AND MODELLING
DAVID STEPHENSON
Department of Civil Engineering, University of the Witwatersrand, I Jan Smuts Avenue,
2001 Johannesburg, South Africa
and
MICHAEL E. MEADOWS
Department of Civil Engineering, University of South Carolina, Columbia,
SC 29208, U.S.A.
ELSEVIE R
All rights reserved. No part of t h i s publication may be reproduced, stored in a retrieval system or
transmitted in any form or by any means, electronic, mechanical, photocopying, recording or other
wise, without the prior written permission of the publisher, Elsevier Science Publishers B.V./Science &
Technology Division, P.O. Box 330, 1000 A H Amsterdam, The Netherlands.
Special regulations for readers i n the USA  This publication has been registed with the Copyright
Clearance Center Inc. (CCC), Salem, Massachusetts. Information can be obtained from the CCC about
conditions under which photocopies of parts o f this publication may be made in t h e USA. A l l other
copyright questions, including photocopying outside of the USA, should be referred t o the publisher,
Elsevier Science Publishers B.V., unless otherwise specified.
PREFACE
CHAPTER 1
I NTRODUCT I ON
HISTORICAL R E V 1 EW
h y d r a u l ic equations.
It was the American hydrologist, Horton, (generally associated with
i n f i l t r a t i o n ) who i n 1934 c a r r i e d out the e a r l i e s t recorded s c i e n t i f i c studies
open channels were set down b y St. Venant in 1871. These equations were
for gradually varied unsteady flow such as flood waves. The idea of
graphical integration using characteristic I ines was first suggested by
Massau in 1889. On the other hand Greco and Panattoni (1977) indicate
that i m p l i c i t solution by f i n i t e differences i s the most e f f i c i e n t method b y
computer, avoiding instabi I i ty and giving rapid convergence. Various
numerical methods of s o l u t i o n of the k i n e m a t i c equations were i n v e s t i g a t e d
by Kibler and Woolhiser (1970). The step length in finite difference
retention u
Definitions
0 t
td
Fig. 1.2 Catchment water b a l a n c e
Travel time (t,) is the time for a p a r t i c l e of water to proceed from the
most remote p a r t of the catchment to the d i s c h a r g e p o i n t . For a plane it
is not equal to time of concentration a c c o r d i n g to k i n e m a t i c theory since
water moves slower than a hydraulic response which travels at wave
occur.
T i m e of excess r u n o f f (t ) i s the time measured from the commencement of
runoff. It is therefore less than the time t from the commencement of
CLASSICAL HYDROLOGY
(1.4)
.2
1.48(.151) ~ + l e 7
(1 3)
C = 7 . 7 ~ 1 0  ~ C ~ ' R ~ ~ ( . 0 1 C( .001
~ ) CN)
~~
(T)
where R is the recurrence interval, S is bed slope in percent, I is
rainfall intensity i n inches p e r h o u r , M i s the f r a c t i o n of watershed w h i c h
i s i m p e r v i o u s a n d CN the S o i l C o n s e r v a t i o n S e r v i c e (SCS) c u r v e n u m b e r .
The assumption of a unique 'C' for any catchment can lead to
s i g n i f i c a n t e r r o r s and u n d e r e s t i m a t i o n o f f l o o d r u n o f f . T h i s i s demonstrated
by F i g u r e 1.3. The runoff rate per unit area for case 'a' is Cil. If the
same C i s used f o r case b, where a higher rainfall i n t e n s i t y occurs, the
0.385 (1.6)
tc = (0.87L3/H)
where t is i n hours, L i s the l e n g t h of catchment i n km a n d H the drop
i n metres, or
t = (11 .6L3/H)0.385 (1 7)
where L i s i n miles a n d H i n f t .
rainfall a n d r u n o f f r a t e s
per u n i t a r e a of c a t c h m e n t
rainfall r a t e il
runoff C i ,
loss = (I C)il=f
1 I
time t
( a ) Medium storm
i
rainfall runof rate
r a i n f a l l r a t e i,
runoff = Ci2
rainfall runoff r a t e
I
I 1 loss. f
Fig. 1.3 E f f e c t of c o n s t a n t C on r u n o f f
9
HYDRODYNAMIC EQUATIONS
a u au
o(+U+V+W)
au au = x  32 + I’(++)
3 ’ ~ a 2 u azu
a t ax ay az ax ax2 ay2 az2
p ( 4a uv  + vav
 + w  av
) a~ = Y  3+ azv
p(++) azv azv
(1.9)
at ax ay a~ aY ax2 ay2 az2
(1.10)
at ax ay az az axz a y 2 a z 2
where p is the mass density of the fluid, u,v,w, are the velocity
components in the x,y,z directions respectively, X,Y,Z, are the body
forces per unit volume, p i s the pressure and p i s viscosity. I n addition
to these t h r e e d y n a m i c e q u a t i o n s we h a v e t h e c o n t i n u i t y equation
__
au + + W
L = o (1.11)
ax aY az
_l a v + v a v + a v + fs so = o (1.12)
g at g ax ax
(1.13)
f o l l o w i n g two e q u a t i o n s (see c h a p t e r 2 ) .
Dynamics S = Sf (1.15)
where i is the input per unit area of surface (e.g. excess rainfall
intensity).
These e q u a t ions are termed the kinematic equations. Equation (1.15)
merely states that the bed slope can be substituted for the energy
gradient in a f r i c t i o n equation.
F o r o v e r l a n d sheet f l o w q p e r u n i t w i d t h these e q u a t i o n s become
(1.16)
m
4 = aY (1.17)
ah
AQ + A = q (1.18)
at
simultaneously for head at the node and flow in the connecting pipes.
11
TENSION
\
CONTROLS
GRAVITY
_.
TIME
I NF I LTRAT I ON
rate. The forces influencing the movement of water into and through the
14
soil are suction and gravity. During the early stages, the upper soil
layer is "thirsty" and infiltration i s dominated b y suction. With time, the
u p p e r centimetre, more o r less, of the s o i l s u r f a c e becomes s a t u r a t e d a n d
the infiltration r a t e reduces to that rate at which water moves through
the saturated soil. At this point, gravity dominates. As long as the
rainfall rate exceeds the instantaneous infiltration rate, or water is
ponded on the surface, infiltration w i l l continue a t the maximum possible
rate, defined b y Horton (1933) as the c a p a c i t y i n f i l t r a t i o n rate. The effect
of rainfall r a t e on the infiltration curve i s next examined. Three general
cases f o r infiltration during a steady rainfall were proposed b y Mein a n d
Larson (1973):
i n f i Itrate.
content.
8; 
8,
I
I
I
I
I I
II TRANSMITTING I I SATURATED
r I ZONE I
a I ZONE
I
a I W I
3
W n
n I I
I
I WElTINGZONE I
I I
BODMAN AND COLMAN GREEN AND AMPT
The c o n s e r v a t i o n of mass e q u a t i o n i s
av+ao=o (1.21)
az at
v =  k dh (1.22)
dz
( 1 .24)
equation.
(1.25)
minor hysteretic loops can occur between the main hysteretic loops. The
h y s t e r e t i c effect i s a t t r i b u t e d to (1) geometric nonuniformity of individual
MOISTURE C O N T E N T
Fig. 1.6 Typical soil suction  moisture relation
have demonstrated the usefulness of the Green and Ampt model for
model l i n g infiltration. As methods for measuring the model parameters
are made easier, it can be expected the model will be more widely
applied.
D a r c y ' s law c a n be w r i t t e n as
v = f
 = k(h + Lf + $f)/Lf (1.26)
n
where f is the infiltration rate and v i s equal to the vertical velocity,
h is the surface ponding depth, Lf is the depth to the wetting front,
and qf i s suction at the w e t t i n g f r o n t .
Several assumptions were n e c e s s a r y to write Darcy's law in the form
of E q . 1.26, namely:
1. There e x i s t s a d i s t i n c t a n d p r e c i s e l y d e f i n a b l e w e t t i n g f r o n t .
r e g a r d l e s s of time a n d depth.
18
3. Above (behind) the wetting front, the soil is uniformly wet and of
constant hydraul ic conductivity k.
4. Below (in front of) the wetting front, the soil moisture content is
r e l a t i v e l y unchanged from i t s i n i t i a l moisture content, 0 _.
These assumptions, when checked against the actual soi I m o i s t u r e p r o f i l e
of Bodman and Colman illustrate the approximate nature of the Green
a n d Ampt m o d e l . T h i s i s s h o w n i n F i g u r e 1.5.
The accumulated infiltration depth, F, can be obtained b y integrat
i n g Eq. 1.26.
f = dF/dt = k(h + Lf + $ f ) / L f (1.27)
o r more d i r e c t l y f r o m
F = (0  Oi)Lf = AOLf (1.28)
soils, assuming each layer has uniform properties. The required soil
properties, i.e. K,, if,
4, a n d Si, a n d the thickness, d, must be k n o w n
for each layer. After computing the infiltration during each time
se = (vJ
b
/ J Ic ) ' " ; f o r J, > jib (1.35)
f = __
11 'b
 (1.36)
r l  1 2
where rl = 2+3/B
H y d r o l o g i c I n f i l t r a t i o n Models
k t
f = f + (fo  fc)e (1.37)
exceed t h e c a p a c i t y i n f i l t r a t i o n rates.
Holton (1961) proposed a conceptual model of infiltration backed by
REFERENCES
CHAPTER 2
ANALYSIS OF RUNOFF
INTRODUCTION
and
5. The s1oDe of the channel i s small.
Conservation of Mass
Fig. 2.1 D e r i v a t i o n of c o n t i n u i t y e q u a t i o n
horizontal measured as an angle, and x and t are the space and time
coordinates i n metres ( f e e t ) a n d seconds. The t o t a l i n f l o w to the section i s
I n f l o w = Q + qiAX (2.1 1
and the t o t a l outflow is
Outflow = Q + 22 AX (2.2)
ax
25
aA
Change i n volume stored = (2.3)
=Ax
(2.4)
ax at
Conservation of Momentum
wt
Fig. 2.2 D e r i v a t i o n of momentum e q u a t i o n
AXP(A .LY + v
at
z),+
at
V A X P ( V a A
ax
+ ZA 2)
ax
S u b s t i t u t i n g the f o l l o w i n g equivalence from c o n t i n u i t y
* a v + v  aA 5 q ,   aA (2.8)
ax ax I at
a l l o w s the r a t e of momentum change to be w r i t t e n as
Eqs. 2. 4 and 2.9 can be made applicable to any cross section for
both overland and open channel flow, though strictly they apply to
S l M P L l F I ED EQUATIONS
(2.10)
28
For prismatic channels, Eqs. 2.4 and 2.10 are often combined into the
s i n g l e equation
a Q + c aQ
 = a2Q (2.11)
at ax DaxZ
where c is the wave celerity in m/s (fps) and D is a hydrograph
dispersion coefficient i n m’/sec (ft*/sec). Because Eq. 2.11 i s of the form
of the classical advectiondiffusion equation, it i s commonly called the
d i f f u s i o n wave model.
The kinematic model further assumes the pressure term is negligible,
r e d u c i n g Eq. 2.10 to
so = Sf (2.12)
which means the e q u a t i o n of motion can be approximated b y a u n i f o r m flow
(2.14)
(2.15)
u n i t w i d t h a n d flow depth i s y .
The quasisteady flow approximation was originally termed the
k i n e m a t i c wave a p p r o x i m a t i o n s i n c e waves can o n l y t r a v e l downstream a n d
(2.17)
(2.19)
m 1
Therefore dx
dt
= c = may (2.20)
R i s i n g H y d r o g r a p h  G e n e r a l Solution
For t h e c a s e o f a long impermeable p l a n e , A = by, Q = bq and R =
y, where q is the flow per unit width, hence Eqs. 2.4 and 2.13 can be
written
(2.15)
and
m
q = aY (2.16)
(2.22)
Eq. 2.22 s t a t e s t h a t t o an o b s e r v e r m o v i n g a t t h e s p e e d

