Water Resour Manage (2010) 24:2091–2112

DOI 10.1007/s11269-009-9540-0

Estimation of Sediment Yield and Areas of Soil

Erosion and Deposition for Watershed Prioritization
using GIS and Remote Sensing

Manoj Kumar Jain · Debjyoti Das

Received: 2 March 2009 / Accepted: 23 November 2009 /

Published online: 2 December 2009
© Springer Science+Business Media B.V. 2009

Abstract A Geographical Information System (GIS) based method is proposed

and demonstrated for the identification of sediment source and sink areas and the
prediction of sediment yield from watersheds. Data from the Haharo sub-catchment
having an area of 565 km2 in the Upper Damodar Valley in Jharkhand State
in India was taken up for the present study due to availability of gauged data
at multiple locations within watershed area. The watershed was discretized into
hydrologically homogeneous grid cells to capture the watershed heterogeneity. The
cells thus formed were then differentiated into cells of overland flow regions and
cells of channel flow regions based on the magnitude of their flow accumulation
areas. The gross soil erosion in each cell was calculated using the Universal Soil
Loss Equation (USLE). The parameters of the USLE were evaluated using digital
elevation model, soil and landuse information on cell basis. The concept of transport
limited sediment delivery (TLSD) was formulated and used in ArcGIS for generating
the transport capacity maps. An empirical relation is proposed and demonstrated
for its usefulness for computation of land vegetation dependent transport capacity
factor used in TLSD approach by linking it with normalized difference vegetation
index (NDVI) derived from satellite data. Using these maps, the gross soil erosion
was routed to the watershed outlet using hydrological drainage paths, for derivation
of transport capacity limited sediment outflow maps. These maps depict the amount
of sediment rate from a particular grid in spatial domain and the pixel value of the
outlet grid indicates the sediment yield at the outlet of the watershed. Up on testing,
the proposed method simulated the annual sediment yield with less than ±40% error.

M. K. Jain (B)
Department of Hydrology, Indian Institute of Technology, Roorkee 247 667 Uttarakhand, India
e-mail: jain.mkj@gmail.com

D. Das
Department of Civil Engineering, Purulia Polytechnic, Purulia, West Bengal 723147, India
2092 M.K. Jain, D. Das

Keywords Soil erosion · GIS · Remote sensing · Sediment yield ·

Transport capacity · Transport limited accumulation · NDVI

1 Introduction

Soil erosion by water is one of the most important land degradation problem and
a critical environmental hazard of modern times worldwide (Eswaran et al. 2001).
Accelerated erosion due to human-induced environmental alterations at global scale
is causing extravagant increase of geomorphic process activity and sediment fluxes
in many parts of the world (Turner et al. 1990; IGBP-BAHC 1997). It causes loss
of fertile top soil cover, delivers millions of tons of sediments into reservoirs and
lakes, resulting in strong environmental impact and high economic costs by its ef-
fect on agricultural production, infrastructure and water quality (Lal 1998; Pimentel
et al. 1995). Not surprisingly soil erosion and sediment delivery have become
important topics on the agenda of local and national policy makers. This has led to an
increasing demand for watershed or regional-scale soil erosion models to delineate
target zones in which conservation measures are likely to be the most effective.
The process of soil erosion involves detachment, transport and subsequent depo-
sition (Meyer and Wischmeier 1969). Sediment is detached from soil surface both by
the raindrop impact and the shearing force of flowing water. The detached sediment
is transported down slope primarily by flowing water, although there is a small
amount of downslope transport by raindrop splash also (Walling 1988). Once runoff
starts over the surface areas and in the streams, the quantity and size of material
transported increases with the velocity of runoff water. At some point downslope,
slope may decrease, resulting in a decreased velocity and hence decreased transport
capacity (Haan et al. 1994; Lal 2001). The sediment is then deposited, starting
with the large primary particles and aggregates. Smaller particles are transported
further downslope, resulting in what is known as enrichment of fines. The amount
of sediment load passing the outlet of a watershed is known as the sediment yield.
Urbanization, agriculture expansion and deforestation predominantly change the
landuse due to which soil erosion takes place. The variables such as climate, soil
type, land cover, topography and anthropogenic activities influence soil erosion and
sediment delivery in watershed (Jain and Kothyari 2000; Lee 2004; Jain et al. 2005).
Detailed assessment of erosion from watershed areas is greatly dependent on their
spatial, economic, environmental, and cultural context (Warren 2002). The informa-
tion on sources of sediment yield within a watershed can be used as perspective on
the rate of soil erosion occurring within that watershed (Jain and Kothyari 2000).
Despite the development of a range of physically based soil erosion and sediment
transport equations, sediment yield predictions at a watershed or regional scale
are at present achieved mainly through simple empirical models as the detailed
data required for application of physically based models are not available at this
scale. Simple empirical models are still widely used for soil erosion and sediment
yield predictions for their simplicity, which makes them applicable even if only
a limited amount of input data is available. Simple methods such as Universal
Soil Loss Equation (USLE; Wischmeier and Smith 1978), Modified Universal Soil
Loss Equation (MUSLE; Williams 1975) or Revised Universal Soil Loss Equation
(RUSLE; Renard et al. 1991), are quite frequently used for estimation of gross
Estimation of Sediment Yield using GIS and Remote Sensing 2093

amount of surface erosion in watershed areas (e.g. Williams and Berndt 1972; Griffin
et al. 1988; Ferro et al. 1998; Jain and Kothyari 2000; Kothyari et al. 2002; Onyando
et al. 2005; Pandey et al. 2007). As stated earlier, not all the soil eroded at a place gets
transported to its downstream position as sediment outflow. The amount of sediment
transported downstream depends on the magnitude of the sediment generation and
transporting capacity of flowing medium. The concepts of sediment delivery ratio
(SDR) (e.g. Onstad 1984; Walling 1988; Ferro et al. 1998; Jain and Kothyari 2000)
and transport limiting sediment delivery (TLSD) were used extensively to model the
process of sediment deposition and transport and, in turn, the outflow (Carson and
Kirkby 1972; Beasley et al. 1980; Desmet and Govers 1995; Morgan et al. 1998; Jain
et al. 2005; Van Rompaey et al. 2005; Verstraeten et al. 2007). It is pertinent to note
that the concepts of SDR and TLSD have some limitations (Polyakov and Nearing
2003; Sander et al. 2007); nowadays the concept of TLSD to model deposition and
outflow of sediment is preferred to those using SDR.
Both of these quantities viz. the surface erosion and sediment yield are found
to have large spatial variability in a watershed due to the spatial variation of
rainfall and watershed heterogeneity. To account for spatial variability, watersheds
have been subdivided either into cells of a regular grid or into units of sub-areas
having approximately homogeneous characteristics and rainfall distribution (Julien
and Gonzales del Tanago 1991; Wilson and Gallant 1996; Jain and Kothyari 2000;
Wu et al. 2005; Jain et al. 2004). The technique of Geographical Information System
(GIS) is well suited for quantification of heterogeneity in the topographic and
drainage features of a watershed (Jain and Kothyari 2000; Jain et al. 2004, 2005).
Therefore, the objective of this study is to use GIS for descritization of the watershed
into small grid size areas, for computation of such physical characteristics of these
areas as slope, land use and soil type which affect the process of soil erosion and
deposition in different sub areas of a watershed. The GIS is also used for partitioning
the watershed grids into overland and channel grids, computation of soil erosion in
individual grids, and determination of watershed sediment yield by using the concept
of transport limiting sediment delivery.

