,!..ll.

:nal ot Hydrology, 43 (1979) 79-97 79

' . - Elsevi<!r Scientific Publishing Company, Amsterdam- Printed in The Netherlands

tl'\STE.\DY STREAMFLOW MODELING GUIDELINES

VULLI L. GUPTA 1 , SYED M. AFAQ', JOHN W. FORDHAM' and JAMES M. FEDERICI'

' Water Resources Center, Desert Research Institute, and Civil Engineering Department,
University of Nevada System. Reno, NV 89506 (U.S.A.) ·
'Department of Environment, Toronto, Ont. (Canada)
) \l',lter Resources Center, Desert Research Institute, University of Nevada System, Reno,
NV 89506 (U.S.A.)
; Accepted for publication March 6, 1979)

ABSTRACT

Gupta, V.L., Afaq, S.M., Fordham, J.W. and Federici, J.M., 1'~79. Unsteady streamflow
rnod<:ling guiddines. In: W. Back and D.A. Stephenson (Guest-Editors), Contemporary
llydrof.(<'ology- The George Burke Maxey Memorial Volume. J. Hydro!., 43: 79-97.

Pr<'S<'Ill<•d lwrein is n case 11tudy of onc~·dimcnsional modclinl( of spatially varied un-

•.:•·:.dy now n~r,im•!
in a river, utilizing the continuity and momentum formulations of
-:.>~tot· y,,nant c•quations.The cmphnsis is on sensitivity analyses of the modeling framework
for obtain in~-( f~uidclines for modeling exercises. Finite-difference schemes were compared
n•l11tive to thdr dficacy in simulating Oow regimes. Model response was examined with
r.·sp<'ct to t~hangcs in six parameters, namely: (1) grid size; (2) lateral inflow or outflow
r~,tcs; ( 3) discrete and composite description of hydraulic elements; ( 4) discrete and com·
posite description of bed slopes; (6) roughness coefficient of stream bed; and (6) mag·
nitude of weighting parameter. ·

iNTRODUCTION

Application of successive approximation and finite-difference methods to

the ;tudy of spatially varied and unsteady flow in channels has a relatively
!.mg history. In previous studies such as estuary modeling (Hann and Young,
1972), hydraulic surge propagation (Chaudry and Contractor, 1973), unsteady
::ow in streams and aquifers (Isaacson et al., 1956; Amein and Fang, 1969;
Car:::ison et al., 1969; Strelkoff, 1970; Contractor and Wiggert, 1972;
i>1cDowdl and Prandle, 1972; Quinn and Wylie, 1972), simulation efforts
were based on simplifications of river geometry and its spatial variability,
channel roughness and other hydraulic properties. Very few studies are
reported in the literature where steep slopes and str~ams of rapidly varying
~ometry were involved. Furthermore, numerical simulation models hitherto
" generally structured with very little input of hydraulic and physiographic

I "ltion obtainable from field data.

• 18 paper describes some of the principal findings related to the modeling
80

of the unsteady flow regime of Truckee River; California-Nevada. Explicit

and implicit finite-difference techniques and their modifications were utilized
for solving the pair of partial differential equations describing the principles
of continuity and momentum.

SCOPE AND OBJECTIVES

The scope of the investigation reported her~in is limited to the main stem
of Truckee River from Tahoe City to Nixon. For modeling purposes, the
stretch of the entire river was considered to be ·made up of five reaches. River
reaches were further represented by a total of ~4 subreaches based on field
survey with regard to cross-sections and other hydraulic elements. The prin-
cipal objectives of the study efforts were as follows.
(1) To develop and test a numerical simulation model for the unsteady
flow regime of Truckee River from Lake Tahoe to Nixon, 172 km in h·ngth,
using a variety of finite-difference techniques.
(2) To compare the relative effectiveness of the finite-difference techniques
in terms of: (a) reproducibility of observed flow events during floods and
low flow regime; (b) propagation of a hypothetical hydraulic surge; and (c)
routing the "Intermediate Regional" and "S~dard Project" flood waves of
the U.S. Army Corps of Engineers. ·
(3) To conduct sensitivity studies to investigate the model behavior due to
the influence of: (a) varying lateral inflow and outflow; (b) variation of
weighting parameter; (c) time-step changes;(~) smoothing the river-bed
roughness coefficient; and (e) relative comparison using discrete and compos-
ite descriptions of channel geometry. ·

