You are on page 1of 4

INTERRELATIONSHIPS BETWEEN THE PEAK

INSTANTANEOUS AND AVERAGE DAILY


DISCHARGES OF SMALL PRAIRIE STREAMS
bv
W. H. Ellis Don M. G r a y
Sir Alexander G i b b and Partners Member C.S.A.E.
W e s t Pakistan W a t e r and Power Rerources Survey Agricultural Engineering Department
Lahore, W e s t Pakistan University of Saskatchewan
Saskatoon. Saskatchewan

INTRODUCTION ratio" would decrease with an increase czding day to the discharge of the
in drainage basin size. For larger maximum day, and ( b ) the ratio of
Discharge records published by such watersheds, the rate of runoff is gener- the daily discharge of the succeeding
agencies as the Water Resources ally high for at least 24 hrs because the day to that of the maximum day. A
Branch of the Departmcnt of Northern runoff producing storm is of consider- plot of these ratios as presented by
Affairs are usually in the form of able duration whereas on small basins Langbein is given in figure 1.
average daily discharges. Such records a cloudburst may cause a flood which
are of little value for studying the peak will give runoff over a few hours re-
instantaneous flows, particularly on sulting in a large Qmax and only a Because of the success and general
small watersheds. The obvious method moderate Q. ccceptance of the methods given bv
of obtaining the peak instantaneous ~ u l l e rand Langbein in resoiving thk
flows on a watershed is by review of Langbein (5) employed a different complex interrelationship between the
the recorded stage hydrographs. Gener- approach to solve the problem by peak instantaneous flows and the aver-
ally, however, these records are not using only flow records. He established age daily flows for the streams they
readily available. Further, the task of a relationship between the peak in- studied, it was assumed s i m i I a r
analysis is extremely time consuming stantaneous discharge and maximum mzthods could be successfully em-
if a long period of record is involved. daily discharge in terms of ( a ) thc ployed to establish the properties for
An alternative method to the above ratio of the daily discharge of the pre- Prairie Streams. T o accom~lish this
which is much more expedient, is to (0 objective, discharge hydrogriphs were
utilize the average daily discharges by collected from 22 small watersheds,
applying a conversion factor to these a
0.9
which varied in size from 12 - 1000
flows to derive the peak instantaneous 0.8 square miles, located within the central
discharges. a
region of Canada. All watersheds se-
0.7
CURENT PROCEDURES P lected were equipped with automatic
5 0.6
water level recorders which had been
Two approaches are reported in 0.5 in operation at least five years. In
the literature for resolving the relation 5 total, about 260 hydrographs, including
between the peak instantaneous dis- 8 0.4
Y
U both rainfall and snowmelt runoff
charge with another flow parameter a
0.3
events were studied.
and/or a geomorphic property of the g 0.2
watershed. Fuller ( 3 ) in 1914 studied Details of the watersheds as to their
P station description, location and size
the relationship between the peak in- 0.1

stantaneous discharges and the average


L
7 k co ! i i
0.1 o.2 0.3 0.4 o.s O.6 0.l o.s o.r 1.0
1' are given in table 1 and figure 2 re-
spectively. As shown on the figure, the
24-hr flows of floods on large basins
located in the Eastern United States. RATIO OF SUCCEEDING DAY TO MAXIMUM DAY are widely separated gee-
He developed a relationship between Figure I. Discharge at and the t i m e of crest in rela-
tion t o daily mean discharges (redradn from
graphically, and hence they represent a
these variables of the form, Langbein) wide range in physiographic conditions.
Qmax - Q = 2A-".30 ,......,....,... 1
Q
where Qmax = maximum recorded
instantaneous discharge (cfs) ,
Q = largest average rate of flow
over a 24 consecutive hour period
which was recorded during the
same period of years as Qmax
(the two discharges did not neces-
sarily occur during the same
flood), and
A = drainage basin area in sq.
miles.
In establishing the "excess ratio",
(Qmax - Q ) /Q, of equation 1, Fuller
rationalized that for all floods, Qmax
would be greater than Q. In addition, Flgure 2. Location o f Hydrometric Stations.
it was postulated that the "excess