2 The Study Area

The study area (Haharo sub-catchment) is a head water watershed in Damodar-

Barakar river system in the Upper Damodar Valley (UDV) area, Jharkhand, India.
Downstream of the study watershed, the Damodar-Barakar river system has a
network of five reservoirs, namely Maithon and Tilaiya on the Barakar River,
Panchet, Tenughat and Konar on river Damodar. Among these, Panchet is the
largest dam constructed across the Damodar River. Geographically the Haharo sub
watershed is located between 85◦ 00 E to 85◦ 24 E longitude and 23◦ 45 N to 23◦ 49
N latitude. The total area of the sub-catchment up to the gauging point on the
main stream is 565 km2 . Figure 1 depicts Drainage map of Haharo watershed with
gauging stations and sub watersheds boundaries. Physiographically the watershed
falls into three main landscapes, viz. the hilly and undulating area in the north-
west, gently undulating and rolling uplands, and valley lands. Most of the area of the
watershed comprises of gently sloping uplands except for hilly and undulating areas.
Elevation of the watershed ranges from 300–830 m above Mean Sea Level. Most of
2094 M.K. Jain, D. Das

Fig. 1 Drainage map of Haharo watershed with gauging stations and sub watersheds boundaries

the uplands are subjected to sheet erosion. The climate of the study watershed is
sub-humid tropical. The annual rainfall ranges between 1,160 and 1,400 mm. About
63% of annual rainfall is concentrated in the three monsoon months of July, August
and September (Rajan and Gopala Rao 1978). The predominant land use in the
watershed is agricultural and forest. The Haharo sub-catchment is characterized by
overgrazing, degraded forest cover and undulating topography coupled with erratic
and intense rainfall. The surface soil texture in the study watershed is mainly loamy
type and particle size classes are fine loamy type. Depth of soil varies from shallow to
very deep and having parent material as granite-Gneiss and sand stone (NBSS&LUP
1996).
In Haharo sub-catchment rainfall, run-off and sediment data were available at
(1) Sirma (area = 169.91 km2 ) (1) Barkagaon (area = 131.87 km2 ) (2) Bisrampur
(area = 124.58 km2 ) and (3) Simratari (area = 18.15 km2 ). These observations were
made by Damodar Valley Corporation (DVC) since 1979. At all four locations,
rainfall was observed with ordinary rain gauge, runoff was measured with a stage
level recorder and sediment samples were collected with the help of USDH-48 depth
integrating sampler. The sediment samples were filtered at the gauging site itself
after allowing it to settle for some time. The total volume of runoff multiplied by
the sediment concentration produce the total sediment volume. The bulk density
of sediment was taken as 1.4 g/cc as most of the sediment comprise of fine loamy
Estimation of Sediment Yield using GIS and Remote Sensing 2095

particles (Samuel 1995). In order to accommodate the bed load, 20% of suspended
silt load was added to the measured suspended load as suggested by Varshney (1977).

3 Methodology

In a larger sized watershed, part of the soil eroded in upland areas gets deposited
within the watershed before reaching the outlet. Such watersheds are, therefore,
divided into smaller sub-areas to account for spatial heterogeneity (Jain and Kothyari
2000). The grid or cell approach of sub-division has been used extensively for
discretization of watershed into homogenous sub-areas (Hadley et al. 1985; Beasley
et al. 1980; Wicks and Bathurst 1996; Kothyari et al. 2002; Jain et al. 2005; Pandey
et al. 2007). The grid or cell approach is quite adaptive to collection of input data
on a regular pattern with the use of remote sensing and GIS and it accounts for the
variation in topographic characteristics over a watershed in detail (Jain and Kothyari
2000; Jain et al. 2004, 2005). Therefore, a grid-based procedure for discretization of
the watershed is adopted in the present study. Grid thus formed can be categorized as
having lying on overland areas and those lying in channel areas. Such a differentiation
is necessary because the process of soil erosion and delivery in them are widely
different. In the present study, such a differentiation is achieved using the concept
of channel initiation threshold (ESRI 1994). The methods such as the USLE have
been found to produce realistic estimates of surface erosion over small size ar-
eas (Wischmeier and Smith 1978; Ferro et al. 1998; Jain and Kothyari 2000; Kothyari
et al. 2002; Jain and Goel 2002; Lee 2004; Onyando et al. 2005; Pandey et al. 2007).
The USLE is therefore adopted for estimation of gross erosion rates in the different
discretized cells of the watershed. The eroded sediment is routed from each cell
to the watershed outlet using the concept of TLSD described later. The USLE for
estimation of soil erosion within a grid (or cell) is expressed as

SEi = RKi LSi Ci Pi (1)

where SEi = gross amount of soil erosion in cell i (MT ha−1 year−1 ); R = rainfall
erosivity factor (MJ mm ha−1 h−1 year−1 ); Ki = soil erodibility factor in cell i
(MT ha h ha−1 MJ−1 mm−1 ); LSi = slope steepness and length factor for cell i (di-
mensionless); Ci = cover management factor i (dimensionless) and Pi = supporting
practice factor for cell i (dimensionless).
Estimation of gross soil erosion from grid sized area of the watershed requires
estimates of various factors appearing in Eq. 1. Wischmeier and Smith (1958)
and Wischmeier (1959) after evaluation of correlations between soil erosion and a
number of rainfall parameters, defined the R-factor as the product of total storm
energy and maximum 30-min intensity divided by 100 for numerical convenience,
known as the EI30 index. On an annual basis, the R factor is the sum of values
of EI30 values of the storms in an individual year. Realistic estimation of monthly
or annual rainfall erosivity EI30 values requires long term pluviographic data at
15 min intervals or less (Wischmeier and Smith 1978). In many parts of the world,
especially in developing countries, spatial coverage of pluviographic data is often
difficult to obtain (Yu et al. 2001; Cohen et al. 2005; Shamshad et al. 2008). Monthly,
seasonal and annual rainfall data are usually available for longer periods and are
generally used to calculate R-factor. Rainfall erosivity estimation using rainfall data
2096 M.K. Jain, D. Das

with long time intervals have been attempted by several researchers for different
regions (e.g. Morgan 1995; Millward and Mersey 1999; Mati et al. 2000; Grimm
et al. 2003; Natalia 2005; Shamshad et al. 2008). Using the data for storms from
several rain gauge stations located in different zones, linear relationships were
established between average annual rainfall and computed EI30 values for different
zones of India and iso-erodent maps were drawn for annual and seasonal EI30 values
(Ram Babu et al. 2004). Following equation was developed for Damodar Valley area
in Jharkhand India by Ram Babu et al. (2004) and used in the present study