dx
dt am’
m1
(2.23
a
dt
=‘ e . (2.24
characteristic begins at the upslope end of the plane and travels the
length of the plane durin.g the time tC. In this case, the depth at the
upstream end is zero, yo = 0, for all t. Therefore, as long as the
rainfal I intensity remains constant, once this initial characteristic has
reached the downstream end of the plane, the depth profile along the
plane will remain constant r e g a r d l e s s of how long the rainfall persists,
i.e., an equilibrium depth profile will be e s t a b l i s h e d . The time required
0 i
Time of Concentration
One can solve for the time of concentration from Eq. 2.27 u s i n g the
conditions that at t =. tc, x  x = L . S u b s t i t u t i n g a n d r e a r r a n g i n g to
solve for concentration time tc which i s equal to time to e q u i l i b r i u m t
m'
tC = ( L/ccieml ) 1 /m (2.30)
t = (6.9/ie0'4)(nL/So0*5) O e 6
(2.31 )
t = ( 0 . 9 2 8 / i ~ O . ~( )~ L / S ~ O * ' ) O e 6 (2.32)
Equilibrium Depth P r o f i l e
y ( x ) = ( i e x / a ) 1/m (2.33)
which f o r Manningkinematic flow i n S I u n i t s becomes
F r o m Eq. 2.34
0.6
06 )
0.023( 6 . 9 ~ 1 (50)
Y(L) =
[ ] = 0.0034 metres
(0.01
or y ( L ) = 3.4 mm
The Receding H y d r o g r a p h
f u t u r e time. T h i s p r i n c i p l e i s i l l u s t r a t e d i n F i g u r e 2.4.
34
A L
x = x + A x (2.37)
1
m1
= x + amyl (ttd) (2.38)
1
where x 1 was the position for point B1. Note that if the storm duration
Y, = i ieX1/"1 1/ m (2.39)
x = x (m1 ) / m ( t  t ) (2.40)
1 + amIi,xl/a] d
A t the downstream end of the p i a n e x = L a n d q = ay
1
= ieL.
After substituting these identities into Eq. 2.40, we obtain the f o l l o w i n g
relationship between discharge and time for the recession hydrograph
(tt ) = 0 (2.41 )
d
35
(2.42)
The d e p t h a t p o i n t B1 w i l l m o v e a t a c o n s t a n t r a t e and w i l l r e a c h t h e e n d
of the p l a n e a t t i m e t,. T h i s time i s e v a l u a t e d a s
L  x,
t:; = td t
dx/dt (2.43)
I n c o r p o r a t i n g Eqs. 2.26, 2.27, and 2.30, Eq. 2.43 becomes
o1 iml m aiml m
e t c  e td (2.44)
t, = td t
.m1 m
a mi
e td
which can b e s i m p l i f i e d to
1
t, = td 1 +  [(tc/tdIm  1 } (2.45)
m
The d i s c h a r g e at the e n d of the p l a n e will r e m a i n c o n s t a n t b e t w e e n
td 5 t S t, and w i l l be
mi
q = a(i t (2.46)
e d
After t,, the recession proceeds according to Case I and E q . 2.41 applies
First determine the equation for the rising graph. The coefficient
in Eq. 2.29 is
c1 = Soe5/n = (0.01)0.5/0.023 = 4.35
D i s c h a r g e , m3/sec
Time, Minutes Depth, mm
5
0.0 0.0 0.0 x 10
1 .o 0.42 1 .o
2.0 0.83 3.2
3.0 1.25 6.3
4.0 1.67 10.2
5.0 2.08 14.8
6.0 2.50 20.0
7.0 2.92 25.9
8.0 3.33 32.4
9.0 3.42 33.7
10.0 3.42 33.7
Fig. 2.5 K i n e m a t i c h y d r o g r a p h s h a p e f o r s i m p l e p l a n e w i t h td = tC
37
FR I C T I ON EQUAT I ON
(2.47)
q = JS 5/3 (2.48)
n Y
1
21
Q = 7.7(R/k)'/6A(SRg (2.49)
(2.51 )
1
__ = 2 log (  + 7 2.5 1
14.8R Re f
Ji
38
where Re is the Reynolds number, for pipes VD/v, or 4VR/V for non
higher values of the Darcy friction coefficient f for low Reynolds number
a n d any r e l a t i v e roughness k/R, the S t r i c k l e r e q u a t i o n assumes f depends
only on the r e l a t i v e roughness k/R. The S t r i c k l e r a n d M a n n i n g equations
can therefore be expected to underpredict roughness for low Reynolds
numbers. Higher values of n should therefore be used for o v e r l a n d flow
than f o r channel flows.
In general, the value of n and hence flow depth has to be deter
mined by trial (assuming the ColebrookWhite equation to apply a n d not
Strickler's). It i s therefore p r o b a b l y easier to use the Darcy equation f o r
this purpose but since an explicit equation is required for analytical
solutions to the kinematic equations a n d the v a r i a t i o n i n n i s less t h a n
the v a r i a t i o n in f with y, the M a n n i n g e q u a t i o n i s p r e f e r r e d .
Table 2.2 i n d i c a t e s values of n a n d f w i t h v a r y i n g water depths i n a
wide channel with a slope of 0.0025 a n d absolute roughness k = 0.0125 m.
The values of f are calculated from the ColebrookWhite equation using a
f i r s t estimate of Re from M a n n i n g ' s e q u a t i o n , a n d then n i s recalculated
ve loc i ty d i st ri b u t ion.
Resistance to rainfall induced o v e r l a n d flow over natural and man
made surfaces is influenced by several factors including surface rough
ness, raindrop impact, vegetation, wind and infi Itration. Although there
have been many laboratory and field investigations to determine the
in/hr for 8 minutes and the second was 3.55 in/hr for 8 minutes. This
41
1. I
~ I
i 2.6~ . C o m p a r i s o n o f T u r b u l e n t and L a m i n a r K i n e m a t i c W a v e
Solu ions w i t h Observed R i s i n g H y d r o g r a p h s
minimum kinema ic flow number as Run No. 136. The runoff s u r f a c e was
an asphalt p l a n e w i t h the following physical characteristics:
L = 72 f t ; M a n n i n g n  v a l u e = 0.024; and So = 0.01.
Blandford and Meadows (1983) analyzed these events with a finite
element formulation of the kinematic overland flow model and obtained
the results shown in Figure 2.7. For Run No. 136, the predicted r i s i n g
REFERENCES
I z z a r d , C.F., 1946. H y d r a u l i c s o f R u n o f f f r o m D e v e l o p e d S u r f a c e s , H i g h 
w a y Research B o a r d , P r o c e e d i n g s o f t h e 2 6 t h A n n u a l M e e t i n g pp.
1 291 50.
M o r r i s , E.M., a n d W o o l h i s e r , D.A., A p r i l 1980. U n s t e a d y o n e  d i m e n s i o n a l
f l o w o v e r a p l a n e : p a r t i a l e q u i l i b r i u m a n d recession h y d r o g r a p h s .
W a t e r R e s o u r c e s R e s e a r c h , 1 6 ( 2 ) pp 355366.
Overton, D.E., 1972. A V a r i a b l e Response O v e r l a n d F l o w M o d e l , Ph.D.
D i s s e r t a t i o n . Dept. o f C i v i l E n g i n e e r i n g , U n i v . o f M a r y l a n d .
O v e r t o n , D.E. a n d Meadows, M.E., 1976 S t o r m w a t e r M o d e l l i n g , A c a d e m i c
P r e s s , New Y o r k .
St. V e n a n t , A.J.C. B a r r e de, 1848. E t u d e s T h e o r i q u e s e t P r a t i q u e s s u r
l e Mouvement d e E a u x C o u r a n t e s . ( T h e o r e t i c a l and P r a c t i c a l S t u d i e s
o f Stream F l o w ) , P a r i s .
U.S. A r m y C o r p s of E n g i n e e r s 1954. D a t a R e p o r t , A i r f i e l d D r a i n a g e I n v e s t 
igations, Los Angeles District, O f f i c e o f t h e C h i e f of E n g i n e e r s ,
A i r f i e l d s Branch Engineering Division, M i l i t a r y Construction.
Woolhiser, D.A. 1981. Physical Il y b a s e d m o d e l s of watershed runoff,
pp. 189202 in S i n g h , V.P., (Ed.) R a i n f a l l Runoff Relationships.
W a t e r R e s o u r c e s P u b l i c a t i o n s , C o l o r a d o , 582 pp.
Woolhiser, D.A. a n d L i g g e t t , J.A., 1967. U n s t e a d y o n e  d i m e n s i o n a l flow
over a plane. The r i s i n g h y d r o g r a p h . Water Resources Research,
3 ( 3 ) pp 753771.
OBSERVED
4r
c
L
\
e
w
(3
0 4 8 12 16 20 24 28 32
TIME( m in 1
CHAPTER 3
D E S I G N PARAMETERS
m
Energy 9 = CLY (3.2)
where x is the direction of flow, t is time, i is the excess rainfall
rate if, f is the loss rate and q is the discharge rate per unit
catchment width. For the present a l l u n i t s must be assumed consistent.
Later units will be introduced in order to render the numerical values
more meaningful. It will be assumed in the following analysis that i
and f are uniform in time’ a n d space f o r the duration of the storm t
d‘
q is the flow rate per unit width of plane, y i s flow depth, CL is a
coefficient a n d m i s an exponent.
(3.3)
(3.5)
dy/dt = ie a t the wave front point (and downstream of it). One also
has from the c o n t i n u i t y equation b y e x p a n d i n g the a q / a x term

3 Y + 3 dy = i
*
at ay dx e
By comparing w i t h aatv + dx

dt ax
 i
e
one must h a v e 
dx = a t the wave f r o n t
dt ay
45
m 1
a n d from 3.2, d x / d t = may (3.7)
Since y = iet,

dx = mcr(iet) m1 (3.8)
dt
which i s the speed of the wave front at any time t 5 tC. Also during
equilibrium the discharge r a t e at any point x from the upstream water
shed i s q = i x ( l i n e ABlC3 i n Fig. 2.4). Hence from 3.2
e
x = aym/i (3.9)
An expression for the discharge rate after the storm stops, which
is assumed to be after the time to equilibrium (t2t zt ) , i s obtained
d c
by considering the water depth profile along the catchment a g a i n . A f t e r
the rain stops the effect of all upstream depths travels down to the e x i t
at a speed dx/dt given by 3.7. To predict when the depth at the e x i t
is 'y', imagine a series of waves travelling from the water p r o f i l e c u r v e
in a downstream direction at a constant speed d x / d t = aq/ay = mayrn'
Integrating, x = x + maym'(ttd) (3.10)
lI/m l / m
= q / i e i mq a (tt 1 (3.11)
d
since y = (q/a)l/m. (3.12)
I n p a r t i c u l a r a t the e x i t ,
L = q / i e + mq 11 / m a t / m b t d ) (3.13)
which i s an imp1 i c i t expression for the falling limb of the hydrograph.
The f u l l h y d r o g r a p h shape i s thus as i n F i g . 3.1.
that the falling limb of the hydrograph is obtained from the implicit
(3.13)
The total depth of excess rain has been kept constant in each case
in Fig. 3.1 so that i = p/td where P is the depth of precipitation.
46
Flow
q
I c
n u d
+
Time 1
At = (tctd)/m (3.18)
This is the duration of the flat top of the hydrograph I in Fig. 3.1.
It should be noted that the f a l l i n g l i m b s of the hydrograph in Fig. 3.1
omit losses after rain stops. If infiltration ( f ) continues the h y d r o g r a p h
will look l i k e those in F i g . 3.2. It i s generally necessary to model such
I D

