American Journal of Engineering and Applied Sciences

Original Research Paper

Uncertainty Analysis of Time of Concentration Equations

based on First-Order-Analysis (FOA) Method
Asghar Azizian

Department of Water Engineering,

Faculty of Engineering and Technology, Imam Khomeini International University, Qazvin, Iran

Article history Abstract: The time of concentration (Tc) is one the most important time
Received: 10-02-2018 parameters to predict the response of a catchment to a given rainfall and
Revised: 12-02-2018 plays a key role in the hydrologic design and rainfall-runoff modeling.
Accepted: 19-03-2018 There are a huge number of empirical/semi-empirical equations for
estimation of Tc and depending on several parameters such as rainfall
Email: [email protected]
attributes, topographic and land cover map scale, DEM resolution and
streams delineation threshold causes significant uncertainties in the Tc
value. How to quantitatively evaluate the uncertainties in model parameters
and the resulting uncertainty impacts on model outputs has always been a
question which has attracted much attention. In this study, the method
based on the First-Order-Analysis (FOA) is used to analyze the uncertainty
and the contribution of each parameter on the output of 47 Tc formulas in
Kasilian and Amameh watersheds. The results show that among the 47 Tc
equations, equations which are based on watershed’s characteristics, rainfall
attributes and land cover-related coefficients such as Overton-Meadows,
ASCE, Akan, Kinematic-Wave, McCuen et al. and Izzard have relatively
high uncertainty and the average CV of these equations is about 45%. In
addition, equations that are based on only geomorphological parameters have
relatively low uncertainty (the average CV is about 16%). Further analysis of
the effects of parameter uncertainties on the Tc equations reveals that the
uncertainty associated with rainfall attributes and land cover-related
coefficients have great impacts on results of the Tc equations and the
uncertainty caused by these factors in humid regions relative to dry/ semi dry
regions is different. Moreover, in the geomorphological-based equations, the
uncertainty caused by streams delineation threshold is approximately 3-6
times of scale effects’ uncertainty.

Keywords: Time of Concentration, Uncertainty Analysis, First Order Analysis,

Scale Effects

Introduction semi-empirical and analytical formulas. So far, several Tc

equations have been developed that each of them resulted
Most of the hydrological analyses require time- from studies performed in specific regions and hence,
dependent parameters which among them time of researchers are often confused by the many equations and
concentration (Tc) is frequently used parameter (McCuen et often select a method without evaluating and comparing
al., 1984; Wong, 2009). It is defined as the time required for its accuracy and uncertainty with other formulas (McCuen
runoff to travel from the most remote point to the outlet of a et al., 1984; Wong and Asce 2005). Most of the developed
catchment (McCuen et al., 1984; Haan et al., 1994) Tc equations usually are based on two different
https://en.wikipedia.org/wiki/Time_of_concentration. This approaches. In the first approach, Tc is the time needed for
parameter reflects the speed at which the watershed water to travel from the hydraulically or physically most
responds to rainfall events (Povlovic and Moglen, 2008) distant point from the catchment’s outlet (Kirpich, 1940;
and therefore play an important role in estimating peak Singh, 1988; Fang et al., 2008). Equations are based on
floods and hydrologic designs. According to the this approach depend on geomorphological characteristics,
significance of time of concentration, researchers and such as flow path length, flow path slope, catchment area,
hydrologist have developed a huge number of empirical, rainfall attributes and land cover-related- coefficients (e.g.,

© 2018 Asghar Azizian. This open access article is distributed under a Creative Commons Attribution (CC-BY) 3.0
license.
Asghar Azizian / American Journal of Engineering and Applied Sciences 2018, 11 (1): 327.341
DOI: 10.3844/ajeassp.2018.327.341

manning’s roughness coefficient, curve number, and all independent and dependent variables are the Second
retardance factors). In the second approach, Tc is the time Moment Variables (SMV), which means that the behavior
distance between the end of effective rainfall and the of any SMV is completely described by its mean and
inflection point of the hydrograph’s falling limb (McCuen standard deviation (Bobba et al., 1996). According to the
et al., 1984; Fang et al., 2007; de Almeida et al., 2016). simplicity of this method, it has been applied to many
Formulas fall into the first approach, due to depending on environmental simulators, e.g. hydrological models, flood
several parameters and factors such as map scale, DEM levee design, dam overtopping assessment and groundwater
resolution, streams delineation threshold and land cover- pollution models (Cheng et al., 2013). As mentioned, a few
related-coefficients, lead to a great level of uncertainty. studies have carried out to determine the uncertainty and
How to assess the uncertainties in hydrological model sensitivity of Tc equations and most of them have focused
parameters and their impacts on the uncertainty of model only on some formulas. Thus, the aim of this study is to
simulations has always been a topic of great interest determine the uncertainty of about 47 Tc equations and
(Zhang et al., 2014). The quantitative evaluation of identify the contribution of all input parameters, such as
parameter uncertainty and its influence on the uncertainty rainfall intensity, land cover-related-coefficients, the scale
of hydrological model simulations is critical in reducing of topographic and land cover maps, DEM resolution and
the uncertainty of these simulations and in assessing their streams delineation threshold, on the output uncertainty.
effectiveness (Christiaens and Feyen, 2000; Bastola et al.,
2008; Hughes et al., 2010). So far, limited studies have Materials and Methods
been carried out to determine the uncertainty and
sensitivity of Tc formulas. For example, USWRC (1981) Structure of Tc Equations
reported that the coefficient of variation of Tc varied from The accuracy of hydrological designs is sensitive to the
2 to 155%. Kosari et al. (2010) by focusing on four Tc accuracy of the estimated Tc (Salimi et al., 2016). In this
formulas (Kirpich, BransbyWilliams, California, and Research 47 empirical/semi-empirical and analytical
SCS) assessed the sensitivity of them to input parameters equations, presented in Table 1, are used to investigate the
and concluded that in low slopes the effect of slope uncertainty analysis. As it can be seen, these equations
parameter is more considerable than other parameters, but depend on several inputs, but the geomorphological
by increasing of slope the importance of other parameters parameters, rainfall attributes, and land cover coefficients
such as river length and SCS curve number will be play an important role. The average adopted values of land
increased. Wong and Asce (2005) by assessing seven Tc cover-related-coefficients and parameters in the selected
formulas on two surfaces: concrete and grass, stated that equations are summarized in Table 2.
Chen and Wong formula has the best accuracy for both
surfaces. Grison et al. (2008) analyzed 19 observed Extraction of Geomorphological Parameters Using
hydrograph in Pequeno catchment, Brazil and found that QGIS
due to the uncertainties in determining the end of rainfall Estimation of Tc requires some watershed’s
event and the inflection point of hydrograph’s falling parameters that should be obtained from topographic
limb, there is about 30% uncertainty in estimating Tc maps or Digital Elevation Models (DEMs). By
value. McCuen (2009) by using Monte Carlo method development of Geographical Information System (GIS)
found that the uncertainty of the velocity method for small over the last years, especially in hydrological modeling,
and large watersheds is about 39 and 59%, respectively. several algorithms have been developed for automatic
Besides, he showed that the accuracy of computed Tc is extraction of geomorphological characteristics. In order to
largely controlled by the accuracy of roughness extract the watershed’s characteristics in this study, DEMs
coefficient. Azizian and Shokoohi (2014; 2015a) with different scales (SRTM 90m, ASTER 30m and
investigated the scale effects and streams delineation topographic map with a scale of 1:25000) and different
threshold on geomorphological parameters on the resolutions (varies between 30 and 300 m) are used in
performance of KW-GIUH model and found that the QGIS environ as an open source software. To extract
coefficients of variation of river length, river mean slope streams network it is necessary to determine precisely a
and time of concentration is about 18.8, 19.3 and 21%. threshold, which is the percent of watershed total cells
There are several statistical and mathematical methods poured into the target cell. In some GIS extensions such
that can be used to propagate input uncertainties through as ArcHydro and HEC-GeoHMS, the default value for
the model into output uncertainties, of which the First- streams delineation threshold is equal 1% of catchment
Order-Analysis (FOA) proposed by Tung and Mays area or 1% of maximum flow accumulation grid
(1981) is one of the most common methods. The (Azizian and Shokoohi, 2014; 2015a). Therefore,
procedure is based on first-order-analysis terms in the focusing on this value, the thresholds of 0.25, 0.5, 1, 2
Taylor series expansion of the dependent variable about its and 3% are employed for the extraction of streams
mean value with respect to one or more independent network. Meanwhile, d8 algorithm used as the best
variables. The major assumption in this procedure is that efficient method for flow tracing (Tarboton, 1991).