CANADIAN AGRICULTURAL ENGINEERING, FEE. 1966


In an attempt to minimize the effect of consistent flow parameter than the Stichling and Blackwell ( 6 ) and
regional influences in the investigations, exccss ratio. Within regional groups, Durrant and Blackwell ( 1) found that
the watersheds were separated to the there was a definite trend for the for these streams use of the "effective
regional groups listed in table 1. peak ratio to decrease with drainage area" of a basin produced more suc-
basin size. cessful relationships with flow dis-
charges than use of the gross area. Ef-
TABLE I. DESCRIPTION
--
O F HYDROMETRIC STATIONS fective area, as defined by Stichling and
Area a/ Blackwell ( 6 ) is, "the gross area of
Group Code Name Location WRB No. A4 Ae the basin less the areas which do not
.-
Foothills 1.1 North Branch near Twin SAD,, 12 1i contribute to the mean annual flood".
Drywood R. Butte, Alta. That portion of the gross area which is
1.2 North Fork above U.S. St. I IAAo:., 62 62 excluded includes marshes and sloughs
Milk R. with no connecting channel to the
1.3 Stinson Ck. near Pekisko. 5BL, 96 96 stream and areas upstream from "on-
I . Highwood R. near Dicbel's SBL,, 330 300 stream" lakes which have the capacity
to detain runoff waters until the peak
Cypress Hills 2.1 Lyons at International I I AB,, 66 47 flow has been observcd.
Coulee boundarv
2.2 Woodpile near Internation- I lAB,:, 68 51 Plots were made between the peak
Coulee al boundary
ratios of all storms and effective areas
2.3 Rattle Ck. at Ranger I 1ABBl 78 64
Station, Sasl.. for watersheds in groups 2, 3 and 4. In
2.4 East Fork near lnterna- I I AB,:, III 82 this analysis, the values for the cffec-
Battle Ck. tional boundary tive areas were taken from the works
2.5 Manyberries at Brodin's 5AF10 127 120 of Durrant and Blackwell ( 1 ) . For
Ck. Farm watersheds which were used but not
2.6 Sage Ck. at Q Ranch near I IAA,, 174 149 listed in the reference, the values were
Wildhorse, Altn. determined by applying a regional
coefficient, the average ratio of the
Wood Mountains 2a.l Horse Ck. near Interna- I I AE,:, 74 71 effective area to gross area for water-
tionill boundary sheds in the region, to the gross area
2a.2 East Branch at Interna- 1 IAE, 534 190 of the watershed in question. For
Poplar R. tional boundary
-- --
watcrsheds listed in the Foothills
group, the effective areas had not been
Central 3.1 Little Boggy at Cole, SasL. 5MD, 60 15' determined, hence, the gross areas
Ck. were used. It is believed that for
3.2 Indianhead near Indian 5JL, 203 45 watcrsheds in this group the differences
Ck. Head, Sask.
3.3 Boggy Ck. above Junction 5JF, 175 56 between the two areas would be small
Qu'Appelle K. bccause of the high degree of drainage
3.4 Qu'Appelle R. above 5JG, 932 222 development and pronounced relief in
Buffalo Pound L. the region.
Manitoba Escarpment 4.1 Scott Ck. near Laurier, 5LJZh 26 15b A typical array of peak ratio values
Man.
4.2 McKinnon C1. near 51,Jz7 30 17b
and effcctive areas is shown by those
McCreary, Man. watersheds in the Cypress Hills and
4.3 Garland R. neal 51,Gz 30 22b Wood Mountain region (see figure 3 ) .
Garland, Man.
4.4 Edwards Ck. near 51>J2" 67 38"
Dauphin, Man. . a 8

45 Mink R. near 5LJ,, 82 47


.
cd

Ethelbert, Man. KC":


Pine River, Man. - SUMMER
0- SNOWMELT
4.6 Pine R. near 5LG1 84 48b 1
-- -- -.

