Eldho-Geetha-Wrm-Pub-Feb08 - Continuos CN

See
discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/227314115
SCS-CN-based Continuous Simulation Model

for Hydrologic Forecasting
Article in Water Resources Management · February 2007

DOI: 10.1007/s11269-006-9149-5
CITATIONS READS
39 343
5 authors, including:
Surendra Mishra T.I. Eldho

Indian Institute of Technology Roorkee Indian Institute of Technology Bombay
118 PUBLICATIONS 2,263 CITATIONS 167 PUBLICATIONS 1,226 CITATIONS
SEE PROFILE SEE PROFILE
A. K. Rastogi Rajendra P. Pandey

Indian Institute of Technology Bombay National Institute of Hydrology
44 PUBLICATIONS 238 CITATIONS 81 PUBLICATIONS 657 CITATIONS
SEE PROFILE SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Quantitative Assessment of vulnerability to drought View project
Hydrological droughts View project
All content following this page was uploaded by T.I. Eldho on 09 July 2014.
The user has requested enhancement of the downloaded file.

Water Resour Manage (2008) 22:165–190
DOI 10.1007/s11269-006-9149-5
SCS-CN-based Continuous Simulation Model

for Hydrologic Forecasting
K. Geetha & S. K. Mishra & T. I. Eldho &

A. K. Rastogi & R. P. Pandey
Received: 19 March 2006 / Accepted: 22 December 2006 /

Published online: 1 February 2007
# Springer Science + Business Media B.V. 2007
Abstract A new lumped conceptual model based on the Soil Conservation Service Curve
Number (SCS-CN) concept has been proposed in this paper for long-term hydrologic
simulation and it has been tested using the data of five catchments from different climatic
and geographic settings of India. When compared with the Mishra et al. (2005) model
based on variable source area (VSA) concept, the proposed model performed better in all
applications. Both the models however exhibited a better match between the simulated and
observed runoff in high runoff producing watersheds than did in low runoff producing
catchments. Using the results of the proposed model, dominant/dormant processes involved
in watershed’s runoff generating mechanism have also been identified. The presented
model is found useful in the continuous simulation of rainfall–runoff process in watersheds.
Keywords hydrological forecasting . long-term hydrologic simulation . saturation excess

overland flow . streamflow . rainfall–runoff . variable source area . curve number .
antecedent moisture
Nomenclature
P total rainfall
Ia initial abstraction
Pe effective rainfall
Ft infiltration at any time ‘t’ in SCS-CN-based model
K. Geetha : T. I. Eldho (*) : A. K. Rastogi

Department of Civil Engineering, Indian Institute of Technology,
Bombay, Powai, Mumbai 400 076, India
e-mail: eldho@civil.iitb.ac.in
S. K. Mishra
Indian Institute of Technology, Roorkee, 247 667 Uttaranchal, India
R. P. Pandey
National Institute of Hydrology, Roorkee, 247 667 Uttaranchal, India
166 K. Geetha, et al.
ROt rainfall excess (surface runoff) at any time ‘t’ for 5-days from beginning of
storm
DRt subsoil drainage at any time ‘t’
PRt percolation at any time ‘t’
DSPt deep seepage at any time ‘t’ in SCS-CN-based model
DPRt deep percolation at any time ‘t’
SROt surface runoff at any time ‘t’, if t>5-days
THRt throughflow at any time ‘t’ in SCS-CN-based model
BFt base flow at any time ‘t’ in SCS-CN-based model
TROt total runoff at any time ‘t’ in SCS-CN-based model
EVPt potential evaporation at any time ‘t’
EVt evaporation at any time ‘t’
TRt transpiration at any time ‘t’
ETt actual evapotranspiration at any time ‘t’ in SCS-CN-based model
SMS soil moisture store
GWS ground water store
SCS-CN Soil Conservation Service-Curve Number
1 Introduction
The simulation of rainfall-generated runoff is very important in various activities of water

resources development and management such as flood control and its management,
irrigation scheduling, design of irrigation and drainage works, design of hydraulic
structures, and hydro-power generation etc. Ironically, determining a robust relationship
between rainfall and runoff for a watershed has been one of the most important problems
for hydrologists, engineers, and agriculturists since its first documentation by P. Perrault
(In: Mishra and Singh 2003) about 325 years ago. The process of transformation of rainfall
to runoff is highly complex, dynamic, non-linear, and exhibits temporal and spatial
variability, further affected by many and often interrelated physical factors. However an
understanding of various hydrologic variations (spatial and temporal) over long periods is
necessary for identification of these complex and heterogeneous watershed characteristics.
There exists a multitude of watershed models of varying complexity in the hydrologic
literature (Singh 1989, 1995; Singh et al. 2006; Limbrunner et al. 2006; www.usbr.gov/
pmts/rivers/html/index.html). Since the development of Stanford watershed model (Craw-
ford and Linsley 1966), numerous operational, lumped, conceptual models have been
developed. Among them the Boughton model (Boughton 1966, 1968), Kentucky watershed
model (Liou 1970; James 1970, 1972), Institute of Hydrology model (Nash and Sutcliffe
1970; Mandeville et al. 1970), Tank model (Sugawara et al. 1984), HYDROLOG (Poter
and McMahon 1976), MODHYDROLOG (Chiew and McMahon 1994), HRUT (Yao et al.
1996) are worth citing.
The variable source area (VSA) concept coupled with concepts of topographic index and
curve number (described below) has been quite useful in adaptable runoff estimation (Choi
et al. 2002). Beven and Kirkby (1979) suggested that a conceptual model that can simulate
the variable source areas could be used for long-term water yield estimation which
incorporates soil moisture replenishment, depletion, and redistribution for the dynamic
variation of areas during and after the storm contributing to direct runoff. The TOPMODEL
(Beven and Kirkby 1979; Beven et al. 1984, 1995) which is a continuous hydrologic model
does however not account for the spatial variability of land use and soil characteristics.
SCS-CN-based continuous simulation model for hydrologic forecasting 167
The other widely used simple technique is the Soil Conservation Service Curve Number
(SCS-CN) method and is now known as Natural Resource Conservation (NRCS) (USDA
1986). It was first developed by the United States Department of Agriculture (USDA) in
1954 to transform rainfall to direct surface runoff. Ponce and Hawkins (1996) critically
examined the SCS-CN method and indicated several advantages over other methods. Since
1996, a significant amount of literature is published and several recent articles have
reviewed the methodology in detail, and the work still continues, as for example, the work
of Michel et al. (2005). Recent advances in computer environments and availability of
Geographic Information System (GIS) tools have made a drastic change in the hydrological
modeling and development with the application of SCS-CN technique (Nageshwar et al.
1992; Arnold et al. 1993; Heaney et al. 2001). Though the model has its own limitations
(Ponce and Hawkins 1996; Choi et al. 2002; Mishra and Singh 2003; Michel et al. 2005), it
has been widely used in numerous models, including AGNPS (Heaney et al. 2001) for
single events; long term hydrological impact assessment model, L-THIA (Harbor 1994;
Bhaduri 1998); SWAT model (Spruill et al. 2000); CELTHYM (Choi et al. 2002) etc.
Besides the application of the SCS-CN model in long-term simulation (Huber et al. 1976;
Williams and LaSeur 1976; Knisel 1980; Soni and Mishra 1985; Woodward and Gbuerek
1992; Mishra et al. 1998; Mishra and Singh 1999, 2004a; Mishra 2000), this method has
also been employed for determination of infiltration and runoff rates (Mishra 1998; Mishra
and Singh 2002a,b, 2004b).
The original SCS-CN method was designed to predict direct (surface) runoff, as an
infiltration-excess overland flow model. It assumes that only one process is responsible for
producing streamflow over the entire watershed area. The SCS-CN supports Horton’s overland
flow mechanism to compute the surface runoff, in contrast to the VSA-Based model that
assumes saturation-excess overland flow mechanism from dynamic source areas. In VSA-
Based model, the catchment is assumed to be partitioned as source areas and non-source areas
in which infiltration occurs through non-source areas.
The original SCS-CN method computes the direct runoff by considering only the available
rainfall on the current day without taking into account of the effect of the moisture available
prior to the storm. On the other hand, the curve numbers are sensitive to antecedent conditions
(Ponce 1989). The original method does not contain any expression for time and ignores the
impact of rainfall intensity and its temporal distribution. There is no explicit provision for
spatial scale effects also. The other demerits of the original method are the absence of clear
guidance of how to vary the antecedent moisture conditions and the fixing of initial
abstraction ratio 0.2, pre-empting a regionalization based on geological and climatic settings
(Choi et al. 2002). Besides, since the conventional method was developed for agricultural
sites, it works best on these sites, fairly on range sites, and poorly on forest sites. Here in this
paper, an attempt has been made to improve the existing SCS-CN model by eliminating the
demerits of the existing model.
With the improved understanding available on both the variable source area and SCS-CN
concepts, it is essential to apply these models to long-term (daily) data from different climatic
and geographic settings and compare their performance. Thus, the main objectives of this paper
are to (a) conceptualize and develop a lumped model based on the SCS-CN technique modified
for accounting of the antecedent moisture effect and (b) compare this model performance with
another lumped conceptual model based on variable source area (VSA) theory (Mishra et al.
2005) on the data of five Indian watersheds of Cauvery in Karnataka, Narmada in Madhya
Pradesh, and Ulhas in Maharashtra. The modified SCS-CN-based lumped model considers
various hydrologic components involved in the runoff generation mechanisms and takes into
account of temporal variations of curve number.
The proposed model differs from the original model, as in daily flow simulation, the
original SCS-CN method is used to compute the direct surface runoff considering the
rainfall of the current day utilizing the CN-values corresponding to antecedent 5-day AMC,
allowing unrealistic sudden quantum jumps in CN-variation. Secondly, the value of initial
abstraction coefficient is fixed as 0.2, which has shown to be varying in literature (For
example, Mishra and Singh 2003). The proposed long term hydrologic model obviates
these limitations and is capable of simulating, other than direct surface runoff, the total
streamflow and its components such as surface runoff, throughflow, and base flow which is
conceptualized to have two different moisture stores, i.e. soil moisture store and ground
water store. This continuous simulation model considers a daily time step interval for
analysis. Thus, the present version is a significant enhancement over the previous ones
utilizing original SCS-CN method.
This long term hydrologic model is capable of simulating streamflow and its
components such as surface runoff, throughflow, and base flow and is also conceptualized
to have two different moisture stores, i.e. soil moisture store and ground water store.
2 Existing SCS-CN Model
The original SCS-CN method was documented in Section 4 of the National Engineering
Handbook (NEH) in 1956. The document has since been revised subsequently in 1964,
1965, 1971, 1972, 1985, 1993 (In: Mishra and Singh 2003), and 2004 (SCS 2004). The
method which is derived to compute the surface runoff from rainfall in small agricultural
watersheds is based on water balance equation and the two hypotheses, respectively, as
follows (Mishra and Singh 1999, 2003):
P ¼ Ia þ F þ Q ð1Þ
Q F
¼ ð2Þ
P Ia S
Ia ¼ lS ð3Þ
where P=total precipitation; Ia =initial abstraction; F=cumulative infiltration; Q=direct