R = 81.5 + 0.38R N (340 ≤ R N ≤ 3500 mm) (2)

where R N is the average annual rainfall in mm. For the present study, Eq. 2 is used
to compute annual values of R-factor by replacing R N with actual observed annual
rainfall in a year.
There are many relationships available for estimation of the LS factor (e.g.
Wischmeier and Smith 1978; Moore and Burch 1986a, b; McCool et al. 1987, 1989;
Moore and Wilson 1992; Desmet and Govers 1996). Among these, the one that is
best suited for integration with the GIS is the theoretical relationship proposed by
Moore and Burch (1986a, b) and Moore and Wilson (1992) based on unit stream
power theory, given as
   
Asi n sin αi m
LSi = (3)
22.13 0.0896
where Asi is the specific area at cell i (=Aup/ wn ), defined as the up slope contributing
area for overland grid (Aup ) per unit width normal to flow direction (wn ); αi is the
slope gradient in degrees for cell i. It has been shown that the values of n = 0.6, m =
1.3 give results consistent with the RUSLE LS factor for slope lengths <100 m and
slope angles <14 degrees (Moore and Wilson 1992). Exponents n and m can also be
obtained through calibrated if data are available for a specific prevailing type of flow
and soil conditions. In the present study value of n = 0.6 and m = 1.3 is used.
The LS factor for a cell area is computed with Eq. 3 in ArcGIS using upslope
contributing area and slope gradient computed from DEM of the study watershed.
For generation of DEM, topographic information on drainage and contours of the
study area was prepared using the Survey of India topographic maps at 1:50,000
scale. Digitized contours were interpolated at 30-m pixel resolution using ‘topo to
raster’ command in ArcGIS to create hydrologically correct DEM. The generated
DEM was further conditioned to maintain continuity of flow to watershed outlet
(Jenson and Domingue 1988). Location of the outlet of the watershed was marked
on DEM and area of the watershed was delineated using eight direction pour point
algorithm (ESRI 1994). The DEM is then further analyzed to distinguish overland
and channel grids. The concept of channel initiation threshold is used for defining
cells as channel cells. For channel grid areas, the value of Aup is considered to be
equal to the value of the threshold area corresponding to the channel initiation (Jain
and Kothyari 2000). A grid is considered to lie in the overland region if the size of the
area from which it receives the flow contribution is smaller or equal to the specified
threshold area for initiation of a channel. While grids receiving flow contribution
from area of more than threshold value are considered to form the channel grids
(ESRI 1994). The grids with no flow accumulation lie on the watershed boundary.
Different values of channel initiation threshold would result in stream networks with
Estimation of Sediment Yield using GIS and Remote Sensing 2097

different total stream lengths, and consequently with different drainage densities
(Wang and Yin 1998). Jain and Kothyari (2000) considered equivalence of the total
stream length generated using a given threshold and total stream length estimated
from 1:50,000 scale topographic maps (digitized in vector from) for choosing value of
channel initiation threshold. For the present study, the value of channel initiation
threshold is chosen such that total stream length generated using threshold and
channel network seen in satellite data and Google earth image (digitized in vector
form) are equivalent. Accordingly, a channel initiation threshold value of 0.45 km2
is adjudged appropriate to define channel cells. The use of Eq. 3 in the estimation
of the LS-factor allows the introduction of the three-dimensional hydrological and
topographic effect of converging and diverging terrain on soil erosion (Panuska et al.
1991).
The values for the factors K, C and P are computed for different grids on overland
and channel region as per Wischmeier and Smith (1978), Singh et al. (1981, 1992),
Haan et al. (1994) using the information about landuse, soil texture and management
practices adopted in the watershed. To this end, soil map of the watershed was
digitized from soil survey report prepared by National Bureau of Soil & Landuse
Planning (NBSS&LUP 1996). Figure 2 shows the digitized soil map of the Haharo
sub-catchment. Details such as fraction of sand, silt, clay and organic matter and
other related parameter information for different mapping units were taken from

Fig. 2 Soil map and assigned soil erodibility factor (K) for Haharo watershed
2098 M.K. Jain, D. Das

Table 1 Soil type statistics and values of soil erodibility factor (K) for Haharo watershed
Type of soil % of watershed area K factor
Sirma Barkagaon Bisarampur Simratari
Loamy 2.58 3.98 10.41 14.11 0.0527
Silt loam 82.61 96.02 89.59 85.89 0.0395
Silty clay loam 14.81 − − − 0.0382

NBSS&LUP (1996) for Haharo sub-catchment. K-values for mapped soil categories
were then calculated for each of the mapping units using the Wischmeier and Smith
(1978), Singh et al. (1981, 1992), Haan et al. (1994) procedure, and these are given in
Table 1.
The C factor layer was prepared using the LANDSAT TM satellite data corre-
sponding to November 1st, 1992 (path 140 to 141 and Row 43 to 44) downloaded
from GLCF (NASA) site. The image was geometrically corrected and analysed
using ERDAS Imagine image processing software (ERDAS 2005). To discriminate
the vegetation from other surface cover types, the Tassel Cap and Vegetation
Index (VI) transformations were performed in ERDAS. Then a stratified supervised
classification using Maximum Likelihood Method was carried out to generate the
land use/land cover map for the watershed. The classified image was further verified
for locations and extensions of various land cover classes using limited ground

Fig. 3 Land use land cover map of the Haharo watershed

Estimation of Sediment Yield using GIS and Remote Sensing 2099

Table 2 Landuse statistics and values of cover management factor (C) of Haharo watershed
Land use % of watershed area Cover management
Sirma Barkagaon Bisarampur Simratari factor (C)
Agriculture 27.05 24.88 33.33 48.57 0.40
Water body 1.36 0.88 1.13 2.89 1.00
Waste land 25.07 14.08 3.65 4.10 0.65
Settlement 1.88 2.26 4.68 9.17 0.80
Degraded forest 44.65 57.89 57.21 35.27 0.03

truth information, Google earth image and Survey of India topographic maps.
Finally a landuse map was generated with 5 classes Viz. degraded forest, agriculture,
settlement, water body and wasteland. Generated landuse map of the watershed
is given in Fig. 3. Based on the land cover categories, the attribute values for the
C factor were assigned to individual cells from the tabulated values suggested by
Wischmeier and Smith (1978), Singh et al. (1981, 1992), Haan et al. (1994). Table 2
summarizes the land cover statistics and C-factor used for Haharo sub-catchment.
Since in the study area, no major conservation practices are followed except low
height bunds in few pockets of the agricultural lands, the P factor was taken equal to
1 for all land use and land cover categories as majority of land areas have not been
provided with any conservation support.

