet (Pandey et al. 2009). Hikaru et al. (2000) demonstrated by successful application of USLE to
mountainous forests in Japan. Tripathi et al. (2001) estimated sediment yield from a small watershed of India using
MUSLE and GIS, and the estimated values were very close to the observed values. Based on the reported
advantages and applicability of method, the present investigation has been taken up to assess the applicability of the
MUSLE for the Meghadrigedda watershed of Visakhapatnam, Andhra Pradesh, India; where there is difficulty in
identifying suitable models for estimation of soil erosion and sediment yield at the watershed. The basin also has
problems of irregular and discontinuous runoff and sediment data availability.
Fuzzy set theory was originally introduced by Zadeh (1965) and later it became popular in control engineering
problems. Many researchers concluded that fuzzy logic is very effective in handling uncertain data and when data
also associated with certain vagueness in nature. Despic and Simminivic (2000) presented a general methodology
for numerical evaluation of complex qualitative criteria based on the theory of fuzzy set. Mimicking the reservoir
operator has been used for modelling the reservoir operation by some researchers. Ross thimothy. (1997) used
Fuzzy Model for real time operation. Anjaneya Prasad and Mohan (2006) adopted fuzzy Logic based model for
Multi-Purpose Reservoir operation.

2. Objectives

To develop a model, to predict the sediment yield with greater reliability in watersheds with particularly when
deficiency of record in sediment data. To validate sediment yield model by comparing predicted values and
observed values. To develop a fuzzy logic based model and to compare the results obtained with MUSLE and also
recorded sediment data.
 P. Sundara Kumar et al. / Aquatic Procedia 4 (2015) 1291 – 1298 1293

3. Study Area

The geographic location of the Meghadrigedda reservoir catchment is located in the north eastern part of
Visakhapatnam district of Andhra Pradesh State and lies between latitudes of 17.43’N-17.57’N and longitudes
83.02’E-83.17’E. The geographical area of the Meghadrigedda reservoir catchment has 363 Sq. Km. The Reservoir
is spread about 6.9 sq.km. Major streams/rivers feeding the reservoir are Meghadrigedda, NarvaGedda and
Borramma gedda. Meghdrigedda is flowing from north-west to south-east direction about 17 km , Borrammagedda
is flowing from west to east for about 7 km and Narvagedda is flowing from south west to north east for about
6.5 km. Physiographic characteristics of the catchment: Catchment area is consists of Hilly area , undulating terrain
and plains. Hill portion is located in the north west and north east and south-west, undulating areas is located at the
foot hill portions and plain areas are located in north and central portions. Major portion of the catchment area is
covered by agriculture land which consists of nearly about 51% of the total area, Hill area is covering about 17%
and water bodies are consists of 11% in the total area presented in the Table 1.

Table 1. Land use/Land cover statistics are presented in the table below

Sno LU/LC Type Area in Sq.KM Percentage

1
Agriculture Land 186.13 51.28
2 Built-up Land 13.32 3.67
3 Hill 61.65 16.98
4 Plantation 30.28 8.34
5 Waste Land 31.32 8.63
6 Water Bodies 40.30 11.10
Total 363.00 100.00

4. Materials and Methods

In the present study, MUSLE equation is used to estimate sediment yield for the Meghadrigedda watershed.
Runoff factor is a major input into the MUSLE model. It is computed using the runoff and peak runoff rates
measured at the outlet of the study area using SCS-CN method. The sediment yields estimated by MUSLE for
different events during the years 2010-2013 are compared with the observed sediment yield data collected from the
stream ungauged station located at the outlet of the watershed. The model performance is evaluated on the basis of
test criteria recommended by the ASCE Task Committee (1993) and graphical performances criteria as suggested by
Haan et al. (1982).

4.1 SCS-CN Method of Estimating Runoff Volume

SCS-CN method developed by Soil Conservation Services (SCS) of USA in 1969 is a simple, predictable and
stable conceptual method for estimation of direct runoff depth based on storm rainfall depth. It relies on only one
parameter, CN. Currently it is a well-established method and widely accepted in USA and many other countries.

Direct Runoff Q
P  0.2S 2 (1)
 P  0. 8 S 
Where, P is the total rainfall, Q is the direct runoff, S the potential maximum retention potential maximum
retention S of watershed is related to a CN, which is a function of land use, land treatments, soil type and antecedent
moisture condition of watershed. The CN is dimensionless and its value varies from 0 to 100. The Q-value in mm
can be obtained from CN by using the relationship of equation (1).