328
Asghar Azizian / American Journal of Engineering and Applied Sciences 2018, 11 (1): 327.341
DOI: 10.3844/ajeassp.2018.327.341

Table 1: Summary of the selected time of concentration equations

Name Formula Remark
Arizona DOT (ADOT 1993) Tc = 0.0097956 A0.1 (1000 L)0.25 L0.25
ca S
−0.2 Data of agricultural basins
ASCE (Morgali and Linsley, 1965) Tc = 7.2983 L0.6 n 0.6 i −0.4 S −0.2 Analysis of the kinematic wave (L<0,09 km)
Tc = 7.354 L0.6 n0.6 S −0.3 ( i − K s )
−0.4
Akan and Houghtalen (2003) + Obtained from simultaneously solve the kinematic
3.1K s1.33 H f Pf (1 − Si ) * i −2.33
wave equation and green-ampt infiltration method
Basso (Eslamian and Mehrabi, 2005) Tc = 0.957 L1.155 H m−0.385 N/A
Bransby Williams (ASDOT, 1995) Tc = 0.605 L (100 S )−0.2 A−0.1 Specially recommended to rural basins
California Culverts Practice (CDH, 1960) Tc = 0.95 L1.155 H −0.385 Data of small mountain basins in the USA
Carter (1961) Tc = 0.0977 L0.6 S −0.2 Data of an urban basin in the USA (A < 20.72 km2) and
(S<0.005)
Chow (1988) Tc = 0.1602 L0.64 S −0.32 Data of 20 rural basins in the USA (0.01–18.5 km2) and
(0.0051<S<0.09)
Corps of Engineers (Linsley et al, 1977) Tc = 0.191 L0.76 S −0.19 Data of 25 rural basins in the USA (A≤12 km2)
Desbordes (Silveira, 2005) Tc = 0.0869 A0.3039 S −0.3832 Aimp
−0.4523
N/A
DNOS (Silveira, 2005) Tc = 0.419k A L S−1 0.2 0.2 −0.4 Data of 6 rural basins in the USA (A<0.45 km2) and
(0.03<S<0.1)
Dooge (1973) Tc = 0.365 A0.41S −0.17 Data of 10 rural basins in Ireland (145 - 948 km2)
Epsey (Hotchkiss and McCallum, 1995) Tc = 6.89 L0.36 S −0.18 Data of 11 rural basins in the USA
Epsey and Winslow (McCuen, 1998) Tc = 0.45 Φ L0.29 S −0.11i −0.6 Obtained from 17 rural and urbanized basins range
from 1 to 35 mi2
Egelson (1962) Tc = 0.275 n L R −0.667
S −0.5 Calibrated from data for catchments less than 8 mi2
FAA (1970) Tc = 0.3788(1.1 − C ) L0.5 S −0.332 Data of airports’ drainage
Flavell (1983) Tc = 2.31 A0.54 Obtained from observed times of hydrograph rise on
basins in the south west of Western Australia
Morgali and Linsley (1965) Tc = 7.354 n 0.6 L0.6 S −0.3i −0.4 For small urban areas with drainage areas less than 10
or 12 acres and useful for drainage is basically planar
Ragan and Duru (1972) Tc = 7.275 n0.6 L0.6 S −0.3i −0.4 Based on the kinematic wave equation
Ribeiro (1961) Tc = 0.267 (1.05 − 0.2 p ) −1 LS −0.04 Data of 7 rural basins in the USA and a rural basin in
India (A<19000 km2) and (0.03<S<0.1)
Giandotti (1934) Tc = (4 A + 1.5 L) / (0.8 H m ) Data of basins in central and northern Italy (170-70000 km2)
Haktanir and Sezen (1990) Tc = 0.7473 L0.841 Data of 10 basins in Turkey (11-9867km²)
Izzard (1946) Tc = 85.5 ( i / 36286 + Cr ) i −0.667 L0.33 S −0.333 Based on a series of laboratory experiments by the Bureau of
public roads. This method is designed for applications in
which the product of intensity and flow length is than 500
Johnstone and Cross (1949) Tc = 0.4623 L0.5 S −0.25 Data of 19 rural basins in the USA (64.8 – 4206.1 km2)
Kerby (1959) Hathaway (1945) Tc = 0.6061 N 0.47 L0.47 S −0.234 Analysis of overland flow in experimental surfaces
(L<0.37 km)
Kinematic wave Tc = 7.35 n 0.6 i −0.4 L0.6 S −0.2 Analysis of overland flow in experimental surfaces
(Kibler and Aron, 1983) (L<0.03 km)
Kirpich-Tennessee (Kirpich, 1940) Tc = 0.0653 L0.77 S −0.385 Data of small watersheds in Tennessee and Pennsylvania
(0.004 – 0.453 km2) and (0.03<S<0.1)
Kirpich-Pennsylvania (Kirpich, 1940) Tc = 0.01104 L0.77 S −0.5
McCuen et al. (1984) Tc = 2.2535 i −0.7164 L0.5552 S −0.207 Starting from data of 48 urban basins in the USA (0.4-6
km2) and (0.0007<S<0.03)
Papadakis and Kazan (1986) Tc = (2.1539 n 0.52 0.5
L ) / (i 0.38
S 0.31
) Data of 84 small rural basins in the USA (A <5 km2)
Pasini (1914) Tc = 0.108 A0.332 L0.332 S −0.5 Data of rural basins in Italy
Pickering (Mata-Lima et al. 2007) Tc = 0.9482 L1.155 H −0.385 Equivalent to Kirpich’s
Picking (Silveira 2005) Tc = 0.0883 L0.667 S −0.332 Data of rural basins
Pilgrim and Mac Dermott (1982) Tc = 0.76 A0.38 Developed from 96 basins in eastern New South Wales
Overton and Meadows (1976) Tc = 22.92 n L S 0.8 0.8 −0.4
P2
−0.5 N/A
SCS Lag (Mockus, 1961) Tc = 0.057 ( (1000 / CN ) − 9 )
0.7
L0.8 S −0.5 Data of 24 rural basins in the USA (A <8 km2)
SCS Ave Velocity (NRCS, 1972; 1986) Tc = 0.278 ∑ L / KS 0.5 This method relates watershed slope and surface to flow velocity