9
iL
0.5
0 I I J 1 5
t/f,
For small catchments the maximum peak runoff rate occurs when
the duration of excess rain equals the concentration time, t . For p l a i n
rectangular catchments the concentration time is a function of excess
rainfall rate
hyetogroph
{I I
\>(
ntial
abstrac
tion
t
4
~t+4
Fig. 3.3 Excess flow hyetograph d e r i v e d from IDF curve
(3.25)
Long Catchments
If a local, intense storm turns out to be the design storm, the areal
reduction factor applied to point rainfall intensity relationships may
be less significant. The factor is generally closer to unity the smaller
the lateral extent of the storm, but on the other h a n d shorter duration
storms have a more significant reduction factor (less than long storms).
These facts will not b e revealed u s i n g the Rational method w i t h rainfall
propor t iona I losses.
Before equilibrium is reached the runoff per unit width at the
mouth of the catchment at any time t after the commencement of excess
r a i n or runoff is
m
q = a (iete) (3.27)
where t td  u / i
 (3.28)
m1 l / m
<tc = ( L / a i e ) (3.29)
(3.30)
(3.31)
(3.32)
q/aam is plotted against td in Figs. 3.4 to 3.6 (the full lines) for
different values of the dimensionless parameters U = u/a and F = f/a.
For a l l cases of F > 0 the l i n e s e x h i b i t a peak r u n o f f a n d the correspond
ing storm duration td for an infinitely long catchment. For most catch
ments it i s necessary to e s t a b l i s h whether t is less t h a n t i.e. whether
the peak occurs before the catchment has reached e q u i l i b r i u m .
I n fact, t = td  t (3.33)
Therefore f o r t =
c te
(3.35)
U
previously established. If the peak lies to the left, read the revised
design storm duration td corresponding to the p e a k , and the correspond
m
i n g peak flow parameter q / a a on the l e f t h a n d o r d i n a t e .
M o d i f i c a t i o n for P r a c t i c a l U n i t s
ately both the Manning resistance equations and the IDF relationships
are empirical and the coefficients depend on the u n i t s employed. I n the
Manning form of equation (3.241, q is in m2/s if iete is in metres.
a is 6 / n in S.I. u n i t s where S is the dimensionless slope a n d n i s the
M a n n i n g roughness.
It is most convenient to work with td in hours and i and a in
mm/h. The numbers are then more realistic. I n equation (3.32) if q is
in m3/s/m, a i n ms units, a in mm/h and t in hours then the leff
d
h a n d side should be replaced by
q
  l ~ 1~ 0 ~~ m
~ 
aam 5/3
oa
= I O ~ Q
5/3 (3.36a)
Baa
where Q i s total runoff r a t e off a catchment of width B metres. Note t h a t
the right hand side of (3.32) is in hm i f td is i n h, so no c o r r e c t i o n
is made to the above factor to convert a to secs, only to convert mm
to m. This is what the left hand axis of Figs. 3.4 to 3.6 represent if
a is in mm/h. It is referred to as the runofffactor, QF. Similarly the
left hand side of equation (3.35) is L/aaml in homogeneous units, or
if a is in mrn/h, L in metres and a i n ms units, then it should be
replaced b y :

L a looom
  L
a 3600000 m 236aa
/ 3 (3.37a)
63q ( 3.36b )
aa ’I3
52
... LENGTH W l O A
T t 9.8
Y. Y
u= 0.004 q.0
c= 0.zL.l0
P= 0.890
3.6
3.2
L
1
2.8
0
I
v
2.q R
LL
I
t
2.0 9
W
I
1.6
I .I
I1.B
0. I 0.9
a. I 6.0
0 I 2 3
5JDRl bCRAlIOU IN HA5
q.O
3.6
3.2
L
.l
2.8 fy:
0
I
U
2.q [L
r
I
I
7.0 D
z
4
1.6
1.2
0. H
0.q
0.0
E I 2
51ORR DURATION IN HR5
I .a 4.0
1.9 3.6
0.9 3.2
LL
L
13 I
2.9
K 0.7
~y:
0 n
I I
LJ V
21.6 2.Y a:
L
LL
L
I
I
g 0.s
LL
2.0 irr
z
7 w
I
ct:
O.Y I .6
0.3 I .2
0.2 0.9
0. i 0.q
a.n 0.0
687a a‘’’
Figs. 3.4 to 3.6 are plotted in terms of the dimensioned expressions for
w i d t h 0 = 450 m
slope S = 0.01
M a n n i n g roughness n = 0.1
inland region, MAP = 620mm/annum
20 y e a r r e c u r r e n c e i n t e r v a l storm
r a i n f a l l f a c t o r a = (7.5 + 0.034 x 620)20°‘3 = 70mm/h
The c o r r e s p o n d i n g p r e c i p i t a t i o n r a t e i s :
i = a / ( c + t )’ = 70/(0.24 + 1.3)o’89 = 48mm/h
d
The e q u i v a l e n t r a t i o n a l c o e f f i c i e n t C i s
1.60/(450 x 800 x 48/3600000) = 0.33
EFFECT OF CANALIZATION
The charts presented are for the case of overland flow. It fre
q u e n t l y occurs t h a t runoff reaches channels, a n d t h u s flows to the mouth
of the catchment faster than if overland. The critical storm duration
may thus be s h o r t e r a n d the peak flow h i g h e r t h a n w i t h no c a n a l i z a t i o n .
An estimate for the concentration time of a catchment with a wide
rectangular channel down the middle may be made using this chapter
if overland flow time can be neglected. The effective catchment width
is taken as b, the stream width, and both rainfall rate i and losses
f and u should be increased by the factor B/b where B is the true
catchment width. The charts herein can then be applied as in the
example.
I n many situations both o v e r l a n d flow a n d stream flow are signifi
cant and the problem cannot be solved as s i m p l y as herein. The h y d r o 
logist must then resort to trial and error methods using dimensionless
h y d r o g r a p h s f o r catchment  stream systems as presented l a t e r .
‘channel
(paved or roofed).

mm 
inches ~
mm inches mm/h inches/h
Paved up to 5 0.2 0 0 0 0
Clay up to 5 0.2 15 0.8 2  5 0.1  0.2
Loam up to 7 0.3 20 1.2 5  15 0 . 2  0.6
Sandy up to 10 0.4 30 1.5 15  25 0.6  1
Dense up to 15 0.6 5 0.2 5  15 0 . 2  0.6
vegetation
58
REFERENCES
CHAPTER 4
Lighthill and Whitham (19551, kinematic overland flow models have been
utilized increasingly in hydrologic investigations. The first application
to watershed modelling was by Henderson and Wooding (1964). The
conditions under which the kinematic flow approximat ion holds for surface
runoff were first investigated by Woolhiser and Liggett (1967); they
found it is an accurate approximation to the ful I equations for most
Q = bay" (4.1
where Q is the discharge, y i s the depth of flow, b the w i d t h a n d ci , m
a r e constants.
The latter conclusion can be established by normalizing the
momentum equation by the steady uniform discharge Qn. The momentum
e q u a t i o n t h e n becomes
60
Q Qn (4.3)
which means that graduallyvaried flow may be approximated by a
uniform flow formula such as Manning’s equation. If one writes
Manning‘s equation for a wide rectangular crosssection such as an
overland flow plane, since the hydraulic radius can be approximated
by the depth of flow, one obtains the following expression (SI units)
Q = 1 b y y 2/3s f (4.4)
or Q/b = A S y5I3
(4.5)
Governing Equations
(4.7)
and
61
vi
av

at + v z
av
+ g=
ay 
 g(S0  5f )  e
Y
(4.8)
UNIFORM RAINFALL
x F
L x
(4.9)
C being the Chezy coefficient which equals Jf/sg where f i s the Darcy
friction factor. By writing Eqs. 4.7 and 4.8 in dimensionless form, the
number of parameters are reduced from f i v e to two with obvious advant
ages. Woolhiser and Liggett (1967) first presented the following dimen
sionless equations
' H + " E + H  au
 = 1 (4.10)
aT ax ax
and
(4.11)
where
H = y/yo, U = v/vo, X = x/L, T = tvo/L (4.12)
and y and v are the normal depth and velocity, respectively, at the
end of the p l a n e f o r a given steady r a i n f a l l excess r a t e , i The normal
e'
i z i n g parameters a r e r e l a t e d b y :
i L = v y (4.13)
0 0
62
and
2 2
vo/coYo = so (4.14)
The two independent parameters in Eqs. 4.10 and 4.11 are the normal
flow Froude number, Fro, vo/ J ( g y o ) , and the kinematic flow number,
k (Woolhiser a n d L i g g e t t , 1967).
k =  (4.15)
2
yoFro
0 .I 2 3 .4 .5 .6 .7 .8 9 1.01.1 1.21.31.41.51.6
T

aH a H 2/3 = 1
aT + x (4.18)
U = T1/ z (4.20)
Thus, the r i s i n g h y d r o g r a p h i s g i v e n b y
(4.21)
I .5
Q* :$
AC k =I0
/ PA RAME TER :Fro
.I
'0
1 .I .2 .3 .4.5 .6 .7 .8 .9 1.01.1 1.2 1.31.41.51.6
T
k
'
m
KINEMATIC APPROX.
10
FULL S A I N T VENANT
Kinematic F l o w Number
1 .2s0.4L0.2
k = 1 . 7 ~lo6 "_ .0.8
(4.23)
e
and for r a i n f a l l intensity in in/hr a n d l e n g t h i n feet
kFr2 = ~ (4.25)
0
YO
tC = L/vo (4.28)
66
y o = i t (4.29)
e c
which, when s u b s t i t u t e d i n t o Eq. 4.25 yields
2 oL
kFo = 
i t
(4.30)
e c
Using the definitions of Eqs. 4.28 and 4.29, and Manning's equation,
one o b t a i n s the desired expression, f o r r a i n f a l I i n mm/hr,
(4.30a)
and
460Sb*3 L O e 4
(4.30b
kF: = 0.6i0.2
e
for r a i n f a l l i n i n / h r .
2
I n general, kFo values are high for smooth, steep, long planes with
low rainfall rates. This result is similar to the expression for k, except
that the effect of roughness on the Froude number suggests the k i n e m a t i c
model may be more a p p l i c a b l e to u r b a n watersheds w i t h smooth impervious
surfaces.
To illustrate the hydrological applicability of these results,
consider an asphalt parking lot with the following characteristics:
L = 50 meters; So = 0.005; n = 0.022. For an a v e r a g e excess intensity
of 50 m m / h r , k = 200 a n d k F r 2 = 31.
(4.31)
w i t h the d e f i n i t i o n of the t o t a l d e r i v a t i v e of Q
dQ  aQ dx
 +
aQ
(4.32)
dt ax dt
By r e w r i t i n g E q . 4.31 as
a Q
+   aA dQ d t = (4.33)
ax at dA dx
to an observer moving w i t h wave speed, c,
(4.34)
(4.35)
This result follows from the d e f i n i t i o n of the total derivative, Eq. 4.32,
a n d the e q u a t i o n of c o n t i n u i t y , Eq. 4.31.
For most c h a n n e l s where the flow i s inbank
(4.36)
(4.38)
depths move at faster rates, it follows that the leading limb of the
hydrograph w i l l steepen a n d the recession limb will develop an elongated
tail. Eq. 4.38 also shows that kinematic waves are propagated down
stream only, i.e. Eq. 4.38 is a forward characteristic. Kinematic flow
does not e x i s t where there a r e b a c k w a t e r effects.
Crest Subsidence
(4.39)
M a n i p u l a t i n g t h i s equation y i e l d s

d Y = V + Q dt = o (4.40)
dx ax at dx
which establishes that theoretically, the kinematic wave crest does not
subside as the wave moves downstream.
These results show that a kinematic wave can alter in shape but
does so without crest subsidence. Further, the maximum discharge rate
occurs with the maximum depth of flow. ( T h i s i s the assumption implicit
in the slopearea method for e s t i m a t i n g flood discharges from h i g h water
marks).
cantly from the single valued rating, the conclusion can be d r a w n that
the main body of a hydrograph moves k i n e m a t i c a l l y . I n w h i c h case, the
kinematic model (or the diffusion model) should be sufficient for most
and discharge first set forth by Leopold and Maddock ( 1 9 5 3 ) that the
flow in many streams i s essentially kinematic. The fact that the channel
c h a r a c t e r i s t i c s of natural streams seemed to constitute an interdependent
system which could be described by a series of graphs h a v i n g a simple
geometric form suggested the term "hydraul ic geometry". Subsequent
studies have verified and expanded on this initial work with the r e s u l t
that hydraulic geometry equations may be used to estimate general
channel characteristics at any locat ion within the drainage system.
v = kQm (4.41~)
where w is width, d i s depth, v i s crosssectional mean v e l o c i t y , Q is
discharge, and a, b, c, f, k, and m a r e best fit constants. It follows
that since width, depth, and mean velocity are each functions of
discharge, then b + f + m = 1.0; and ack = 1.0. Betson (1979) noted
that a f o u r t h r e l a t i o n s h i p also can be presented
(4.41d)
A = nQp
where A is the crosssectional area of flow. Betson also noted that
f = p  b and m = 1  p. The relationship in Eq. 4.41 are for indi
vidual stations in that they relate channel measures to concurrent
discharge.
The result in Eq. 4.40 frequently does not agree with nature.
Q = vA (4.42
The momentum e q u a t i o n can be r e w r i t t e n as follows:

Q aQ Q2 aA + 
 1 
aQ  
Q 
aA + ay = g(s 5 )  
qi (4.43
A2 ax A3 ax A at A2 at ax o f A
28 aQ Q2 aA + 1 a B + a v =
s sf (4.44)
ax ax gA at ax
gA2 gA3
70
Exponents
c = 38 (4.45)
2A
D r a w i n g on these two relationships and the definition for Froude number
Fr2 2 = __
g28
(4.46)
gy gA3
the v a r i o u s terms in Eq. 4.44 c a n be r e w r i t t e n as
and
(4.47c)
and
71
_1 av = (0.75 Fr
2 aY
 (4.48b)
g at ax
w h i c h allows the momentum equation to be w r i t t e n as
(1  0.25 Fr2) a
ax
=
'0  'f
(4.49)
Wave Speed
C r e s t Subsidence
modified diffusive wave equation, Eq. 4.49. For the following develop
ment, a rectangular cross section is assumed. As with the derivation
of most overland and open channel flow equations, this assumption
greatly simp1 i f i e s the mathematics, yet does not alter appreciably the
f i n a l form of the equations b e i n g developed.
72
(1  0.25 Fr2) a
ax
= S
o
 