STUDY AREA

The study area encompasses the Truckee River basin which is situated on
the eastern slopes of the Sierra Nev~a Mountains and has a drainage area of
7925 km 2 • Drainage area with runoff potential, however, is 2764 km1 out of
which 1310 km 2 directly drain into Lake Tahoe. The river originates from
Tahoe City located on the northwestern rim of the lake where a dam regulates
the outflow from the lake. The flow is further regulated by a number of
reservoirs. The setting of the study _area is illustrated in Fig.1. The drainage
network of Truckee River system is characterized by two types of stream
channels. The first type occurs in the upper reaches of Truckee River and
tributary stream draining into Lake Tahoe with bed slopes of the order of
95 m/km, whereas the second type occurs in middle and lower reaches below
Reno where the bed slopes average about 2 m/km. The general layout of
Truckee River system, indicating the gage locations and regulation aspects,
' is illustrated in Fig.l. ·
 ··-

' -•-o-...
- - J.
'
Q
'.

0 Dllt.a not SUit4lbU

0 Dllt.a not IIW.t.flble
- - - Di'lllllnicn
- lnflaw

fi~t. 1. Truckee River system.

METHODOLOGY

The two ba?ic laws of fluid mechanics generally found to be applicable to

the• description of unsteady flow regime in rivers are: (a) continuity of un-
sU>hdy flow or conservation of mass; and (b) dynamics of unsteady flow or
conservation of momentum. Expressed in mathematical terms, these two
r•rinciples lead to Saint-Venant's equations (Chow,'1969; Strelkoff, 1970),
st;1ted as:
i•ytot + V(aytax) + D(ov/ox) = qJT (1)
:md
•'

.l au
+ 2 au + ay = (S0 _ 8 ) _~. q(V- ut)
g at · g ax ax r : Ag
(2)

\::.;.:tiplying eq.l by T and substituting T =oA/oY and V"' Q/A, the resulting
:'·xpressi.on becomes: · · .
7.'i3y/i>t) + aQtax =q (3)
:;Jbstituting V= Q/A and simplifying, eq.2 becomes:

_!__ aQ·:... Q oA + Q aQ _ Q2 .aQ + ay = ·so-Sc _ (V-ut)q

Ag aL 2 at 2 ax A 2g. ox ax . Ag <4 >
 82

Replacing iJA/3t by T(3y/3t) and 3A/3x by T{3y/3x), eq.4 can be re-arranged

as:
1_ aQ _ _y_aQ _ Q2 T ay +a).= _ _ (V-ui)q
80 8 (5)
Ag at Ag ax A 3g ax ax f Ag
Substituting q- aQ;ax for T(ay;at) and F 2 .for Q2 T/A 3g, eq.5 can be:.
simplified as: · ··
1._ aQ + 2v aQ + (l- F 2 ) ay = 80 _8 ~ u1q (6)
Ag at Ag ax ax 1 Ag
In eqs.l-6, A = waterway cross-seetional area8; D = hydraulic depth expressed
as the ratio of waterway cross-sectional area to the width of free water surface;
F = Froude number; g =acceleration due to graVity; Q =flow rate in the·
waterway; q = rate of lateral inflow or outflow; 8 0 = bed slope of stream; Sr =
slope of hydraulic gradient; T == width of free vrater surface; t = time; U1 =
component of lateral flow velocity in the direction of streamflow in the river.

Finite-difference formulations

In the most general form, equations of continuity and momentum are .

represented by the following weighted average approximations. Finite-differ-
ence formulation of eq. 3 can be stated as:
YjjH - Y; j 1 . . (7)
T;,j 'dt ' + dx O(Qi+l,j...J-Qi-t,}+1)+(1-8)(Q;...J,}-Qi-1,j)=Qm
2
Similarly, finite-difference formulation of eq. 6 can be expressed as:
. .
1 [Q· · - Q·;]
I,J+l
A;,J8 - · tlt
J, + 2V·;I,
A;,jK 2/lx
~f}
(Q;...t,J+t- Q;.l,)+l)
+

(1-8) (Q;+s,j- Q;.l,j)] +

l...,F't.J]
[ 2 Ax
r
LO(Y;... a,}+l- Yi-1,}-+1) +

(1-8) (Y;H,j-Y;-1 ,j)] == So-[8St1,J+~ +(l-8)St~,1 J (8)