329
Asghar Azizian / American Journal of Engineering and Applied Sciences 2018, 11 (1): 327.341
DOI: 10.3844/ajeassp.2018.327.341

Table 1: Continued
Simas and Hawkins (2002) Tc = 0.322 A 0.594 L−0.594 S −0.15 S scs
0.312 Data of 168 basins in the USA (0.001 - 14 km2)
Schaake et al. (1967) Tc = 0.0828 L0.24 S −0.16 A imp
−0.26 N/A
Sheridan (1994) Tc = 2.20 L0.92 nine flatland watersheds located in Georgia and Florida
and ranging in size from 2.62 to 334.34 km2
Temez (1987) Tc = 0.3 L S0.76 −0.19 Data of natural basins in Spain
Van Sickle (1962) Tc = 0.0081 L0.13
t L0.13
m S
−0.065 Calibrated from data collected in Houston, with drainage
areas less than 36 mi2
Ventura (Mata-Lima et al., 2007) Tc = 4 A0.5 L0.5 H −0.5 Data of rural basins in Italy
Williams (1922) Tc = 0.272 A0.4 L D −1S −0.2 Data of basins in India (A <129,5 km2)
Woolhiser and Liggett’s Tc = 7.3015 n0.6 L0.6 i −0.4 S −0.5 Based on the theory of kinematic wave
(1967, Wong and Asce, 2005)
Yen and Chow’s (1983) Tc = 1.2 n 0.6 L0.6 S −0.5 Based on the theory of the kinematic wave
Zomorodi (2005) Tc = 0.1101 n 0.75 L0.75 S −0.375 Modified version of NRCS method
Note: Tc: time of concentration (hr), A: catchment area (Km²), C: runoff coefficient, CN: SCS Curve-number, D: equivalent diameter
of the catchment (Km), H: quota difference between the ends of the main channel (m), Hm: mean altitude in the catchment (m), i:
rainfall intensity (mm/h), K: coefficient of the type of surface, L: flow path length (Km), Lca: mean length starting from the
concentration spot along the L up to the spot where L is perpendicular to the centroid of the catchment (m), N: retardance coefficient,
n: Manning’s roughness coefficient, p: relation between the vegetation cover and the total area of the basin, S: flow path slope
(m/m), Sscs: maximum capacity of retention (mm), Cr: retardance coefficient (ranges from 0.007 to 0.06), I: percent of impervious
area, P2: maximum 24 h rainfall with return period of 2 (mm), Pf: soil suction head (mm), Si: saturated moisture content, Hf: soil
porosity and Ks: saturated hydraulic conductivity (mm/hr), N/A: Not available data

Table 2: Adopted values of the coefficients and parameters in the selected equations
Formula Coefficient Adopted value
Akan Pf 150.200
Akan Hf 0.300
Akan Ks 3.400
Akan Si 0.250
DNOS k 4.450
Epsey-Winslow Φ 0.150
FAA C 0.190
Izzard Cr 0.014
Kerby-Hathaway N 0.330
Kinematic Wave/Eagleson/ASCE/Papadakis-Kazan/Overton-Meadows/Yen n 0.023
-Chow’s/Zomorodi/Woolhiser-Liggett’s/Ragan-Duru/Morgali-Linsley/Akan
Overton-Meadows P2 14.500
Ribeiro p 0.240
Schaake/Desbordes/Epsey Aimp 0.110
SCS Lag/Simas-Hawkins CN 78.200

Uncertainty Analysis Method specifically, FOA method enables one to estimate the
mean and variance of a random variable that is
An uncertainty analysis is not the same as a sensitivity functionally related to several other variables, some of
analysis. An uncertainty analysis attempts to describe the them are random. By using first-order analysis, the
entire set of possible outcomes, together with their combined effect of uncertainty in a model formulation, as
associated probabilities of occurrence (Loucks et al., well as the use of uncertain parameters, can be assessed.
2005). In the design and analysis of hydro systems, many Consider a random variable y that is a function of k
quantities of interest are functionally related to a number random variables (Mays and Tung, 2002):
of variables, some of which are subject to uncertainty. A
rather useful technique for the approximation of y = g ( x1 , x2 ,..., xk ) (1)
uncertainties is the First-Order-Analysis (FOA) of
uncertainties, sometimes called the delta method. This This function can be a deterministic equation such as
method is very popular in many fields of engineering rational formula, time of concentration equations,
because of its relative ease in application to a wide range Manning's equation or it can be a complex model that
of problems (Mays and Tung, 2002). FOA method is used must be solved on a computer (Mays and Tung, 2002).
to estimate the uncertainty in a deterministic model The objective is to treat a deterministic model that has
formulation involving parameters that are uncertain. More uncertain inputs in order to determine the effect of the

330
Asghar Azizian / American Journal of Engineering and Applied Sciences 2018, 11 (1): 327.341
DOI: 10.3844/ajeassp.2018.327.341

uncertain parameters x1, x2,..., xk on the model output y. (e.g., the 2-year, 2-h rainfall intensity) and land
equation (1) can be expressed as y = g(x) where x = x1, cover-related-factors (e.g., retardance and Manning's
x2,...xk. Through a Taylor series expansion about k random roughness coefficient)
variables, ignoring the second and higher order terms, the • Case II: Considering the uncertainty associated with
approximate form could be obtained via Equation 2: geomorphological characteristics and their
contribution on the output uncertainty.
k
 ∂g  • Case III: Considering the uncertainty associated
y ≈ g (x) + ∑   ( X i − xi ) (2)
i =1  ∂xi  x
with rainfall attributes (e.g., the 2-year,
2-h rainfall intensity) and land cover-related-factors
The derivatives [∂g/∂xi] are the sensitivity and their contribution on the output uncertainty.
coefficients that represent the rate of change of the
function value g(x) at x = x . Assuming that the k Study Area
random variables are independent and then the variance In this research, Kasilian and Amameh watersheds
of y is approximated as: with different climate and topographic conditions are
used. Both of them are located in the north of Iran in
σ y2 = Var[ y ] = ∑ ai2σ x2 i
(3) regions with humid and dry climate, respectively.
Kasilian watershed, in the central Alborz mountain
And the coefficient of variation is Ωy: chain, is a mountainous area covered by forest. This
watershed with an area of 67 Km2 and a perimeter of
 k
0.5 37.8 Km is drained by the Kasilian River with a length
 2
2
 x of 17 Km. Amameh watershed, in the south Alborz
Ω y =  ∑ ai2  i  Ω xi  (4)
 τ =1  µ y   mountain chain, is one the sub-basins of Jajrood
 