Q2
c2A2R
(4.50)
T a k i n g t h e p a r t i a l d e r i v a t i v e w i t h respect to time
( 1  0.25 F 2~ a) ()
 ay =
Q2_ 1 2
 _ aQ 2 aA 1 aR (4.51)
at ax c2A2R Q a t A a t R at J
From c o n t i n u i t y
(4.52)
or
(4.53)
(4.54)
F o r a p r i s m a t i c section
dA =
_ B
dy
such t h a t

a A * B d y g B a
ax dA ax (4.55)
(4.56)
(4.57)
(4.58a)
(4.58b)
Combining similar terms and recognizing that the coefficient terms are
merely S
0’
s [ 2 aQ  3 qij
o a a t A a x A
The w h o l e e q u a t i o n t h u s becomes
(4.59)
Multiplying by Q/2
(4.60)
c = 3Q (4.35)
2A
(4.61)
(4.63)
by
Q _  4 ( 1  0.25 Fr')
ax
(4.64)
'n
where Q is the uniform flow at a given depth. This expression is
rendered more useful if the spatial derivative is replaced by some
ay  1 
ay (4.65)
ax c a t
Eq. 4.64 can be w r i t t e n as
2
( 1  0.25 Fr )
(4.66)
at
I t must be noted that Eq. 4.66 i s not strictly correct since the k i n e m a t i c
r e l a t i o n s h i p was included.
A typical looped rating curve is shown in the Figure 4.5. Com
parison with the associated discharge hydrograph illustrates that as
a flood hydrograph passes a point, the maximum discharge is first
observed, then the maximum depth, and finally a point where the flow
is uniform. The u n i f o r m 'flow occurs when the flood wave i s essentially
horizontal and therefore has a slope, dy/dx, that i s very small relative
to the bed slope. T h i s obviously will occur close to the r e g i o n of max
imum depth. The occurrence of uniform flow is illustrated graphically
as the point of i n t e r s e c t i o n of the looped rating curve with the single
v a l u e d uniform flow r a t i n g curve.
It should be noted that the scale i s exaggerated for clarity. The
three points in question are more likely to occur much closer together
than i n d i c a t e d b y the f i g u r e .
The usefulness of the looped rating curve compared with a single
valued rating curve is determined by how wide the loop is relative to
TIME
11+12+&=
01+02 s2s1
(4.69)
2 2 At
where
KZ + 0.5At
(4.72b)
c1 = KKZ t 0.5At
KKZ  0.5At
'2 = KKZ + 0 . 5 ~ 1
(4.72~)
and
At
(4.72d)
'3  KKZ + 0 . 5 ~ t
K and t must h a v e the same time unit, and the first three coefficients
sum to 1.0.
E s t i m a t i o n of Model P a r a m e t e r s
= $Q) (4.74a)
and
2 (4.74b)
=  Q(10.25Fr )]
BSoA x c ( Q)
REFERENCES
CHAPTER 5
(5.1
a v a i l a b l e comput i n g equipment.
METHOD O F CHARACTER I ST I C S
By m a k i n g the s u b s t i t u t i o n
2
c = SY (5.3)
into Eqs. 5.1 and 5.2 and then by writing first the sum, and then t e
difference, of the two new equations, we o b t a i n the two equations
(5.4a
qi
(,,)a0
ax + a0
at = g(so  S f )  ( v+c) 
A (5.4b
z _;  v + c (5.5)
then
dx
 a(vk2.c)
+ a(V'2.c) = d (v*2c)
(5.6)
ax
~
dt at dt
which gives the desired set of ordinary differential equations to r e p l a c e
the partial differential equations. The characteristic roots (directions)
are given by Eqs. 5.5, and along each direction the respective. total
derivatives i n Eqs. 5.6 hold. The r e s u l t i n g equations can be r e w r i t t e n :
c + . dx
_  v + c (5.5a)
' dt
d (v+2c)
  qi
dt
g(S0  Sf)  ( v  c ) 
A
(5.6a)
c . d_dxt = v  c (5.5b)
'
d ( v_ 2_
_ c)  'i
g(S0  Sf)  ( v + c ) 
A
(5.6b)
dt
where c
+ and c symbolically designate forward and backward character
i s t i c respective1y.
AI
UPSTREAMI t DOWNSTREAM
BOUNDARY BOUNDARY
x
x=0 X=:L
F i g . 5.1 Zones of solution domain defined b y c h a r a c t e r i s t i c s .
(Woolhiser a n d L i g g e t t , Water Resources Research,
3, 755, 1967, American Geophysical U n i o n ) .
These zones are formed by the intersection of the forward and backward
characteristics ernanat ing from the upstream (x=O) a n d downstream (x=L)
end of the channel reach at the initial time. The solution in Zone A
requires only the initial values (the beginning state of the system at
all x); w h i l e the solution i n Zones B a n d C r e q u i r e s both initial values
and a boundary condition. This is because these zones lie above the
backward and forward characteristics, respectively. Zone B requires the
downstream boundary condition, a n d Zone C , the upstream. Finally, Zone
D, which lies above both characteristics requires the initial values and
both b o u n d a r y conditions.
Numerical I n t e g r a t i o n of C h a r a c t e r i s t i c Equations
a n d XM = XR + ( t M  tR) (V  c ) ~
These two equations can be e a s i l y solved for the two unknowns XM and
tM. Once these a r e known Eqs. 5.6 a r e solved b y the same approach.
be solved for vM, cM hence yM. The boundary condition, inflow, initial
values and downstream rating curve must be known. A more stable and
accurate solution can be obtained with a nonlinear formulation (Overton
a n d Meadows, 1976; a n d Mahmood a n d Yevjevich, 1975).
When solving problems using the method of characteristics, check
whether the, flow is subcritical or supercritical. When the flow is sub
critical, v<c and the forward characteristic has a positive slope dx/dt
in the (x,t) plane while the backward characteristic has a negative
slope, as shown in F i g u r e 5.4a. When, however, the flow is supercritical,
v>c and both characteristics have a positive slope in the (x,t) plane,
F i g u r e 5.4b.
++
F i g . 5.4 C h a r a c t e r i s t i c L i n e s f o r S u b c r i t i c a l a n d S u p e r c r i t i c a l Flows
F I N I T E DIFFERENCE METHODS L
Difference Quotients
df(x)
~  lim f(x+Ax)  f ( x )
dx Ax (5.14)
a x>o
Thus the finite difference quotient i s an approximation to the continuous
d e r i v a t i v e as l o n g as A x i s kept small.
Several difference quotients can be defined to approximate partial
derivatives. To illustrate some of them, consider a function of two
independent variables, say U(x,t). With reference to the f i n i t e difference
grid in Figure 5.5, the most commonly used difference quotients are
defined as follows. The forward difference approximation to the first
p a r t i a l d e r i v a t i v e f o r U w i t h respect to x is

2U = U(x+Ax,t)  U(x,t)
(5.15)
ax Ax
Physically, one can t h i n < of an observer standing at the point (x,t),
looking ahead ( f o r w a r d ) to the p o i n t (x+ Ax,t), and using the elevation
(function value) d i f f e r e n c e between the two p o i n t s d i v i d e d b y the d i s t a n c e
to evaluate the slope ( v a l u e of the d e r i v a t i v e ) . The b a c k w a r d difference
approximation i s
au 
  u(x,t)  u(xAx,~)
ax A X
(5.16)
au
 = U(x+Ax,t)  U(xAx,t)
(5.17)
ax 2 ax
88
NUMER I C A L S O L U T I O N
There are two basic finite difference schemes used in solving the
streamflow routing equations. They a r e the e x p l i c i t and imp1 i c i t schemes.
Explicit schemes utilize initial value and left hand side (upstream)
boundary information and solve for the remaining grid p o i n t s one a t a
time. They are subject to stability limitations on the allowable grid
interval size which means explicit schemes typically have large data
requirements. However, explicit methods often result in linear algebraic
equations from which the unknowns can be evaluated directly without
iterative computations. Implicit schemes utilize initial value and both
left and right hand side boundary information, and solve for the
unknown grid points at the next time level simultaneously. Therefore,
imp1 i c i t schemes often require matrix techniques. Implicit methods
typically involve nonl inear algebraic finite difference equations whereby
the s o l u t i o n i s attained by iteration. Both schemes can be and h a v e been
used in solving the governing equations for overland and open channel
flow.
89
The use of the first two methods was summarised by Liggett and
Woolhiser (1967). They reviewed different explicit finite difference
schemes. T h e schemes w e r e :
a) method of c h a r a c t e r i s t i c s
b) u n s t a b l e method
c) diffusion method
d) LaxWendroff method
e) leapfrog method
s o l u t i o n of the equations.
The method of characteristics employs the fact that flow conforms
to certain relationships along characteristic curves and therefore the
It is the most acurate method for the same initial point spacing of all
methods. Its accuracy is a consequence of following the characteristic
curves which describe the path of the disturbances in the flow. It also
covers the x  t p l a n e f a s t e r t h a n a n y o t h e r m e t h o d w i t h t h e same i n i t i a l
point spacing. The main disadvantage of the method of characteristic
the St. Venant equations. The main reason for nor being used is that
f i n i t e element programs a r e expensive to r u n a n d accuracy and s t a b i l i t y
c r i t e r i a can become tedious to a p p l y .
Explicit finite difference schemes h a v e been w i d e l y used in the p a s t
for the solution of the onedimensional St. Venant equations. They differ
from each other in the way they define their discharge and depth
gradients, but they all express the flow properties at a certain time
( d ) L a x Wendroff method
Uses f o r d i f f u s i n g scheme
f o r the f i r s t time i n t e r v a l
a n d the leapfrog scheme
f o r subsequent t ime i n t e r v a l s
X AX X X+hX
( c ) Leapfrog method
E x p l i c i t Scheme
(5.18)
(5.19)
(5.20)
(5.21 )
(5.22)
93
A t. Once this row of values has been determined, advance the com
putations to time level t t 2At. The values at time level t + A t become
the initial values for determining the unknowns at this advanced time
level. The solution proceeds in this fashion until all the grid points
in t h e s o l u t i o n d o m a i n h a v e been determined.
To ensure stability, the g r i d sizes Ax and at a r e chosen to satisfy
the c o n s t r a i n t
(5.23)
d i f f e r e n c e scheme i s i I l u s t r a t e d ( A m e i n a n d F a n g , 1969).
pstream Downstream
t Boundary Boundary
: o n d i t . i o n 2 m / Condition
A t
1
The fol lowing approximat ions to the derivative terms are made:
(5.24)
(5.25)
(5.26)
(5.27)
(5.28)
Sf = 1,
1 [Sf(l) + S f ( 2 ) + Sf(3) + Sf(4)]
(5.29)
( d e p t h 1.
The resulting set of algebraic finite difference equations is non
linear and must be solved using an iterative rootfinding scheme. Amein
a n d Fang (1969) found that the Newton scheme c o u l d b e used to linearize
the equations w h i c h they then solved u s i n g m a t r i x techniques.
The solution procedure is to solve for all the unknowns at one
advanced time level before proceeding to the next. All values are
determined simultaneously, and must satisfy al I boundary conditions.
Therefore, this method avoids the stability requirements of the explicit
method meaning that larger x and t grid interval sizes can be used
w h i c h r e q u i r e s less i n p u t d a t a .
95
to ask the questions: "How well is the natural system modelled by the
differential equations?", a n d , "How well is the s o l u t i o n to the d i f f e r e n t i a l
equations represented by the computational algorithm?". In the analysis
here more attention is paid to the second question. The first question
U(x+Ax,t) = U ( x , t ) + Ax%
ax
+ %
2!
2
L
ax
2u
+ * ' *
(5.31)
where the derivatives are evaluated at x,t. Dividing Eq. 5.31 by ,x,
and rearranging, gives the series equivalent to the forward difference
quotients, Eq. 5.8
(5.32)
first power of Ax; we cal I this first order error (or approximation).
Similarly, it can be shown that the backward difference quotient has
first order error, and the centered difference has second order error.
Consider the f o l l o w i n g p a r t i a l d i f f e r e n t i a l e q u a t i o n
aa
 + c  = o
aa
(5.33)
at ax
One f i n i t e d i f f e r e n c e a p p r o x i m a t i o n to t h i s equation is