3.1 Sediment Transport and Outflow

The eroded sediment from each grid follows a defined drainage path from a particu-
lar cell to the outlet of watershed as shown in Fig. 4. The rate of sediment transport
from each of the discretized cell depends upon the transport capacity of the flowing
water (Meyer and Wischmeier 1969). The sediment outflow from an area is equal

Fig. 4 Schematic showing

discretized grid cells in a
watershed
2100 M.K. Jain, D. Das

to soil erosion in the cell plus contribution from upstream cells if transport capacity
is greater than this sum. However if transport capacity is less than the sum of soil
erosion in the cell and contribution from upstream cells, then the amount of sediment
exceeding the transport capacity gets deposited in the cell and sediment load equal
to transport capacity is discharged to next downstream cell (Meyer and Wischmeier
1969).

3.2 Mean Annual Sediment Transport Capacity

The rate of transport of the sediment is governed by the transporting capacity of the
flowing water. Most geomorphologic models assume that overland flow is transport
limited accumulation and sediment flux Qs is mainly predicted by the following
equation (Desmet and Govers 1995)
Qs = K L p Sq (4)
where K is soil erodibility, L is the upslope distance (m), S is the local slope gradient
(m m−1 ) and p, q are constants. For three-dimensional landscapes (Kirkby and
Chorley 1967; Carson and Kirkby 1972), this equation becomes
Qs = K S p Aqs (5)
Prosser and Rustomji (2000) made a review on the constants p and q, and found that
the median value obtained in experimental studies is 1.44 for both constants. This
concept was further studied by Verstraeten et al. (2007) and based on their hypoth-
esis the following equation for mean sediment transport capacity was proposed and
the same is adopted in this study.
TCi = KTCi R Ki A1.44
si Si1.44 (6)

where TCi is the transport capacity in cell i (kg/m2 /year); KTCi is the transport
capacity coefficient in cell i and reflects vegetation component within the transport
capacity; and Si is the slope gradient of cell i.
Since KTCi is strongly depended on land use land cover types, it is in order to
co-relate it with vegetative index value of the area to get the spatial distribution of
transport capacity coefficient. In the present study, the transport capacity coefficient
is linked empirically to normalized difference vegetation index (NDVI) derived
from satellite data. The NDVI of an area depends on presence of vegetation in
the area. It is well known that the reflectance of an area depend on land cover
present and vegetation reflects most in near infrared region (NIR). Using this
property, many indices have been developed in past and can give a perspective
on presence of vegetation in a cell. These indices have also been used to estimate
factors which depend on land cover present. For example, VanderKnijff et al. (2002)
assessed monthly cover management factor values for Italy using Advanced Very
High Resolution Radiometer imagery (AVHRR) by relating Normalized Difference
Vegetation Index (NDVI) with cover management factors (C). Dwivedi et al. (1997)
derived the erosion risk by classifying satellite image according to the percentage of
bare soil. Mathieu et al. (1997) and Haboudane et al. (2002) improved this scheme by
estimating vegetation cover from Landsat TM and SPOT-4 data and combining them
with slope generated from a digital elevation model (DEM) to produce erosion rate
maps. A comparison of methods for producing maps of vegetation related variables
Estimation of Sediment Yield using GIS and Remote Sensing 2101

for soil erosion studies using coarse and medium satellite imagery were applied by
Leprieur et al. (2000) and Symeonakis and Drake (2004), who founded that NDVI
was the most useful.
The NDVI is the image transformation based on the normalized difference
between Near-Infra Red (NIR) and Visible Red (VR) (Rouse et al. 1973; Sabins
1997), and expressed as
 
NIR − VR band4 − band3
N DV I = = (7)
NIR + VR band4 + band3
The possible range of values for NDVI is between −1 and 1, but the typical range
is between about −0.1 (NIR less than VIS for a not very green area) to 0.6 (for a
very green area) (Kidwell 1990). Typically higher +ve value for a pixel in NDVI
image indicate more vigorous or dense vegetation and vice versa. In the proposed
new relation, KTC for a cell sized area is hypothesized as an exponential function of
NDVI. Mathematically KTCi is expressed as
 
−N DV Ii
KTCi = β ∗ exp (8)
1 − N DV Ii
Where NDVIi is the NDVI value for cell i and β is a scaling factor which can be
determined through calibration. Functional behavior of proposed Eq. 8 is studied for
possible range of NDVI values and plot of computed KTC with NDVI is depicted in
Fig. 5 as illustration. Figure 5 also depicts effect of scaling coefficient β on computed
values of KTC . As can be seen from the figure, the computed values of KTC increases
as value of NDVI changes from +1 to −1 indicating increase in value of KTC as
vegetation density or vigor reduces. Also notice the scaling effect of parameter β to
cater wide range of situations.

3.3 Transport Limited Accumulation

Sediment is routed along the runoff paths towards the river taking into account
the local transport capacity, TC of each pixel. If the local TC is smaller than the
sediment flux, then sediment deposition occurs. This approach assumes that sediment
transport is not necessarily restricted to a transport limited system. If TC is higher
than the sediment flux, then sediment transport will be supply limited. Thus, by

Fig. 5 Proposed functional

relation of KTC with NDVI
2102 M.K. Jain, D. Das

introducing transport capacity coefficient (KTC ), a more realistic representation

of overland flow sediment transport can be simulated. Because much sediment is
routed to these locations from steeper hill slopes adjoining the thalwegs, heavy
sedimentation occurs because of lower transport potential. The predicted sediment
delivery values need to be interpreted as sediment delivery towards the complete
length of the river in the watershed. For grid based discretization system transport
limited accumulation can be computed as:
  
Touti = min SEi + Tini , TCi (9)

Di = SEi + Tini − Touti (10)

where SEi = annual gross soil erosion in cell i, TCi = transport capacity of cell i,
Tini = sediment inflow in cell i from upstream cells, Touti = sediment outflow from
the cell i, and Di = deposition in cell i.
Use of Eqs. 1, 6, 9 and 10 produces different maps of gross erosion, sediment
transport capacity, sediment outflow and sediment deposition, whereby a distinction
is made between gross erosion, net erosion, and net sediment deposition. Finally,
priority areas are identified by grouping the net erosion estimated on cell basis
into the following scales of priority: Slight (0 to 5 t ha−1 year−1 ), Moderate (5 to
10 t ha−1 year−1 ), High (10 to 20 t ha−1 year−1 ), Very High (20 to 40 t ha−1 year−1 ),

Fig. 6 Map depicting variation of KLSCP values

Estimation of Sediment Yield using GIS and Remote Sensing 2103

Severe (40 to 80 t ha−1 year−1 ) and Very Severe (> 80 t ha−1 year−1 ) erosion classes
following the guidelines suggested by Singh et al. (1992) for Indian conditions. Such
maps are of immense use for identification of critical soil erosion and deposition areas
in the watershed.