watershed with an area of 37.2 Km2 and a perimeter of
where, ai = (∂g / ∂x) x . 29.1 km. The average slope of Kasilian and Amameh
watersheds is about 16.4 and 13.0%, respectively. In
By considering the general form of Tc equations, order to extract watersheds’ characteristics in this study,
shown in Equation 7, the total uncertainty of each Tc SRTM 90m, ASTER 30m DEMs and topographic maps
equation could be obtained via Equation 6: with a scale of 1:25000 are used to create DEMs with
different cell sizes (30 to 300m). Furthermore, for
Tc = a.S b Lc Ad nei f H g H mh (5) estimation of land cover-related-coefficients such as n,
C, K, N, and Cr map with scales of 1:50000, 1:100000,
ΩT2c = b 2Ω S2 + c 2Ω 2L + d 2Ω 2A + e 2Ω 2n + f 2Ωi2 + g 2Ω 2H + h 2Ω 2H m (6) 1:250000 and different recommended tables (e.g., Chow,
1964; McCuen, 2009) are used. The location of Kasilian
where, a, b, c, d, e, f, g, h are the exponents of each input and Amameh watersheds and their Digital Elevation
parameters. The dependency of Tc equations to different Models (DEMs) are shown in Fig. 1.
parameters which generally includes rainfall intensity, scale
of topographic and land cover maps, DEM resolution, DEM Results and Discussion
creation method (scale effects) and streams delineation Scale Effects (Map Scale and DEM Resolution) on
threshold impose significant uncertainties to the input the Geomorphological Parameters
parameters and the output of Tc equations. Several studies
have been carried out to show the scale effects on the Different thresholds (0.25-3%), various maps (Topo,
catchment’s geomorphological parameters (Wolock and SRTM, and ASTER DEMs) and several DEM
Price, 1994; Pradahan et al., 2008; Hancock, 2005; Azizian resolutions (25-300 meter) are used for extraction of
and Shokoohi, 2015a; 2015b) and therefore this issue is one geomorphological parameters that required for Tc
equations. Because of the huge number of extracted data,
of the main uncertainty sources. Moreover, land cover maps
only parameters that obtained in Kasilian catchment (at
with different scales lead to different values for land cover- different scales) at thresholds of 0.25 and 3% are shown
related-coefficients which are used in the most Tc equations. in Fig. 2-8. Findings illustrate that in all stream orders
Besides, rainfall attributes variations, due to playing an the mean slope of rivers and sub-basins that obtained
important role, should be considered as another uncertainty from Topo DEM are lower than that of from SRTM and
sources. In the present study, uncertainty analysis has been ASTER DEMs. For instance, the main channel slope
carried considering the following cases: based on Topo and SRTM DEMs at the threshold of
0.25% is about 5.3 and 6.5%, respectively. In addition, at
• Case I: Considering the uncertainty associated with the threshold of 3% the mentioned values are 1.9 and
all geomorphological characteristics (e.g., flow path 3.3%, respectively. Generally, the mean slope of the
length, flow path slope, area), rainfall attributes main channel varies between 1.9 and 6.5%.

331
Asghar Azizian / American Journal of Engineering and Applied Sciences 2018, 11 (1): 327.341
DOI: 10.3844/ajeassp.2018.327.341

Fig. 1: Location and the DEM of the study areas

0.20
Sc-Topo Sc-SRTM 0.34
Sb-Topo Sb-SRTM

Mean slope of subbasins (m/m)

Mean slope of stream (m/m)

0.16 0.30

0.26
0.12
0.22

0.18
0.08
0.14

0.04 0.10
1 2 3
Stream Order
Fig. 2: Effect of data resolution on the mean slope of streams and sub-basins of i-th order streams at threshold 0. 5% (DEM-50m)

0.35
Sc-Topo Sc-SRTM
0.13 Sb-Topo Sb-SRTM
Mean slope of subbasins (m/m)
Mean slope of stream (m/m)

0.31
0.10
0.27
0.08

0.23
0.05

0.03 0.19

0.00 0.15
1 2 3
Stream Order

Fig. 3: Effect of Data resolution on the mean slope of streams and sub-basins of i-th order streams at threshold 3% (DEM-50m)

332
Asghar Azizian / American Journal of Engineering and Applied Sciences 2018, 11 (1): 327.341
DOI: 10.3844/ajeassp.2018.327.341

L-Topo
10 L-ASTER
L-SRTM

Mean river length (km)

8

0
1 2 3 4
Stream Order

Fig. 4: Effect of data resolution on the mean length of i-th order streams at threshold 0.25% (DEM-50m)

L-Topo
4.5 L-ASTER
Mean river length (km)

L-SRTM

3.5

2.5

1.5
1 2 3
Stream Order

Fig. 5: Effect of data resolution on the mean length of i-th order streams at threshold 3% (DEM-50m)

DEM100m
10
DEM50m
Mean length of stream (km)

DEM25m
8 DEM300m

0
1 2 3 4
Stream Order

Fig. 6: Effect of DEM resolution on the mean length of i-th order streams at threshold 0.25%

333
Asghar Azizian / American Journal of Engineering and Applied Sciences 2018, 11 (1): 327.341
DOI: 10.3844/ajeassp.2018.327.341

DEM 25m
0.21
DEM 50m

Mean slope of stream (m/m)

DEM 100m
0.18
DEM 500m

0.15

0.12

0.09

0.06

0.03
1 2 3 4
Stream Order

Fig. 7: Effect of DEM resolution on the mean slope of i-th order streams at threshold 0.25%