In eqs. 7 and 8, i and j denote the position of the variable with reference to
space and time grids, respectively; 0 is the weighting parameter; QIN is the
rate of lateral inflow or outflow; and the remaining symbols are as defined.
· before. The effect of lateral flow in the finite-difference formulation of .
momentum principles, eq. 8 has been igD.ored. Collection of field data, as :
discussed in the subsequent sections, did not inciude infonnation related t()
the orientation of tributaries and other situations of lateral flow regime. 'fhe
dependent variables in the foregoing formulations are discharge, Q, and depth,
y. In eqs. 7 and 8, values of 0 of 0, 0.5 and 1 lead to explicit, Crank- Nicholson's
implicit, and fully implicit formulations.
 83

Finite-difference formulations for the weighted average implicit schemes,

0 < 0 < 1, were also structured.
A eomputer program was developed, in FORTRAN IV, with the capability
of handling the foregoing finite-difference formulations. A program listing
is presented by Moin (1974).

F!I':LD SURVEY

Efforts toward systematic fulfillment of the study objectives were based

on two catt>gories of data, namely, field data and historic hydrologic data. In
summer 1973, field surveys in the study area were conducted to obtain in-
fNmation related to: (a) cross-sections at 25locations on the main stem of
river and 13 locations on the tributaries; (b) concurrent discharge measure-
ml'nts using current meter; and (c) river-bed level survey using information of
hcneh-marl< lfwels supplied by the U.S. Coast and Geodetic Survey. Field survey
data w~'n' proc(~ssed to yield the following:
{a) lwtl slop<'S for each of the subreachcs and reaches.
(h) Mmming's roughness coefficient values for each subreach ealculatcd
from the flow measurement bed slope and corresponding waterway area.
(c) JP'aphical relationships among the hydraulic parameters: depth-area,
depth-top width, depth-wetted perimeter, top width-area and depth-dis·
charge.
(d) regression equations for each of the graphical relationships stated in (c).
Length and slope of the five main reaches of the river are presented in
Table.I.

TABLE I

Reaches of the main stem of the Truckee River

No. Station Elevation Length Slope

(km) (km)

1 Tahoe City Gage 1.896

53.13 0.0061
2 Farad Gage 1.572
36.63 0.0059
3 Reno Gage 1.354
10.87 0.0020
4 Vista Gage 1.333
29.06 0.0018
5 Derby Dam 1.281
41.78 0.0019
6 Nixon Gage 1.202

Field survey information was processed to yield empirical relationships

relating pairs of hydraulic elements of major concern to modeling endeavors.
For each of the 25 subreaches and five main reaches, regression. equations
were developed in the following form:
A =K 1 yn,, Q = K2 yn, (9), (10)
P = K 3 Y"a, T = K 4 yn4 (11), (12)
where A = waterway area; K 1-K4 = coefficients obtainable from regression
I

\1
I

\
TABLED
\ ~on equations of hydraulic elements - composite reaches

I No. Reach Length

(km)
Area-depth
A .. K 1 Y"t
Discharge-depth
Q .. K,Y"a
Wetted perimeter-depth
P= K,Y"a
--
Top width-depth
T"" K 4 Y"•
I n, n, n·4
K. K, K, K.
"•
1 Tahoe-Farad 53.13 21.12 1.68 21.80 2.62. . 39.56 38.63 . ·0.62
\ 2 Fa.--ad-Reno 36.63 37.48 1.72 25.94 2.65 63.59
0.69
0.66 64.79 0.65
3 Reno-VISta 10.87 38.24 1.54 11.25 2.96 56.00 0.63 57.16 0.62
4 Vuta-Derby Dam 30.04 26.62 1.74 15.79 2. 77 37.34 0.79 38.71 0.77
5 De:rby Dam-Nixon 40.80 21.42 1.80 12.23 3.13 36.94 0.79 37.85 0.77
 86

llooplhlo • • -
,...,, . '" ,.,,., ..,..
..• 1.1 u

. / ..
,.._,,
/.:~7-
,
.. lO

l' lj _,,/':_...
(1-·-1- 3t ~
I 250
£
/; I c
•• j

! 00-ttli ,. i
k
200

.. J
.!

1f
---- -- -
.5 lllybwly 150 R1----·-
§ r ...... ..i:
..
·w
5
} J
---
II

"' II .

s Dlpttllll- s .,.,U.lll-

Fig. 2. Depth-top width relationship. Fi1. 3. Depth-area relationship.

Depth in Mf'ten

.. ...

Sl

. ~

Sl
*!
c

~
.s J!
.! !
.! to
.t
J ll
'

Fig. 4; Depth-wetted perimeter relationship.