where the l a s t term indicates a second order of approximation. On
inspection it appears that Eq. 5.34 is consistent with Eq. 5.33 as A t
0. However, for this particular solution, stability considerations
r e q u i r e that
Ax
c 5  (5.36)
A t
97
Substituting this inequality into Eq. 5.34 transforms the error term into
(5.37)
which indicates a small error term, but one that can become s i g n i f i c a n t
if At  > 0 faster than Ax2  > 0. Since A x and At a r e f i n i t e and a r e
not approximately zero, Eq. 5.34 approximates Eq. 5.33 with second o r d e r
accuracy but with a term introducing artificial (numerical) dispersion.
This example was chosen because it illustrates how k i n e m a t i c models
can simulate a dispersing hydrograph. Eq. 5.33 is merely the kinematic
wave equation for no lateral inflow which, theoretically, cannot predict
hydrograph dispersion. Eq. 5.34 is one of the finite difference models
solution i s solution is
stable
___)
I
due t o numerical diff sion
Ax
b, at
Ax s h a l l be close to ()
At c r
to minimise d i f f u s i o n e r r o r s and obtain
optimal accuracy.
c) the difference scheme’shal I be convergent. T h i s c o u l d be a s c e r t a i n e d
by running the scheme with different Ax‘s and At’s and comparing
w i t h a n a l y t i c a l r e s u l t s i n a simple case.
must therefore h o l d :
Lix dx
 >  (5.39)
At = dt
This is referred to as the "CFL condition" after Courant, Friedrichs
a n d Lewy (1928), o r s i m p l y the Courant c r i t e r i o n f o r s t a b i l i t y .
t t
(a) ax
At
> dx
dt
I t has been noticed, however, that even if one satisfies the CFL
t
T
at
11 I i+l
4 t
dx AX
r = (3lQ / t (5.40)
EFFECT OF FRICTION
(5.44)
+ 1  0
[(Q"/K")~ + (Q" /K" 1'3 (5.45)
4
j j j+l j+l
Strelkoff (1970) indicates that the d i r e c t e x p l i c i t scheme i s i n h e r e n t l y
unstable. He indicates the Lax type scheme should satisfy the Courant
criterion. For implicit schemes he suggests that to ensure stability in
friction
KO
At < A g / T (5.46)
 at t = k  1 t = k  1
ax
il i i+l
INDEX
+ p o i n t s used f o r c a l c u l a t i n g d i s c h a r g e a t time t = k  1
0 p o i n t s used f o r c a l c u l a t i n g d e p t h a t time t = k  1
REFERENCES
CHAPTER 6
DIMENSIONLESS HYDROGRAPHS
UNIT HYDROGRAPHS
quest ion.
The hydrographs are intended for use b y d e s i g n e n g i n e e r s where n o t
only the hydrograph peak flow rate but the shape of the hydrograph is
Computer models can account for any time and space variation of
r a i n f a l I and catchment c h a r a c t e r i s t i c s as described l a t e r . T h e i r use e n t a i I s
substantial computer time and the model has to be used in conjunction
with various storm i n p u t s to ensure c r i t i c a l storm input. I n t h i s section,
runoff hydrographs off catchments of fixed shapes and with spatially
pa ramet ers.
The k i n e m a t i c equations h a v e been used to p r e p a r e the hydrographs
presented by Constantinides a n d Stephenson (1982). Computer solution of
the finite difference form of the equation of motion and the flow
resistance e q u a t i o n was performed f o r numerous s i t u a t i o n s . With the use of
dimensionless parameters the number of variables i s reduced c o n s i d e r a b l y
and a few graphs present a r a n g e of hydrographs covering the r a n g e of
parameters normal l y encountered.
Runoff hydrographs off three model catchments a r e presented, these
b e i n g the f o l l o w i n g :
( a ) A s l o p i n g p l a n e catchment
( b ) A c o n v e r g i n g surface catchment
( c ) A Vshaped catchment w i t h stream
L i s t of Symbols
x space a x i s a l o n g o v e r l a n d p l a n e (m o r f t )
z space a x i s a l o n g channel (m o r f t )
L length of o v e r l a n d p l a n e (m o r f t )
n roughness coefficient of o v e r l a n d p l a n e s
n roughness coefficient of channel o r stream
0 a n g l e d e s c r i b i n g c o n v e r g i n g surface catchment (radians)
r r a t i o d e s c r i b i n g c o n v e r g i n g s u r f a c e catchment
w w i d t h of o v e r l a n d flow i n c o n v e r g i n g surface catchment ( m or f t )
H depth of channel ( m or f t )
b w i d t h of channel (m o r f t )
yo depth of o v e r l a n d flow ( m or f t )
qo discharge p e r u n i t w i d t h of o v e r l a n d flow (m’/s or ft’/s)
ys depth of channel flow (m o r f t )
Q discharge of channel flow (m3/s or f t 3 / s )
Q discharge of c o n v e r g i n g s u r f a c e ( m 3 / s or f t 3 / s )
Kinematic equations
aQ
 +  aA 
ax at  q~ (6.1 1
and q = uym (6.2)
EXCESS RA I NFALL
rate) after the storm. Fig. 6.2 d e p i c t s the assumed excess r a i n f a l l input
and Fig. 6 . 1 shows the assumed rainfall input and loss d i s t r i b u t i o n f o r
o b t a i n i n g the excess r a i n f a l l d i s t r i b u t i o n shown i n Fig. 6.2.
109
(mm/h
U
VfC
i l 
* u ted t (h)
I
+4
i
* ld ).
*te
Rainfall intensit:
(mm/h)
t (hi
t
1 
1 
t
time runoff stops
(6.3a)
x =x
LO
P qo
= 
i eLo
T = mte

tc 0
mt
TD = ed
~
(6.9)
tCO
F = fc
 (6.10)
I
e
where t is the time of concentration of a sloping plane in kinematic
co
theory a n d i s g i v e n b y :
_ 1_ _
ap
+ aP = 1 f o r T < T
4 D
aT ax
= F for T (6.12)
> TD
&
Pmax versus' T
D
(6.13)
r e s p e c t to t a n d s u b s t i t u t i n g i n e q u a t i o n (6.13) yields:
a+Q o 5
wo
ll/rn
= i w for t j t
ax 7 0 0 at
= w
e o
f for t > td
ed
o c (6.16)
In addition to dimensionless variables defined in equations (6.8) to
(6.10) the following dimensionless variables are defined (Singh, 1975):
(6.17)
(6.18)
where (1r2 )/2 is the area of the catchment and r the ratio of bottom
segment to t h e t o t a l catchment r a d i u s .
For the converging surface tCO i s defined as the time of equilibrium
f o r a s l o p i n g catchment of l e n g t h L ( 1  r ) , i.e.
la
Lo(lr) 1 /m
~
ti0  Oieml] (6.19)
(6.20)
(6.21)
a Q ~+
__
bays
= 2q0L (6.22)
az at
A basic assumption in equation (6.22) is that the natural depth of the
channel is always greater than the water depth i n the channel. Another
assumption i s that the channel area i s small compared to t h e p l a n e a r e a .
The u n i f o r m f l o w r e s i s t a n c e e q u a t i o n f o r t h e c h a n n e l may b e w r i t t e n :
m
Qs = baSYs (6.23)
R = 0.05 F = 0 .0 0