4 Analysis and Discussion of Results

4.1 Generation of the Erosion Potential Map

The land use, soil, slope steepness and management parameters are the main factors
governing soil erosion potential at particular location to the erosive power of rainfall.
Maps for values of the USLE parameters viz. K, LS, C and P factors were integrated
in ArcGIS using raster calculator to form composite map of terms KLSCP. The map
of composite term KLSCP represents soil erosion potential of different grid cells.
High values of this term indicate a higher potential of soil erosion in the cell and
vice versa. Figure 6 shows the areas of varying KLSCP values and hence the soil
erosion potential in the different cells of the watershed of Haharo sub-catchment.
The information shown in Fig. 6 could be utilized for identification of the sediment
source areas of the watersheds.

Table 3 Comparison between observed and predicted values of sediment yield

Name of Year Annual Annual Observed Predicted % Error
gauging rainfall rainfall sediment sediment
station (mm) erosivity (R) yield (t) yield (t)
Sirma 1982 649 328.12 17,950.1 17,137.1 4.53
1983 849 404.12 17,129.0 21,106.7 −23.22
1984 1120 507.10 33,570.5 26,486.8 21.10
1985 723 356.24 19,247.4 18,607.1 3.33
1986 1321 583.48 39,633.2 30,475.0 23.11
Barkagaon 1980 983 455.04 37,111.3 24,115.8 35.02
1981 773 375.24 24,151.8 19,886.7 17.66
1982 566 296.58 15,474.4 15,718.0 −1.57
1983 770 374.10 17,785.3 19,826.4 −11.48
1984 1186 532.18 27,391.7 28,203.7 −2.96
Bisarampur 1979 699 347.12 26,487.4 19,919.0 24.80
1980 983 455.04 39,009.4 26,112.0 33.06
1981 769 373.72 13,416.5 21,446.6 −59.85
1982 567 296.96 20,876.9 17,041.6 18.37
1983 849 404.12 20,836.2 23,188.5 −11.29
1984 1097 498.36 26,731.3 28,598.0 −6.98
Simratari 1979 699 347.12 3,739.2 3,470.0 7.20
1980 983 455.04 5,403.2 4,549.0 15.81
1981 841 401.08 4,572.6 4,008.6 12.33
1982 579 301.52 4,718.2 3,014.1 36.12
1983 849 404.12 4,641.1 4,039.8 12.96
1984 1098 498.74 5,132.1 4,984.2 2.88
1985 733 360.04 3,690.6 3,604.2 2.34
1986 1211 541.68 2,626.0 5,416.0 −106.25
2104 M.K. Jain, D. Das

4.2 Estimation of Gross Soil Erosion

Maps depicting gross amount of soil erosion from different descritized cells of the
Haharo sub-catchment were computed by multiplication of the erosion potential
map produced by integration of KLSCP maps with annual values of rainfall erosivity
factor R (Eq. 2). Computed values of annual R factor are presented in Table 3 from
1979 to 1986 for different sub-watersheds. Figure 7 illustrates the gross soil erosion
for year 1982. Such maps indicate the gross amount of soil erosion from each cell.
It is worth mentioning here that all erosion produced in a cell does not necessarily
move from the area and only a part of it finds its way to watershed outlet depending
on the transport capacity of the flowing water.

4.3 Computation of Spatially Distributed Transport Capacity

As described earlier, all erosion produced in a grid cell does not get transported to the
next downstream cell or outlet. The amount of eroded soil particles finding its way
from upstream cells to downstream cells and finally to the watershed outlet depends
on the transporting capacity of the flowing water. An annual value of spatially
distributed transport capacity for all cell areas was computed using Eq. 6. Here, mean
annual transport capacity varies from year to year as a function of annual rainfall.
The transport capacity coefficient (KTC ; Eq. 6) which reflects vegetation component

Fig. 7 Map depicting gross soil erosion for year 1982

Estimation of Sediment Yield using GIS and Remote Sensing 2105

within the transport capacity was determined through proposed Eq. 8. To this end,
Normalized Difference Vegetation Index (NDVI) layer was prepared in ERDAS
imagine software from the Landsat satellite image after doing geometric correction
and re-projection to same coordinate system as of other  maps.
 Using the NDVI layer,
the spatially distributed transport capacity coefficient KTCi is computed using Eq. 8
whereby the effect of spatial variation of vegetation on sediment transport dynamics
is accounted. An upslope contributing area per unit length of contour (Aup ) and slope
maps were prepared in ArcGIS. Transport capacity of overland flow was calculated
for each year from the relationship stated in Eq. 6 by multiplying the R factor of each
year in ArcGIS. The value of scaling factor β appearing in Eq. 8 was determined
through calibration by minimizing error between observed and computed sediment
yield at Sirma gauging site only. Data at other three gauging sites were used to
test the performance of proposed procedure. Calibrated value of the scaling factor
β equal to 1 (i.e. no scaling, see Fig. 5) is found to produce good match between
observed and computed sediment yield at Simra gauging site. Figure 8 shows the
transport capacity map for the year 1982 as illustration. As can be seen from this
and other such figures not shown here, the areas showing higher transport capacity
coincide with steep head water areas and channel areas in the watershed. Smaller
values of transport capacity are mainly found to be associated with the overland
regions that surround the confluence of the main stream with just lower order streams
and flatter land areas found in the cultivated valley lands in the watershed.