 86

\ analysis of logarithms of the respective pairs of hydraulic elements; n 1 -n 4 =

exponents obtainable from regression analysis of logarithms of the respective
pairs of hydraulic elements; P = wetted perimeter; Q = discharge; and T '=
width of the free water surface. The resulting equations for the five main
reaches are presented in Table II. A sampling of graphic relationships of
hydraulic elements at selected subreaches of the river are illustrated in
Figs. 2-4.

DATA ANALYSIS AND MODELING EXERCISE

Data analysis in terms of stressing the model consisted of the following

tasks:
(1) Routing of the flood wave which occurred during January-February,
1963.
(2) Study of flow regime for the Farad-Vista reach of the river for a five-
day duration, November lD-14, 1973.
(3) Routing of synthesized hydrographs for "intermediate regional" and
"standard project" flood events estimated by the U.S. Army Corps of ·
Engineers, Sacramento, California for the Truckee Meadows reach from
Mayberry to Vista. ·
( 4) Effects of surge propagation due to hypothetical flood event wherein
the flow rate was increased from 14 to 280 m 3 /s in a time span of 20 min.
Identification of the specific reaches where each of the aforementioned
modeling exercises were conducted is presented in Table III. Pertinent findings
of the various modeling efforts are described below.·

SIMULATION OF A HISTORIC FLOOD EVENT

Recorded streamflow data (U.S. Army Corps of Engineers, 1970; Young

and Harris, 1966) relative to the 1963 flood occurrence in the Truckee River
basin was utilized in investigating the modeling capabilities toward reproducing
the observed flood flow. For the five main reaches of the river, the model was
operated separately, and in each instance the observed hydrograph at the
boundary was compared with the computed hydrograph. In the absence of
gaged data for describing the lateral flow regime, several trials for various
assumed values of lateral inflows were made. By trial-and-error approach,
QxN was altered in a variety of ways such that a reasonable agreement between
the observed and computed hydrographs was found. Findings of this phase
of modeling exercises are presented below.

(l) 'J'ulwe .. Fartld rt~acll. )1'ur thiH 63.3 km long rouch, Hmoothml or cumpOHit(l
values of hytlraulic elements were (~hOI!Cil. l!!ven though the topugruphy uppearY
to warrant it, use of discrete sections within this reach would not havt~ been
advantageous because of hydrologic operation at Lake Tahoe in terms of .
storage in the lake and releases to the river. The lateral inflow regime was,
after several trials, hypothesized to vary with time.
TABLE III

Modeling exercises

Task Tahoe-Farad Farad-Reno Reno-Vista Vista-Derby Dam Derby Dam-Nixon

1963 flood simulation )( X X X X

1973 short-term flow
simulation· X
Fully implicit scheme X X X X X
Weighted average scheme X X
'Composite-hydraulic
element model X X. X X X
Discrete-hydraulic
element model X X
Intermediate regional
flood-routing X X
Standard Project Flood
routing X X
Surge propagation X
88

••
40

".,g'C ""
E
0
30
.:; 10 ~
~ to
2
i
IS
&;
zo lS

10

Fig. 5. Farad hydrograph, 1963 flood event.

A sample of trial exercises for Tahoe-FSJ-ad reach is illustrated in Fig. 5

wherein a reasonable agreement of observed and computed profiles is apparent
with respect to peak time and crest segme~t. Recession segment registered a
relatively poor agreement. Modeling exercises for this reach were found to be
very cumbersome due to: (a) unavailability of tributary flow data; and (b)
difficulties in describing the regulation at Lake Tahoe in mathematical terms.

--- t..)burved
----- · ... -.. Computed, t..l!ll•r•1 lntluw
(llhwr•to on~ l'omt>olll to lleollont)
l'OfiiiNI•dt Nl> IAtorol lnllow I

•1
"

Fig. 6. Reno hydrograph, 1963 flood evel\t, discrete and eompoalte sections.
 /
TABLE III

Modeling exercises
Reno-Vista Vista-Derby Dam Derby Dam-Nixon
Tahoe-Farad Farad-Reno
Task
X X
X X
)(
1963 flood simulation
1973 short-term flow
X
.. simulation X X
X X
Fully implicit scheme X
X X
Weight~d average scheme
·eompo!lite-hydraulic X X
X X
element mO(lel X