runoff; S=potential maximum retention or infiltration; l=initial abstraction coefficient. But
l varies in the range of 0 to ∝ and is assumed as a standard value of 0.2 in usual practical
applications (Mishra and Singh 1999, 2004b).
2.1 Computation of Surface Runoff
Combination of Eq. 1, 2, and 3 leads to the following popular form of the SCS-CN method
to compute daily direct (surface) runoff ROt Eq. 4 with time t as subscript as
Pe2t
ROt ¼ ð4Þ
Pet þ St
where Pe(t) is the effective rainfall at any time ‘t’ excluding the initial abstraction;
PeðtÞ ¼ Pt IaðtÞ ð5Þ
25; 400
St ¼ 254 ð6Þ
CNt
where CNt is the curve number at any time‘t’. Here, Pe(t) ≥0; else ROt =0. In Eq. 6 St varies
from 0 to ∝ , which can be mapped onto a non-dimensional curve number CN with a more
appealing range of 0 to 100 as in Eq. 6; and all dimensions are in millimetres.
3 Formulation of Continuous Simulation Model
The present model formulation incorporates the SCS-CN concept revised for rainfall-
dependent initial abstraction and quantification of flows adopting various flow paths in
streamflow generation, such as (1) Surface runoff, (2) Throughflow and (3) Base flow
(Fig. 1a). This algorithm operates on daily time basis and, therefore, requires daily data of
rainfall and evaporation as input to explain the physical behavior of the catchment. The
observed runoff is used for model evaluation. A complete description of individual
components of the proposed model as follows:
3.1 Initial Abstraction
Initial abstraction is considered as a short term loss before ponding such as interception,
infiltration, surface storage (Ponce and Hawkins 1996; Mishra and Singh 2003). Here it is
Fig. 1 Schematic diagram of P-Ia

ETt
SCS-CN-based lumped concep-
tual rainfall-runoff model EVt SROt
TRt
Ft
SMS
DRt
THRt
PRt
GWS
DSPt
BFt
TROt
DPRt
assumed that this loss is a fraction of the possible retention in the soil and is computed as:
IaðtÞ ¼ lSt if t 5 days ð7Þ
Here, l is taken as 0.2. Physically this means that for a given storm, 20% of the potential
maximum water retention is the initial abstraction before runoff begins (Singh 1992).
Otherwise
a
Pt
IaðtÞ ¼ l1 St ð8Þ
Pt þ St
Here l1 and a are the coefficient and exponent of the initial abstraction which are to be
optimised.
3.1.1 Antecedent Rainfall
In literature, the term antecedent varies from previous 5 to 30 days (SCS 1971; Singh 1992;
Mishra and Singh 2003). However no explicit guideline is available to vary the soil
moisture with the antecedent rainfall of certain duration. Since the NEH-4 (SCS 1971) uses
5-day rainfall based on the exhaustive field investigations, this duration of 5 days was
retained.
In this model, for the first 5 days beginning from the starting day of simulation (June 1–
June 5, in this study), curve number CN is taken as CN0 and as the day advances, CN
varies with respect to antecedent moisture amount, AM, based on the antecedent rainfall
(ANTRF) as:
ANTRFt ¼ Pðt1Þ þ Pðt2Þ þ Pðt2Þ þ Pt3 þ Pðt4Þ þ Pðt5Þ ð9Þ
where t is the day under consideration and P is the rainfall of the respective day.
3.1.2 Antecedent Moisture
The initial moisture available in the soil prior to storm plays a vital role in the estimation of
runoff (Mishra and Singh 2002a) as curve number CN variability is primarily attributed to
antecedent moisture amount rather than the antecedent moisture conditions (SCS 1971;
Mishra and Singh 2003) which may lead to sudden jumps in daily curve number values.
This model assumes that the current space available for water retention St is constant for
first 5 days of simulation and hence CNt =CN0. Using Eq. 6 St can be computed from curve
number CN0 of the first day which is determined by optimisation. When the number of
days exceeds 5, the antecedent moisture, representing the initial moisture available in the
watershed on the day under consideration (AMt), can be computed as follows:
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
AMt ¼ b ANTRFt ð10Þ
Here, b is the coefficient of antecedent moisture to be determined by optimisation. Then,
St is modified as
ðSt Þ2
St ¼ ð11Þ
ðAMt þ St Þ
Here in this model, it is considered that daily antecedent moisture amount AMt varies
with respect to antecedent rainfall (Eq. 10) and hence the daily possible water retention of
the soil is computed using Eq. 11.
3.2 Rainfall Excess
The amount of rainfall (P) reaching on the ground after the initial losses (Ia) is termed as
effective rainfall (Pe) and this is available for initiating various other processes in the
hydrologic cycle. The effective rainfall (Pe) is assumed to be partitioned as surface runoff
or rainfall excess (RO) and infiltration (F) as stated in Eq. 1. Using the daily effective
rainfall (Pet), the daily rainfall excess ROt can be computed by using Eq. 4 for the first 5-
days of simulation, only if rainfall P exceeds initial abstraction (Ia), it is zero otherwise.
3.2.1 Routing of Rainfall Excess
When the number of days exceeds 5, to transform the surface runoff that is produced at the
outlet of the basin, the rainfall excess ROt (Eq. 4) is routed using a single linear reservoir
concept, as follows (Nash 1957; Mishra and Singh 2003):
SROt ¼ C0 ROt þ C1 ROðt1Þ þ C2 SROðt1Þ ð12Þ
where
ð1=K Þ
C0 ¼ ð13aÞ
2 þ ð1=K Þ
C1 ¼ C0 ð13bÞ
2 ð1=K Þ
C2 ¼ ð13cÞ
2 þ ð1=KÞ
where K is the storage coefficient.
3.3 Infiltration
The amount of water reaching the ground after initial abstraction and not produced as direct
surface runoff is assumed to infiltrate into the upper soil. It is modelled as:
Ft ¼ Pt IaðtÞ ROðtÞ ð14Þ