Moreover, because of scale effects and different identifying the probability distributions the coefficient of
thresholds, the main channel length and mean slope of variation and mean values of all input parameters can be
sub-basins vary between 6.43-14.04 km and 22.1-31.2%, easily calculated. To keep the paper in reasonable extend
respectively. The slope of both overland and channel only the results that obtained in Kasilian catchment
parts affects the travel time of raindrop and flood wave presented in this section. The uncertainty of all Tc
velocity moving towards the basin outlet. Therefore, it formulas and the contribution of each parameter on the
can be expected that, due to the greater slope and smaller total uncertainty are shown in Fig. 8-11.
channel length of SRTM DEMs, the time of
Case 1
concentration will be greater than that estimated from
Topo and ASTER DEMs. Besides, when DEM resolution Results indicate that, based on CV value, Tc
gets coarser the information content of all maps with equations could be categorized into two different
different scales reaches to a constant value and this issue groups. Group1 contains the equations which depend
results the same watershed’s characteristics. As it can be on all parameters, while in group2 Tc formulas are
seen, different map scales and streams delineation only based on watershed’s characteristics. In the first
threshold have considerable effects on the values of group, the minimum, mean and maximum values of
geomorphological parameters and therefore one can uncertainty are about 36.7, 49.9 and 61.4%,
conclude that these factors are one the most important respectively. In this group, Overton-Meadows and
sources of uncertainty in estimating of Tc values. Yen-Chow equations result the highest and lowest
uncertainty, respectively. One of the most important
Uncertainty Analysis of Tc Equations factors that affect the CV value in this group is the
dependency to different parameters with a wide range
For estimation of uncertainty based on FOA method,
of variations. For example, the CV value of rainfall
Coefficient of Variation (CV), variance and mean values
intensity and Manning's roughness coefficient in
of each input parameter should be identified. According
Kasilian catchment is about 82 and 50%, respectively
to the statistical analysis, the probability distribution of and this issue leads to higher uncertainty. In contrast
flow path Length (L), flow path Slope (S) and the to group1, the uncertainty of second group’s equations
difference between the elevations of endpoints of river is relatively low and the maximum and minimum
(H) follow as Gama distribution, while rainfall intensity, values of CV in this group are about 28.4 and 0.5%,
Manning's roughness coefficient and most of the other respectively. In this group, Pilgrim-McDermott and
land cover-related-coefficients follow as log-normal Flavell formulas because of depending only on
distribution. These findings show broad agreement with catchment Area (A) and according to the low variation
McCuen (2009) studies that assumed all input of this parameter in different map scales, result low
parameters of velocity method follow a gamma uncertainty. Moreover, the exponent of watershed’s
distribution. Needless to say that the variation of the characteristics plays an important role on the uncertainty
catchment area and perimeter values in different data of Tc equations. In the next sections, the relationship
resolution and DEM resolution aren’t significant and between some parameters’ exponents and their
therefore they assumed as constant values. After contributions on the total uncertainty will be discussed.

334
Asghar Azizian / American Journal of Engineering and Applied Sciences 2018, 11 (1): 327.341
DOI: 10.3844/ajeassp.2018.327.341

Case 2 respectively). In the second group’s equations

geomorphological parameters, especially flow path
Findings reveal that the importance of
length and flow path slope, play an important role on
geomorphological parameters, such as flow path the output uncertainty. As shown in Fig. 9, the
Length (L) and flow path slope (Sc or Sb), on the total percentage of uncertainty that caused by flow path
uncertainty, is different in both groups (Fig. 9 and
length in the second group varies between 29 and
10). In the first group, the contribution of 100%. Also, in about 66% of all Tc equations (31
geomorphological characteristics is relatively low. For formulas) flow path length contributes more than 30%
instance, in Epsey-Winslow, ASCE and Eagleson
on the total uncertainty. Compared to flow path
equations the uncertainty associated with flow path length, flow path slope parameter leads to lower
length is about 1.0, 7.9 and 13.9% (Fig. 9), contribution and almost in about 50% of all Tc
respectively and the uncertainty caused by flow path
formulas it contributes lower than 20% (Fig. 10).
slope isn’t significant (about 0.5, 1.0 and 7.8%,
80
Max=61.4
70 Group1
Coefficient of Variation (%)

60
Group2
Min=36.7

50

Max=28.4
40

30

20
Min=0.5

10

0
Morgali-Linsley
KinematicWave

VanSickle
Dooge

Simas-Hawkins
Desbordes

Williams
Zomorodi

SCS-Vel

Woolhiser-Ligget

Pilgrim

Giandotti

Pasini
Yen-Chow

ASCE

Epsey-Winslow

ArizonaDOT
FAA

Chow
HEC
Papadakis-Kazan

Rgan-Duru

Izzard
Eagleson
McCuen
Akan

Haktanir-Sezen
Sheridan
Kirpich-Ten

Kirpich-Pen
DNOS
SCS-Lag

Epsey

Ventura
Kerbay-Hathaway

Picking

Basso
Pickering
California
Ribeiro
Overton-Meadows

Schaake

Johnstone-Cross
Carter

Temez
Flavell

BransbyWilliam

Fig. 8: Coefficient of Variation of all Tc Equations

Max= 100
120
Contribution of Length Parameter (%)

100

80

60
Min=29

40
Max= 16

20
Min=1

0
Akan
DNOS
Izzard
McCuen
Papadakis-Kazan

Kirpich-Pen

Kirpich-Ten

Haktanir-Sezen
Sheridan
Morgali-Linsley

SCS-Lag

Kerbay-Hathaway

Ventura

Epsey

Picking

Pickering

Basso
KinematicWave
Overton-Meadows

Schaake

Johnstone-Cross

Williams
Simas-Hawkins
Carter
VanSickle

Temez
Woolhiser-Ligget

BransbyWilliam
Zomorodi
Pasini

SCS-Vel

Giandotti
Epsey-Winslow

ASCE

Yen-Chow

FAA

ArizonaDOT
Chow

HEC
Rgan-Duru

Eagleson

California

Ribeiro

Fig. 9: The Contribution of Flow Path Length on the Total Uncertainty of Tc Equations

335
Asghar Azizian / American Journal of Engineering and Applied Sciences 2018, 11 (1): 327.341
DOI: 10.3844/ajeassp.2018.327.341

140

Max=100.0
Contribution of Slope Parameter (%)
120

100

80

60

Min=29.4
Mx=20.2
40
Min=0.5

20

Izzard
KinematicWave

Zomorodi

Temez

VanSickle
Carter

SCS-Vel

Pasini
Dooge
Ribeiro

ASCE

Morgali-Linsley

BransbyWilliam

Woolhiser-Ligget

Kerbay-Hathaway

Picking

Epsey

ArizonaDOT

SCS-Lag
Epsey-Winslow

McCuen

Akan

Rgan-Duru
Overton-Meadows
Papadakis-Kazan
Williams
Eagleson

Simas-Hawkins

HEC

Yen-Chow
DNOS

Chow

Johnstone-Cross
Kirpich-Ten
FAA

Kirpich-Pen

Desbordes
Schaake
Fig. 10: The Contribution of flow path slope on the total uncertainty of Tc equations

120
U= 45.576ln(E) + 89.45
R² = 0.7691
100
Uncertainty by Length (%)

80

60

40

20

0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
Exponent
(a)

18

16
Uncertainty by Length (%)

14
U = 16.656E1.8548
12 R² = 0.8554

10

0
0 0.2 0.4 0.6 0.8 1 1.2
Exponent
(b)

Fig. 11: Variation of Flow Path Length Exponent and its Contribution on the Output Uncertainty a) Group 1 b) Group 2

336
Asghar Azizian / American Journal of Engineering and Applied Sciences 2018, 11 (1): 327.341
DOI: 10.3844/ajeassp.2018.327.341

