Equations for Estimating Synthetic

Unit-Hydrograph Parameter Values

for Small Watersheds in Lake County,
Illinois

U.S. GEOLOGICAL SURVEY

Open-File Report 96–474

Prepared in cooperation with

LAKE COUNTY STORMWATER MANAGEMENT COMMISSION
Equations for Estimating Synthetic
Unit-Hydrograph Parameter Values
for Small Watersheds in Lake County,
Illinois
By CHARLES S. MELCHING and JASON S. MARQUARDT

U.S. GEOLOGICAL SURVEY

Open-File Report 96-474

Prepared in cooperation with

LAKE COUNTY STORMWATER MANAGEMENT COMMISSION

Urbana, Illinois
1997
U.S. DEPARTMENT OF THE INTERIOR
BRUCE BABBITT, Secretary

U.S. GEOLOGICAL SURVEY

Gordon P. Eaton, Director

The use of firm, trade, and brand names in this report is for identification purposes only and does
not constitute endorsement by the U.S. Geological Survey

For additional information write to: Copies of this report can be purchased
from:

U.S. Geological Survey U.S. Geological Survey

District Chief Branch of Information Services
221 N. Broadway Ave. Box 25286
Urbana, Illinois 61801 Federal Center
(217) 344-0037 Denver, CO 80225
CONTENTS
Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Purpose and Scope. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Description of Study Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Synthetic Unit-Hydrograph Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Clark Unit-Hydrograph Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Soil Conservation Service Dimensionless Unit-Hydrograph Method. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Linear-Reservoir Method. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Previous Relations Between Synthetic Unit-Hydrograph Parameters and Watershed Characteristics in Illinois. . . . 11
Determination and Evaluation of Synthetic Unit-Hydrograph Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Storm Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Hydrograph-Parameter Determination by Manual Computation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Hydrograph-Parameter Determination by Rainfall-Runoff Model Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Comparison of Calibrated Hydrograph-Parameter Values with Results of Previous Relations . . . . . . . . . . . . . . . . . 16
Equations for Estimating Synthetic Unit-Hydrograph Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Equation Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Equation Verification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Application Limits for the Estimation Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Application Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
Summary and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
References Cited . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

FIGURES
1-2. Maps showing:
1. Location of Lake County in northeastern Illinois and U.S. Geological Survey rain gages . . . . . . . . . . . . . . . . 4
2. Locations of the watersheds in Lake County, Ill., utilized to develop, verify, and illustrate equations for
estimation of synthetic unit-hydrograph parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
3. Soil Conservation Service dimensionless curvilinear unit hydrograph and equivalent triangular unit
hydrograph . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
4. Graph showing time of concentration for storms on nine watersheds in Lake County, Ill., for the Clark unit-
hydrograph method measured and computed as a function of watershed area, percentage of impervious
cover, and depth of effective precipitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
5. Graph showing watershed-storage coefficient for storms on nine watersheds in Lake County, Ill., for the Clark
unit hydrograph measured and computed as a function of watershed area, percentage of impervious cover,
and main channel slope. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
6. Graph showing unit-graph lag for storms on nine watersheds in Lake County, Ill., for the Soil Conservation
Service dimensionless unit hydrograph measured and computed as a function of watershed area, percentage of
impervious cover, and depth of effective precipitation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
7. Graph showing hydrograph-time lag for storms on nine watersheds in Lake County, Ill., measured and
computed as a function of watershed area, percentage of impervious cover, main channel slope, and
depth of effective precipitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
8-18. Graphs showing measured direct-runoff hydrographs and direct-runoff hydrographs simulated with the
Clark unit-hydrograph method in the U.S. Army Corps of Engineers Flood Hydrograph Package HEC-1 with
the time of concentration and storage coefficient determined from the estimation equations for
Lake County, Ill., for the storms of:
8. April 26, 1995, on Bull Creek . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

Contents III
 9. April 26, 1995, on Indian Creek. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
10. November 5, 1994, on Green Lake Ditch. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
11. November 27, 1994, on Green Lake Ditch. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
12. April 26, 1995, on North Branch of the Chicago River . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
13. May 23, 1995, on North Branch of the Chicago River. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
14. April 26, 1995, on Skokie River at Lake Forest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
15. April 26, 1995, on Skokie River near Highland Park. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
16. April 26, 1995, on Squaw Creek . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
17. April 26, 1995, on Flint Creek . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
18. May 23, 1995, on Flint Creek. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

TABLES
1. Rainfall data-collection sites in and near Lake County, Ill.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2. Geomorphologic and land-cover characteristics of watersheds in Lake County, Ill., selected for
determination of equations for estimation of synthetic unit-hydrograph parameters. . . . . . . . . . . . . . . . . . . . . . 7
3. Average values of measures of calibration quality for the calibrated Clark unit-hydrograph method and the
calibrated Soil Conservation Service dimensionless unit-hydrograph method for all storms on selected
watersheds in Lake County, Ill., utilized to develop and verify the equations for estimation of synthetic
unit-hydrograph parameters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
4. Range and mean of unit-hydrograph peak factors of the type applied in the Soil Conservation Service
dimensionless unit-hydrograph method determined by manual hydrograph analysis for all storms on
selected watersheds in Lake County, Ill., utilized to develop and verify the equations for estimation of
synthetic unit-hydrograph parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
5. Time of concentration for the Clark unit-hydrograph method estimated with the method of Graf and
others (1982b) compared with the minimum, mean, and maximum values determined from calibration
for all storms on selected watersheds in Lake County, Ill., utilized to develop and verify the equations for
estimation of synthetic unit-hydrograph parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
6. Watershed-storage coefficient for the Clark unit-hydrograph method estimated with the method of Graf
and others (1982b) compared with the minimum, mean, and maximum values determined from calibration
for all storms on selected watersheds in Lake County, Ill., utilized to develop and verify the equations for
estimation of synthetic unit-hydrograph parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
7. Time-to-peak for the Snyder unit-hydrograph method estimated with the method of Singh (1981) compared
with the minimum, mean, and maximum values determined from calibration for all storms on selected
watersheds in Lake County, Ill., utilized to develop and verify the equations for estimation of synthetic
unit-hydrograph parameters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
8. Standard error and multiple correlation coefficient for logarithmic data in the equations for estimating
synthetic unit-hydrograph parameters for Lake County, Ill. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
9. Percentage error in the estimated peak discharge and time-to-peak discharge for the verification storms on
selected watersheds in Lake County, Ill., simulated with the Clark unit-hydrograph method utilizing estimated
values of time of concentration and watershed-storage coefficient. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
10. Direct-runoff hydrograph characteristics and parameters determined from manual analysis of hydrographs
on watersheds in Lake County, Ill., for storms utilized to develop and verify the equations for estimation of
synthetic unit-hydrograph parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
11. Parameters for the Clark and Snyder unit-hydrograph methods determined from calibration of the
U.S. Army Corps of Engineers (1990) Flood Hydrograph Package HEC-1 and measures of calibration
quality for data from watersheds in Lake County, Ill., for storms utilized to develop and verify the
equations for estimation of synthetic unit-hydrograph parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
12. Unit-graph lag for the Soil Conservation Service dimensionless unit hydrograph determined from
calibration of the U.S. Army Corps of Engineers (1990) Flood Hydrograph Package HEC-1 and
percentage error in calibrated-hydrograph peak discharge for data from watersheds in Lake County, Ill.,
for storms utilized to develop and verify the equations for estimation of synthetic unit-hydrograph parameters 37