3.4 Evapotranspiration
The amount of water goes back or lost to the atmosphere is in the form of
evapotranspiration ETt and can be obtained by the summation of daily evaporation from
the water bodies and transpiration from the soil zone in the watershed.
3.4.1 Evaporation
The daily evaporation EVt is computed as follows:
EVt ¼ PANC EVPt ð15Þ
where EVPt is the potential evaporation based on the field data and PANC is the pan
coefficient, assumed as 0.8 for June–September and 0.6 for October–November in this
study (Project Report 1978).
3.4.2 Transpiration
Transpiration from the soil zone is considered as a function of water content available in the
soil store above the wilting point of the soil (Putty and Prasad 1994, 2000; Mishra et al.
2005). The transpiration is computed as:
TRt ¼ C1 ðSabs St qw Þ ð16Þ
where C1 =coefficient of transpiration from soil zone, θw =wilting point of the soil, Sabs =the
maximum possible water retention, and St =possible water retention on tth day. The total
actual evapotranspiration is taken as the sum of evaporation and transpiration as follows:
ETt ¼ EVt þ TRt ð17Þ
3.5 Drainage
The term drainage is used as the outflow from a linear reservoir (Nash 1957) only when the
moisture content in the soil zone increases and exceeds the field capacity θf (Putty and
Prasad 2000; Mishra et al. 2005) as:
DRt ¼ C2 ðSabs St qf Þ ð18Þ
where C2= subsoil drainage coefficient, Sabs =maximum potential water retention, St =
possible water retention on tth day, DRt =drainage rate at time‘t’, and θf =field capacity of
the soil.
3.6 Throughflow or Interflow
The outflow from the unsaturated soil store is partitioned into two components: (1)
subsurface flow in lateral direction and (2) vertical percolation into ground water zone. The
former component representing the through flow is taken as a fraction of the above drainage
rate (Putty and Prasad 1994, 2000):
THRt ¼ C3 DRt ð19Þ
where THRt =throughflow at time ‘t’ and C3 =unsaturated soil zone runoff coefficient.
3.7 Percolation
The outflow in the vertical direction from the unsaturated zone meets the ground water
store due to the permeability of the soil. This percolated amount of water is considered as a
part of drainage, and it is estimated as (Putty and Prasad 1994, 2000; Mishra et al. 2005):
PRt ¼ ð1 C3 Þ DRt ð20Þ
where PRt=percolation at time ‘t’.
3.8 Deep Seepage
The saturated store is considered as a non-linear reservoir and from this saturated store,
outflow occurs at an exponential rate in the form of deep seepage. As the saturated store is
considered as a non-linear store, the formulation for the deep seepage is made as an
exponential function of the percolation. This is modeled as follows:
DSPt ¼ ðPRt ÞE ð21Þ
where DSPt =deep seepage at any time ‘t’ and E=exponent of ground water zone. Deep
seepage can travel in lateral direction as well as vertical direction through the saturated
store. This seepage is again bifurcated into two components: (1) active ground water flow
(base flow) and (2) inactive ground water flow (deep percolation) into the aquifers.
3.9 Base Flow
The base flow of a watershed is the ground water release from a catchment in a stream. This
active ground water flow which is also known as delayed flow can be modeled as outflow
from a non-linear storage in the form of base flow (BFt) as follows:
BFt ¼ BCOEF DSPt ð22Þ
where BCOEF=ground water zone runoff coefficient
3.10 Deep Percolation
The inactive ground water flow into aquifers is termed as deep percolation, occurs from the
saturated ground water zone in vertical direction, and is considered as a loss from the
saturated store which is modeled as:
DPRt ¼ ð1 BCOFFÞ DSPt ð23Þ

where DPRt =deep percolation at any ‘t’ and BCOEF=ground water zone runoff
coefficient. Here, it is worth emphasizing that the proposed model considers deep seepage
which is partitioned into two components, base flow and deep percolation. On the other
hand, both the Putty and Prasad (1994, 2000) and Kentucky (James 1972; Singh 1989)
models do not account for deep seepage and deep percolation.
3.11 Total Stream Flow
The total stream flow (TROt) on a day t, is obtained as the sum of the above three
components, surface runoff, throughflow, and base flow (Eqs. 4 or 12, 19, and 22).
TROt ¼ ROt þ THRt þ BFt if t 5 days ð24aÞ
TROt ¼ SROt þ THRt þ BFt if t > 5 days ð24bÞ
3.11.1 Water Retention Budgeting
The computation of daily water retention storage or soil moisture budgeting is essential in a
daily hydrologic simulation. This SCS-CN-based model represents a soil-water balance
model. The current space available for retention of water St is again modified by taking into
account the evapotranspiration loss, drainage from the soil moisture zone, and daily
infiltration to the unsaturated store as:
St ¼ Sðt1Þ Fðt1Þ þ ETðt1Þ þ DRðt1Þ ð25Þ
where S(t−1) is the previous day potential maximum retention (mm); ET(t−1) is the previous
day evapotranspiration (mm); DR(t−1) is the drainage on the previous day; F(t−1) is the
previous day infiltration (mm), computed using water balance equation:
Fðt1Þ ¼ Pðt1Þ Iaðt1Þ ROðt1Þ ð26Þ
Here, if Pe(t) ≥0, F≥0.
4 Development of Continuous Simulation Model
The proposed long-term hydrologic simulation model (Fig. 1) is developed for describing
watershed hydrology by considering temporal and spatial variations of various processes
involved in the runoff generation mechanism and also by incorporating modified soil
conservation service curve number (SCS-CN) technique as well as storage concepts to
represent the catchment response in a better way.
This modified SCS-CN based lumped model that captures the relevant catchment
features requires 13 parameters, viz., CN0, l1, a, b, K, C1, C2, C3, Sabs, θf, θw, BCOEF, and
E to derive an acceptable model output. Commonly, a close fit between calculated and
observed variables is possible for models with a high number of parameters even if the
model assumptions are false (Grayson et al. 1992; Beven 1997). In contrast to the studies
showing less number of parameters is needed to establish rainfall–runoff relationship
(Jakeman and Hornberger 1993), here the number of parameters involved in the model is
comparatively large, but it is at the gain of significant higher efficiency and it generates not
only streamflow but also its components, a distinctive feature. It is notable that the
presented model requires easily available rainfall and evaporation in order to generate
streamflow and its components.
Here the parameters involved in the model are determined using non-linear Marquardt
algorithm (In: Mishra and Singh 2003), coupled with trial and error, utilising the objective
function of minimising errors between the computed and observed data or maximising
model efficiency (Nash and Sutcliffe 1970). The ranges and initial values selected for the
optimizations are presented in Table 1. Notably, initial estimates of the parameters fixed by
trial-and-error need not be provided.
A comparison has also been made between the present model based on SCS-CN concept
and a model based on variable source area (VSA) theory (Mishra et al. 2005). This 3-
component VSA-based model assumes three different stores of moisture as interception
Table 1 Ranges and initial estimates of parameters (SCS-CN-Based model)
Sl. No. Parameters Range/Initial value Catchment
Hemavati Hridaynagar Manot Mohegaon Kalu
1 CN0 Range 0–100 0–100 0–100 0–100 0–100