There are a huge number of reasons that affect the parameter contributes to 76.8 and 71.6% in estimating
contribution of each parameter on the output uncertainty, Tc value by Zomorodi and Eagleson formulas.
but the most important one is the value of parameters’ Besides, in some equations such as Epsey-Winslow
exponent especially flow path length’s exponent. The and Izzard, because of the higher variation in rainfall
findings from the research illustrate that by increasing intensity especially in humid regions, the effects of
the exponent of each parameter the uncertainty of Tc land cover coefficients are not considerable and the
equations and the contribution of that parameter maximum contribution is lower than 14%.
increases. For example, in group 2 there is a significant In comparison to geomorphological-based formulas
relationship between the exponent of flow path length (group 2), the uncertainty caused by rainfall intensity and
parameter and its contribution on the total uncertainty. Manning's roughness coefficient is more considerable. It
As shown in Fig. 11, by increasing the flow path length’s should be noted that the contribution of rainfall intensity
exponent from 0.2 to 1.155, the uncertainty associated and Manning's roughness (group 1) on the output
with this parameter increases to reach the maximum uncertainty are completely different in both catchments
value (100%). In addition, in the group1, although the and affected by climate conditions and land cover
exponent varies between 0.3 and 1, the maximum value variety. Kasilian catchment, due to being located in
of flow path length’s contribution is lower than 15% humid climate, experiences remarkable variations of
and this means that the watershed’s features play a rainfall intensity and land cover types rather than Amameh
minor role in the estimation of Tc value with the first catchment. The results, shown in Fig. 13, express that the
group’s equations. average uncertainty associated with rainfall intensity for
equations of group2 is about 25% higher than Manning's
Case 3 roughness in Kasilian catchment, while in Amameh
As mentioned, land cover-related-coefficients play catchment the uncertainty caused by Manning's roughness
a key role in estimating T c and other hydrological coefficient is about 2 times of rainfall intensity.
processes. The values of these coefficients depend on As an overall view, one can deduce that the most
the scale of map and field observations and hence, for effective parameters which contribute on the
a specific region different values could be obtained. uncertainty of Tc equations in humid regions are:
Investigation of the effects of these coefficients on the rainfall intensity (i), Manning's roughness coefficient
output uncertainty demonstrates that if there is no (n) and other land cover coefficients, flow path
accurate estimation for them it leads to significant Length (L), the difference elevation between endpoint
sensitivity and uncertainty in estimating Tc value. The of main river (H), flow path slope (S c and S b) and
effect of land cover coefficients on the uncertainty of catchment Area (A), respectively. However, in dry
Tc equations is shown in Fig. 12. It can be concluded and semi-arid regions, because of the low variations
from this figure that inaccurate estimation of land of rainfall intensity, Manning's roughness and other
cover coefficients leads to higher uncertainty. For land cover coefficients are more sensitive and cause
instance, the uncertainty in manning’s roughness considerable uncertainty in estimating Tc value.

Zomorodi 76.8
Eagleson 71.6
DNOS 67.4
Yen-Chow 66.8
Kerbay-Hathaway 45.3
Desbordes 33.8
Schaake 32.7
Simas-Hawkins 16.6
SCS-Lag 14.9
Epsey-Winslow 13.9
FAA 12.2
Izzard 10.8
SCS-Vel 6.4
Ribeiro 0.6
0 10 20 30 40 50 60 70 80 90 100
Uncertainty Associated with LandCover Coefficients(%)

Fig. 12: Contribution of land cover coefficients on the output uncertainty

337
Asghar Azizian / American Journal of Engineering and Applied Sciences 2018, 11 (1): 327.341
DOI: 10.3844/ajeassp.2018.327.341

Uncertainty Associated with Scale Effects (Data delineation threshold, shown in Fig. 14, is about 3-6
Resolution and DEM Resolution) and Streams times greater than other factors. Thus, the streams
Delineation Threshold delineation threshold emerges to be the biggest
contributor to the uncertainty when only the
All of the Tc equations depend on geomorphological
geomorphological parameters are considered. Hence, the
characteristics and hence affected by scale effects and
value of this factor needs to be more considered with
streams delineation threshold. The question brings up
precision for accurate estimation of Tc especially in
here is that ″ which of them makes more sensitivity and
ungauged catchments. Needless to say that in the first
considerable uncertainty in estimating Tc value? ″. The
group’s equations, because of the importance of rainfall
results of the analysis demonstrated that in all Tc
attributes and land cover coefficients, the scale effects
equations which only depend on geomorphological
and other factors is not remarkable and could be ignored.
parameters the uncertainty associated with streams

ASCE
60

50
Overton-
Akan
40 Meadows

30

20
10
Papadakis- Woolhiser-
0
Kazan Ligget

KinematicWave Morgali-Linsley

Rgan-Duru

Rainfall_Intensity Manning_Roughness
(a)

ASCE
70
60
Overton-
Akan 50 Meadows
40
30
20
10
Papadakis- Woolhiser-
0
Kazan Ligget

KinematicWave Morgali-Linsley

Rgan-Duru

Rainfall_Intensity Manning_Roughness
(b)

Fig. 13: Uncertainty associated with rainfall intensity and Manning's roughness coefficient (a) Kasilian Catchment (b) Amameh Catchment

338
Asghar Azizian / American Journal of Engineering and Applied Sciences 2018, 11 (1): 327.341
DOI: 10.3844/ajeassp.2018.327.341

69%

Data_Resolution
19%
DEM_Resolution
12% Threshold

Fig. 14: Uncertainty associated with the scale effects and streams delineation threshold