0,
5 R=O .05
F=O. 50
Smax versus T
D
S
F i g . 6 .8 Dimensionless r u n o f f h y d r o g r a p h s f o r the c o n v e r g i n g s u r f a c e c a t c h m e n t
R = 0.05 F = 0.50
118
(6.24)
Q = Q , / ~ L ~ L ~ ~ ~ (6.25)
2 = z/L!j (6.26)
where t is the same as for the sloping plane, i.e. equation (6.11).
LO
Substituting for Q ,z,t,qo, and rn in equation (6.24) and rearranging
yields:
(6.27)
where G = (c2L5
5
)
0.6 ba 0.6
2L0
(6.28)
H > y s a t a1
C a tc hme n t
9
d
Q 
5
9
d
61
d
3
d
8
d
R
d
3
d
Fig. 6 .ll Dimensionless runoff hydrograph for the Vshaped catchment with stream
G = 0.5 F = 0.5
I
Fig. 6 .12 Dimensionless r u n o f f h y d r o g r a p h s for the Vshaped catchment b v i t h stream
G = 2.0 F = 0.0
G=2.0
F i g . 6 .13 D i m e n s i o n l e s s r u n o f f h y d r o g r a p h s f o r t h e V  s h a p e d catchment w i t h s t r e a m
G = 2.0 F = 0.5
123
X
w
0
0 2 4 6 8 10
EXCESS STORM DURATION (HOURS]
a n d b e d slopes
Problem
So Iu t ion
G = Z(1350)
2 (308.9)
Figs. 6.10 and 6.11 with G = 0.5 are used for choosing the critical
Outlef
Scale I :7500
___r  .
0 .6 ~ 0.4
( 3 . 6 ~ 1O b )
3600
0 1
var j.a bX e
t
CO
a Qs I F a c t o r s t.o d i m e n o j . o n
r u n o f F hydrograph
M u 1t l p l y
a x is
dirnensionl.ess
source guess excess IDF's hydrographs
1
11.66 2.70 1 P.620 I 2.942
~~..,.._.......I_ 
3.0
2 5
n
E
I .o
0.5
REFERENCES
CHAPTER 7
DES I GN PRACT I C E
makes any other basis for design difficult. There is little information
available on instantaneous precipitation rates, storm cell size and cell
movement. Time average precipitation r a t e or precipitation depth can be
predicted from intensitydurationfrequency curves (e.g. Van Wyk and
Midgley, 1966) o r equations such as that of Bell (1969). The most common
method of abstracting data from rainfall records is to select a duration
and calculate the maximum storm precipitation in that period. The so
defined storm may include times of low rainfall intensity immediately
preceding a n d succeeding a more intense p r e c i p i t a t i o n r a t e .
Such simp1 i f i c a t i o n s in data render runoff calculation simplistic.
Even when employing numerical models it is simplest to use a uniform
intensity hyetograph for every point on the catchment. Although time
varying storms are sometimes used, the precipitation pattern is seldom
r e l a t e d to the maximum possible r u n o f f rate.
Warnings have been made against simplification in rainfal I
patterns. For example, James and Scheckenberger (1983) indicated that
storm movement can affect' the runoff hydrograph significantly. Eagleson
(1978) has expounded on the s p a t i a l variability of storms a n d Huff (1967)
studied the time v a r i a b i l i t y of storms.
Although much research has been done on storm variability,
relatively little has been published on the resulting effects on runoff
hydrographs (Stephenson, 1984). Research appears to h a v e concentrated on
models of particular (monitored) storms over particular catchments. The
design engineer or hydrologist does not have sufficient guidance as to
what storm pattern to design for. Presumably certain rainfal I sequences,
STORM PATTERNS
The position of the peak intensity could be varied and was observed
mm/h
i
'
1
Spat i a I d i s t r i b u t i o n
The nature of storm cells within a potential rain area has been
documented by many researchers e.g. Waymire and Gupta (1981). The
persistence of storms observed in the northern hemisphere has not been
found i n c o u n t r i e s south of the equator however ( C a r t e , 1979). The l a r g e r
air mass within which storm cells occur is referred to as the synoptic
area (see Fig. 7.2). The synoptic area can last for 1 to 3 days and
the size is generally greater than 104km2. W i t h i n the synoptic area are
large mesoscale areas (LMSA) of lo3 to lo4 km2 which have a l i f e of
several hours. Sometimes small mesoscale areas (SMSA) of lo2 to lo3 k m 2 can
e x i s t simultaneously. Within the mesoscale areas or sometimes on their
own, convective cells, which are r e g i o n s of cumulus convective precipi
tation, exist. These may have a n a r e a extent of 10 to 30 km2 a n d h a v e
an average l i f e of several minutes to half an hour. These cells a r e of
concern to the hydrologist involved in stormwater design. By comparing
the storm cell size with the catchment size he can decide whether the
133
l a p p i n g small storms.
Synoptic a r e a 7
The shape of the storm cell has significance for catchments larger
than the cell. Scheckenberger (1984) indicates that the cells are ellip
tical which may be related to storm movement. The rainfall intensity
is highest at the centre a n d decreases outwards. The intensity has been
shown to decrease exponentially, radially outwards from the focus, in
various localities as in Fig. 7.3 (Wilson et al., 1979). Generally the
variability i n i n t e n s i t y does not necessarily cause h i g h e r r u n o f f intensities
but on small catchments near the centre of the c e l l the average precipi
tation can be higher than for a larger catchment, and as a rule, the
r a i n f a l l depth increases the smaller the storm area.
S t o r m movement
NUMERICAL MODELS
Fig. 7.3 I l l u s t r a t i o n of s p a t i a l d i s t r i b u t i o n of p r e c i p i t a t i o n i n t e n s i t y
Kinematic equations
x = x/L
T = t/tc
I = i / i
e a
Q = q/iaL
where L i s t h e l e n g t h o f o v e r l a n d flow, i a i s t h e t i m e and s p a c e a v e r a g e d
excess rainfall rate and tc is the time to equilibrium, or time of
concentration, for an average excess rainfall i
a
. Subscript c r e f e r s to
time of concentration, d to storm duration, a to t i m e and s p a c e a v e r a g e
and p to peak. Then the following expression for t can be derived:
Numerical Scheme
AT/AX to ensure s t a b i l i t y .
Woolhiser (1977) documented various numerical schemes including
very accurate methods such as LaxWendroff ' s . Brakensiek (1967)
suggested 3 schemes: f o u r p o i n t , i m p l i c i t and e x p l i c i t . His second scheme
(implicit) is adopted here as it is accurate and r a p i d for the examples
chosen.
I M 'l M *
X
Fig. 7.5 XT g r i d employed i n numerical solution
aQ  Q1Q2
ax AX
Q 1 + Q Q Q
aQ  2 3 4
2T 2AT
f i n i t e difference a p p r o x i m a t i o n to the d i f f e r e n t i a l e q u a t i o n i s t h u s :
137
Q +€I0 . 4 Q2 Q +Q Q
(3( ) 3 4 2
+
_
2AT
1
_
5/3
AX
 .
~ (
Q3+Q4
2
)
0.4
'I
T 1 T
P
Fig. 7.6 Temporally v a r y i n g storm
138
Q
.
Constant exce'
0 1 2 T
I . _ . . . _ . _ . a
0 1 2 1
a uniform storm however, the peak can be higher. Fig. 7.8 is for a
storm of constant volume peaking at its termination (TP = 1) and for
durations represented by Td = 0.4 to 1.2. These hydrographs are for
storms of equal volume so that the shorter duration storms are of a
higher intensity than longer d u r a t i o n storms. Depending on the IDF c u r v e
then a short duration storm may or may not result in a higher runoff
rate than for one of duration equal to the concentration time of the
catchment.
It should be recalled that all other hydrographs plotted are for
a specified excess rate of precipitation. That is, if the hyetograph is
uniform so are the abstractions. In practice, losses will be higher at
of rain will appear as runoff n?ar the end. This tends to make the
excess rain versus time graph concave upwards if the hyetograph was
a straightlined triangle. This effect is not modelled here but all the
effects result in a higher peak than for a uniform input. Scheckenberger
in fact indicates peaks up to 30% greater than for uniform storms due
to the sum of these effects.
140
Spatial v a r i a t i o n s
I
peak intensity
'P
=2
I,
_
=1
0 1 2 1
Fig. 7.10 Simulated dimensionless hydrographs caused by steady semi
infinite storms of varying distribution down catchment (Fig.
7.9).
141
cate that the runoff n e v e r exceeds that for a rectangular spatial distri
bution of rainfall. The resulting dimensionless time to equilibrium is
nearly unity for all cases, implying t h e same t i m e o f concentration holds
for uneven distribution as for uniform distribution of rain. There is
therefore not a chance of a shorter duration storm with a higher
intensity contributing to a greater peak than the uniform storm (unless
the intensityduration curve is a b n o r m a l l y s t e e p ) s i n c e t h e t i m e t o e q u i l 
i b r i u m i s n o t r e d u c e d r e l a t i v e to a u n i f o r m storm.
X=x/L
cotchmcnt lcngth
F i g . 7.11 C a t c h m e n t w i t h a s t o r m m o v i n g d o w n i t
9 1 1
1 2 1
Moving storms
Q
1
cS=
0 1 ' 2 7
g r e a t e r t h a n f o r a s t a t i o n a r y storm.
For storms of limited extent travelling up the catchment, the peak
flow was observed ' t o be less t h a n for a stationary storm a n d the f a s t e r
the speed of t r a v e l of the storm the s m a l l e r the peak r u n o f f .
143
REFERENCES
CHAPTER 8
CONDUIT FLOW
function of water depth. The sides of the channel (and top in the case
of closed c o n d u i t s ) increase f r i c t i o n drag. As far as the form of the
basic kinematic equations is concerned the mathematical expressions
become more complicated, and numerical solutions are necessary in the
m a j o r i y y o f cases.
The c o n t i n u i t y equation remains
where the first term is the rate of rise, the second prism storage and
t h e t h i r d wedge s t o r a g e .
The dynamic e q u a t i o n reduces to
M
Q = aAR (8.3)
where Q is the discharge rate, a IS a function of conduit roughness,
q is inflow per unit length, B is the surface width, A is the cross
sectional area of flow and R is the hydraulic radius A/P where P is
the wetted perimeter. Employing Manning's f r i c t i o n equation,
LIZ
a = K S /n a n d M = 2/3 (8.4)
1
where K1 = l(S.1. u n i t s ) and 1.486 (itsec units)
n = M a n n i n g ' s roughness coefficient
A = D 2 a . a
( 0 cossin)
4 2 2 2
a n d P = DO
2
Thus i f one takes 0 as the variable, the continuity equation becomes
aA a o
_ aa

ax
+
aa at = q ;
and
’; (1 + s i n 2 ~  cos2 0 ao i@ =
(8.7)
2 %)at+ ax
a = o + ( q  G Q) 8 Gt
AX
2 1 ~ 2 ( l + s i n 2 5 cos’g) (8.8)
2 2
a n d i n terms of t h e new , s i n c e 61 = aARZ3
0 .a
. a cossin 2 3
a D2 0  2 2
Q =
4
(3 cossln)
2 2
{ z(l
o’} (8.9)
i
e
= a
b + t d (8.10)
designed.
148
*
\ drain
L s u b c o t c h m e n t boundary
4 ‘drain number
Program description
input format is described below. Data is read in free format and can
F i r s t l i n e of d a t a :
M, A, B, E , IN, IR, 11, G.
Second and subsequent lines of data (one l i n e f o r each l e n g t h of pipe):
x(I), s(I), z(I), C(I), SO(I), EO(I), I B ( I ) .
The i n p u t symbols a r e e x p l a i n e d below:
The pipe data are next read in line by line for M pipes.
As the program stands, 98 individual pipes are permitted,
and any number of legs subject to the maximum number of
pipes.
between.
The order in which pipes are tabulated should be obtained
as f o l l o w s :
151
L.OOO1
L.0002
L.0003
L.0004 5
L.0005 10
L.0006
L.0007
L.0008
L. 0 0 0 9
L.OO1O
L.OO1l
L.0012 12
L.0013
L.0014
L.0015 13
~ . 0 0 1 6I 5
L.0017
L.OOI~
L.0019
L .0520
1.0021
L.0022
L.5023
L.5024
L.0025 2 0
~.0026
L.OU27
~.502a
L.0029
L.0030
L.UJ31 23
L.0032
L.UJ33
L.0024
L.0533
L.3036
L.i)037
L.UO3J
L.0039
L.0040
L.0041
L.0042 30
L.0063
L.JU44
L.0045
L.0040
L.JJ47
L.304d 32
L.0049
L.305S 35
L.J05l 40
L.>il52 45
L.dOb3 5u
L.3054
L.3055
L.5056 100
L.3057
L.5058 110
L.0054 120
L.0060
Led061
L.SO62
LSJt3
L.0064 201)
L.0065 44(3./4. I
L.0066
L.0067
L.0068
L.0069 290
L.0070 300
L.0071
L.0072 350
L.0073
L.0074 6 0
L.0075
L.0076 400
L.5077 7 0
L.0078
L.0079 6 0
L.00eO S I O P
L.0081 END
L.5001 S T U d M SEWER C € S I G N
L.0002 P I P E LENGTh C I A C R A O E DSFLC/S STORM S AREA
~.0003 I 100. 576 .0020 244 1016. 20006.
L.OO0'4 2 150. 514  0 0 4 C .155 Y11. 20000.
L.JJU> 3 200.  6 4 3 0040 .162 206d. 40000.
L.0306 4 100. 415 .0020 102 772. 10000+
~.0007 5 100. .574 .0040 .342 20od. 40000.
~.u006 6 200 .613 .0040 .u17 2068. 10000.
L.00U'i 7 2CO.  2 5 3 .0020 .o9b 2068. 40000.
1.i) J 10 @ 100.  5 0 5 .0050 1.287 2068. 20000.
LaOJ11 DATA @ .0751440..0010 2563 301) 60
152
After drawing out a plan of the catchment with each pipe, the
longest leg possible is marked, starting from the outfall, then success
ively shorter legs on first the longest, then successively shorter pipes.
Now the pipes are numbered in the reverse over, starting at the top
of the shortest leg etc. Proceed down each leg w i t h the numbering until
a junction is reached. Never proceed past a branch which has not been
tabulated previously. In this way all pipes leading into a pipe will
have had their diameters calculated before the next lower pipe is
designed.
Sample Input
TRAPEZOIDAL CHANNELS
A = Y (b + y/S1 + Y/SI)
P = b + y 1 J(1 + 1/S12) + i ( l + 1/S2’))l
In particular for a vertical sided rectangular channel of limited width
b, employing the M a n n i n g equation.
A yb,
P = b + 2 y
Q  ayb yb )23
(b+2y
= ~ i ( y b ) ~/ (’ b~+ 2 y I Z 3
as well as their relative time positions are important for the accurate
( 8 .1 1 )
(a) Pipe
I
(b) Trapezoid
m1
q = a a R (8.12)
0 = 1
n
5 112 a n d m = 5/3 (8.13)
1
q =  s
4 .5/3
(8.14)
p2/'3
The geometry of the c o n d u i t s i s d e s c r i b e d b y e q u a t i o n s 8.15  8.18
I56
(8.15)
(8.16)
A = by + y2 t a n (90  0 (8.17)
Trapezoid
P = b + 2 y sec (90  0) (8.18)
For the t r a p e z o i d ,
Q = q/qc (8.23)
A = a/b2 (8.24)
P = p/b (8.25)
Y = y/b (8.26)
a n d f o r both sections
x = x/L (8.27)
T = t/t (8.28)
k
where q i s t h e maximum f l o w c a p a c i t y of t h e p i p e ( m 3 / s ) , 7
0 . 335285~h d 8 / 3
m
qc is a discharge variable, being a function of
f r i c t i o n coefficients
0 , m a n d bottom w i d t h of t r a p e z o i d , b (m’/s) i.e. qc = 5 112 b 8 1 31
n tk ’
i s a t i m e c o n s t a n t ( 5 ) a n d L i s t h e l e n g t h of t h e c o n d u i t (rn).
To define the discharge and time constants appropriately the
(8.29)
157
Rearranging yields:
(8.30)
t = Ld2
(8.31)
qm
For t h e t r a p e z o i d :
t = LbZ
(8.32)
qc
where the d i m e n s i o n a l c o n t i n u i t y equation i s :