Initial value – – – – –
2 BCOEF Range 0–1.0 0–1.0 0–1.0 0–1.0 0–1.0
3 K Range 0.1–2.0 0.1–5.0 0.1–1.0 0.001–10 0.1–1.0
Initial value 0.5 0.5 0.5 0.5 0.5
4 l1 Range 0.0–5.0 0–5.0 0–10.0 0–10.0 0–1.0
Initial value 0.2 0.02 0.02 0.02 0.05
5 a Range 0–10.0 0–5.0 0–10.0 0–10.0 0–1.0
Initial value 0.01 0.01 0.01 0.01 0.01
6 b Range 0–3.0 0–3.0 0–1.0 0–10.0 0–1.0
Initial value 0.01 0.01 0.01 0.01 0.001
7 E Range 0–1.0 0–1.0 0–2.0 0–2.0 0–2.0
8 C1 Range 0–1.0 0–1.0 0–1.0 0–1.0 0–1.0
9 C2 Range 0–1.0 0–1.0 0–1.0 0–1.0 0–1.0
Initial value 0.05 0.1 0.1 0.1 0.05
10 C3 Range 0–1.0 0–1.0 0–1.0 0–1.0 0–1.0
Initial value 0.05 0.05 0.05 – –
11 Sabs Range 500–1500 300–1000 300–1000 300–1000 500–1500
12 θf Range 100–1000 100–500 100–500 100–1000 100–1200
13 θw Range 40–120 40–120 40–120 40–120 40–120
Fig. 2 a Drainage map of Hemavati catchment. b Index map of Narmada at Manot, Banjar at Hridaynagar,
and Burhner at Mohegoan watersheds. c Drainage map of Kalu catchment
Fig. 2 (Continued)
store, soil moisture store, and ground water store. The model quantifies daily as well as
annual streamflow and its components i.e. source area runoff, throughflow and base flow.
The main difference between the SCS-CN-based model and VSA-based model is that the
former is an infiltration-excess model and assumes the surface runoff to be produced from
the entire catchment whereas the latter produces it from source areas only, i.e. certain
dynamic contributing areas varying with storm intensity. Furthermore, the former assumes
the surface runoff to be produced due to infiltration-excess overland flow, similar to
Hortonian overland flow whereas the latter produces the surface runoff due to saturation-
excess overland flow. The latter model is a modification of the Kentucky watershed model
(James 1972) and SAHYADRI model (Putty and Prasad 1994, 2000), as these models do
not account for the deep seepage and deep percolation.
5 Case Study Watersheds
The above proposed model is applied to different catchments falling under different
climatic and geographic settings of India. The daily monsoon (June–November) data of
study catchments (Fig. 2a, b, c), Hemavati, a tributary of River Cauvery in Karnataka State,
Hridaynagar, Manot, and Mohegaon catchments, tributaries of River Narmada in Madhya
Pradesh, and Kalu catchment, a tributary of River Ulhas, in Maharashtra State of India,
observed at five gauging stations, are used for the analysis. Details of these catchments are
presented in Table 2.
178
Table 2 Data used for model calibration and validation
Description Catchment
Hemavati Hridaynagar Manot Mohegaon Kalu
River Cauvery Narmada Narmada Narmada Ulhas

State Karnataka Madhya Pradesh Madhya Pradesh Madhya Pradesh Maharashtra
District Chikmanglur Durg Shahdol Mandla Thane
Area 600.00 3,370.00 5,032.00 4,661.00 224.00
Latitude 12°55′ to 13°11′ N 21° 42′ N 22°26′ to 23°18′ N 22° 32′ N 19° 17′ to 19° 26′N
Longitude 75°29′ to 75°51′ E 80° 50′ E 80°24′ to 81°47′ E 81° 22′ E 73°36′ to 73° 49′ E
Topography Low land, semi hilly and both flat and undulating Hilly Both flat and undulating lands Hilly
hilly lands
Land use Forest – 12% Forest – 65% Forest – 35% Forest – 58% Forest – 50%
Coffee plantation – 29% Agriculture – 29% Cultivation – 52% Agriculture – 42% Cultivation – 50%
Agriculture land – 59% Degraded land water Waste land – 13%
bodies – 6%
Soil Red loamy soil and red Black to mixed red soil Red, yellow, and medium Red and yellow silty loam and silty Silty loam and sandy
sandy soil black soil clay loam loam
Elevation (m) above 1,240–890 600–372 450–1,110 900–509 1,200
m.s.l
Average annual rainfall 2,972.00 1,178.00 1,596.00 1,547.00 2,450.00
(mm)
Calibration–Monsoon 1974–1976 (3 years) 1981–1985 (5 years) 1981–1985 (5 years) 1981–1985 (5 years) 1990–1992 (3 years)
Period
Validation–Monsoon 1977–1978 (2 years) 1986–1989 (4 years) 1986–1989 (4 years) 1986–1989 (4 years) 1993 (1 year)
Period
K. Geetha, et al.
(a) Hemavati catchment: River Hemavati is a tributary of River Cauvery, originating in