IV Contents
CONVERSION FACTORS

Multiply By To obtain

inch (in.) 25.4 millimeter

inch per hour (in/h) 25.4 millimeter per hour

foot (ft) 0.3048 meter

mile (mi) 1.609 kilometer

acre 2.471 hectare

square mile (mi2) 2.59 square kilometer

cubic foot per second (ft3/s) 0.0283 cubic meter per second

foot per mile (ft/mi) 0.1894 meter per kilometer

Contents V
VI Contents
Equations for Estimating Synthetic Unit-Hydrograph
Parameter Values for Small Watersheds in Lake
County, Illinois
By Charles S. Melching and Jason S. Marquardt

Abstract tiple linear regression of the logarithms of the

parameters and characteristics.
Design hydrographs computed from design
Separate sets of equations were developed
storms, simple models of abstractions (intercep-
with watershed area and main channel length as
tion, depression storage, and infiltration), and syn-
the starting parameters. Percentage of impervious
thetic unit hydrographs provide vital information
cover, main channel slope, and depth of effective
for stormwater, flood-plain, and water-resources
precipitation also were identified as important
management throughout the United States. Rain-
characteristics for estimation of synthetic unit-
fall and runoff data for small watersheds in Lake
hydrograph parameters. The estimation equations
County collected between 1990 and 1995 were
utilizing area had multiple correlation coefficients
studied to develop equations for estimation of syn-
of 0.873, 0.961, 0.968, and 0.963 for TC, R, UL,
thetic unit-hydrograph parameters on the basis of
and TL, respectively, and the estimation equations
watershed and storm characteristics. The synthetic
unit-hydrograph parameters of interest were the utilizing main channel length had multiple correla-
time of concentration (TC) and watershed-storage tion coefficients of 0.845, 0.957, 0.961, and 0.963
coefficient (R) for the Clark unit-hydrograph for TC, R, UL, and TL, respectively.
method, the unit-graph lag (UL) for the Soil Con- Simulation of the measured hydrographs for
servation Service (now known as the Natural the verification storms utilizing TC and R obtained
Resources Conservation Service) dimensionless from the estimation equations yielded good results
unit hydrograph, and the hydrograph-time lag (TL) without calibration. The peak discharge for 8 of
for the linear-reservoir method for unit- the 11 storms was estimated within 25 percent and
hydrograph estimation. Data from 66 storms with the time-to-peak discharge for 10 of the 11 storms
effective-precipitation depths greater than was estimated within 20 percent. Thus, application
0.4 inches on 9 small watersheds (areas between of the estimation equations to determine synthetic
0.06 and 37 square miles (mi2)) were utilized to unit-hydrograph parameters for design-storm
develop the estimation equations, and data from 11 simulation may result in reliable design hydro-
storms on 8 of these watersheds were utilized to graphs; as long as the physical characteristics
verify (test) the estimation equations. The syn- of the watersheds under consideration are within
thetic unit-hydrograph parameters were deter- the range of those for the watersheds in this study
mined by calibration using the U.S. Army Corps of (area: 0.06-37 mi2, main channel length: 0.33-
Engineers Flood Hydrograph Package HEC-1(TC, 16.6 miles, main channel slope: 3.13-55.3 feet
R, and UL) or by manual analysis of the rainfall per mile, and percentage of impervious cover:
and runoff data (TL). The relation between syn- 7.32-40.6 percent). The estimation equations are
thetic unit-hydrograph parameters, and watershed most reliable when applied to watersheds with
and storm characteristics was determined by mul- areas less than 25 mi2.