Fig. 8 Map depicting transport capacity for different grids for the year 1982
2106 M.K. Jain, D. Das

4.4 Computation of Transport Limited Sediment Accumulation and Outflow

The gross erosion from each grid was routed downstream (Eq. 9) following hy-
drological drainage paths using DEM to generate map of sediment deposition and
accumulated sediment yield limited by transport capacity (sediment yield). This
process was repeated for all years of data used in the analysis. Such maps give
amount of sediment transported from the system at every grid and are useful for
determination of sediment flowing out of the watershed at any location. Transport
limited sediment outflow and sediment deposition maps were prepared for all years
used in analysis. Figure 9 depicts the sediment yield map for the year 1982 as
illustration. The pixel value of the sediment outflow map denotes the amount of
sediment leaving the current cell to the next downstream cell. The pixel value of
the cell at the watershed outlet denotes the sediment coming out of the watershed.
Comparison of predicted sediment yield with the observed sediment yield for all
year from 1979 to 1986 for each watershed is shown in Table 3. It is seen from
Table 3 that the percentage of error in the estimated sediment yields from the
observed values varies in the range of (−) 23.22% (over prediction) to (+) 36.12%
(under prediction) excluding data point for the year 1981 of watershed no 4/3 and
for the year 1986 of watershed no 4/4 showing larger variations (Table 3) and are
ascribed to probable uncertainties in observations. Other source of uncertainty could
be attributed to large spatial and temporal variability in rainfall, dynamic nature of

Fig. 9 Sediment outflow from each grid for the year 1982
Estimation of Sediment Yield using GIS and Remote Sensing 2107

vegetation which influence greatly transport capacity and crop management factor
for overland regions and need to be further investigated in future studies. As seen
from Table 3, for most of the years, the error between observed and computed
sediment yield lie within ±40% indicating reasonably accuracy in computation of
sediment yield (ASCE 1975; Foster 1982; Hadley et al. 1985; Wu et al. 1993; Wicks
and Bathurst 1996).

4.5 Identification of Priority Areas (Sediment Source and Sinks)

Maps for deposition of sediment for different years were derived using Eq. 10,
helpful in identification of areas vulnerable to silt deposition in the watershed. Also
net erosion maps for different years were calculated by subtracting the deposition
rates for each grid cell from the gross erosion rates for corresponding grid cell.
Negative values on the net erosion map are the areas where sediment deposition
occurs (i.e. true sediment deposition), whereas positive values correspond to grid
cells with net sediment erosion. Finally, the net erosion estimated on a cell basis
for the watershed was grouped into the following scales of priority: Slight (0 to
5 t ha−1 year−1 ), Moderate (5 to10 t ha−1 year−1 ), High (10 to 20 t ha−1 year−1 ),Very
High (20 to 40 t ha−1 year−1 ), Severe (40 to 80 t ha−1 year−1 ) and Very Severe
(> 80 t ha−1 year−1 ) erosion classes as per the guidelines suggested by Singh et al.
(1992) for Indian conditions. Figure 10 shows the net erosion/deposition map for the

Fig. 10 Map depicting soil erosion/sediment deposition areas for year 1982
2108 M.K. Jain, D. Das

Table 4 Areas under different classes of soil erosion in Haharo watershed for different years
Net erosion/deposition % area of watershed
(t ha−1 year−1 ) 1979 1980 1981 1982 1983 1984 1985 1986
Deposition 0.83 0.83 0.83 0.83 0.83 0.83 0.83 0.83
0–5 85.04 80.1 83.27 86.90 82.73 77.77 82.89 75.06
5–10 7.93 10.55 8.88 6.79 9.19 11.65 9.22 12.82
10–20 3.16 4.50 3.61 2.82 3.74 5.28 3.62 6.22
20–40 1.78 2.19 1.94 1.62 1.99 2.38 1.95 2.64
40–80 0.92 1.22 1.04 0.80 1.08 1.36 1.03 1.52
> 80 0.33 0.59 0.42 0.24 0.45 0.73 0.44 0.90

year 1982 as illustration, helpful in identification of areas vulnerable to soil erosion

and deposition. As can be seen from this and other such figures (not shown), most
of the deposition of sediment resulted by the sides of some of the stream reaches
in valleys and also on flatter land areas found in the cultivated valley lands in the
watershed at the grids where transport capacity was low. Also it is seen on these
figures that agricultural and waste lands on steep slopes mostly drained by first order
streams falls under high net erosion zone i.e. more than 5 t ha−1 yr−1 , indicating that
they have undergone severe erosion due to high slope, undulating topography and
possibly faulty method of cultivation practices. Statistics about percentage of area in
Hararo watershed falling in different erosion classes identified above for different
years is given in Table 4. As can be seen from Table 4, area under slight erosion class
is ranging from 75% to 87% in different years. Area covered by moderate, high, very
high, severe and very severe erosion zones taken together vary from 12% to 25% in
different years and can be termed as critical erosion prone areas requiring immediate
attention from soil conservation point of view. Depending upon priority levels, the
watershed area should be treated with suitable vegetative and structural measures.
Such maps are extremely important in planning conservation measures; the areas
producing more sediment receive priority for their implementation.

5 Summary and Conclusions

Scientific management of soil and water is very important to arrest erosion and
enhancing the agricultural production. Soil erosion is the major cause of the loss
of fertility, diminishing crop production and land degradation. Deterioration of soil
in the study area can be controlled effectively by adopting watershed treatment
measures if spatial distribution of soil erosion is known. Various thematic layers rep-
resenting different factor of USLE were generated and overlaid to compute spatially
distributed gross soil erosion maps for the Haharo watershed (India). The concept of
transport limited accumulation was used in ArcGIS for generating maps for transport
capacity, gross soil erosion was routed to the watershed outlet using hydrological
drainage paths resulting in generation of transport capacity limited sediment outflow
maps. Such maps provide the amount of sediment flowing from a particular grid
in spatial domain. A comparison of the observed and computed sediment yield
reveals the proposed method to compute sediment yield with reasonable accuracy.
Further, maps for deposition of sediment were also generated for identification of
areas vulnerable to silt deposition in the watershed. The deposition of sediment was
Estimation of Sediment Yield using GIS and Remote Sensing 2109

found to occur at grids where transport capacity was low, mostly lying by the sides of
some of the stream reaches. Superimposition of sediment deposition map over gross
erosion map led to areas vulnerable to soil erosion and deposition. Such maps are
important in planning conservation and control measures. The specific conclusions
are given as below:

1. Proposed empirical relation for computation of land vegetation dependent

transport capacity factor can be used to compute spatially distributed value of
parameter KTC whereby the effect of spatial variation of vegetation on sediment
transport dynamics is accounted.
2. Areas showing higher transport capacity coincide with steep head water areas
and channel areas in the watershed and smaller transport capacity values are
mainly associated with the overland regions that surround the confluence of the
main stream with just smaller order streams and flatter land areas found in the
cultivated valley lands in the watershed.
3. The proposed method produces satisfactory estimates of sediment outflow from
watershed with ±40% deviation from observations.
4. Spatially computed soil removal from most of the watershed area is limited to
0–5 t ha−1 year−1 except to few pockets which produce more sediment yield.
5. Deposition of sediment resulted at grids where transport capacity was low,
mostly lying by the sides of some of the stream reaches in valleys and at places
where steep slope converges into flatter slopes.

Acknowledgement The authors wish to sincerely thank anonymous reviewers whose comments
greatly improved the quality of this paper.