Ballaiarayanadurga in the Western Ghats in Mundgiri taluk of Chikmanglur district in
Karnataka State (Fig. 2a). It traverses a total length about 55.13 km. upto Sakleshpur.
Agriculture and plantation are the major industries of the basin. Soils in the forest area
and coffee plantations are greyish due to high humus content.
(b) Narmada upto Manot: River Narmada (Table 2) rises from Amarkantak plateau of
Maikala range in Shahdol district in Madhya Pradesh at an elevation of about 1059 m
above mean sea level. The length of the river Narmada from its origin up to Manot is
about 269 km (Fig. 2b). It has continental type of climate classified as sub-tropical
and sub-humid. It is very hot in summer and cold in winter.
(c) Banjar up to Hridaynagar: The Banjar River (Fig. 2b), a tributary of Narmada in its
upper reaches, rises from the Satpura range in Durg district of Madhya Pradesh near
Rampur village at an elevation of 600 m and the elevation drops from 600 to 372 m at
Hridaynagar gauging site (Table 2). Climate of the basin can be classified as sub-
tropical sub-humid and about 90% of the annual rainfall is received during monsoon
season (June–October).
(d) Burhner up to Mohegaon: The Burhner river (Mohegaon catchment), a tributary of
Narmada River rises in the Maikala range, south-east of Gwara village in the Mandla
district of Madhya Pradesh at an elevation of about 900 m. It flows in westerly
direction for a total length of 177 km to join the Narmada near Manot. The elevation
at Mohegaon gauging site drops to 509 m. and climate of the basin can be classified
as sub-tropical and sub-humid.
(e) Kalu catchment: River Kalu is a tributary of Ulhas River in the Thane District of
Konkan Region in Maharashtra (Fig. 2c). It originates near Harichandragad in Murbad
Taluka of Thane District at an elevation of 1200 m above mean sea level (Table 2).
Most of the rains are received during June–October. Existing crop pattern of the
cultivation covers 46% paddy, 16% nanchani vari, 3% pulses, and 35% grass.
6 Model Application
The above described model employs the revised SCS-CN model incorporating antecedent
moisture amount, AMA, to calculate the space available for water retention; and updates the
water retention store daily using evapotranspiration, drainage from soil moisture store, and
infiltration to soil moisture store. Its application requires daily data of rainfall and
evaporation. It has been applied to the daily data of monsoon period (June–November) of
the above catchments. The details of study watersheds and the record of hydrologic data
used in this analysis are presented in Table 2.
The optimal estimates of model parameters (Mishra et al. 2005) were obtained by
using non-linear Marquardt algorithm (In: Mishra and Singh 2003) coupled with trial-and-
error. Table 3 presents the model efficiency along with runoff coefficients of the catchment.
It is seen that Hridaynagar catchment shows the least runoff coefficient of 0.262 and
behaves as dry catchment (Gan et al. 1997). The catchments of Kalu and Hemavati can
be classified as high runoff producing catchments with the coefficient values of 0.964 and
0.803, respectively. The runoff coefficients for Manot and Mohegaon catchments are
0.475 and 0.362, respectively, describing them to lie in the intermediate category of dry
and wet.
Table 3 Estimates of model parameters and model efficiency (SCS-CN-based model)
Sl. No Parameter Hemavati Hridaynagar Manot Mohegaon Kalu

Data Length – Monsoon Period (June–November)
Cal – 3 years Cal – 5 years Cal – 5 years Cal – 5 years Calib – 3 years
Val – 2 years Val – 4 years Val – 4 years Val – 4 years Valid – 1 year
1 CN0 35.0 69.0 61.0 30.0 51.0

2 K 1.039 1.007 0.123 0.016 0.548
3 l1 0.551 1.822 4.937 5.566 0.289
4 a 4.174 1.633 6.851 3.074 0.794
5 b 0.651 0.681 0.990 5.627 0.720
6 C1 3.75 4.2 8.2 17.36 2.775
7 C2 0.173 0.041 0.096 0.020 0.046
8 C3 0.898 0.135 0.305 0.790 0.200
9 BCOEF 0.195 0.62 0.19 0.45 0.740
10 E 0.63 0.91 1.49 1.04 1.35
11 Sabs 1,350.0 500.0 500.0 700.0 1,200.0
12 θf 800.0 350.0 350.0 500.0 950.0
13 θw 100.0 80.0 90.0 80.0 90.0
14 Efficiency 84.43% 58.84% 74.36% 73.38% 67.86%
Calibration
15 Efficiency 87.70% 47.73% 66.51% 37.0% 84.0%
Validation
16 Runoff 0.803 0.262 0.475 0.372 0.964
Coefficient
The model yields maximum efficiency of 84 and 88% in calibration and validation,
respectively, in Hemavati catchment whereas Hridaynagar catchment produces the least
efficiencies of 59 and 48%, respectively. The highest efficiency reveals that the model is
efficacious to high runoff producing Hemavati catchment. The lower the efficiency, the
higher the error between observed and simulated runoff .The other catchments, like Manot,
Mohegaon, and Kalu, exhibit 74, 73, and 68% efficiencies, respectively, in calibration, and
67, 37, and 84%, respectively, in validation. It follows that, except Hridaynagar, all other
catchments are more amenable to the above model than the others showing lowest runoff
coefficient or dry catchment. Notably, as expected, the efficiencies of all catchments, except
Hemavati and Kalu, are higher in calibration than in validation, but reverse holds for the
others. The average relative error (R.E.) values for Kalu and Hemavati are computed as 5
and 9%, respectively, and for Hridaynagar, Manot, and Mohegaon catchments are,
respectively, 26, 19, and 17% (Table 4). These values generally exhibit a satisfactory
model performance. The R.E.-values indicating negative values as seen in Table 4 imply
that the model overestimates the runoff values.
Figure 3a–e presents daily variations of estimated and observed runoff with respect
to daily average rainfall for all the catchments. Also an average relative error is calcu-
lated for the study period and presented in Table 4. The relative error computed are 9,
25, 19, 17, and 5% in Hemavati, Hridaynagar, Manot, Mohegaon, and Kalu catchment,
respectively.
Table 4 Annual rainfall, observed runoff, simulated runoff and relative error (SCS-CN-based lumped
model)
Sl.No. Year Rainfall (mm) Observed Runoff Simulated Runoff Relative Error
(mm) (mm) (%)
Hemavati
1 1974 2,827 2,421 2,248 7
2 1975 2,399 1,604 1,733 −8
3 1976 2,434 1,804 1,865 −3
4 1977 2,783 2,831 2,285 19
5 1978 2,795 1,987 2,191 −10
Average 2,650 2,129 2,064 9
Hridaynagar
1 1981 1,433 308 248 19
2 1982 1,400 306 284 7
3 1983 1,755 364 411 −13
4 1984 1,179 287 268 7
5 1985 1,187 237 210 11
6 1986 1,632 584 361 38
7 1987 834 173 99 43
8 1988 1,370 557 269 52
9 1989 1,025 280 172 39
Average 1,313 344 258 25
Manot
1 1981 1,045 374 389 −4
2 1982 1,002 366 389 −6
3 1983 1,257 529 536 −1
4 1984 1,190 601 592 2
5 1985 1,122 701 482 31
6 1986 1,136 691 451 35
7 1987 1,249 752 485 36
8 1988 1,267 625 565 10
9 1989 1,056 265 391 −48
Average 1,147 545 476 19
Mohegaon
1 1981 1,148 324 391 −21
2 1982 1,049 332 323 3
3 1983 1,396 463 472 −2
4 1984 1,180 494 492 0
5 1985 1,103 553 398 28
6 1986 1,182 459 306 33
7 1987 1,050 364 355 3
8 1988 1,153 544 419 23
9 1989 826 221 317 −43
Average 1,121 417 386 17
Kalu
1 1990 3,348 3,528 3,339 5
2 1991 3,169 3,058 3,194 −5
3 1992 1,903 1,601 1,477 8
4 1993 3,129 2,944 2,974 −1
Average 2,887 2,783 2,746 5
200 0
Rainfall in mm
160
Hemavati 1974-1978 100
Runoff in mm
R.E.9%
120
Runoff (E) 200
80 Runoff (O)
Rainfall
300
40
0 400
1 71 141 211 281 351 421 491 561 631 701 771 841 911
a Time in Days
100 0
Rainfall in mm
75
Runoff in mm
Runoff (E) 100