Conclusion Tc value, especially in ungauged watersheds. In

addition, to reduce the confidence interval of estimated
The important issues focused on in this study are the Tc, the parameters leading to greater uncertainty need to
assessment and uncertainty analysis of time of be estimated more accurately.
concentration (Tc) equations, the uncertainty associated
with different map scales and different inputs. Since
uncertainty is considered to be a serious problem in
Acknowledgment
hydrological model parameters and their impacts on the I thank IKIU University and anonymous reviewers
uncertainty of model simulations, this study investigates for their useful comments.
the factors that cause uncertainty on the estimation of Tc
values in Kasilian and Amameh catchments. The results Ethics
of the analyses show that based on Coefficient of
Variation (CV), all Tc equations could be categorized The present study was approved by IKIU University.
into two groups. The first group’s equations due to
depending on several inputs, such as rainfall attributes, References
watershed’s characteristic and land cover related
coefficients, result remarkable uncertainty and the Akan, A.O. and R.J. Houghtalen, 2003. Urban
minimum and maximum uncertainty of equations fall hydrology, hydraulics and storm water quality:
into this group is about 36.7 and 61.4%, respectively. Engineering applications and computer modeling.
Compared to the first group, the uncertainty of second 1st Edn., John Wiley and Sons, Hoboken, ISBN-10:
group’s formulas is slightly low and the average and 0471431583, pp: 373.
maximum CV value is about 16.5 and 28.4%, ADOT, 1993. Highway Drainage Design Manual
respectively. The major factors causing uncertainty in Hydrology. 1st Edn., Arizona Department of
the estimation of Tc value are factors due to maps scale, Transportation, Phoenix.
which are used for extraction of watershed’s parameters ASDOT, 1995. Alaska Highway Drainage Manual.1st Edn.,
and land cover coefficients, DEM resolution, streams Alaska State Department of Transportation, Alaska.
delineation threshold and rainfall attributes. Findings of Azizian, A. and A.R. Shokoohi, 2014. DEM resolution
this study indicate that the rainfall attributes play an and stream delineation threshold effects on the
important role in estimation of Tc especially in humid results of geomorphologic-based rainfall runoff
region with a wide range of rainfall variations. But in
models. Turkish J. Eng. Env. Sci., 38: 64-78.
dry and semi-dry regions the contribution of land
Azizian, A. and A.R. Shokoohi, 2015a. Effects of data
cover-related coefficients on the uncertainty of Tc
equations is higher than rainfall attributes and other resolution and stream delineation threshold effects
factors. This research also shows that, for equations on the results of a Kinematic Wave based GIUH
are only based on geomorphological parameters, the model. J. Water S.A, 4: 61-70.
uncertainty associated with streams delineation Azizian, A. and A.R. Shokoohi, 2015b. Investigation of
threshold is about 3-6 times of DEM resolution and the effects of DEM creation methods on the
map scale (data resolution). Needless to say that in the performance of a semi distributed model:
second group’s equations, because of the importance TOPMODEL. J. Hydro. Eng.
of rainfall attributes and land cover coefficients, the Bastola, S., H. Ishidaira and K. Takeuchi, 2008.
effects of these factors is not remarkable and could be Regionalization of hydrological model parameters
ignored. The results of this research can be used as an under parameter uncertainty: A case study involving
appropriate guidance for engineers and researchers TOPMODEL and basins across the globe. J.
who are looking for a best formula for estimation of Hydrol., 357: 188-206.

339
Asghar Azizian / American Journal of Engineering and Applied Sciences 2018, 11 (1): 327.341
DOI: 10.3844/ajeassp.2018.327.341

Bobba, A.G., V.P. Singh and L. Bengtsson, 1996. Haktanir, T. and N. Sezen, 1990. Suitability of two-
Application of first-order and monte carlo analysis parameter gamma and three-parameter beta
in watershed water quality models. Water Resour. distributions as synthetic unit hydrographs in
Manag., 10: 219-240. Anatolia. Hydrol. Sci. J., 35: 167-184.
CDH, 1960. California Culvert Practice. 2nd Edn., Calif. Hancock, G.R., 2005. The use of digital elevation
State Print. Office, Sacramento, pp: 119. models in the identification and characterization of
Carter, R.W., 1961. Magnitude and frequency of floods catchments over different grid scales. Hydrol. Proc.,
in suburban areas. US Geologi. Survey, Reston, VA. 45: 231-243.
Chen, Y., Y. Hsu, K.T. Kuo, 2013. Uncertainty in the Haan, C.T., Barfield, B.J., and Hayes, J.C. 1994. Design
methods flood discharge measurement. Water Hydrology and Sedimentology for Small
Catchments. Elsevier, Amsterdam, Netherland.
Resour. Manag., 27: 153-167.
Hathaway, G.A., 1945. Design of drainage facilities.
Chow, V.T., 1964. Handbook of Applied Hydrology. 1st
Trans. Am. Soc. Civ. Eng., 110: 697-733.
Edn., McGraw-Hill, New York. Hotchkiss, R.H. and B.E. Mccallum, 1995. Peak
Chow, V.T., D.R. Maidment and L.W. Mays, 1988. discharge for small agricultural watersheds. J.
Applied Hydrology. 1st Edn., McGraw-Hill, New Hydraulic Eng., 121: 36-47.
York. Hughes, D.A., E. Kapangaziwiri and T. Sawunyama,
Christiaens, K. and J. Feyen, 2000. The influence of 2010. Hydrological model uncertainty assessment in
different methods to derive soil hydraulic properties southern Africa. J Hydrol., 378: 221-232.
on the uncertainty of various model outputs of a Izzard, C.F., 1946. Hydraulics of runoff from developed
distributed hydrological model. Phys. Chem. Earth surfaces. Proceedings of the 26th Annual Meetings
Pt B, 25: 679-683. of the Highway Research Board, National Research
de Almeida, I.K., A.K. Almeida, J.L. Steffen and S.T. Council, (NRC’ 46), Washington.
Alves, 2016. Model for estimating the time of Johnstone, D. and W.P. Cross, 1949. Elements of Applied
concentration in watersheds. Water Resources Hydrology. 1st Edn., Ronald Press, New York.
Manage., 30: 4083-4096. Kerby, W.S., 1959. Time of concentration for overland
Dooge, J.C.I., 1973. Linear Theory of Hydrologic flow. Civ. Eng., 26: 60-60.
Systems. 1st Edn., Agricultural Research Service. Kibler, D.F. and G. Aron, 1983. Evaluation of Tc
Egelson, P.S., 1962. Unit hydrograph characteristics for methods for urban watersheds. Proceedings of the
sewered areas. j.hydr.Div, ASCE. 88 (HY2) Conference on Frontiers in Hydraulic Engineering,
Eslamian, S. and A. Mehrabi, 2005. Determination of (FHE’ 83), New York, pp: 553-558.
experimental relations in estimation of concentration Kirpich, Z.P., 1940. Time of concentration of small
time in mountainous watershed basins. J. Natural agricultural watersheds. Civil Eng., 10: 362-368.
Resources Agric. Sci., 12: 23-34. Kosari, M.R., M.A. Saremi Nayeeni, M. Tazeh and F.M.
Fang, X., D.B. Thompson, T. Cleveland and P. Pradhan, Rahim, 2010. Sensitivity analysis of four
2007. Variations of time of concentration estimates concentration time estimation methods in watershed
using NRCS velocity method. J. Irrigation Drainage basins. J. Khoshkboom, 1: 43-55.
Eng., 133: 314-322. DOI: 10.1061/(ASCE)0733- Linsley, R.K., M.A. Kohler, J.L. Paulhus, M.F. Serra and
9437(2007)133:4(314) F.D. Aparicio, 1977. Hidrologia para ingenieros.
Fang, X., D.B. Thompson, T.G. Cleveland, P. Pradhan McGraw-Hill, pp: 386.
and R. Malla, 2008. Time of concentration estimated Loucks, D.P., J.R. Stedinger, J.P.M. Dijkman and M.T.
using watershed parameters determined by Villars, 2005. Water resources systems planing and
automated and manual methods. J. Irrigation mangement an introduction to methods, models and
Drainage Eng., 134: 202-211. application. Unesco Publishing
Mata-Lima, H., H. Vargas, J. Carvalho, M. Gonçalves
FAA, 1970. Advisory circular on airport drainage. US
and H. Caetano et al., 2007. Comportamento
Federal Aviation Administration
hidrolَgico de bacias hidrogr‫ل‬ficas:integraç‫م‬o de
Flavell, D.J., 1983. The rational method applied to small métodos e aplicaç‫م‬o a um estudo de caso. Revista
rural catchments in the south west of western Escola de Minas.
Australia. Civ. Eng. Trans. Inst. Engrs. Aut., 25: Mays, L.W. and Y.K. Tung, 2002. Hydrosystems
121-127. Engineering and Management. 1st Edn., Water
Giandotti, M., 1934. Previsione delle piene e delle magre Resources Publications, Highlands Ranch, ISBN-10:
dei corsi d’acqua. Ministero LL.PP., Memorie e 1887201327, pp: 530.
studi idrografici 8(2) Servizio Idrografico Italiano, McCuen, R., S. Wong and W. Rawls, 1984. Estimating
Rome (in Italian) urban time of concentration. J. Hydraulic Eng. Am.
Grison, F., M. Kobiyama and R. Da Silva, 2008. Society Civil Engineers, 110: 887-904.
Conceptual investigation of time of concentration. McCuen, R.H., 2009. Uncertainty analyses of watershed
From Headwaters Ocean. time parameters. J. Hydraulic Eng., 14: 490-498.