Discrete-hydraulic X
X
element model
Intermediate regional X
X
flood-routing
Standard Project Flood X
X
routing
X
Surge propagation

rI
 89

(2) .Farad-Reno reach. Except for a small stretch in the vicinity of Farad, this
reach of the river lies in Truckee Meadows, characterized by withdrawals and
accretions at several locations. Initially, a no-lateral inflow situation was
assumed and the resulting hydrograph is illustrated in Fig. 6. Significant
disparity was evident between the computed and observed flow profiles
during the rising limb and at the crest of the flood wave. This disparity is
attributable to the total neglect of the lateral inflow. After several trial runs,
nn adequate lateral flow regime was derived and included to improve the
a~n'emtmt between the observed and the computed flood hydrogrnphs as
illustrated in Fig. 6. These runs were calculated using composite hydraulic
elements. Similarly, Fig. 6 also shows the results based on discrete hydraulic
elements and the associated properties. The phrase "composite" refers to
consideration of the river in the entire reach to be prismatic. In contrast, the
phrase "discrete" refers to the consideration of actual cross-sections of the_
subreaches based on the field survey. Comparison of results, as illustrated in
Fig. 6, indicates that there is practically no gain in the modeling capability
using discrete descriptions of the hydraulic elements instead of using compos-
ite or average properties.

(3) Reno--Vista reach. Alignment of the river in this reach is important .

from floodplain delineation considerations since it flows through the urbanized
area of Reno-Sparks. This reach was modeled as being made up of fifteen . ·
subreaches, with composite hydraulic geometry, constant n of 0.05, and a
uniform bed slope of 0.002. Steamboat Creek is the only major tributary to
the river within the Reno-Vista reach and was superimposed on the Vista
hydrograph in Fig. 7. ·

..
'\j
O:llpJtld

I !
~-

4 ~
•t ii5
~
·:_-..,_

tO •
.. Fig. 7. Vista hydrograph, 1963 flood ~ent.
90

•
i • 'It
ti
I g
3
{ 4
.~
.t:.
It
i5

O·~o~~--~~~~l~0--~40~~~~~~--~,o---a~o---~~__j
Tiaa (Hn.J,

Fig. 8. Derby Dam hydrograph, 1963 flood· event.

(4) Vista-Derby Dam reach. This reach of the river is characterized by rugged
topography and alignment, no tributary flow inputs, and a major flow diversion
via Truckee Canal at Derby Dam. There is no documented operational policy
relative to flood control practice at Derby Dam. Consequently, several runs
were made in modeling efforts with a variety of assumed storage policies at
Derby Dam. Results of one of these policies were found to yield a close match
between the computed and observed hydrographs and are illustrated in F'ig. 8.

(5) Derby Dam-Nixon reach. This reach of the river is characterized by flat
bed slopes, relatively wide floodplain, a number of flow diversions and accre-
tions for which no systematic data are available. After several runs, lateral flow
regime criteria were found to yield a reasonably close agreement between the
observed and computed hydrograph profiles. The resulting hydrograph is
compared against the historic hydrograph illustrated in Fig. 9.
Simulation efforts throughout the five reaches were characterized by trial-
and-error approaches relative to the lateral flow regime. If information with
regard to the linkages of surface water and subsurface water in the river basin
were available, trial-and-error approaches might not be warranted.

Simulation of bi-hourly streamflow .

Modeling exercises were conducted to investigate the simulation capabilities

for the Farad-Reno reach of the river. Bi-hourly flows for the time span of
November lQ-14, 1973 were modeled using a fully implicit finite-difference
scheme with discrete sections, discrete bed slopes, and discrete values of
 91

Fig. 9. Nixon hydrograph, 1963 flood event.

Manning's roughness coefficient. Field survey data were thus utilized to the
fullest possible extent. In addition to lateral inflows, flow withdrawals from
the river were also incorporated. The routed hydrograph for Reno is illustrated
in Fig.lO. A comparison of observed and computed hydrographs reveals
reasonable agreement relative to peaks and troughs in terms of timing and mag-
nitude. The profiles of the rising and recession limbs; however, do not match
satisfactorily .

.
~
!J Q:lapJt.al
71
g ~
0
'"'
.!\ ~
0
$

t
4

-a f
Z5

0
~0--,~~~.~0--~60--~--~.-~~~--~~~~~.
Tine (lira • )