Hridaynagar 1981-1990 Runoff (O)
50 R.E 25% Rainfall
200
25
0 300
1 161 321 481 641 801 961 1121 1281 1441 1601
b Time in Days
200 0
160
100
Rainfall in mm
Manot 1981-1990
R.E.19%
Runoff in mm
120
Runoff (E) 200
Runoff (O)
80
Rainfall
300
40
0 400
1 161 321 481 641 801 961 1121 1281 1441 1601
c Time in Days
200 0
160
Rainfall in mm
100
Runoff (E) Mohegaon 1981-1989
Runoff in mm
120
R.E.17%
Runoff (O) 200
80 Rainfall
300
40
0 400
1 161 321 481 641 801 961 1121 1281 1441 1601
d Time in Days
500 0
400
Rainfall in mm
Kalu-1990-1993 150
R.E.5%
300
Runoff in mm
Runoff (E)
300
Runoff (O)
200
Rainfall
450
100
0 600
1 61 121 181 241 301 361 421 481 541 601 661 721
e Time in Days
Fig. 3 (Continued)
As shown in Fig. 4a–b, the least deviation in RE-values is observed in Kalu catchment,
i.e. 5% in 1990 in calibration, and 1% in 1993 in validation, respectively. The model
performs satisfactorily in this catchment, except for few peaks, where the computed runoff
is lower than the observed. It is largely because the limitation of the optimized function that
is minimized based on a large number of other data points than the peaks. The sample
RFig. 3 a–e Daily variations of rainfall, observed runoff (O), estimated runoff (E) and average relative error
(%) of different catchments for SCS-CN based lumped conceptual model
300 0
250 50
400 0
Kalu 1990
R.E.5.38% 350 50
200 100
300 100
Rainfall in mm
Kalu 1993
Runoff in mm
Runoff (E)
Rainfall in mm
150 Runoff (O) 150 250 R.E. 0.98% 150
Runoff in mm
Rainfall
200 200
Runoff (E)
100 200
150 Runoff (O) 250
Rainfall
100 300
50 250
50 350
0 300 0 400
1 21 41 61 81 101 121 141 161 181 1 21 41 61 81 101 121 141 161 181
Time in Days Time in Days
a Calibration b Validation
100 0
50 0
80 40
40 80
Hridaynagar 1981 Hridaynagar 1988
R.E.19.48% 60 R.E.51.66% 80
Rainfall in mm
30 160
Rainfall in mm
Runoff in mm
Runoff in mm
Runoff (E) Runoff (E)

Runoff (O) 40 Runoff (O) 120
20 Rainfall 240
Rainfall
10 320 20 160
0 400 0 200
1 21 41 61 81 101 121 141 161 181 1 21 41 61 81 101 121 141 161 181
Time in Days Time in Days
c Calibration d Validation
Fig. 4 a–d Calibration and validation of results of Kalu and Hridaynagar exhibiting the best and the worst
performance respectively (SCS-CN-based lumped conceptual model)
calibration and validation results yielding the maximum errors on the data of Hridaynagar
watershed is also presented in Fig. 4c–d, respectively. The resulting errors are 19% in 1981
in calibration, and 52% in validation in 1988, indicating a poor agreement between the
observed and simulated runoff values. Also various component processes involved are also
quantified and given in Table 5 & Fig. 5a–i. It is apparent that the initial abstraction takes a
maximum value in Hridaynagar catchment, indicating less amount of rainfall reaching the
ground to generate runoff.
While comparing the runoff components, base flow is insignificant in all watersheds
except Kalu catchment, whereas surface runoff is dormant only in Hridaynagar catchment.
The runoff produced in the lateral direction (throughflow) is maximum in Hemavati
Table 5 Percent estimates of hydrological components for different catchments (SCS-CN-based model)
Sl. No. Components Hemavati Hridayanagar Manot Mohegaon Kalu

(%) (%) (%) (%) (%)
1 Rainfall (R) 100 100 100 100 100

* *
2 Initial abstraction (Ia) 4.26 47.07 0.52 42.11 17.70
3 Effective Rainfall (Pe) 95.74 52.93 99.48 57.89 82.3
4 Infiltration (Ft) 58.72 44.48 78.55 45.76 35.33
5 Drainage (DRt) 44.42 17.82 35.48 25.33 35.05
* *
6 Percolation (PRt) 4.57 15.38 24.67 5.35 28.06
* *
7 Deep seepage (DPSt) 5.21 14.17 51.44 5.17 55.59
*
8 Surface runoff (ROt) 37.02 8.45 20.92 12.13 46.97
* *
9 Throughflow (THRt) 39.85 2.44 10.81 19.98 7.00
* * * *
10 Base flow (BFt) 1.02 8.76 9.76 2.32 41.15
11 Simulated runoff (TROt) 77.89 19.65 41.49 34.43 95.12
* * *
12 Deep percolation (DPrt) 4.19 5.41 41.67 2.85 14.44
*
Dormant process
catchment, and minimum in Hridaynagar catchment. Maximum amount of surface runoff is

derived from the Kalu watershed for a given amount of rainfall, and least from the
Hridaynagar catchment. Deep percolation is dormant in Hemavati, Hridaynagar, and
Mohegaon catchments. Based on the results (Table 5), the dominancy/dormancy of the
processes in each catchment can be identified.
7 Model Comparison
This section compares the application of the two models, viz., the proposed SCS-CN-Based
lumped conceptual model (Fig. 1) and the long term hydrologic VSA-Based model using
storage and source area concepts (Mishra et al. 2005).
Tables 6 and 7 compare the model efficiencies and average RE-values due to the above
models. Both the models show a satisfactory performance (the higher model efficiencies
and less relative error) on high runoff producing catchments, like Hemavati and Kalu
catchments. The catchment in dry region, like Hridaynagar, indicates the lowest efficiency
in both model applications. The comparison based on average relative error (Tables 6 and
7) indicates the model based on the modified SCS-CN technique performs significantly
better than the model based on variable source area (VSA) theory.
It is seen from Table 8 that Hridaynagar catchment generates monsoon runoff 26%
of the average monsoon rainfall as observed whereas simulated ones are 20 and 19%
derived from SCS-CN-based and VSA-based models, respectively. It implies that the
losses are high in the Hridaynagar catchment. The other low runoff producing catch-
ments, Manot and Mohegaon catchments, respectively produce 47 and 37% runoff. The
runoff computed from SCS-CN-based and VSA-based models in Manot are 42 and
44%; and in Mohegaon are 34 and 30% of average monsoon rainfall. The wet catch-
ments, viz., Kalu and Hemavati, produce high amount of runoff as observed and are,
respectively 96 and 80% of average monsoon rainfall. The corresponding SCS-CN
60 50
Surface runoff 45 Throughflow
50 40
Surface runoff (%)
35
Throughflow (%)
40
30
30 25
20
20 15
10
10
5
0 0
1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9
a Time in years b Time in years
50 90
45 Base flow 80
40 70
35 60
Infiltration (%)
Base flow (%)
30
50
25
40
20
30
15
10 20
Infiltration
5 10
0 0
1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9
c Time in Years d Time in Years
60 35
Drainage 30 Percolation
50
25
Percolation (%)
40
Drainage (%)
20
30
15
20
10
10 5
0 0
1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9
e Time in Years f Time in Years
Fig. 5 a–i Annual variations of various processes in runoff generation mechanism (SCS-CN-Based lumped
conceptual model)
based model simulated values for these catchments are 95 and 78%, whereas the VSA-
based model yields 81 and 71%, respectively, with respect to average monsoon rainfall. Thus,
the results due to the SCS-CN-based model are closer to the observed than the VSA-based
model, indicating the better performance of the proposed SCS-CN-based model in long-term
hydrologic simulation.
120 50
Simulated runoff 45
100
40
Simulated runoff (%)
Deep Percolation
80 35
30
Deep Percolation
60 25
20
40
15
20 10
5
0 0
1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9
g Time in Years h Time in Years
120
Observed runoff Hemavati
100
Hridaynagar
Observed runoff(%)
80 Manot
Mohegaon
60 Kalu
40
20
0
1 2 3 4 5 6 7 8 9
i Time in Years
Fig. 5 (Continued)
Table 6 Data length and model efficiency (%) with runoff coefficient
Catchment Area Data Length Efficiency (%) Runoff

(km2) (Monsoon Period) Coefficient
Calibration Validation SCS-CN-based VSA-based lumped

(Years) (Years) lumped conceptual conceptual model
model
Calibration Validation Calibration Validation
Hemavati 600.00 3 2 84.43 87.7 78.37 68.92 0.803