Abstract 1
INTRODUCTION and watershed and storm characteristics had been
developed for application on a national (or even global)
Design of stormwater management facilities basis if suitable relations for local conditions are not
and other hydraulic structures (culverts, bridge water- available. The relations for estimating the time of
ways), determination of flood plain boundaries, and
concentration and traveltime for the SCS Technical
assessment of the safety of structures in rivers typically
Release 55 (TR55) (Soil Conservation Service, 1986)
involve the application of a design hydrograph. These
are examples of national relations. Application of
design hydrographs are computed on the basis of
design storms of a specified probability of occurrence national relations may result in substantial errors in a
determined from standard references, such as the specific region.
U.S. Weather Bureau Technical Paper Number 40 The Clark (1945) unit-hydrograph method as
(TP40) (Hershfield, 1961) or the Illinois State Water implemented in the U.S. Army Corps of Engineers
Survey Bulletin 70 (Bulletin 70) (Huff and Angel, (1990) Flood Hydrograph Package HEC-1 and the
1989). Abstractions from rainfall resulting from inter- SCS dimensionless unit-hydrograph method as imple-
ception, depression storage, and infiltration are then mented in TR55 are the most commonly applied
determined on the basis of available data from the synthetic unit-hydrograph methods in Illinois. In addi-
literature and considering the effects of the soil type, tion to utilizing a unit-graph-lag parameter, the SCS
land cover/land use, and antecedent moisture condi- dimensionless unit-hydrograph method applies a fixed
tions. Typically, the Soil Conservation Service (SCS, relation between unit-hydrograph peak discharge and
now known as the Natural Resources Conservation Ser- watershed area and time-to-peak. This peak factor has
vice) (1985) curve-number method is applied been found to be substantially high in a number of areas
to determine the abstractions. By subtracting the
in the United States (Woodward and others, 1995).
abstractions from the design rainfall, the precipitation
excess, which approximately equals the direct runoff Lake County, Ill., is undergoing rapid urbaniza-
(effective precipitation) resulting from the design tion and management of the resultant increases in
storm, is obtained. The precipitation excess is then stormwater runoff and flooding is an important activity
transformed into a hydrograph at the outlet of the in the county. Design hydrographs are needed for
watershed utilizing a synthetic unit hydrograph. If a stormwater management planning, flood plain delinea-
large area is studied, it is subdivided into a number tion, and stormwater-mitigation structure design in
of subwatersheds and the runoff hydrographs from Lake County. The U.S. Geological Survey (USGS)
each of these subwatersheds is routed to the watershed has operated an extensive network of rainfall and
outlet with hydrologic or hydraulic routing methods. streamflow gages in and near Lake County since
For example, Snider (1971) recommended that a single December 1989 in cooperation with the Lake County
SCS dimensionless unit hydrograph should not be used Stormwater Management Commission (LCSMC). The
for watersheds greater than 20 mi2. data from this network are sufficient to develop rela-
The primary advantage of utilizing synthetic tions between hydrograph parameters, and watershed
unit hydrographs is that the complete unit hydrograph and storm characteristics for stormwater management
may be determined with the specification of one or on small watersheds (less than 25 mi2) in Lake County.
two hydrograph parameters. Further, many studies Therefore, the USGS, in cooperation with the LCSMC,
have shown that relations between these hydrograph began a study to develop relations between hydrograph
parameters and watershed and storm characteristics
parameters (unit-graph lag, time of concentration,
may be developed. Synthetic unit hydrographs may
hydrograph-time lag, and watershed-storage coeffi-
be derived for ungaged watersheds utilizing the
relations between hydrograph parameters, and cient), and watershed (area, main channel length and
watershed and storm characteristics as long as the slope, percentages of impervious cover, forest cover,
ungaged watersheds are hydrologically similar to the and open water) and storm (effective-precipitation
gaged watersheds for which the relation was devel- depth, duration, and intensity) characteristics for use
oped. Hydrologic similarity includes similarity in in stormwater management in Lake County, Ill. The
topography, geomorphology, soil types, land cover/ applicability of the peak factor utilized in the SCS
land use, and climate. For some synthetic unit hydro- dimensionless unit hydrograph for small watersheds
graphs, the relation between hydrograph parameters, in Lake County also was evaluated.

2 Equations for Estimating Synthetic Unit-Hydrograph Parameter Values for Small Watersheds in Lake County, Illinois
Purpose and Scope time step of 5 minutes. This network of rain gages
was installed to develop rainfall-runoff relations for
This report (1) describes the procedures utilized simulation of streamflow for watersheds in Lake
to develop and test the relations between hydrograph County as described by Duncker and others (1995).
parameters, and watershed and storm characteristics, When the project to develop rainfall-runoff relations
and (2) illustrates the accuracy and application of the was completed on September 30, 1993, four of the
relations developed for computation of synthetic unit rain gages were discontinued and a new rain gage was
hydrographs for small, ungaged watersheds in Lake installed in Highland Park, Ill. Thus, data from a total
County, Ill. The development of the relations included of 24 rain gages in and near Lake County (fig. 1) were
detailed analysis of rainfall and runoff data to identify available at various times (table 1) to assist in the devel-
storms suitable for hydrograph-parameter determina- opment of the estimation equations. The areal extent of
tion; computation of hydrograph-time lag by manual the rain-gage network results in reliable rainfall-depth
analysis; determination of unit-graph lag, time of and temporal distribution data for determination of
concentration, and watershed-storage coefficient by hydrograph parameters for watersheds with streamflow
calibration of HEC-1; and determination of the gages in Lake County.
relations by multiple nonlinear regression. Testing of Streamflow data are available at 14 gages on
the relations included comparison of hydrograph- streams draining watersheds, primarily in Lake County,
parameter values determined for verification storms during the period of detailed rainfall data (December
to values estimated with the relations and comparison 1989-September 1993). These streamflow gages are
of hydrographs computed utilizing the estimated listed below.
values of the hydrograph parameters to measured
hydrographs for verification storms. The accuracy of
the relations is inferred from the results of HEC-1 Station
calibration, nonlinear regression, and relation verifica- number Station name
tion.
05527940 Tempel Farms Ditch near Old Mill
Creek, Ill.
Description of Study Area 05527950 Mill Creek at Old Mill Creek, Ill.
05528030 Bull Creek near Libertyville, Ill.
The objective of this study is to develop relations 05528040 Terre Faire Ditch at Libertyville, Ill.
between hydrograph parameters, and watershed and 05528230 Indian Creek at Praire View, Ill.
storm characteristics for computation of synthetic unit 05528475 Green Lake Ditch at Buffalo Grove, Ill.
hydrographs for small, ungaged watersheds in Lake 05528500 Buffalo Creek near Wheeling, Ill.
County. Therefore, the areas studied are the watersheds 05534500 North Branch Chicago River at
in Lake County for which detailed rainfall and runoff Deerfield, Ill.
data are available. The network of rainfall gages oper- 05535000 Skokie River at Lake Forest, Ill.
ated by the USGS in and near Lake County during the 05535070 Skokie River near Highland Park, Ill.
study period is shown in figure 1 and the station num- 05535500 West Fork of North Branch of Chicago
bers, names, and periods of record of the rain gages are River at North Brook, Ill.
listed in table 1. The watersheds utilized to develop the 05547755 Squaw Creek at Round Lake, Ill.
estimation equations are shown in figure 2. The factors 05549835 Lakeview Plaza Ditch at
affecting the selection of these watersheds are Lake Zurich, Ill.
described below. 05549850 Flint Creek near Fox River Grove, Ill.
The network of rain gages in and near Lake
County was started in December 1989 with the Five of these watersheds and streamflow gages
installation of eight gages at various locations were not considered in this study for the following
throughout the county. The network increased to reasons. Tempel Farms Ditch drains a small 0.492-mi2
14 rain gages in February 1990, to 18 rain gages in watershed consisting of 100 percent pervious land
April 1990, and finally to 23 rain gages in May 1991. cover in the form of agricultural pasture. As a result
All installations included tipping-bucket rain gages of the pervious land cover, and corresponding high
capable of measuring rainfall depths of 0.01 in. at a interception and depression storage in the watershed,