4.2 Estimation of sediment yield using MUSLE

1294 P. Sundara Kumar et al. / Aquatic Procedia 4 (2015) 1291 – 1298

The original USLE is based on soil loss by rainfall energy with slope angle and slope length, which are used as
a proxy for the flow detachment process. The MUSLE model (Williams, 1975) improved USLE model (Wischmeier
and Smith, 1965) by adding a runoff factor to the driving force. There is usually no provision for deposition in this
model. Therefore, MUSLE is a sediment yield prediction model, for estimating sediment yield from a specified land
in a specified cropping and management system. The MUSLE equation (3) is applicable to the point where overland
flow enters the streams and all those point are summed up to give the total amount of sediment delivered to the
stream network within watershed. It computes the sediment yield for a given site, as a product of seven major
variables (William 2005). MUSLE can be adopted for computation of sediment yield from a single storm event.

Sy
11.8
A
 
VQ Q p 0.56 .K .LS .C.P (3)
3 3
Where, Sy is sediment yield in tones :A is Area (ha) : VQ is Runoff volume (m ) : QP is Peak flow rate (m /sec) and
K is the soil erodibility factor, which is the erosion rate per unit of erosion index for specified soil in cultivated
continuous fallow, having 9% slope and 22.13 m. C is the cropping management factor defined as the ratio of soil
loss from a field with a specified cropping and management to that from the fallow condition for which the factor K
is evaluated; and finally, P is the erosion control practice factor, which is the ratio of soil loss with contouring, strip
cropping, or terracing to that with straight row farming, up and down slope. The data about K, LS, C, and P are
adopted from the literature (Arekhi and Kaur 2007). The soil erodibility, crop management, and soil erosion control
practice factors, which are more sensitive to temporal variations than other watershed parameters, were assumed to
be constant as study period is short as advised by Kinnell (2005). The average weighted values of 0.09 tonnes per
hectare, Mj−1 mm−1, 3.36, 0.02, and 1 are thus allotted to the watershed factors of K, LS, C, and P, respectively.
Subsequently, all the parameters are substituted in the MUSLE equation in order to derive event-wise sediment
yields. Furthermore, the MUSLE model is validated by comparing the estimated sediment yields with the observed
sediment yield for forty storm events occurring from 2010 to 2013. The results of application of the MUSLE model
for the storms are shown in Table 3.The Modified Universal Soil Loss Equation (MUSLE) has been widely used to
estimate the sediment yield from a single storm event (Williams and Berndt, 1977). MUSLE method has improved
accuracy of soil erosion prediction over USLE and RUSLE (Williams 1975a, b; Williams and Berndt 1977).

4.3 Sediment Yield Model using Fuzzy Logic Model

Many researchers have adopted fuzzy logic based model for reservoir management problems and also to
address uncertainty and vagueness in variables. In the present study a fuzzy logic based model has been developed
to estimate the sediment yield. In the developed model peak rate of flow, rainfall and runoff volume were considered
as input to model and sediment yield as output from the model. All these variables which are uncertain and also
sometimes vague in nature particularly where measurements are involved were considered as fuzzy variables. The
data from 2010 to 2013 events has been considered to develop during calibration of the model. Membership
functions have been derived based on the input and output variables from the part of data base for calibration of the
model. Various membership forms have been tried for all the input and output variables before arriving to final
membership forms for input and output variables. It was found that combination of Triangular, Trapezoidal and
gumbell form of memberships were found to perform well for peak flood variable and Triangular and Trapezoidal
forms were found to perform well for all other variables. A typical membership form for Peak flow and sediment
yield were shown in Fig 1 and 2.

Fig 1. Membership form of Input Variable “Peak Flow”

 P. Sundara Kumar et al. / Aquatic Procedia 4 (2015) 1291 – 1298 1295

Fig 2. Membership form of Output Variable “Sediment Yield”

Fig 3. Graph showing coefficient of determination

Table 2.0. Results of application of MUSLE model at Meghadrigedda watershed

Predicted
Runoff Peak Observed
Sediment Estimation
Storm date Volume discharge sediment
Yield error %
(m3) (m3/sec) yield (t/ha)
( t/ha)

9/1/2010 2054580 41.322144 9754.1756 7608.257 22.00

9/3/2010 4849914.6 97.5425 25524.834 19909.37 22.00
9/4/2010 1534044.9 30.853033 7032.0084 5484.9665 22.00
9/5/2010 4193348.9 84.337513 21687.479 16916.233 22.00
6/13/2010 1891392.4 38.04008 8890.7044 6756.9353 24.00
9/19/2010 3116208.1 62.673831 15552.577 11819.959 24.00
9/20/2010 4899416 98.538083 25809.3 9643.095 62.64
9/24/2010 1772801.9 35.654965 8268.7562 4630.5035 44.00
9/29/2010 2309451 46.448163 9809.0964 6572.0946 33.00
9/30/2010 2658379.6 53.465888 16840.26 7731.9 54.09
7/3/2011 2064905.7 41.529817 2655.3429 2150.8278 41.22
7/4/2011 3345580.5 67.287017 6455.0471 5228.5882 44.33