Hridaynagar 3,370.00 5 4 58.84 47.73 38.46 46.05 0.262
Manot 5,032.00 5 4 74.36 66.51 26.28 55.60 0.475
Mohegaon 4,661.00 5 4 73.38 37.00 −13.93 51.86 0.372
Kalu 224.11 3 1 67.86 84.00 69.70 78.18 0.964
− Negative efficiency
Table 7 Annual average rainfall, observed & simulated runoff and relative error (%)
Sl. No. Catchment Average Average Model based on Model based on

rainfall observed SCS-CN concept VSA concept
(mm) runoff
(mm) Average Average Average Average
simulated Relative Simulated Relative
runoff (mm) Error (%) runoff (mm) Error (%)
1 Hemavati 2,650 2,129 2,064 9 1,875 12

2 Hridaynagar 1,313 344 258 26 243 29
3 Manot 1,147 545 476 19 501 22
4 Mohegaon 1,121 417 386 17 339 19
5 Kalu 2,887 2,783 2,746 5 2,332 16
8 Conclusion
A new SCS-CN-based continuous simulation model for hydrologic forecasting has been
presented here and applied to the data of watersheds of River Cauvery in Karnataka,
River Narmada in Madhya Pradesh, and River Ulhas in Maharashtra in India. This model
can be used in the computation of annual hydrologic components as well as total runoff
values. When compared with the available VSA-based model, the SCS-CN-based model
better represented the hydrologic behavior of the catchments with different soil, vege-
tation, and climate than did the VSA-based model. Both the models however performed
satisfactorily on high runoff producing Kalu and Hemavati watersheds. Besides, initial
abstraction was dormant in Hemavati and Manot catchments; deep percolation was neg-
ligible in Hemavati, Hridaynagar and Mohegaon catchments; base flow was insignificant
in all catchments except Kalu watershed; the least runoff producing Hridaynagar catch-
ment generated negligible surface runoff, throughflow, and base flow; the wet Kalu catch-
ment produced relatively less throughflow compared to runoff; and percolation and deep
seepage were dormant in Hemavati and Mohegaon catchments. As shown in this study, the
present model can be used as a continuous simulation model for hydrologic forecasting of
watersheds.
Table 8 Comparison of annual average runoff with respect to annual average rainfall
Sl. No. Catchment Observed Simulated runoff (%) Category of

runoff (%) watershed
Model based on Model based on
SCS-CN concept VSA concept
1 Hemavati 80.36 77.89 70.76 Wet

2 Hridaynagar 26.21 19.65 18.54 Dry
3 Manot 47.49 41.50 43.71 Normal
4 Mohegaon 37.20 34.43 29.57 Normal
5 Kalu 96.43 95.12 80.79 Wet
References
Arnold JG, Engel BA, Srinivasan R (1993) Continuous time, grid cell watershed model, application of
advanced information technologies. Effective management of natural resources. ASAE Publication, 04–
93. American Society of Agricultural Engineers 267–278
Beven KJ (1997) TOPMODEL: a critique. Hydrol Process 11:1069–1085
Beven KJ, Kirkby MJ (1979) A physically based variable contribution area model of basin hydrology.
Hydrol Sci Bull 24 1(3):43–69
Beven KJ, Kirkby MJ, Schofield N, Tagg AF (1984) Testing a physically based flood forecasting model
(TOPMODEL) for three U.K. catchments. J Hydrol 69:119–143
Beven KJ, Lamb R, Quinn P, Romanowicz R, Freee J (1995) In: Singh VP (eds) TOPMODEL in Computer
models of watershed hydrology. Water Resources Publication, Colorado, USA, pp 627–668
Bhaduri B (1998) A geographic information system based model of long-term impact of land use change on
non-point source pollution at watershed scale. Ph.D. thesis, Purdue University, West Lafayette, IN
Boughton WC (1966) A mathematical model for relating runoff to rainfall with daily data. Civil Eng Trans
Inst Eng Aust CES(1):83–97
Boughton WC (1968) A mathematical model of estimating runoff. J Hydrol N Z 7(2):75–100
Chiew FHS, McMahon TA (1994) Application of the daily rainfall–runoff model MODHYDROLOG to 28
Australian catchments. J Hydrol 153:383–416
Choi J-Y, Engel BA, Chung HW (2002) Daily streamflow modeling and assessment based on the curve-
number technique. Hydrol Process 16:3131–3150
Crawford NH, Linsley RK (1966) Digital simulation in hydrology: Stanford watershed model IV. Tech report
no.39. Stanford University, Palo Alto, CA
Gan TY, Dlamini EM, Bifu GF (1997) Effects of model complexity and structure, data quality, and objective
functions on hydrologic modeling. J Hydrol 192:81–103
Grayson RB, Moore ID, McMahon TA (1992) Physically based hydrologic modelling-2, is the concept
realistic? Water Resour Res 28(10):2659–2666
Harbor JM (1994) A practical method for estimating the impact of land use change on surface runoff, ground
water recharge, and wet land hydrology. J Am Plan Assoc 60(1):95–108
Heaney JP, Sample D, Wright L (2001) Geographical information systems. Decision support systems, and
urban storm water management. U.S. Environmental Protection Agency, Edison, NJ
Huber WC, Heaney JP, Bedient BP, Bender JP (1976) Environmental resources management studies in the
Kissimmie river basin. Rep. no. ENV-05-76-3, Dept of Env Engrg Sc Uni. of Florida, Gainsville, FL, May
Jakeman AJ, Hornberger GM (1993) How much complexity is warranted in a rainfall–runoff model? Water
Resour Res 29(8):2637–2649
James LD (1970) An evaluation of relationships between streamflow patterns and watershed characteristics
through the use of Opset: a self calibrating version of Stanford watershed model. Res Rep 36. Water Res
Inst, University of Kentucky, Lexington
James LD (1972) Hydrologic modeling parameter estimation and watershed characteristics. J Hydrol 17:283–
307
Knisel WG (1980) CREAMS, A field scale model for chemicals, runoff, and erosion from agricultural
management systems. Conservation rep. no. 26. USDA, ARS, Washington DC
Limbrunner, Vogel, Chapra (2006) In: Singh VP, Frevert DK (eds) A parsimonious watershed model in a
book “Watershed Models.” CRC, Boca Raton
Liou EY (1970) Opset: program for computerized selection of watershed parameter values for the Stanford
watershed model, Res. Rep. 34. Water Res Inst, University of Kentucky, Lexington
Mandeville AN, O’Connell PE, Sutcliffe JV, Nash JE (1970) River flow forecasting through conceptual
models. Part III. The ray catchment at Grendon Underwood. J Hydrol 11:109–128
Michel C, Vazken A, Charles P (2005) Soil conservation service number method: how to mend among soil
moisture accounting procedure? Water Resour Res 41(2)
Mishra SK (1998) Long-term hydrologic simulation using SCS-CN method. Tech report. National Institute
of Hydrology, Roorkee-247 667. UP, India
Mishra SK (2000) SCS-CN based long term hydrologic simulation. Tech. rep. National Institute of
Hydrology, Roorkee-247 667. UP, India
Mishra SK, Singh VP (1999) Another look at the SCS-CN method. J Hydrol Eng 4(3):257–264
Mishra SK, Singh VP (2002a) SCS-CN-based hydrologic simulation package, Ch. 13. In: Singh VP, Frevert
DK (eds) Mathematical models in small watershed hydrology and applications, Water Resources
Publications, P.O. Box 2841, Littleton, CO 80161, pp 391–464
Mishra SK, Singh VP (2002b) SCS-CN method. Part I. Derivation of SCS-CN based models. Acta Geophys
Pol 50(3):457–477
Mishra SK, Singh VP (2003) Soil Conservation Service Curve Number (SCS-CN) methodology. Kluwer,
Dordrecht, The Netherlands, ISBN 1-4020-1132-6
Mishra SK, Singh VP (2004a) Long-term hydrological simulation based on the soil conservation service
curve number. Hydrol Process 18(7):1291–1313
Mishra SK, Singh VP (2004b) Validity and extension of the SCS-CN method for computing infiltration and
rainfall-excess rates. Hydrol Process 18:3323–3345
Mishra SK, Goel NK, Seth SM, Srivastava DK (1998) An SCS-CN-based long-term daily flow simulation
model a hilly catchment. Int Symp hydrology of ungauged streams in hilly regions for small hydro
power development. Mar 9–10, New Delhi, India
Mishra SK, Geetha K, Rastogi AK, Pandey RP (2005) Long-term hydrologic simulation using storage and
source area concepts. Hydrol Process 19:2845–2861
Nageshwar RB, Wesley PJ, Ravikumar SD (1992) Hydrologic parameter estimation using Geographic
Information System. J Water Resour Plan Manage 118(5):492–512
Nash JE (1957) The form of the instantaneous unit hydrograph. IASH Pub no. 45, 34:114–121
Nash JE, Sutcliffe JV (1970) River flow forecasting through conceptual models. Part I. A discussion of
principles. J Hydrol 10:282–290
Ponce VM (1989) Engineering hydrology: principles and practices. Prentice Hall, Englewood Cliffs, NJ
Ponce VM, Hawkins RH (1996) Runoff curve number: has it reached maturity? J Hydrol Eng 1(1):11–19
Poter JW, McMahon TA (1976) The Monash model user manual for daily program HYDROLOG. Dept of
Civil Eng Monash University vic. Res, Rep. 2/76,41
Project Report (1978) History and completion report of Ramganga project, vol I. Irrigation Department, Uttar
Pradesh, India
Putty Y, Prasad R (1994) Streamflow generation in Western Ghats. In: 6th nat symp on hydrology, Shillong
Putty MRY, Prasad R (2000) Understanding runoff processes using a watershed model–a case study in the
Western Ghats in South India. J Hydrol 228:215–227
Soil Conservation Service (SCS) (1971) Hydrology national engineering handbook, section 4, Chapter 10.
Soil Conservation Service, USDA, Washington DC
Soil Conservation Service (SCS) (2004) http://www.wcc.nrcs.usda.gov/hydro/hydro-techref-neh-630-html
Singh VP (1989) Hydrologic systems-watershed modeling, vol II. Prentice Hall, Englewood Cliffs, NJ,
076323
Singh VP (1992) Elementary hydrology. Prentice Hall, Englewood Cliffs, NJ, 076323
Singh VP (1995) Computer models of watershed hydrology. Water Resources Publications, Littleton, CO
Singh VP, Frevert DK, Rieker JD, Leverson V, Meyer S, Meyer S (2006) Hydrologic modeling inventory:
cooperative research effort. J Irrig Drain Eng 132(2):98–103
Soni B, Mishra GC (1985) Soil water accounting using SCS hydrologic soil classification, case study.
National Institute of Hydrology, Roorkee (India)
Spruill CA, Workman SR, Taraba JL (2000) Simulation of daily and monthly discharge from small
watersheds using the SWAT model. Trans of American Society of Agricultural Engineers 43
(6):1431–1439
Sugawara M, Watanbe I, Ozaki E, Katsuyama Y (1984) Tank model with its snow component. Res. Notes of
Nat Res. Center for Disaster Prevention no. 65. Science and Technology Agency, Ibaraki-Ken, Japan
USDA (1986) Urban hydrology for small watersheds. Tech release 55, Natural Resources Conservation
Service, US Dept of Agriculture
Williams JR, LaSeur V (1976) Water yield model using SCS curve numbers. J Hydraul Eng 102
(HY9):1241–1253
Woodward DE, Gbuerek WJ (1992) Progress report ARS/SCS curve number work group. In: Proceedings
ASCE, Water Forum, 92. ASCE, New York, pp 378–382
Yao H, Hashino M, Yoshida H (1996) Modeling energy and water cycle in a forested head basin. J Hydrol
174:221–234
View publication stats