Introduction 3
 88° 07′ 30″ 88° 87° 52′ 30″

Lake
County
23 24
ILL

45

41

Mil
l
42° 30′ WISCONSIN
2 ILLINOIS

Cr
Old Mill

Des
Creek
12
22 3
Spring
Grove Fox 21
Lake
Lindenhurst

20
42° 22′ 30″
Fox Gurnee
19 Waukegan
Lake 11
4
Round Grayslake
Lake
MC HENRY

Squa

5
r
Bull C

12
w

6
LAKE

Lake
Cr

Libertyville
Michigan
Lake
18 Forest
Wauconda
42° 15′ 17
Pla
Ind

14 14
in

Lake
es
ian

Cr 1
13 Zurich 16
Fox 15 Highland Park
River
Rive

Cr 7 Deerfield
Grove 10
Flint
r

LAKE
COOK 8
Buffalo Wheeling 9
Grove

Base from U.S. Geological Survey 0 3 6 MILES

1:100,000 Digital Data
Albers Equa-Area Conic Projection 0 3 6 KILOMETERS
Standard paralleles 33˚ and 45˚, central meridian -89˚

EXPLANATION
1 U.S. GEOLOGICAL SURVEY RAIN GAGE
AND SITE NUMBER--See table 1

Figure 1. Location of Lake County in northeastern Illinois and U.S. Geological Survey rain gages.

4 Equations for Estimating Synthetic Unit-Hydrograph Parameter Values for Small Watersheds in Lake County, Illinois
Table 1. Rainfall data-collection sites in and near Lake County, Ill.
[Site identifiers correspond to those in figure 1; present refers to September 10, 1996]

Site Station Period

identifier number Station name of record

1 040874126 Southwest Fork of South Branch of Ravine 10 at Highland Park, Ill. 12/07/93-present
2 05527800 Des Plaines River at Russell, Ill. 12/05/89-present
3 05527940 Tempel Farms Ditch near Old Mill Creek, Ill. 07/11/91-present
4 05528000 Des Plaines River near Gurnee, Ill. 12/06/89-present
5 05528030 Bull Creek near Libertyville, Ill. 12/04/89-present

6 05528040 Terre Faire Ditch at Libertyville, Ill. 05/08/91-present

7 05528475 Green Lake Ditch at Buffalo Grove, Ill. 04/25/91-present
8 05528500 Buffalo Creek near Wheeling, Ill. 04/17/90-present
9 05534500 North Branch Chicago River at Deerfield, Ill. 12/10/89-present
10 05535070 Skokie River near Highland Park, Ill. 12/04/89-present

11 05547755 Squaw Creek at Round Lake, Ill. 12/06/89-11/21/93

12 05548280 Nippersink Creek near Spring Grove, Ill. 03/13/91-present
13 05549835 Lakeview Plaza Ditch at Lake Zurich, Ill. 04/08/91-04/20/94
14 05549850 Flint Creek near Fox River Grove, Ill. 12/06/89-present
15 421113088042200 Lake Zurich Wastewater Treatment Facility at Lake Zurich, Ill. 02/13/90-present

16 421215087573400 Vernon Hills Rain Gage at Praire View, Ill. 04/17/90-present

17 421428088012900 Diamond Lake Wastewater Treatment Facility at Diamond Lake, Ill. 02/13/90-present
18 421533088084600 Wauconda Wastewater Treatment Facility at Wauconda, Ill. 02/12/90-present
19 422118088014700 Grayslake Wastewater Treatment Facility at Grayslake, Ill. 02/01/90-present
20 422315088091800 Fox Lake Rain Gage at Fox Lake, Ill. 04/23/90-present

21 422459087520700 Waukegan Airport at Waukegan, Ill. 04/17/90-present

22 422553088015300 Lindenhurst Wastewater Treatment Facility at Lindenhurst, Ill. 12/06/89-12/03/93
23 423451088052400 Paddock Lake Wastewater Treatment Facility at Paddock Lake, Wis. 02/06/90-12/06/93
24 423526087551800 Kenosha Airport Rain Gage at Kenosha, Wis. 02/06/90-present