Fig.10. Reno hydrograph, 1973 short-term flow study; time step • 2 hr.
92

Intermediate regional and standard project flood waue propagation

This aspect of modeling was conducted for Truckee Meadows reach of the
river, 21.82 km long, stretching from Lawton near Mayberry bridge to Vista,
Nevada. The study reach includes the urbanized area of Reno-Sparks. There
have been several flood events in the past in this area. Consequently, flood-
plain studies for Truckee Meadows have received the attention of the U.S.
Army Corps of Engineers, Sacramento District, California. Two measures of
flood flow hazard were suggested in their studies (U.S. Army Corps of
Engineers,.l970, 1971), Intermediate Regional Flood (IRF) and Standard
Project Flood (SPF). IRF is associated with a recurrence interval of 100 years,
and such estimates are usually based on frequency analysis of streamflow
records. SPF, however, implies the· consequence of a combination of the most
severe meteorologic and hydrologic conditions.
In view of the significance of the urban area, a study was conducted with
the objective of testing the model capabilities for routing the two flood waves
separately. Fully implicit scheme, discrete sections, discrete values for bed ·
slope and roughness coefficients were utilized. Flood routing exercises require
the hydrograph description for the flood inflow. Available data (Young and
Harris, 1966; U.S. Army Corps of Engineers, 1970) include peak rates of run-
off and timing of the peak occurrence.s.

16
_:_ ___ Viata ~"""

- - - llono ll)drogr.P,

- - - MoyiJeny ll)drogr.P,

~~~-b--~~~~~~~~~~~~~
~0 60 70 80 90 0
Tlono (~Q~n)

_Fig. 11. Propagation of the Standard Project Flood wave, Truckee Meadows reach.

Utilizing the findings of Gupta and Moin (1974), profiles for inflow hydro-
graphs were synthesized and applied at Lawton, one each for IRF routing
and SPF routing. Results of the flood routing exercises are presented in
Table IV. The computed hydrographs for the SPF are illustrated in Fig.11.
 93

Comparison of estimated flow by U.S. Corps of Engineers and computed flows

Station Intermediate Regional Flood (m 3 /s) Standard Project Flood (m•/s)

estimated* computed % djff. estimated* computed % diff.

------------· ----------------'-----------
He no 512.5 632.1 3.8 982.6 963.9 -1.9
Vista 608.8 677.9 -6.1 1116.7 1083.1 -2.9

*U.S. Corps of Engineers (1970, 1971).

l!ydraulic surge propagation

Madding capabilities for routing a transient hydraulic phenomenon such

r<s a sevt>re surge wave were studied, on a limited scale, for the Farad-Reno
r<:.J.ch. Fully implicit model with discrete cross-sections, discrete values of bed
sk'?es ;md roughness coefficients, was used in modeling the hydraulic bore
pwpugation.
A uniform flow rate of 14 m 3 /s was assumed at Farad, and this flow rate
was arbitrarily assumed to increase from 14 to 269m 3 /sat Farad in a time
span of 20 min. The resulting hydrograph rise at Reno is illustrated in Fig.12.
Time delay in the surge development at Reno was found to be 4.4 hr. This
exercise attests to the modeling capabilities in analyzing a transient hydraulic
phenomenon such as a severe flow surge resulting from a possible failure of
upstream dams of regulation structures.

4.0

s
u 3.0 1
~ g
:g
t
l
l.t i
$

1.8

Fig. 12. Reno hydrograph, propagation of hydraulic surge.

SENSITIVITY STUDIES

Six parameters were considered throughout the modeling exercises. The

influence of changing the parameters upon the computed hydrographs was
investigated. The parameters are: (1) mesh size; (2) lateral inflow and/or
outflow rates; (3) discrete and composite description of hydraulic elements;
(4) discrete and composite description of bed slope; (5) discrete and compositt
description of Manning's roughness coefficient; and (6) weighting parameter, e.

(1) Mesh size. Mesh size refers to the magnitude of time and space steps.
Shorter distances and time increments tend to yield accurate results ~ut may
not necessarily be economical in the simulation exercises. Model behavior in
simulating the hi-hourly flow within Farad-Reno reach during November 10-
14, 1973 was studied using three different values of time steps of the finite-
difference scheme. These were 0.5, 1 and 2 hr. Results from the model are
illustrated in Fig.l3, and it was observed that the model behavior is relatively
independent of the time step used in the computations.

( 2) Lateral flow .regime. In all the modeling exercises, lateral flow regime was
found to be the most influential parameter. Trial-and-error approaches by
imposing several different rates of lateral flow were used to obtain a reason-
able match between the computed and observed hydrographs. It cannot be
inferred that the rates of lateral flow so used are unique values.
Extreme sensitivity of the model to the lateral flow regime can be readily
inft•rred from t~xamining }t'ig.6, where the results of tht~ 1963 flood event for
}t'a.rad-Reno reach are illustrated. By ignoring lateral flow, the peak was
found
to be underestimated by about 32%. Computed peak rate was found to lag
behind the observed peak by about 3.5 hr. Superimposition of lateral flow,

'11.-St~p•l.O . ..

1't.&ap• a.o ...

'11.- SUp • 0.5 llr.

Tu. Oint.)
Fig.l3. Reno hydrograph, 1973 abort-term Oow study, effect of time step.
 95

tw trial and error, has yielded a much closer agreement, as can be seEm in
Fif{.6. Consideration of composite sections will tend to damp out the varia-
tions in dt•pths. Flow modeling E'X('rciRes in the Truckee Meadows reach and
)<'arad--Reno reach were based on discrete descriptions, and in all other reaches
composite descriptions were found to be sufficient.

(4) Discrete and composite stream bed slopes. Examination of Fig.6

indicates that the model behavior is relatively the same whether varying or
constant channel bed slope was used.

(5) Dbwrrtc or composite Manning's roughness coefficient. Examination of

Fil~· 6 also indicates that the model behavior is the same whether avProged
and c\1mposite values for n or varying and discrete values are used. Discrete
values were used in modeling Truckee Meadows reach, and in all other instances
composite values were used.

(6) Weighting parameter, 0. Magnitude of weighting parameter 0 was found

to be influential in controlling the computed hydrograph profile. The 1963
flood wave propagation for Vista-Derby Dam reach of the river was inves-
tigated by setting 0 = 0.5, 0. 75 and 1.0. The resulting hydrographs at Derby
Dam are illustrated in Fig.14. Model behavior was found to be relatively
insensitive to the weighting parameter. It was found that the hydrograph
corresponding to 0 = 0. 751ies between those for 0 • 0.5 and 1.0.

-·-

For 8 :r. 0,75 the ~ tieo

between ~plio ror 0 a O.S
~ and." 1.0
u
%0
"'g
.!S
~ 15
2
~
Z5
10