aQ aA 
a x + aT
  (8.33)
f o r the p i p e :
(8.34)
w h e r e t h e m a x i m u m c a r r y i n g c a p a c i t y of a p i p e c a n b e s h o w n to b e
F o r t h e t r a p e z o i d t h e u n i f o r m f l o w e q u a t i o n r e d u c e s to:
Q q c = 1S z  ( A b ‘ ) 5 / 3 (8.37)
(Pd)2/3
Defining q as in equation 8.38 reduced equation 8.37 to the dimension
C
l e s s f l o w e q u a t i o n f o r t h e t r a p e z o i d g i v e n i n e q u a t i o n 8.39:
‘c
 tn ,t b8/3 (8.38)
For the t r a p e z o i d :
A 5/3
Q =  (8.39)
2
,3
/
Pipe
1
A =  cos
4
1
(12Y)  (1  Y )
2
. (‘f  Y2)1/2 (8.40)
1
P = cos (12Y) (8.41)
L
t = (8.42)
Channel
A = Y + Y2 tan (900) (8.43)
P = 1 + 2Y sec (900) (8.44)
(8.45)
one a uniform and the other a triangular time distribution. These two
time distributions were chosen as they represent extreme cases, i.e. a
natural runoff hydrograph, from overland flow, would have a shape
between these two extremes depending on the rainfall and catchment
c h a r a c t e r i st i cs.
In addition to the shapes the hydrographs were assumed to have
a variety of durations and intensities. Fig. 8.5 illustrates the inflow
I
h y d r o g r a p h s i n t h e i r dimensionless form.
QIM ‘IM
where t d i s d u r a t i o n .
_
dT
 C1
(8.46)
a n d for t rapezoi ds :
aA
2P
 ’
2
cos (900) + Y sin (900) (8.48)
2. 4
PIPE
2. 0 
DEPTH/DIhMETER
(0
4
4
t
N
4
0
...
.LP/XP
0
0
C
(0
9
N
d
Computer S i m u I a t ion
b) The ratio of peak at the outlet over peak at the inlet increases
with increasing storm duration ( f o r a constant inflow factor) or i n other
words inflow hydrographs of smal l e r storm duration undergo higher
discharge attenuation than hydrographs of longer duration. The reason
for this is the same as in a), i.e. lower volumes spread more than
bigger volumes resulting in lower depths of flow and thus lower dis
charges.
h i g h e r i n f l o w f a c t o r s i m p l y h i g h e r volumes of water.
constant inflow duration) than for high inflow factors, the reason b e i n g
the same as f o r o b s e r v a t i o n b ) .
Inflow h y d r o g r a p h d u r i t i o n / t c
g/
8
9
' Uniform lnpui
/'
6
KI nemal I c
r u u t 1 n g R",t
be used TlrneSt11fi ,
 '
1.2
\\ PIPE
1.1
1 .o
0.9
.
I
," 0.8
0.7
0.6
I I 1 I
0.5
0 0.2 0,Q 0.5 0.8 0
iinnffllooww ppeeaakk dd >> ss cc hh aa rr gg ee // qq m
m
inflow peak ratio and obtaining the dimensionless duration for a peak
ratio of 0.9, the peak ratio of 0.9 corresponding to a 10% peak
attenuation. The r e s u l t s a r e summarised i n Figs. 8.8  8.10.
Two assumptions are currently popular for calculating the time lag
of a hydrograph to be routed by time shift methods. The time lag is
Method 1 (TLp = L/(qm/am).) The time constant (tk) for the pipe is
tLP  a rn
 
t dZ
k
where a /d2 i s the dimensionless flow a r e a for a pipe discharging at
rn
maximum c a p a c i t y . S u b s t i t u t i n g f o r arn/dz
‘LP
 = 0.7653 (8.50)
t
k
166
This gives the dimensionless lag time for a pipe and plots i n Fig.
(8.51)
2.5
Time s h i f t method
i.c
_' i .5
Unlforrn I n p u t
1.0
i
_I
0 7 4 6 8
Inflow peak d i r c h a r g e / q c
TRAPEZOID
A N G L E  90'
\\\