no storms produced more than the target level of direct- small size and high imperviousness of the watershed.
runoff volume for hydrograph analysis (discussed in Because the goal of this study was to develop estima-
the “Determination and Evaluation of Synthetic Unit- tion equations suitable for watersheds with areas less
Hydrograph Parameters” section). Mill Creek drains than 25 mi2, data from Indian and Flint Creeks could
a 59.6 mi2 primarily rural watershed with substantial have been omitted from the analysis. However, data
wetland areas along the main stem. The wetlands and from these watersheds were retained to include a wider
semipermanent debris jams along Mill Creek result in range of watershed conditions in the regression analy-
substantial attenuation of the runoff hydrograph for sis. Thus, the study area consists of nine watersheds in
this watershed, which is not representative of condi- and near Lake County, Ill. (fig. 2).
tions on small (less than 25 mi2), ungaged watersheds Lake County lies entirely within the Wheaton
in Lake County. The Buffalo Creek watershed was Morainal Region (Leighton and others, 1948). Under
eliminated from consideration because of a flood- the more natural, nonurban conditions present in 1954,
control reservoir about 1 mi upstream from the Mitchell (1954, p. 335) noted that the Wheaton
streamflow gage. The West Fork of the North Branch Morainal Region is characterized by flat slopes, long,
of Chicago River watershed was eliminated from narrow basins, and large storage in lake and swamp
consideration because of two off-line flood-control areas. In the mid 1980’s, rapid urbanization began in
reservoirs in the watershed that substantially alter rural areas in Lake County. Since then the presence of
the natural rainfall-runoff process for larger storms. impervious areas and drainage structures (storm sewers
Finally, Lakeview Plaza Ditch drains a 0.0055-mi2 and swales) has substantially affected the rainfall-
watershed consisting of nearly 100 percent impervious runoff process and resulting hydrographs in Lake
area in the form of a commercial shopping mall and County. The primary geomorphologic and land-cover
adjacent parking lot. The hydrograph parameters for characteristics of the watersheds utilized for determina-
this watershed are not useful for development of rela- tion of the parameters for synthetic unit hydrographs
tions for general application because of the extremely and development of relations for estimating these

Introduction 5
 88° 07′ 30″ 88° 87° 52′ 30″
42° 30′ WISCONSIN Lake
ILLINOIS 41
County
ILL
Old Mill
Creek

Lindenhurst
Fox
Lake
45

Fox
42° 22′ 30″ Waukegan
Lake
Round Gurnee
Lake
Grayslake

Squ
McH ENR Y

aw
12
Lake
LAKE

TFD
Bu Michigan
Cr

Nor
ll Cr

th

Sko
Libertyville Lake

B ra

kie
Wauconda Forest
42° 15′

nch
Des
India n

Ch
Pl a

ica
Fox River Cr ine

go
s Highland
Grove Lake Park
Zurich

Ri
ve
R

r
iv
er
Fl in t Cr Buffalo GLD Deerfield
LAKE Creek
COOK Buffalo Wheeling
Grove
Riv

42° 07′ 30″ 14

er

Base from U.S. Geological Survey 0 1 2 3 4 5 MILES

1:100,000 Digital Data
Albers Equal-Area Conic Projection 0 1 2 3 4 5 KILOMETERS
standard parallels 33˚ and 45˚, central meridian -89˚

EXPLANATION
WATERSHED BOUNDARY--For Skokie River,
northern watershed is Skokie River at Lake
Forest, and the entire watershed is Skokie
River near Highland Park
TFD TERRE FAIRE DITCH
GLD GREEN LAKE DITCH

Figure 2. Location of the watersheds in Lake County, Ill., utilized to develop, verify, and illustrate equations
for estimation of synthetic unit-hydrograph parameters.

6 Equations for Estimating Synthetic Unit-Hydrograph Parameter Values for Small Watersheds in Lake County, Illinois
parameters are listed in table 2. The watersheds utilized characteristics of the watersheds and storms from
in this study include a representative range of the which they were determined. Therefore, unit hydro-
watershed characteristics likely to substantially affect graphs may be estimated for ungaged watersheds
hydrograph shape for watersheds in Lake County. with geomorphology, soils, land cover/land use, and
The land-cover percentages listed in table 2 were climate similar to that for the gaged basins. Many
determined on the basis of remotely sensed thematic synthetic unit-hydrograph methods have been proposed
mapping and side-looking-airborne-radar imagery. The in the hydrologic literature. In this report, only three
procedure for determining the percentage of impervi- synthetic unit-hydrograph methods are considered: the
ous area was specially calibrated to data from the Green Clark (1945) unit-hydrograph method, the SCS dimen-
Lake Ditch watershed and verified for the Bull Creek sionless unit-hydrograph method (Snider, 1971), and
watershed as described in Duncker and others (1995, p. the linear-reservoir method. The first two methods are
17). The percentage of forest area for the Skokie River commonly applied for hydrologic design and analysis
and North Branch Chicago River watersheds appear in Illinois. The third method is frequently applied for
relatively high. However, these high percentages reflect small watersheds in Du Page County, Ill., and was
Forest Preserve land along each river and the large res- found to result in reliable unit hydrographs for water-
idential estates with extensive wooded areas in each sheds smaller than 5 mi2 (Rao and others, 1972). Rela-
watershed. tions between hydrograph parameters for two synthetic
unit-hydrograph methods and characteristics of Illinois
watersheds have been developed in previous studies.
Acknowledgments
These previous studies also are discussed below.
Twelve municipalities in and near Lake County
allowed gage installations at facilities such as
wastewater-treatment plants, fire stations, and airports. Clark Unit-Hydrograph Method
Their cooperation and assistance are appreciated. The
authors appreciate the private citizens throughout Lake The movement of flow through a watershed is
County who allowed installation of gages on their dominated by the processes of translation and attenua-
property and recorded data for the study. tion. Translation is the movement of flow downgradient
through the watershed in response to gravity. Attenua-
tion results from the frictional forces and channel stor-
SYNTHETIC UNIT-HYDROGRAPH age effects that resist the flow. Clark (1945) noted that
the translation of flow through the watershed can be
METHODS
described by a time-area curve, which expresses the
Synthetic unit-hydrograph methods are utilized curve of the fraction of watershed area contributing
to describe the entire unit hydrograph for a gaged runoff to the outlet of the watershed as a function of
watershed with one or two hydrograph parameters. time since the start of effective precipitation. The
These hydrograph parameters can be related to the time-area curve is bounded in time by the watershed