a/Ag is the gross drainage area in square miles and Ae the effective area in square miles.
b/Indicates an estimated effective area.
Peak Ratio - Area Relation,shil~s Prairie hydrologists have long rec-
ognized the error of using the gross
In the initial phases of the study, drainage area in rational flood flow
plottings were made using an excess formuae when applied to Prairie
ratio (q, - q2)/q2.': (patterned after streams characterized by large amounts
the works of Fuller) and a peak ratio, of depressional storage. Under these
qp/q2, against gross drainage area. conditions, the area of the watershed
Gross drainage area as used hcrein is which contributes to flow during any
the total area of the watershed en- given runoff cvent varies with the
closed within the topographic divide. magnitudc of the flood as different I
S I00 150 190
In these plottings, it was found that the channels and depressions become con- EFFECTIVE AREA. A e ( Sq. mi. )
peak ratio was a much more stable and nected to the main stream system. Figure 3. Peak ratio arrays-Group 2.
*Note, in the excess ratio and peak ratio, the peak instantaneous discharge, qp. and the
average daily flow for the same day of occurrence of the peak, q2 are those for individual Perhaps, the most striking feature of
storm events. the arrays was the large scatter of the
2 CANADIAN AGRICULTURAL ENGINEERING. FED. 1966
data which represents a wide range in 2 are situated in the lee of the Rockies
values for the peak ratio with different and do not receive intense cloudbursts
storms. In addition, several other whcrc A, -= effective arc ;q. but are affected primarily by long-
features were exhibited by the plots. miles, and C,n = coeffic ~d duration cyclonic-type storms, and ( b )
exponent. the watersheds are densely forested
( a ) For watersheds within a selec- thus the streams are of the permanent
ted regional grouping there was a trend Values of the coefficient, C, - type with high base flows. On water-
for the peak ratio to decrease as the !m!ient, n, for the different j 1 sheds in the Central Plains, beeause of
size of the basin increased. This groups are tabulated in table their poor relief and drainage develop-
characteristic is to be expected as
large watersheds have high base flows.
In addition, the hydrograph from large TAB1.F. I T . PEAK RATIO-EFFEC LREA RELA'TIONSHIPS OF WATERSHEDS
watersheds have long time bases so I N DIFFER EGlONAL GROUPS
--
that qp is more nearly equal to q,, and - -- --

the ratio closer to unity. - -- Group ~on Range of Valld~ty


- -
--
-
--
-
-- -
-
-
--
-
-
- .
1 qp!qz ~ ~ 22 - 0 60 - 300 mi2
( b ) For watersheds within a given 2 and 2a q 11/92 Ae -0 46 50 - 200 m12
re~ionalgrouping, the peak ratios from 3 qp.jq.2 - . . A, -0 36 45 - 225 m12
ramfall and snownlelt events were 4 qpi:q2 = 1 1 A~ -0 30 15 - 50 r n 1 2
diffcrcnt. In general, it was found that - -. -
- - -- -- -- -

for the watersheds in the Foothills


group, the peak ratios of snowmelt Di.\cnssion ment, storage masks the flashiness
events tended to be higher than those caused by the intense storms and at-
from rainfall. Conversely, an opposite 'I'hc coefficient, C, of equation 2 tenuates the peak flows.
trend was apparent in the arrays for can hi: considered as an index of the
watersheds in the Cypress Hills and flashiness of the streams. As shown in To indicate the interrelationship bc-
Wood Mountain group (figure 3 ) and table 2. those watersheds which fall in tween the width of the hydrograph and
the peak ratios, the average ratio.
those in the Central Plains. Unfor- Cypress Hills and Wood Mountain q , / q 2 for each watershed was plotted
tunately, there were not sufficient data (Groups 2 and 2a) are very flashy in against the average time in hours during
available either on watersheds within comparison with the others. Streams of which the stage was above 75 percent
the same regional grouping or for this region are of the intermittent type of the peak stage as measured above
watersheds in different physiographic carrying little or no base flow during base flow. T,, (see figure 5 ) . As
regions to permit a complete study of the summer months. In addition, the
this phenomena. area is covered by relatively-sparse,
vegetal cover, the soil mantle is thin
Average peak ratios of summer and the relief very pronounced. In thc
storm events and effective area region, runoft-producing storms during
the summer months are primarily short,
As a further reduction to the data, intense. thunderstorms. The combina-
the averagc values of the peak ratios tion of physiographic and climatic
of summer storm events were deter- elements of the region tcnd to produce
mined and plottcd with effective area hydrographs with a short period of
as shown in figurc 4. As indicated on rise and rapid recession. In contrast,
this figure, the average peak ratios watersheds in the Foothills although on Figure 5. A v e r a g e peak r a t i o a n d t h e a v e r a g e
duration.
from watersheds within a given physio- steep topography do not tend to be
graphic region can be related to cffec- nearly as flashy. This can be attributed shown in the figure, there is a distinct
tivc area in equational form as. to the fact that. ( a ) most of the basins trend for q,/q, to dccrease as the
width of the hydrograph increases
I Further, when T,, is greater than ap-
I l l I
N - proximately 24 hrs the ratio tends
<a.
-
-
Key:
FOOTHILLS
toward a limit of unity.
u- CYPRESS HILLS I The exponent of area, n. in equation
I I ------- CENTRAL P L A I N S
0 5 I
1 --- MAN. ESCARPMENT
2, is an index of the relative importance
of area as opposed to other factors in
contributing to the peak ratio. Thc
strcams in Groups 2 and 2a and 3
havc the largest exponents being -0.46
and -0.36 respectively whereas the ex-
ponents of the equation for Groups 1
and 4 are smaller. Gray ( 4 ) has shown
that although area is only one of the
many factors of the geomorphic regime
that establishes hydrograph shape, it
is often predominant inasmuch as
within a given physiographic region it
reflects many other watershed factors
10 I00 400
(lengths and slopes of streams) which
EFFECTIVE AREA, A, sq. mi. affect the time distribution of surface
Figure 4. Regional t r e n d s i n p e a k ratio-Area relation. runoff.