Predicted
Runoff Peak Sediment Observed
Estimation
Strom date Volume discharge Yield sediment
error %
(m3) (m3/sec) yield (t/ha)
( t/ha)
1296 P. Sundara Kumar et al. / Aquatic Procedia 4 (2015) 1291 – 1298

8/5/2011 642978.29 12.931714 145.57307 117.91419 0.01

9/6/2011 1421155.7 28.582582 3887.2226 3148.6503 19.00
10/7/2011 48113.63 0.9676714 34339.395 16139.516 53.00
10/8/2011 903609.85 18.173591 27067.432 12721.693 39.65
9/11/2011 6320633.7 127.12191 7725.1948 3630.8416 -50.84
9/12/2011 5110787 102.78922 24164.684 19602 18.88
8/30/2012 1668374.2 33.554693 9704.1773 8801.6889 9.30
8/31/2012 4618493.5 92.88811 5262.9934 4773.535 9.30
9/27/2012 2045174.4 41.132976 7389.9637 6702.697 9.30
9/28/2012 1184336.4 23.819622 15344.014 12470.08 18.73
9/29/2009 1603580.2 32.251541 13189.309 10718.952 18.73
9/17/2012 3078869.6 61.922871 19346.785 15723.132 18.73
9/20/2012 2689771.9 54.097257 11218.365 3557.4 -46.58
9/21/2012 3786823.3 76.161386 23140.532 20132.263 -14.94
9/22/2012 2327846.4 46.818135 26531.223 23082.164 -14.94
9/16/2012 4443320.8 89.365 232.0428 354.78 34.60
9/17/2012 5020292.7 100.96918 266.04312 304.33 12.58
9/19/2012 1594592.9 32.070788 73.638227 87.99 16.31
2/8/2013 1261261 25.366746 56.628762 32.09 -76.47
5/12/2013 8472291.9 170.39651 478.07554 566.08 15.55
6/24/2013 4870843.3 97.963423 257.18887 156.45 -64.39
7/24/2013 1545904.5 31.091556 71.124642 82.78 14.08
7/14/2013 1448370 29.129921 66.118131 44.45 -48.75
7/28/2013 1049070 21.099116 46.071941 51.03 9.72
8/25/2013 1495560 30.079017 68.535535 45.34 -51.16
9/10/2013 1553640 31.247134 71.523366 87.4 18.17
10/23/2013 1738770 34.970507 81.134712 76.9 -5.51
10/25/2013 1364880 27.450753 61.864472 47.34 -30.68
10/27/2013 1339470 26.939702 60.575981 45.78 -32.32
Total 7.67

Fig 4. Graph Showing Sediment Yield from MUSLE model and Fuzzy Logic model”
 P. Sundara Kumar et al. / Aquatic Procedia 4 (2015) 1291 – 1298 1297

Fig 5. Graph Showing Measured Sediment Yield and Fuzzy Logic model

5. Results and Discussions

The discharge and sediment yield data from Meghadrigedda watershed were collected for storm events occurring
from January 2010 to 2013. All the required information for the application of the MUSLE model such as
L=3600m, Tp=13.82 hrs, S=0.0004 and area of the Meghadrigedda basin were extracted with the help of GIS
database. Comparison of predicted and estimated values has been carried out and is reported in table 2.0.The good
coefficient of determination value (0.75) indicates that good relation exists between observed and estimated values
as shown in Fig.3. The percentage deviation of the storm in estimated yield from the observed values and calculated
values varied in the range of 0.01% to 53.0% it has presented in table 2.0. The average value of the estimated error
for the studied storm was estimated 7.67% for the MUSLE model. The average value of estimated error is within
20%. Hence the value can be considered as the acceptable with level of accuracy for the simulation as per the
recommendation of Bingner et al (1989). In other words the model is acceptable for the model process considering
the natural phenomena Das (2000). The fuzzy simulation model developed has been calibrated with 30 events of
the data base and the balance 10 events were used for validating the model. However the results of both calibration
and validation has been given with that of MUSLE model was found to give god correlation between the fuzzy logic
based model and MUSLE model. The correlation obtained between the both model sediment yield was found to
0.97. In order to verify the performance of Fuzzy Logic based model the sediment yield results were also compared
with actual measurement of sediment and it has resulted in correlation of 0.937. Fig 4 and Fig 5 shows the
correlation obtained from Fuzzy Logic based model and measured sediment separately. Hence it can be concluded
that the fuzzy logic based model for sediment yield is performing well and can be adopted as a means to estimate
the sediment yield in basin of similar hydrological in nature.