Table 2. Geomorphologic and land-cover characteristics of watersheds in Lake County, Ill., selected for determination of
equations for estimation of synthetic unit-hydrograph parameters
{mi2, square mile; mi, mile; ft/mi, foot per mile]
Drainage Main channel Impervious Forest Wetland
area Length Slope area area area
Watershed (mi2) (mi) (ft/mi) (percent) (percent) (percent)

Bull Creek 6.3 6.4 3.13 13.9 7.48 6.80

Terre Faire Ditch .077 .33 55.3 27.7 .00 2.00
Indian Creek 35.7 11.6 13.6 15.8 3.48 4.22
Green Lake Ditch .06 .60 14.0 40.6 .00 .00
North Branch Chicago River 19.7 13.5 3.24 21.3 32.5 .77
Skokie River
at Lake Forest 13.0 10.8 5.58 29.4 24.0 .15
near Highland Park 21.1 16.6 5.29 34.4 30.1 .24
Squaw Creek 17.2 7.8 4.79 7.32 3.73 7.32
Flint Creek 37.0 12.9 7.99 8.83 8.97 5.09

Synthetic Unit-Hydrograph Methods 7

time of concentration (TC). Thus, TC is a hydrograph time required for the last drop of effective precipitation
parameter of the Clark unit-hydrograph method. Atten- at the hydraulically most distant point in the watershed
uation of flow can be represented with a simple, linear to reach the channel network. From linear system the-
reservoir for which storage is related to outflow as ory and the conceptual model of pure translatory flow,
the two definitions of time of concentration are equiva-
lent. However, the subtle differences in the definition of
S = RO, (1)
time of concentration between the Rational method and
the Clark unit-hydrograph method imply that the time
where,
of concentration estimation equations commonly
S is the watershed storage,
applied in the Rational method may not be appropriate
R is the watershed-storage coefficient, and
for the Clark unit-hydrograph method. In most applica-
O is the outflow from the watershed.
tions of HEC-1, TC is determined from values cali-
Therefore, Clark (1945) proposed that a synthetic unit brated with measured rainfall and runoff data either
hydrograph could be obtained by routing 1 in. of direct directly, by scaling from hydrologically similar water-
runoff into the channel in proportion to the time-area sheds, or from relations, such as those developed in this
curve and routing the runoff entering the channel study.
through a linear reservoir.
Numerous researchers have found that the actual
time-area curve for the watershed need not be deter- Soil Conservation Service Dimensionless
mined to obtain a reasonable unit hydrograph. For Unit-Hydrograph Method
example, Turner and Burdoin (1941) and O’Kelly
(1955) found that reasonable unit hydrographs were In the SCS dimensionless unit-hydrograph
obtained when simple geometric shapes were substi- method, all the hydrograph ordinates are given by
tuted for the actual time-area curve. Experience with ratios between instantaneous discharge and peak
the Clark unit-hydrograph method at the U.S. Army discharge and between time and time-to-peak as illus-
Corps of Engineers, Hydrologic Engineering Center, trated in figure 3. The unit-hydrograph peak discharge
indicates that a detailed time-area curve usually is not also is directly related to the time-to-peak from consid-
necessary for accurate synthetic unit-hydrograph esti- eration of the volume of the unit hydrograph. This is
mation (Ford and others, 1980). In most instances, the best illustrated for the SCS dimensionless, triangular
dimensionless time-area curve included in HEC-1 unit hydrograph shown in figure 3. The volume of the
(U.S. Army Corps of Engineers, 1990) is satisfactory unit hydrograph in cubic feet is
for obtaining a reliable synthetic unit hydrograph.
In Illinois, HEC-1 (U.S. Army Corps of Engi-
neers, 1990) typically is utilized to compute the Clark V = (1 in.)(A mi2)(1 ft/12 in.)(5,280 ft/1 mi)2 =
unit hydrograph. The hydrograph parameters required 2,323,200 A, (2)
for HEC-1 computations of the Clark unit hydrograph
are TC and R. The time of concentration for the Clark where V is volume of direct runoff in cubic feet and A
unit hydrograph is slightly different than the typical is watershed area in square miles.
definition applied in stormwater management, such as The volume of runoff under the SCS dimension-
in the Rational method (Kuichling, 1889). In the typical less, triangular unit hydrograph is
definition, the time of concentration (tc) is the travel-
time required for the first drop of effective precipitation
at the hydraulically most distant point in the watershed V = 0.5{(qp ft3/s)[(Tp hr) + (Tr hr)]}(3,600 s/1 hr)
to reach the watershed outlet. In the Clark unit-
hydrograph method, TC is the time from the end of
V = 1,800 qp(Tp + Tr),
effective precipitation to the inflection point of the
recession limb of the runoff hydrograph. The inflection
point on the runoff hydrograph corresponds to the time where
when overland flow to the channel network ceases and qp is unit-hydrograph peak discharge in cubic feet
beyond that the measured runoff results from drainage per second,
of channel storage. Therefore, Clark’s TC is the travel- Tp is the time-to-peak discharge in hours, and

8 Equations for Estimating Synthetic Unit-Hydrograph Parameter Values for Small Watersheds in Lake County, Illinois
Synthetic Unit-Hydrograph Methods