38 CANADIAN AGRICULTURAL ENGINEERING, FEE. 1966


From the preceding discussions it bear a close resemblance to those of discharge for Prairie streams using
should not be construed that large the more amenable streams studied by published average daily flows for; (a)
values of the peak ratio connotate large Langbein (figure 1 ). In general, be- the day preceding the occurrence of
discharges. Generally, the opposite is cause of scarcity and limitations of the the peak rate, (b) the day of occur-
true since large discharges tend to be original data, it was impossible to rence of the peak rate, and (c) the day
morc sustained which cause lower locate contours for values of qp/qz following the occurrence of the peak
values of q~ /q2. Yet, the discharge per greater than 2.5. Other areas of fewest rate.
unit area will be lower on larger basins points, and thus least reliability are
than on smaller areas. delineated on the figure. The practical LITERATURE CITED
Langbein's Approach use of figure 6 is as an aid for de-
termining qp when q,, q2. and q, are I . Durrant. E. F. and S. R. Black-
In accordance with the approach known. well. The Magnitude and Fre-
suggested by Langbein ( 5 ) a graph quency of Floods in the Canadian
was constructed using the following Prairies. Proc, of Symposium No.
flow data; ( a ) the average daily flow SUMMARY 1, Spillway Design Floods. Sub-
for the day preceding the peak rate, committee on Hydrology, NRC
Discharge records collected from 22 Associate Committee on Geodesy
q,, (b) the average daily flow on the watersheds located within the central
day following the peak, q,, and (c) the and Geophysics. The Q u e e n s
region of Canada were analyzed to Printer, Ottawa, 1959.
peak instantaneous discharge, q, and establish relationships between the
peak instantaneous discharge, q , and 2. Ellis, W. H. A Study of Peak
and the average daily discharge, q., Instantaneous and Average Daily
for Prairie streams. From arialysis of Discharges from Small Prairie
the data it was found that the variation Watersheds. Unpublished M.Sc.
in the average peak ratio, q,,/qp, of Thesis. University of Saskatche-
summer storm events for watersheds wan, Saskatoon. 1964.
within selected regional groupings
could bz related to the effective size of 3. Fuller. W. E. Flood Flows. Trans.
the drainage basin. Differences in the ~ m e r Soc.
: Civil Engrs. 77: 564-
relationships for watersheds in different 617. 1914.
physiographic areas could be attributed 4. Gray, Don M. Interrelationships
to such factors as topography, climate of Watershed Characteristics.
and vegetation as they affect the dis- Jour. of Geophys. Res. 66: 1215-
charge characteristics of streams. In 1223. 1961.
addition, it was found that the peak
ratios from rainfall and snowmelt 5. Langbein, W. B. Peak Discharges
events on watersheds within a given from Daily Records. U.S. Geol.
Figure 6 . Peak r a t i o contours i n terms o f d a y before, Survey Bull. p. 145. August,
day of, a n d d a y f o l l o w i n g maximum discharge. regional grouping were not of the same
the average daily flow on the day of magnitude and could not be considered 1944.
occurrence of the peak, q?. These from a homogenous population. This 6. Stitchling W. and S. R. Blackwell.
data. plotted as dimensionless ratios necessitates that a separate relation- Drainage Area as a Hydrologic
are shown in figure 6. ship be established for the different Factor on the Canadian Prairies.
events. Proc. Inter. Union of Geodesy
The most significant feature of the
figure is that it shows that the discharge A graph is presented which can be and Geophysics. Vol. 3, 1957.
characteristics of the streams analyzed used to estimate the peak instantaneous

You might also like