References

American Society of Civil Engineers (ASCE) (1975) Sedimentation engineering. American Society
of Civil Engineering, New York, p 745
Beasley DB, Huggins LF, Monke EJ (1980) ANSWERS—a model for watershed planning. Trans
ASAE 23:938–944
Carson MA, Kirkby MJ (1972) Hillslope form and process. Cambridge University Press, Cambridge,
p 475
Cohen MJ, Shepherd KD, Walsh MG (2005) Empirical formulation of the universal soil loss equation
for erosion risk assessment in a tropical watershed. Geoderma 124:235–252
Desmet PJ, Govers G (1995) GIS-based simulation of erosion and deposition patterns in agricultural
landscape: a comparison of model results with soil map information. Catena 25:389–401
Desmet PJ, Govers G (1996) A GIS-procedure for the automated calculation of the USLE LS-factor
on topographically complex landscape units. J Soil Water Conserv 51:427–433
Dwivedi RS, Kumar AB, Tewari KN (1997) The utility of multi-sensor data for mapping eroded
lands. Int J Remote Sens 18(9):2303–2318
ERDAS (Earth Resources Data Analysis System) (2005) ERDAS field guide. Geospatial Imaging
LLC, Norcross, Georgia, USA
ESRI (Environmental Systems Research Institute) (1994) Cell based modeling with GRID. Envi-
ronmental Systems Research Institute Inc., Redlands, CA, USA
Eswaran H, Lal R, Reich PF (2001) Land degradation: an overview. In: Bridges EM, Hannam
ID, Oldeman LR, Penning de Vries FWT, Scherr SJ, Sombatpanit S (eds) Response to land
degradation. Science, Enfield, pp 20–35
Ferro V, Porto P, Tusa G (1998) Testing a distributed approach for modelling sediment delivery.
Hydrol Sci J 43(2):425–442
2110 M.K. Jain, D. Das

Foster GR (1982) Modelling the erosion processes. In: Haan CT, Johnson H, Brakensiek DL (eds)
Hydrological modelling of small watersheds. ASAE Monograph No. 5. American Society of
Agricultural Engineers, St. Joseph, pp 297–380
Griffin ML, Beasley DB, Fletcher JJ, Foster GR (1988) Estimating soil loss on topographically non-
uniform field and farm units. J Soil Water Conserv 43:326–331
Grimm M, Jones RJA, Rusco E, Montanarella L (2003) Soil erosion risk in Italy: a revised USLE
approach. European Soil Bureau Research Report no. 11, EUR 19022 EN
Haan CT, Barfield BJ, Hayes JC (1994) Design hydrology and sedimentology for small catchments.
Academic, New York
Haboudane D, Bonn F, Royer A, Sommer S, Mehl W (2002) Land degradation and erosion risk
mapping by fusion of spectrally based information and digital geomorphometric attributes. Int J
Remote Sens 18:3795–3820
Hadley RF, Lal R, Onstad CA, Walling DE, Yair A (1985) Recent developments in erosion and
sediment yield studies. UNESCO (IHP), Paris, p 127
IGBP-BAHC (International Geosphere–Biosphere Programme) (1997) Modeling the transport and
transformation of terrestrial materials to freshwater and coastal ecosystems. IGBP Report,
vol. 39. IGBP, Stockholm, Sweden
Jain SK, Goel MK (2002) Assessing the vulnerability to soil erosion of the Ukai Dam catchments
using remote sensing and GIS. Hydrol Sci J 47(1):31–40
Jain MK, Kothyari UC (2000) Estimation of soil erosion and sediment yield using GIS. Hydrol Sci J
45(4):771–786
Jain MK, Kothyari UC, Ranga Raju KG (2004) A GIS based distributed rainfall runoff model.
J Hydrol, Elsevier Science 299(1–2):107–135
Jain MK, Kothyari UC, Rangaraju KG (2005) Geographic information system based distributed
model for soil erosion and rate of sediment outflow from catchments. J Hydraul Eng ASCE
131(7):755–769
Jenson SK, Domingue JO (1988) Extracting topographic structure from digital elevation data for
geographic system analysis. Photogramm Eng Remote Sensing 54(9):1593–1600
Julien PY, Gonzales del Tanago M (1991) Spatially varied soil erosion under different climates.
Hydrol Sci J 36(5):511–524
Kidwell KB (1990) Global Vegetation Index user’s guide. US Department of Commerce/National
Oceanic and Atmospheric Administration/National Environmental Satellite Data and Informa-
tion Service/National Climatic Data Center/Satellite Data Services Division, USA
Kirkby PC, Chorley RJ (1967) Through flow, overland flow and erosion. Bull Int J Hydrol Sci 12:5–21
Kothyari UC, Jain MK, Ranga Raju KG (2002) Estimation of temporal variation of sediment yield
using GIS. Hydrol Sci 47(4):693–705
Lal R (1998) Soil erosion impact on agronomic productivity and environment quality. Crit Rev Plant
Sci 17(3):319–464
Lal R (2001) Soil degradation by erosion. Land Degrad Dev 12(5):519–539
Lee S (2004) Soil erosion assessment and its verification using the universal soil loss equation and
geographic information system: a case study at Boun, Korea. Environ Geol 45(3):457–465
Leprieur C, Kerry YH, Mastorchio S, Meunier JC (2000) Monitoring vegetation cover across
semi-arid regions: comparison of remote observations from various scales. Int J Remote Sens
21(1):281–300
Mathieu R, King C, Bissonnais Y (1997) Contribution of multitemporal SPOT data to the mapping
of a soil erosion index. The case of the loamy plateaux of northern France. Soil Technol 10(1):
99–110
Mati BM, Morgan RPC, Gichuki FN, Quinton JN, Brewer TR, Liniger HP (2000) Assessment of
erosion hazard with the USLE and GIS: a case study of the Upper Ewaso Ng’iro North basin of
Kenya. JAG 2(1):78–86
McCool DK, Brown GR, Foster GR, Mutchler CK, Meyer LD (1987) Revised slope steepness factor
for the universal soil loss equation. Trans ASAE 30:1387–1396
McCool DK, Foster GR, Mutchler CK, Meyer LD (1989) Revised slope length factor for the
universal soil loss equation. Trans ASAE 32:1571–1576
Meyer LD, Wischmeier WH (1969) Mathematical simulation of the processes of soil erosion by
water. Trans Am Soc Agric Eng 12(5):754–758
Millward A, Mersey JE (1999) Adapting the RUSLE to model soil erosion in a mountainous tropical
watershed. Catena 38:109–129
Moore I, Wilson JP (1992) Length slope factor for the revised universal soil loss equation: simplified
method of solution. J Soil Water Conserv 47(4):423–428
Estimation of Sediment Yield using GIS and Remote Sensing 2111