Figure 3. Soil Conservation Service dimensionless curvilinear unit hydrograph and equivalent triangular unit hydrograph.
9
 Tr is the time of recession in hours, which is equal fixed peak factor (484) was checked in two ways for
to 1.67 Tp for the SCS dimensionless, triangu- this study. First, calibrated values of UL are determined
lar unit hydrograph. for each storm on each watershed in HEC-1 simulation.
Therefore, the volume is The corresponding peak discharge for the calibrated
hydrograph is compared to the measured peak dis-
V = 1,800 qp(Tp + 1.67Tp) = 4,800 qpTp. (3) charge to determine the percent error resulting from the
use of the fixed peak factor. Second, values of qpTp/A
Combining equations 2 and 3, the triangular unit- can be determined for each storm on each watershed by
hydrograph peak discharge is manual analysis of the rainfall and runoff data with the
assumption that each direct-runoff hydrograph is the
result of a single period of uniform effective precipita-
qp = 484 A/Tp. (4) tion. Almost all of the direct-runoff hydrographs are the
result of storms with several periods of effective precip-
The SCS dimensionless, triangular unit hydrograph is itation (multiperiod storms). For multiperiod storms,
an approximation of the SCS dimensionless, curvilin- deconvolution techniques must be applied to determine
ear unit hydrograph, as illustrated in figure 3. Thus,
the unit hydrograph. Deconvolution is difficult to apply
equation 4 also expresses the relation between peak
and erratic variations in the computed unit hydrograph
discharge and time-to-peak for the curvilinear unit
may result because of errors in the data and in the deter-
hydrograph. Further, equation 4 is applied in (1) the
mination of the time distribution of the effective precip-
computer program (Soil Conservation Service, 1982)
implementing the SCS Technical Release 20 (TR20), itation (Chow and others, 1988, p. 218). Therefore,
which was utilized to develop the nomographs and deconvolution was not done, and the comparison of the
tables for estimating peak discharges and design fixed peak factor and the measured peak factors is only
hydrographs in SCS TR55 (Soil Conservation Service, approximate because the measured unit hydrographs
1986); and (2) the implementation of the SCS dimen- do not meet the theoretical definition of a unit
sionless unit hydrograph in HEC-1 (U.S. Army Corps hydrograph as resulting from a storm of uniform effec-
of Engineers, 1990). In the SCS dimensionless unit- tive-precipitation intensity.
hydrograph method, the time-to-peak is estimated as

Linear-Reservoir Method
Tp = Td/2 + UL, (5)
The linear reservoir described in equation 1 can
where Td is the duration of effective precipitation and be applied to obtain a unit hydrograph on the basis of
UL is the unit-graph lag. The unit-graph lag is the time the following procedure. The instantaneous unit
from the centroid of effective rainfall hyetograph to hydrograph (the unit hydrograph resulting from 1 in. of
the time of peak discharge. The shape of the SCS effective precipitation falling in an infinitesimal period
dimensionless unit hydrograph is most correct when
of time) resulting for a watershed simulated with a lin-
the duration of effective precipitation equals 0.222 UL.
ear-reservoir model is
However, equation 5 is commonly applied for effec-
tive-precipitation durations substantially different than
0.222 UL to circumvent complex S-curve computations Q(t) = (1/R)e-t/R, (6)
of the appropriate unit hydrograph.
On the basis of the derivation of equation 4 given where Q(t) is the discharge at time t. The derivation of
previously, the value of the peak factor (484) is fixed equation 6 is given in Chow and others (1988, p. 208).
for the shape of the SCS dimensionless unit hydrograph The unit hydrograph corresponding to an effective
shown in figure 3. Changing the peak factor would precipitation with a duration of Td hours may be
require development of a new dimensionless unit
obtained by integrating the convolution integral:
hydrograph for each watershed. Therefore, it is not
practical to develop relations between the peak factor
and watershed and storm characteristics from rainfall Td
Qu ( t ) = ∫0 I ( τ ) ( 1/R )e –( τ – t )/R dτ , (7)
and runoff data in Lake County. The reliability of the

10 Equations for Estimating Synthetic Unit-Hydrograph Parameter Values for Small Watersheds in Lake County, Illinois
where compared the performance of several conceptual unit-
Qu(t) is the unit-hydrograph value at time t, hydrograph models including—
τ is a dummy parameter for integration, 1. single linear-reservoir model,
I(τ) is the intensity of effective-precipitation input 2. multiple linear-reservoir model (the Nash model),
converted to a discharge for the watershed 3. single linear-reservoir/linear-channel model
area, (conceptually similar to the Clark unit
which is hydrograph),
4. Holtan’s 2 reservoir model, and
I(τ) = A(1 in./TD)(5,280 ft/1 mi)2(1 hour/ 5. the instantaneous unit hydrograph from Fourier
3,600 second)(1 ft/12 in.) = 645.33 A/TD transform.
The unit-hydrograph models were compared
for 0 ≤¹ τ ≤¹ Td, and utilizing data from 131 storms on 8 urbanized and
5 rural watersheds in Indiana and Texas ranging in size
from 0.0455 to 19.31 mi2. For watersheds with areas
I(τ) = 0 less than 5 mi2, better results were obtained from the
linear-reservoir model than the other models. Rao and
for τ > Td. others (1972) also applied multiple regression analysis
Integration of equation 7 to a time equal to Td yields to develop relations between the model parameters, and
watershed and storm characteristics including water-
shed area, main channel length and slope,
Qu(t) = 645.33A/Td (1 - e-t/R). (8a)
percentage of impervious cover, and depth and duration
of effective precipitation. The final relations involved
This is the unit hydrograph up to t = Td. For times after only area, percentage of impervious area,
Td, application of S-curve principles (Chow and others, and depth and duration of effective precipitation. Up to
1988, p. 213-218; Viessman and others, 1989,
85 percent of the data variance was explained on the
p. 192-196) to express the drainage from the linear
basis of the multiple-regression relations. Therefore,
reservoir result in the remainder of the unit hydrograph
synthetic unit hydrographs determined from the linear-
reservoir method could be useful for small watersheds
Qu(t) = 645.33A/Td (e-(t-Td)/R - e-t/R). (8b) in Lake County and a strong relation between the
hydrograph-time lag, and watershed and storm charac-
Only the storage coefficient, R, is required for the teristics should be possible.
linear-reservoir method of computing the synthetic unit
hydrograph presented above. For a single linear-reser-
voir model, the storage coefficient is equal to the Previous Relations Between Synthetic
hydrograph-time lag, TL, which is the time difference Unit-Hydrograph Parameters and
between the centers of mass of the direct-runoff Watershed Characteristics in Illinois
hydrograph and the effective precipitation hyetograph.
Therefore, the R for the linear-reservoir method was Graf and others (1982a, b) developed relations
determined manually from the direct-runoff among watershed characteristics, TC, and R for the
hydrograph and effective precipitation hyetograph. Clark unit-hydrograph method. Values of TC and R
The linear-reservoir method is considered in this were determined for 98 watersheds in Illinois ranging
report because linear-reservoir models are applied in in size from 0.45 to 362 mi2 by calibration of HEC-1
Du Page County, Ill., to transform lateral inflows com- for rainfall and runoff data for six to eight large storms
puted in time blocks (that is, uniform inflow for a given per watershed. Multiple regression analysis was
computational time step) in the Hydrological Simula- applied to determine relations among watershed char-
tion Program - FORTRAN (Johanson and others, 1984) acteristics, (TC+R), and R/(TC+R). These combined
into a realistic temporal distribution for hydraulic parameters were utilized to reduce the effects of corre-
routing with the Full Equations Model (Franz and lation between TC and R. The relation among (TC+R)
Melching, in press). Further, Rao and others (1972) and main channel length and slope was determined as