~~~~~~~~~--~--~--~~~--~~
so so 10 eo eo
11me (Hn.l ·
.
Fig.l4. Derby Dam hydrograph, 1963 flood event; effect of varying B.

SUMMARY AND CONCLUSIONS

A simulation model has been developed and tested for its! capability to solve
the unsteady-flow equations using finite-difference schemes of analvsru~_ ThA'
96

· model was applied to simulate the 1963 flood wave movement in the 'I'nlckee
River. The river was modeled as being made up of five main reaches, each of
which has been further subdivided into a number of subreaches of equal length.
The reaches are Tahoe-Farad, Farad-Reno, Reno-Vista, Vista-Derby Dam
and Derby Dam-Nixon. The following list comprises the modeling exercises:
(1) Simulation of the 1963 flood hydrographs using the recorded hydrographs
at the U.S. Geological Survey gaging stations as boundary conditions.
(2) Simulation of short-term flow for the period of November 10-14,1973,
in the Farad-Reno reach of the river.
(3) Routing the IRF and SPF floods in the Truckee Meadows reach from
Mayberry Bridge to Vista using synthesized flood hydrographs.
( 4) Evaluation of hydraulic bore propagation due to a hypothetical situation
wherein a uniform flow of 1.4 m 3 /s in Farad-Reno reach was increased to
269m 3 /sat Farad in a time span of 20 min.
( 5) Sensitivity studies for monitoring the model behavior due to changes in
six parameters, namely: (a) mesh size; (b) lateral inflow and/or outflow rates;
(c) discrete and composite description of the hydraulic elements; (d) discrete
and composite descriptions of bed slopes; (e) discrete and composite descrip-
tion of Manning's roughness coefficient; and (f) magnitude of the weighting
parameter, 8.
The conclusions of the study,listed below, provide useful guidelines for
modeling the unsteady flow regime of livers:
(1) It was found that a computational time step of 1 hr. was best suited
among the three time steps, 0. 5, 1 and 2 hr. examined.
(2) The model was found to be capable of reproducing 1963 flood wave in
the river from Lake Tahoe to Nixon.
(3) The model was found to be capable of simulating short-term flow
hydrographs for the flow regime in Farad-Reno reach for November 10-14,
1973.
(4) Based on synthesized hydrograph descriptions of IRF and SPF flood
events, it was concluded that the Truckee Meadows reach can be modeled
with a high degree of reliability.
(5) The model is capable of handling a transient hydraulic phenomenon such
as a surge propagation in the river.
(6) Sensitivity analyses have indicated that the most influential parameter
·controlling the model behavior is the lateral flow regime. Throughout the
modeling exercises, it was found that lateral flow rates play an important
role in the reproducibility of recorded hydrographs. Based on these findings,
the most critical aspect·. in modeling include the explicit description of
tributary flow inputs and rating curves for ..he river in the immediate vicinity
of the poin'ts of confluence.
(7) It was found that for all values of 0 :.>et\';·een 0.5 and 1.0, the .~omputed
hydrograph is strikingly the same, with vt~cy little difference. This indicates
that any value of 8 between 0.5 and 1.0 is adequate for modeling purposes.
 97