Eldho-Geetha-Wrm-Pub-Feb08 - Continuos CN

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Eldho-Geetha-Wrm-Pub-Feb08 - Continuos CN

Uploaded by

Copyright:

Available Formats

See

SCS-CN-based Continuous Simulation Model

Article in Water Resources Management · February 2007

Surendra Mishra T.I. Eldho

SEE PROFILE SEE PROFILE

A. K. Rastogi Rajendra P. Pandey

SEE PROFILE SEE PROFILE

Quantitative Assessment of vulnerability to drought View project

Hydrological droughts View project

The user has requested enhancement of the downloaded file.

SCS-CN-based Continuous Simulation Model

K. Geetha & S. K. Mishra & T. I. Eldho &

Received: 19 March 2006 / Accepted: 22 December 2006 /

Keywords hydrological forecasting . long-term hydrologic simulation . saturation excess

K. Geetha : T. I. Eldho (*) : A. K. Rastogi

The simulation of rainfall-generated runoff is very important in various activities of water

2 Existing SCS-CN Model

where P=total precipitation; Ia =initial abstraction; F=cumulative infiltration; Q=direct

2.1 Computation of Surface Runoff

3 Formulation of Continuous Simulation Model

3.1 Initial Abstraction

Fig. 1 Schematic diagram of P-Ia

3.1.1 Antecedent Rainfall

3.1.2 Antecedent Moisture

3.2 Rainfall Excess

3.2.1 Routing of Rainfall Excess

SROt ¼ C0 ROt þ C1 ROðt1Þ þ C2 SROðt1Þ ð12Þ

where K is the storage coefficient.

Ft ¼ Pt IaðtÞ ROðtÞ ð14Þ

The daily evaporation EVt is computed as follows:

EVt ¼ PANC EVPt ð15Þ

TRt ¼ C1 ðSabs St qw Þ ð16Þ

ETt ¼ EVt þ TRt ð17Þ

DRt ¼ C2 ðSabs St qf Þ ð18Þ

3.6 Throughflow or Interflow

3.8 Deep Seepage

3.9 Base Flow

3.10 Deep Percolation

DPRt ¼ ð1 BCOFFÞ DSPt ð23Þ

3.11 Total Stream Flow

TROt ¼ ROt þ THRt þ BFt if t 5 days ð24aÞ

TROt ¼ SROt þ THRt þ BFt if t > 5 days ð24bÞ

3.11.1 Water Retention Budgeting

St ¼ Sðt1Þ Fðt1Þ þ ETðt1Þ þ DRðt1Þ ð25Þ

Fðt1Þ ¼ Pðt1Þ Iaðt1Þ ROðt1Þ ð26Þ

Here, if Pe(t) ≥0, F≥0.

4 Development of Continuous Simulation Model

Table 1 Ranges and initial estimates of parameters (SCS-CN-Based model)

Sl. No. Parameters Range/Initial value Catchment

Hemavati Hridaynagar Manot Mohegaon Kalu

1 CN0 Range 0–100 0–100 0–100 0–100 0–100

5 Case Study Watersheds

Table 2 Data used for model calibration and validation

Hemavati Hridaynagar Manot Mohegaon Kalu

River Cauvery Narmada Narmada Narmada Ulhas

(a) Hemavati catchment: River Hemavati is a tributary of River Cauvery, originating in

Table 3 Estimates of model parameters and model efficiency (SCS-CN-based model)

Sl. No Parameter Hemavati Hridaynagar Manot Mohegaon Kalu

1 CN0 35.0 69.0 61.0 30.0 51.0

Runoff (E) 100

Runoff (E) Runoff (E)

Sl. No. Components Hemavati Hridayanagar Manot Mohegaon Kalu