Moore I, Burch G (1986a) Physical basis of the length-slope factor in the universal soil loss equation.
Soil Sci Soc Am J 50:1294–1298
Moore I, Burch G (1986b) Modeling erosion and deposition: topographic effects. Trans ASAE
29(5):1624–1630, 1640
Morgan RPC, Quinton JN, Smith RE, Govers G, Poesen JWA, Auerswald K, Chisci G, Torri
D, Styczen ME (1998) The European Soil Erosion Model (EUROSEM): a dynamic approach
for predicting sediment transport from fields and small catchments. Earth Surf Process Landf
23:527–544
Morgan RPC (1995) Soil erosion and conservation, 2nd edn. Longman, Essex
Natalia H (2005) Spatial modeling of soil erosion potential in a tropical watershed of the Colombian
Andes. Catena 63(1):85–108
NBSS&LUP (1996) Soils of Bihar for optimizing land use. National Bureau of Soil Survey and Land
Use Planning, ICAR, Publication No. 50b, Nagpur, India, ISBN:81-85460-44-2
Onstad CA (1984) Modelling of sediment yields. In: Hadley RF, Walling DE (eds) Ero-
sion and sediment yields: some methods of measurement and modelling. Norwich,
Geobooks
Onyando JO, Kisoyan P, Chemelil MC (2005) Estimation of potential soil erosion for river Perkerra
catchment in Kenya. Water Resour Manag 19:133–143
Pandey A, Chowdary VM, Mal BC (2007) Identification of critical erosion prone areas in the small
agricultural watershed using USLE, GIS and remote sensing. Water Resour Manag, Springer
21:729–746. doi:10.1007/s11269-006-9061-z
Panuska JC, Moore ID, Kramer LA (1991) Terrain analysis: integration into the agricultural non-
point source (AGNPS) pollution model. J Soil Water Cons 46(1):59–64
Pimentel D, Harvey C, Resosudarmo P, Sinclair K, Kurz D, McNair M, Crist S, Shpritz L, Fitton L,
Saffouri R, Blair R (1995) Environmental and economic costs of soil erosion and conservation
benefits. Science 267:1117–1123
Polyakov VO, Nearing MA (2003) Sediment transport in rill flow under deposition and detachment
conditions. Catena 51:33–43
Prosser IP, Rustomji P (2000) Sediment transport capacity relations for overland flow. Prog Phys
Geogr 24(1):179–193
Rajan HG, Gopala Rao (1978) Studies on soils in India. Vikas, New Delhi
Ram Babu, Dhyani BL, Kumar N (2004) Assessment of erodibility status and refined Iso- Erodent
Map of India. Indian J Soil Conserv 32(2):171–177
Renard KG, Foster GR, Weesies GA, Porter JP (1991) RUSLE, revised universal soil loss equation.
J Soil Water Conserv 46(1):30–33
Rouse JW, Haas RH, Schell JA, Deering DW (1973) Monitoring vegetation systems in the Great
Plains with ERTS. In: Third ERTS symposium, NASA SP-351 I, pp 309–317
Sabins FS (1997) Remote Sensing: principles and interpretations, 3rd edn. Freeman, New York
Sander GC, Parlange JY, Barry DA, Parlange MB, Hogarth WL (2007) Limitation of the
transport capacity approach in sediment transport modeling. Water Resour Res 43:W02403.
doi:10.1029/2006WR005177
Samuel JC (1995) Sediment criteria for prioritizing watershed for resource development pro-
grammes. PhD thesis, University of Roorkee, Roorkee, India
Shamshad A, Azhari MN, Isa MH, Wan Hussin WMA, Parida BP (2008) Development of an
appropriate procedure for estimation of RUSLE EI30 index and preparation of erosivity maps
for Pulau Penang in Peninsular Malaysia. Catena 72:423–432
Singh G, Chandra S, Babu R (1981) Soil loss and prediction research in India, Central Soil and Water
Conservation Research Training Institute, Bulletin No. T-12/D9
Singh G, Babu R, Narain P, Bhusan LS, Abrol IP (1992) Soil erosion rates in India. J Soil Water
Conserv 47(1):97–99
Symeonakis E, Drake N (2004) Monitoring desertification and land degradation over sub-Saharan
Africa. Int J Remote Sens 25(2):573–592
Turner BL, Clark WC, Kates RW, Richards JF, Matthews JT, Meyer WB (1990) The earth as
transformed by human action. Cambridge Univ. Press, Cambridge
VanderKnijff J, Jones RJA, Montanarella L (2002) Soil erosion risk assessment in Italy. In: Rubio
JL, Morgan RPC, Asins S, Andreu V (eds) Proceedings of the third international congress man
and soil at the third millennium. Geoforma Ediciones, Logrono, pp 1903–1913
Van Rompaey A, Bazzoffi P, Jones RJA, Montanarella L (2005) Modelling sediment yields in Italian
catchments. Geomorphology 65:157–169
Varshney RS (1977) Engineering hydrology. Nemchand and Brothers, Roorkee, p 639
2112 M.K. Jain, D. Das

Verstraeten G, Prosser IP, Fogarty P (2007) Predicting the spatial patterns of hillslope sediment
delivery to river channels in the Murrumbidgee catchment. Aust J Hydrol 334:440–454
Walling DE (1988) Erosion and sediment yield research—some recent perspectives. J Hydrol
100:113–141
Wang X, Yin ZY (1998) A comparison of drainage networks derived from digital elevation models
at two scales. J Hydrol 210:221–241
Warren A (2002) Land degradation is contextual. Land Degrad Dev 13(5):449–459
Wicks JM, Bathurst JC (1996) SHESED: a physically based, distributed erosion and sediment yield
component for the SHE hydrological modelling system. J Hydrol 175:213–238
Williams JR (1975) Sediment routing for agricultural watersheds. Water Resour Bull 11:965–974
Williams JR, Berndt HD (1972) Sediment yield computed with universal equation. J Hydrol Div
ASCE 98(12):2087–2098
Wilson JP, Gallant JC (1996) EROS: a grid-based program for estimating spatially distributed
erosion indices. Comp Geosci 22(7):707–712
Wischmeier WH (1959) A rainfall erosion index for a universal soil loss equation. Proc Soil Sci Soc
Am 23:246–249
Wischmeier WH, Smith DD (1958) Rainfall energy and its relationship to soil loss. Trans AGU
39(2):285–291
Wischmeier WH, Smith DD (1978) Predicting rainfall erosion losses. Agricultural Handbook no.
537, US Dept of Agriculture, Science and Education Administration
Wu TH, Hall JA, Bonta JV (1993) Evaluation of runoff and erosion models. J Irrig Drain Eng ASCE
119(3):364–382
Wu S, Li J, Huang G (2005) An evaluation of grid size uncertainty in empirical soil loss modelling
with digital elevation models. Environ Model Assess 10:33–42
Yu B, Hashim GM, Eusof Z (2001) Estimating R-factor with limited rainfall data: a case study from
Peninsular Malaysia. J Soil Water Conserv 56(1):101–105