,\CKNOWLEDGMENTS

The authors arc grateful to Rich?rd Cooley for his enthusiastic guidance
throughout the study. Sincere thanks are extended to Thomas Gallagher for
his timely cooperation in the acquisition of field data.
Work r('puri.Pd hemin wns supported in part by the United States Df'part-
nwnt of tlw lntPrior, Offke of Wal~r n.(•:murccs RPscareh, Projr.ct Nn. A-057-
N EV nnd B-mHi-NEV, as authorized under the Water ReRources H.esear<:h Act
of 1964, PI, 88-379, and in part by funds provided by the Desert Research
Institute, University of Nevada System, Reno, Nevada.

REFI·UlENCES

Am(~in, M. and Fang, C.S., 1969. Streamflow routing (with applications to North Carolina
Hiv1~1·).Water Resour. Res. Inst., Univ. of North Carolina, Raleigh, N.C., Tech. Rep.
No. 17.
Amcin, M. and t<'ang, C.S., 1970. Implicit flood routing in natural channels. Am. Soc. Civ.
Eng., J. Hydraul. Div., 96(HY12): 2548-2564.
Chaudry, Y.M. and <bntractor, D.N., 1973. Application of implicit method to surges in
open channels. Water Resour. Res., 9(6): 1605-1612.
Chow, V.T., 1959. Open Channel Hydraulics. McGraw-Hill, New York, N.Y., 680 pp.
Contractor, D.N. and Wiggert, J.M., 1972. Numerical studies of unsteady flow in the
James River. Water Resour. Res. Cent., Va. Polytech. Inst.-Btate Univ., Blacksburg,
Va., Bull. 51, 56 pp.
Garrison, J.M., Granju, J.P.P. and Price, T.T., i969. Unsteady flow simulation in rivers
and reservoirs. Am. Soc. Civ. Eng., J. Hydraul. Div., 95(HY5): 1559-1576.
Gupta, V.L. and Moiu, S.M.A., 1974. Surface runoff hydrograph equation. Am. Soc. Civ.
Eng., J. Hydraul. Div., 100(HY10): 1352-1368.
H.ann, Jr., R.W. and Young, P.J., 1972. Mathematical models of water quality paramete"'
for rivers and estuaries. Water Resour. Inst., Texas A & M Univ., College Station, TexM,
Tech. Rep. 45, 424 pp.
Isaacson, E., Stoker, J.J. and Troesch, 1956. Numerical solution of flood prediction and
river regulation problems, Inst. Math. Sci. New York Univ., New York, N.Y., Rep.
No. IMM-235. .
McDowell, D.M. and Prandle, 1972. Mathematical model of river Hooghly. Am. Soc.
Civ. Eng., J. Waterw., Harb. Coastal Eng. Div., 98(WW2): 225-242.
Moin, S.M. A., 1974. Numerical simulation of unsteady flow hydraulics of Truckee River.
M.S. Thesis, University of Nevada, Reno, Nev., 134 pp.
Quinn, F.H. and Wylie, E.B., 1972. Transient analysis of the Detroit River by the implicit
method. Water Resour. Res., 8(6): 1461-1469.
Strelkoff, T., 1970. Numerical solution of Saint Venant equations. Am. Soc. Civ. Eng.,
J. Hydraul. Div., 96(HY1): 223-252. ·
U.S. Army Corps of Engineers, 1970. Flood plain Information- Truckee River Reno-
Sparks-Truckee Meadows, Nevada. Department of Army, Sacramento District, Calif.,
43 pp. .
U.S. Army Corps of Engineers, 1971. Master Report on Reservoir Regulation for Flood
Contr<"l- Truckee River Reservoirs, Ne-Vada and California. Department of Army,
Sacrau.ento District, Calif., 43 pp.
Young, L.E. and Harris, E. E., 1966. Floods of January-February 1963 in California and
Nevada. U.S. Geol. Surv., Water-Supply ~ap. 1830.A, 472 pp.

r---.----------------------------~-------