340
Asghar Azizian / American Journal of Engineering and Applied Sciences 2018, 11 (1): 327.341
DOI: 10.3844/ajeassp.2018.327.341

Mockus, V., 1961. Watershed lag. US Department of Singh, V.P., 1988. Hydrologic systems: Rainfall-
Agriculture, Soil Conservation Service, ES–1015, runoff modeling, Volume 1, Prentice Hall
Washington, DC. Publication, pp: 960.
Morgali, J.R. and R.K. Linsley, 1965. Computer analysis Simas, M.J. and R.H. Hawkins, 2002. Lag time
of overland flow. J. Hydraulics Division, 91: 81-100. characteristics for small watersheds in the U.S.
NRCS, 1972. National engineering handbook. U.S. Proceedings of the 2nd Federal Interagency Hydrologic
Department of Agriculture, Washington D.C Modeling Conference, (HMC’ 02), Las Vegas.
National Resource Conservation Service. Tarboton, D.G., 1991. On the extraction of channel
NRCS, 1986. Urban hydrology for small watersheds. networks from digital elevation data. Hydrological
Processes, 5: 81-100.
Technical Release 55, Washington, D.C National
Temez, J.R., 1978. Calculo hidrometeorologico de
Resource Conservation Service.
caudales maximos em pequeٌas cuencas naturales.
Overton, D.E. and M.E. Meadows, 1976. Storm Water
Madrid: Ministério de Obras Publicasy Urbanismo
Modeling. 1st Edn., Academic Press, Inc. Ragan. (MOPU). Direccion General de Carreteras.
Papadakis, C. and N. Kazan, 1986. Time of Tung, Y.K. and L.W. Mays, 1981. Risk models for flood
concentration in small rural watersheds. Technical levee design. Water Resour. Res., 17: 833-841.
report 101/08/86/CEE. College of Engineering, USWRC, 1981. Estimating Peak Flow Frequencies for
University of Cincinnati, Cincinnati, OH. Natural Ungaged Watersheds: A Proposed
Povlovic, S.B. and Moglen, G.E. 2008. Discretization Nationwide Test. 1st Edn., U.S. Water Resources
issues in travel time calculation. Journal of Council, Washington, pp: 346
hydrologic engineering, 13(2): 71-79. Van Sickle, D., 1962. The effects of urban development
Pasini, F., 1914. Relazione sul progettodella bonifica on storm runoff. Texas Engineer, 32: 23-27.
renana. Bologna. Italy. Williams, G.B., 1922. Flood discharges and the
Pilgrim, D.H. and G.E. McDermott, 1982. Design floods dimensions of spillways in India. Eng. (London)
for small rural catchments in eastern New South 134: 321-321.
Wales. Civil Engg Trans Inst. Engrs Aust. Wolock, D.M. and C.V. Price, 1994. Effects of digital
Pradahan, N.R., F.R. Ogden, Y. Tachikawa and K. elevation model map scale and data resolution on a
Takara, 2008. Scaling of slope, upslope area and soil topography-based watershed model. Water
water deficit: Implications for transferability and Resources Res., 30: 3041-3052.
regionalization in topographic index modeling. Wong, T., 2009. Evolution of kinematic wave time of
Water Resources Res., 44: 12-21. concentration formulas for overland flow. J. Hydrol.
Ragan, R.M. and J.O. Duru, 1972. Kinematic wave Eng., J. Am. Society Civil Engineers, 14: 739-744.
nomograph for times of concentration. J. Hydr. Div., Wong, T.S.W. and F. Asce, 2005. Assessment of time of
concentration formulas for overland flow. J.
98: 1765-1771.
Irrigation Drainage Eng., 131: 383-387.
Ribeiro, G., 1961. Acerca do calculo da vas‫م‬o de obras
Woolhiser, D.A. and J.A. Liggett, 1967. Unsteady one-
d’arte: Tempo de concentraç‫م‬o, Revista do Clube de
dimensional flow over a plane the rising
Engenharia. 294: 16-19. hydrograph. Water Resour. Res., 3: 753-771.
Salimi, E.T., A. Nohegar, A. Malekian, M. Hoseini and A. Yen, B.C. and V.T. Chow, 1983. Local design storms.
Holisaz, 2016. Estimating time of concentration in U.S. Dept. of Transportation, Federal Highway
large watersheds. Paddy Water Environ., 4: 20-135. Administration, Washington, D.C.
Schaake, J.G., J.C. Geyer and J.W. Knapp, 1967. Zhang, Z.H., L.U. Wenxi, C.H.U. Haibo, C.H. Weiguo and
Experimental examination of the Rational Method. Z.H. Ying, 2014. Uncertainty analysis of hydrological
J. Hydraulics Division, 93: 353-370 model parameters based on the bootstrap method: A
Sharifi, S. and S.M. Hosseini, 2011. Methodology for case study of the SWAT model applied to the Dongliao
identifying the best equations for estimating the time River Watershed, Jilin Province, Northeastern China.
of concentration of watersheds in a particular Sci. China Tech. Sci., 57: 219-229.
Region. J. Irrigation Drainage Eng., 137: 712-719. Zomorodi, K., 2005. Revising the NRCS sheet flow
DOI: 10.1061/(ASCE)IR.1943-4774.0000373 travel time equation for. Water Resources, 336: 1-4.
Sheridan, J., 1994. Hydrograph time parameters for
flatland watersheds. Trans, ASAE, 37: 103-113.
DOI: 10.13031/2013.28059) @1994
Silveira, A.L.L., 2005. Performance of time of
concentration formulas for urban and rural basins.
Brazilian J. Water Resources, 10: 5-23.

341