6. Conclusions
MUSLE model has been successfully used for the estimation of storm-wise sediment yield in the Meghadrigedda
Basin with good coefficient of determination of 0.84 which indicates accurate simulation of sediment yield from the
MUSLE model. The average error value using MUSLE when compared with measured sediment yield was
estimated to be 7.67%. However, the present results can also be used in erosion-based watershed prioritization in the
study area. To regionalize the results of the study area, greater numbers of storms events as well as case studies are
needed. Hence researchers should consider this aspect. In addition, other simple soil erosion and sediment yield
models must be considered with reasonably accurate estimation of system response at the watershed scales, when
scarce information exists. The fuzzy Logic based model developed is performing well for sediment computation
from the inputs peak flow, rainfall and volume of runoff. The resulted correlation with MUSLE model is 0.97 and
with that of measured value it is 0.937.
1298 P. Sundara Kumar et al. / Aquatic Procedia 4 (2015) 1291 – 1298

References

ASCE (American Society of Civil Engineers)., 1970. Sediment sources and sediment yields. Journal of the Hydraulics Division ASCE 96
(HY6):1283–1329.
ASCE Task Committee. 1993. Criteria for evaluation of watershed models. Journal of Irrigation and Drainage Engineering-ASCE 119(3):429–42
Bingner RL, Murphee CE, Mutchler CK., 1989. Comparison of sediment yield models on various watersheds in Mississipi. Trans ASAE
32(2):529–534.
Das, G., 2000. Hydrology and soil conservation engineering. Prentice Hall, India.
Durbovin et al. fozzy model real time reservoir operation., 2002. Journal of water resources planning and management, ASCE 128(1), p 66-73.
Despic., Simminivic., 2000. Aggregation operator for soft decision making in water resources engineering , fuzzy set and systems Elsevier 115,
p11-33
FAO/UNEP (1994) Land degradation in south Asia: its severity causes and effects upon the people. FAO, UNEP and UNEP project, Rome
Haan CT, Johnson HP, Brakensiek DL. 1982. Hydrological modelling of small watersheds. American Society of Agricultural Engineers,
Michigan.
Gottschalk, L.C., 1964. Reservoir Sedimentation. Chapter17-1. In: V.T. Chow, éd. Handbook of Applied Hydrology, McGraw-Hill Book Co.,
New York, NY. USA.
Hikaru, K., Yoichi, O., Toshiaki, S., Kawanami, A., 2000. Application of Universal Soil Loss Equation (USLE) to mountainous forests in Japan.
J For Res 5(4):156–162.
Hrissanthou, V., 2005. Estimate of sediment yield in a basin without sediment data. Catena 64:333–347.
Pandey, A., Chowdary,V,M., Mal, B,C., 2009. Sediment yield modelling of an agricultural watershed using MUSLE, remote sensing and GIS. J
Paddy Water Environ (Springer) 7(2):105–113.
Kirpich, Z. P., 1940. Time of concentration of small agricultural watersheds. Civil Engineering 10 (6), 362. The original source for the Kirpich
equation.
Ross Timothy,J., 1997. Fuzzy logic engineering applications McGraw-Hill's international edition
Tripathi, M,P., Panda, R,K., Pradhan, S., Das, R, K., 2001. Estimation of Sediment yield form a small watershed using MUSLE and GIS. Journal
of the Institute of Engineering I 82:40-45.
UNEP., 1997. World Atlas of desertification, 2nd edition. Arnold, London, p 77.
Walling, D,E., Webb, B,W., 1982. Sediment availability and the prediction of storm-period sediment yields. In: Proceedings of the Exeter
Symposium, 327–337. IAHS Publ. 137. IAHS Press, Wallingford, UK
Williams, JR., 1975a. Sediment routing for agricultural watersheds. Water Resources Bulletin 11:965–975.
Williams, JR., 1975b. Sediment yield prediction with Universal Equation using runoff energy factor. In: Present and prospective Technology for
predicting sediment yields and sources, 244–252, Agricultural Research Service, US Department of Agriculture
Williams, JR, Berndt HD., 1977. Sediment yield prediction based on watershed hydrology. Trans American society of Agriculture Engineers
20(6):1100–1104.
Wischmeier, WH, Smith, DD., 1978. Predicting rainfall erosion losses. USDA Agricultural Research Services Handbook 537. USDA,
Washington, p 57.
M. Anjaneya Prasad., S. Mohan.., 2007. Soft Computing Models using Fuzzy Logic for a Multipurpose Reservoir System Proceedings of
“International Workshop on Integrated Water Resource Management (IWRM- 2007), Organized by Karnataka Environment Research
Foundation, Bangalore, India
Zadeh, L, A.,1965.Fuzzy sets. Information and Control 8 (3): 338.