/
Triangular Input
I I I
6
I n f l o w peak d>scharge/q,
t
LP Ai
~
tk 0 (8.52)
higher but closely resemble the ones from kinematic routing using a
uniform input. Note that t h i s was also the case f o r the p i p e . The reasons
for their resemblance a r e s i m i l a r to those for the p i p e a n d a r e discussed
i n the p r e v i o u s section.
It can be seen from Figs. 8.8 to 8.10 that the dimensionless inflow
duration is much more critical than the dimensionless inflow peak
discharge for determining whether time shift methods can be used. This
is more apparent in the case of trapezoids where the 10% peak attenua
tion curves appear almost vertical for dimensionless inflow peak dis
charge values g r e a t e r than 2.0.
Furthermore, it can be seen that the dimensionless infow duration
decreases with increasing inflow factor. This is expected as inflow
hydrographs with a similar inflow factor need bigger durations than
ones with a higher inflow factor for both inflow hydrographs to have
similar volumes. As was discussed earlier, higher inflow volumes will
imply smal l e r peak attenuation, other parameters being constant, one
exception to this observation being the pipe for inflow factors higher
than 0.8. It can be seen from Fig. 8.8 that as the inflow factor
approaches unity the dimensionless duration (causing a 10% peak
a t t e n u a t i o n to the i n f l o w h y d r o g r a p h ) increases.
This is probably due to the fact that a pipe discharges more when
not f l o w i n g f u l I as a l r e a d y discussed.
It will also be noted that for trapezoids and discharge inflow
establish the necessity of routing while Figs. 8.11 to 8.13 can be used
171
to calculate a lag time for the cases for which time shift routing is
shown to b e ade a u a te .
REFERENCES
Case 1 , 2 and 3
Case 4
i = __
(o.24 70 1 . 5 ) . m 10 = 42.710 = 32.7 mm/h
e
td  t = 1.5  30/42.7 = 0.80h
ted =
F = 10/32.7 = 0.31
Peak f l o w Q = QAie = 0 . 8 5 ~ 2 ~ 6
1x32.7/3.6x106
0 = 15.4m3/s
a r e a c o n t r i b u t e s to the peak.
Not much sense can be made out of comparing the r e s u l t i n g r a t i o n a l
coefficients (ratio of peak runoff rate to rainfall rate times catchment
area). That is because the time of concentration for each case is
different due to differing roughness, rainfall rate etc. In any case it
is irrelevant when it comes to critical storm duration which is shorter
t h a n the time to e q u i l i b r i u m .
179
D E T E N T I O N STORAGE
aA
  aa (9.1 1
at ax
(9.2)
0 2 = c, I I+ c 21 2 + caO1 (9.4)
where cI, c and cg a r e functions of A x and A t. The l a t t e r equation
is referred to as Muskingum's equation used in routing floods along
channels. if X = 0 the routing equation corresponds to level pool or
CHANNEL STORAGE
R is the hydraulic radius A/P where A is the area of flow and P the
wetted perimeter. R can be approximated by depth y f o r wide r e c t a n g u l a r
channels. S is the energy gradient, f is the friction factor and k is
a l i n e a r measure of roughness analogous to the N i k u r a d s e roughness.
Both the roughness. coefficient CY and the exponent m of R or y in
the general flow equation (9.11) affect the peak flow off a catchment.
This is largely due to the attenuating effect of friction resulting in a
larger time to equi l i b r i u m . A rainfal I excess intensityduration relation
ship i s required to e v a l u a t e the effect of each coefficient on peak runoff
r a t e and maximum catchment storage. The f o l l o w i n g expression f o r excess
r a i n f a l I intensity i s assumed:
(9.9)
aa t v . 3ax = 'e
(9.10)
181
q = aym (9.11)
then i t may be shown that tc = (L/aiem')'/m where q i s the r u n o f f r a t e
per unit width of the catchment and y is the flow depth. The rising
0 p . , , ,
1
, , , , . . , , , ,
2 '
,
m
Fig. 9.4 Hydrograph shapes for different values of m in q = ay
a.
_ 1
1 L/a(a/3600000)m1 I
''lm
(9.13)
I c + I P
3600 ( ie/a )
m1
The term L/ua is r e f e r r e d to as the length factor. The constants are
introduced for a in mm/h, and time of concentration i n hour units. The
maximum peak flow factor ie/a is plotted against length factor in Fig.
9.5, since i t i s not easy to solve (9.13) d i r e c t l y f o r i /a
i e / a and s/a
max
2 c=o.9
b=0.25h 1
,/f )
s * .. . I I I , * ] I
00
n 1 I 1 I
L 1.i a
.
rn 1 . . jC
I0
q = i x
e
m
= ay
therefore y = (iex/a)lIm
I n t e g r a t i n g y w i t h respect to x y i e l d s the t o t a l volume on the catchment
1 /m
(
L 1/m
)   
, (9.14)
m+l a a ( a/3600000) m' 3600
minimum
cost
$
T h e c o n t i n u i t y e q u a t i o n becomes ( s e e F i g . 9.7)
(Qi+,Qi) qi + Ai
dhi
=O (9.15)
dt
where the reservoir surface area A. replaces B dx in the open channel
continuity equation where B is the catchment width. q is the reservoir
i n f l o w here. The d y n a m i c e q u a t i o n i s r e p l a c e d b y
Q. = aAm (9.16a)
where A is the (constant) conduit cross sectional area. Since the kine
matic equations omit the dependency of Q o n h e a d d i f f e r e n c e h, t h e l a t t e r
equation assumes the head gradient along the pipe equals the pipe
gradient, i.e. freesurface just full flow. Since A is a constant it is
r e l a t i v e l y easy to r e p l a c e the l a s t e q u a t i o n b y o n e of t h e f o r m
185
Q. = 01 Ah.m (9.16b)
This equation is applicable to free discharge from an orifice o r over a
3
1.4m 3/s
1. Om
In the first data line after the name, the simulation duration and
increment in hours is added at the end of the line. In the p i p e d a t a ,
the first pipes should be from the various reservoirs with the surface
areas of the upstream reservoirs in square metres given at the end of
the pipe data lines. In order to display the reservoir levels in the
biggest reservoir it is necessary to have a supply pipe from a pseudo
fixed head, very large, reservoir to represent a pumped supply feeding
i n t o the a c t u a l b i g g e s t level r e s e r v o i r i n the d i s t r i b u t i o n system.
The selection of 'upper' and 'lower' nodes f o r any pipe, numbered
even if a common factor is fed in. To print out 'old data' in file, it
is necessary to go into revision mode (2) of pipe data input. To get
out of r e v i s i o n mode, type 0 f o r p i p e number to be revised.
Part of the data i s read interactively on the keyboard. The first
lines of data (name, duration, no. of nodes, reservoir data and pump
data) is typed for each run. The pipe and node data can b e typed in
or r e t r i e v e d from a f i l e o r ammended i n a f i l e .
The time increment between iterations for simulatiot mode must
be small enough to avoid large variations in water levels n reservoirs
between iterations. The reservoir surface area and flows will control
this.
Additional pipes can be added in edit ( 2 ) mode a n d dill then be
stored in the data file. Pipes can only be removed by limiting the
number of pipes in the initial lines to e l i m i n a t e those not required at
the end. The other way is to put a very small diameter for a pipe to
be removed from the network. New nodes or reservoirs can be added b y
r e t y p i n g i n data.
188
of pipes in the data file should be specified, The number will auto
m a t i c a l l y be increased when more d a t a l i n e s a r e added.
The l a s t s p e c i f i c a t i o n of any drawoff i s retained if a node happens
to be s p e c i f i e d more than once in input. One should also make sure each
node i s s p e c i f i e d ( a s a N2) at least once to define i t s drawoff.
Data Input
Each line may contain more than one unit of data separated by
commas.
Surface a r e a of r e s e r v o i r , mz
Line 7 O l d ( 0 ) o r new ( 1 ) o r revised pipe data ( 2 ) ;
type 0 , l or 2.
L i n e s 8.. . (one f o r each p i p e i n new p i p e d a t a )
Node 1 no.
Node 2 no.
Pipe l e n g t h rn
Pipe i n s i d e d i a . , rn
Drawoff a t node 2 , m3/s
(Darcy f r i c t i o n f a c t o r i f line 4 is 1 )
Line 9 I f line 7 i s 2, w i l l ask p i p e no. for revision.
L i n e 10 Pipe data for new p i p e s as f o r Line 8 including Darcy
f r i c t i o n factor.
Line 11 No. of pumps or pressure reducing valves (one per
pipe).
L i n e s 12... Pipe no. in which pump or PRV is installed, pumping
head or PRV head loss (  ) in rn.
189
L i s t of Symbols i n P r o g r a m
1 = analysis, 2 = simulation
0 = constant f, 1 = v a r y i n g Darcy f.
0 = o l d data, 1 = new d a t a , 2 = revise old data
0 = no data listing required, 1 = required
h e a d Ios s/Q I Q 1
ZH f o r e a c h SOR
CAF
p i p e diameter (m)
old v a l u e of H ( I )
Darcy friction f a c t o r e.g. 0.012 l a r g e dia. clean p i p e
0.03 small tuberculated pipe
common D a r c y f a c t o r
head a t node or j u n c t i o n I
node counter
n u m b e r of n o d e s
u p p e r node number of p i p e
lower node number on p i p e
n u m b e r of r e s e r v o i r t y p e nodes
i tera t ion
pipe counter
node counter
pipe counter
number of connecting p i p e s
M 2 ( L , M l ( L ) )p i p e n u m b e r c o n n e c t i n g
N$ a l p h a n c r m e r i c n a m e of s y s t e m , up t o 12 c h a r a c t e r s
NO maximum number m a i n i t e r a t i o n s p e r m i t t e d e.g. 4: t 5
N1 m a x i m u m n u m b e r SOR ( s u c c e s s i v e o v e r  r e l a x a t i o n o f
simu l taneous e q u a t i o n s ) i t e r a t i o n s e.g. 4T t 10
N2 counter for main iterations
N3 c o u n t e r f o r SOR i t e r a t i o n s
P n u m b e r of p i p e s
P1 n u m e r o f p i p e s and P R V ' s ( 1 p e r p i p e m a x i m u m )
Q(K) flow in p i p e
Q1 d r a w o f f m3/ s
Q2( I ) d r a w o f f m3 /s
R(k) pump head in m, (or pressure reducing valve head in m
if n e g a t i v e )
S g n2 / 8
190
S ( 2 )I CKij
s3 CHj
S4( I CKijHj
55 o l d Q(K) f o r averaging
T3 drawoff duration, m i n s e.g. 8 h x 60 = 480
T4 simulation duration, m i n s e.g. 24 x 60 = 1440
T5 time increment i n simulation, m i n s e.g. 60
TO tolerance on head in m e.g. 0.0001
T1 t o l e r a n c e o n SOR i n m e . g . 0.01
WSOR f a c t o r e.g. 1.3 (12)
X(K) pipe length m
REFERENCES
1110
1128
1138 NEXT K
1 1 4 0 FOR K = l TO P
1150 S2(Jl<K))=S2<Jl(K))C(K)/hB
StQCK))
1168 S 2 ( J 2 ( K ) ) = S 2 < J 2 < K ) >  C < K ) / A B
S(Q(K))
1170 NEXT K
1188 FOR K = l TO N 1
1190 C 2 = 8
1280 S 3 = 0
1210 N3=N3+1
1220 I F J3+1>J THEN 1380
1238 FOR L=J3+1 TO J
1 2 4 8 S 4 <L)=0
1250 FOR fl3=l TO M l ( L )
1260 Fl=H2<L,M3)
1270
1288 /hB
1298
1388 /AB
1310 NEXT M 3
1320 OE=H<L>
1338
1348
1350
1360 NEXT L
1370 I F C2/S3<=T1 THEN 1390
1388 NEXT K
1398 FOR K = l TO P ! NEW FLOtJS
1400
1410
1428
1430
1440
1450
1468
1470
1488 NEXT K
1450 I F C3/P<=T0 THEN 1510
1588 NEXT I
1510
~~~~ FOR L = l TO ~3 ! RES LEVEL.s
1520 H ( L ) = H ( L )  Q 2 < L ) * T S /~
H(L~
1530 FOR M 3 = 1 TO M l < L >
1540 M=M2<L,M3)
1550 IF J l ( M ) < > L THEN 1578
1560 H(L)=H(L)Q(M>*T5/A(L)
1570 I F J2<M)<>L THEN 1590
1580 H(L)=H(L)+Q(M)*T5/A<L)
1558 NEXT M 3
1680 NEXT L
1610 P R I N T USING "K,DDDDOD,X,K
DDDD . O D " i " T s = " , T6, " H i = "
(1)
1 6 2 8 P R I N T "NODE1 NZ Xm Om 12
m3/s H2m ''
1630 F O R K=I TO r
1 6 4 8 P R I N T CLSING 1668 j J l ( K ? , J 2
iK),X(K),D(K>,Q(K),H~J~~~))
1658 NEXT k
1 6 6 B IMAGE O D D . ~ D D O ~ O D D D D , O D . D D D ,
O D D . DDD, O D D D O . n
1678 NEXT TE.
1688 STOP
ASSIGN# 1 TO X
END
193
lOOm

1
1000mx0.15
0.09
7
958
7 1888
8Bb
.ZSB .678
.l5G
.2@8
.614
.854
E5. 1
53.2
53.2
5 4 708 158 .816 71.4
4 3 988 .1SO .81b 69.2
8 3 788 .ZBB .634 65.2
3 2 688 .2BU .853 76.1
T s = 4 3 2 8 6 H1= 188.88
NODE1 t42 Xn Drn QmS/s H2m
1 5 1588 .JBB .810 5B.2
5 6 888 .280 .689 97.9
5 8 958 .258 .837 95.9
.i'  77 1088
Y 808
.158
.2B8 .888
96.'
96.3
5 4 76P .is(? ,017 53.9
4 3 98B ,150 817 8:3. 3
8 3 768 288 Ei46 88.3
3 2 688 .288 ,863 78.4
Ts=576813 H1= 1 8 0 . 88
HlnDEl N 2 X m Om U N I ~ / C H2m
8 , 3 , 7 6 8 ,. 2 , . 0 3 1 5 1888 .817 95.4
9 'i 5 6 860 .288 ,885 59.6
3,2,668, . 2 , 8 5 8 958 .2513 .,836 97.2
N O . PUMPS/PRVs? rj 7 I8Sb .I58 97.5
1 8 7 888 .280 .Be8 97.5
PIPEN,+HEADm Nlt42 1 ? 5 4 7QB .150 ,816 95.3
1>1 4 3 588 .150 .81G 50.6
8 3 ;fit3 .200, .845 re.@
3 T' .2QB ,861 S8.G
194
CHAPTER 10
K I NEMAT I C MODELL I NG
INTRODUCTION
STORMWATER MODELL I NG
the direct response is mostly through overland flow due to the high
degree of imperviousness. I n contrast, in rural watersheds, an overland
flow component may be nonexistant and direct storm response may be
o n l y near the stream a n d occur as shallow subsurface flow.
model will be based upon the extent to which it can be or has been
verified. Model verification is a function of the data available to test
The modelling process is not new but is nothing more than a modern
expression of the classical scientific thought processes involved in the
design of an experiment. What i s new i s that today a very l a r g e number
of concepts can be evaluated efficiently in a very s m a l l amount of time
at a relatively s m a l l expense u s i n g computers a n d the body of a n a l y t i c a l
L
t
FEEDBACK
FEEDBACK
SYSTEM DEF I N I T I ON
e.g., s t o r m r a i n f a l l and s t o r m r u n o f f .
processes depends on the purpose for the model and how m u c h data a r e
a v a i l a b l e w i t h w h i c h to v e r i f y the model.
In conclusion, the essence of systems analysis as applied to storm
water modelling i s to interrelate rainfall (input) to stormwater (output)
w i t h a r e l i a b l e model in a c o m p u t a t i o n a l l y e f f i c i e n t m a n n e r .
TERM I NOLOGY AND DEF I N I T I O N S
response to known i n p u t .
Two concepts that are frequently confused (misused) are analysis
and simuhtion. The confusion with analysis stems from what it i s being
used to describe. As it relates to stormwater models, analysis is the
procedure used to calibrate a model to the data. It i s an attempt to
MODELL I NG APPROACHES
The c o n t i n u i t y e q u a t i o n becomes
(10.1)
(10.5)
(10.6)
and a x = f u n c t i o n of 5
ox
a z= f u n c t i o n of S
02
This idea for two dimensional flow was used by Orlob (1972). It
will be noticed that qt is always positive while q and q can be
yk = 0.0 (10.8)
I
so qk = i ~ x / 2 (10.9)
I e
One could alternatively assume that the discharge at the origin is
controlled by the depth of water at the origin as assumed for the rest
of the points. For this case we must then use the same equations as
with the other point's. The effect of using the two different boundary
I n i t ia I conditions
After the first time step it may be assumed that the water depth
at al I points, except at the origin in the case of the first boundary
condition, for the case of a n i n i t i a l l y d r y catchment
(10.10)
2000 00
f 1900
f ' 70.0
208
REFERENCES
CHAPTER 11
APPROACHES
models were selected from the range of available models because they
are simple in concept and structure, have been tested extensively, and
are representative of the approaches taken in developing kinematic
watershed models. For these reasons, they should help the r e a d e r to more
f u l l y u n d e r s t a n d k i n e m a t i c m o d e l s and t h e i r a p p l i c a t i o n s .
The reader is reminded that with any watershed model, approx
imations and simplifications are made. Previously, we have seen that
kinematic models are simplifications to the dynamic wave models; and
that their solution, whether analytical or numerical, requires approx
(1973). Input to the model includes daily rainfall, storm rainfall, daily
pan evaporation and a physical definition of the drainage basin
discretized into as many as 50 segments, including overland flow,
channel and reservoir segments. D u r i n g storm days, t h e model generates
a simulated discharge hydrograph based on input data from as many
as three rain gauges. Th e model consists of two m a i n sets o f components:
I n f i l t r a t i o n Component
(11.1)
RAINFALL
I NPUT
I EVAPO
TRANSPIRATION
I COMPUTE:
INFILTRATION
1
, DRAINAGE TO. USE BMS
WITH RGF
BMS TO COMPUTE
t
PS 1
SPILL TO:
DEEPER
STORAGE
z = 0 F
 0.
(11.2)
S I
H + P(3s0i)

dF
dt
= K[l +
F
1 (11.3)
i n which BMS is the initial moisture storage in the soil column; BMSN
is the moisture storage in the soil column at field capacity; PSP i s the
effective value of PS at field capacity; and RGF is the ratio of PS a t
wilting point to that at field capacity. This relationship is shown in
Figure 11.2.
Point potential infiltration (FR) computed by the Philip equation
is converted to effective infiltration over the basin using the scheme
of Crawford and Linsley (1966). Letting SR represent the supply rate
of rainfal I for infiltration and OR represent the rate of generation of
r a i n f a l I excess, the equations a r e
LL
0 RGF X PSP
w
3
B
w
>
5w PSP
LL
LL WILTING FIELD
w
POINT CAPACITY
(BMS = 0 1 (BMS=BMSN 1
SOIL MOISTURE CONTENT
/tF
t I I 1
w, 'R
v, RAINFALL
INFILTRATION

z
a
a
0 25 50 75 100
I m p e r v i o u s A r e a R a i n f a l I Excess Component
O p t i m i z a t i o n Component
JOVERIAND FLOW[
[oEFNDi
I
made in the model program and depends upon the ratio, G , of the
k i n e m a t i c w a v e s p e e d t o Ax/At.
(11.9)
be selected a s
LO
ax =  At (11 . 1 5 a )
eo
Lc a t
Ax =  ( 1 1 .15b)
C
ec
for the overland and channel segments, respectively. When Eqs. 11.14
and 11.15 result in noninteger values, the user must round to the
nearest integer.
Reservoir Segments
method i s l i n e a r storage r o u t i n g
21 8
5 = co ( 11 .16)
 +At o2 = I) + l2 + 
At
( 1 1 .17)
Example Application
The model was applied to the Sand Creek Tributary watershed near
Denver, Colorado. This drainage basin is a 183 acre area of predom
inantly single family residential land use with some multifamily land
use, a church, a r e c r e a t i o n a l center, a f i r e station, a n d two s m a l l p a r k s .
The basin has some storm sewers in i t s upper end but r e l i e s mostly on
street gutters and concrete lined open ditches for flow conveyance.
Detailed records of rainfal I and streamflow are collected at 5 minute
intervals. A stage discharge relation was developed using flow profile
a n a l y s i s a n d d i s c h a r g e measurements made d u r i n g storm r u n o f f .
Two sample runs are discussed. The first run was an optimization
run to calibrate the model on an antecedent period of record. In the
second run, the soi I moisture accounting and i n f i l t r a t i o n parameters were
set to their final values from the first run and ten storm events were
simulated.
Before any simulations were performed, the watershed was delineated
i n t o subbasins (overland flow segments) and a drainage network
(channel segments). A schematic showing how the watershed was
approximated with the overland and channel segments is given in
F i g u r e 11.7.
OFO2, OF03, OF04 and OF09 were then delineated based on this channel
segmentation. It should be noted that overland flow segment F04 does
not h a v e b a l a n c e d lengths of o v e r l a n d flow to CH21. To f u r t h e r s u b d i v i d e
this overland flow segment would also require that segments CH21 and
OF09 be f u r t h e r s u b d i v i d e d .
The unallocated concrete lined ditch was then assigned as channel
segment CH22. O v e r l a n d flow segments OF06 and OF05 were then delineated.
To a v o i d the need to s u b d i v i d e channel segment CH20 which would r e q u i r e
subdividing overland flow segments OF01 a n d OF02, channel segment CH25
was used to bypass channel segment CH20. A junction segment, JTOl,
was required to sum the flow from the two channel segments a t the o u t l e t
of the basin. Finally, the remaining part of the basin was drained by
Гораздо больше, чем просто документы.
Откройте для себя все, что может предложить Scribd, включая книги и аудиокниги от крупных издательств.
Отменить можно в любой момент.