Synthetic Unit-Hydrograph Methods 11

 (TC+R) = 35.2L0.39S-0.78, (9) hydrograph for application in Lake County can be con-
sidered here.
where L is the stream length measured along the main Hydrograph-parameter values for small water-
channel from the watershed outlet to the watershed sheds (less than 25 mi2) in Lake County estimated on
divide in miles, and S is the main channel slope deter- the basis of the previous studies are subject to three
mined from elevations at points 10 and 85 percent of deficiencies in the data and analyses. These deficien-
the distance along the channel from the watershed out- cies are discussed in detail below.
let to the watershed divide, in feet per mile. Regional The rain gages utilized to determine the water-
values of R/(TC+R) were determined for various areas shed-average storm rainfall and the temporal distribu-
of the State. A value of R/(TC+R) equal to 0.7 is appro- tion of rainfall were often located 5 to 25 mi outside of
priate for the study area in Lake County. It was hypoth- the watershed where runoff data were available. There-
esized that these regional values partially account for fore, uncertainties in the temporal distribution of effec-
aspects of watershed geomorphology and land cover/ tive precipitation could substantially affect the
reliability of the hydrograph parameters determined by
land use not considered in the analysis, such as imper-
Graf and others (1982a) from calibration of TC and R
vious area and wetland area.
in HEC-1. Further, uncertainties in effective precipita-
Singh (1981) developed synthetic unit hydro- tion could appreciably affect the estimated storm dura-
graphs for use in dam safety studies in Illinois. The tion in the determination of the unit hydrograph in
Singh synthetic unit hydrograph applied modifications Singh (1981). Problems in determining storm duration
of the Snyder (1938) synthetic unit hydrograph. In the are a primary cause of variations in the S-curve that are
Snyder synthetic unit hydrograph, the peak discharge, difficult to correct. Thus, if the storm duration is incor-
runoff duration, and hydrograph width at various per- rectly identified, the S-curve and averaged and
centages of the peak discharge are related to the time- smoothed unit hydrograph could be substantially
to-peak, and the time-to-peak is related to watershed affected.
characteristics. Singh (1981) determined relations Neither Graf and others (1982a) nor (Singh,
among watershed characteristics and the peak dis- 1981) directly considered the effects of land-cover/
charge, time-to-peak, runoff duration, and several key land-use characteristics that could substantially affect
ordinates of the unit hydrograph. Unit hydrographs the hydrograph parameters, such as percentages of
were determined, converted to a common storm dura- impervious cover and wetland cover. These factors are
tion utilizing the S-curve, averaged, and smoothed for partially considered in the development of the Singh
rainfall and runoff data for four large storms on each of (1981) synthetic unit hydrograph by the division of the
131 watersheds in Illinois ranging in size between 0.07 State into eight regions. However, as discussed later,
and 464 mi2. The State of Illinois was subdivided into the variation of fraction of impervious area and fraction
eight hydrologic regions, and relations between water- of wetland area is substantial within Lake County,
shed characteristics and hydrograph parameters were which constitutes a small portion of the region consid-
developed for each region. For the region including ered by Singh. These factors also are partially
Lake County, data from 20 watersheds, ranging in size accounted for by the regional variation in R/(TC+R) in
between 0.07 and 324 mi2, were utilized. The resulting the analysis of Graf and others (1982b). Graf and
relation for the time-to-peak (tps) is others (1982b) noted that the scattergrams of the esti-
mated and measured values of TC and R showed no
clear separation of the results for the 19 urban water-
tps = 3.0 A0.421S-0.075. (10) sheds studied relative to the results for all other water-
sheds.
The Snyder (1938) synthetic unit hydrograph is not A substantial amount of data from watersheds
studied in detail in this report. However, the Snyder and larger than the largest watershed in Lake County ana-
Clark synthetic unit hydrographs are strongly related in lyzed in this study (37 mi2) was utilized in each previ-
HEC-1, and hydrograph parameters for the Snyder syn- ous study. Fifty-one of the 98 watersheds analyzed by
thetic unit hydrograph corresponding to the optimized Graf and others (1982a and b) were larger than 37 mi2
Clark synthetic unit hydrograph are output in HEC-1. and 63 were larger than 25 mi2. The large amounts of
Therefore, the utility of the Singh (1981) synthetic unit data from large watersheds may appreciably affect the

12 Equations for Estimating Synthetic Unit-Hydrograph Parameter Values for Small Watersheds in Lake County, Illinois
reliability of estimates of TC and R for small water- determine unit hydrographs should include the follow-
sheds. Only 5 of the 20 watersheds analyzed by Singh ing characteristics.
(1981) for the region including Lake County were 1. Storms with a simple storm structure resulting in
larger than 37 mi2, and 8 were larger than 25 mi2. How- well defined hydrographs with distinct peaks.
ever, the average area in this region was 54 mi2, and 2. Storms with uniform distribution of rainfall through-
because area is a key parameter in equation 10, the esti- out the period of effective precipitation.
mated values of tps may be appreciably affected by the
3. Storms with uniform spatial distribution over the
values for the large watersheds.
entire watershed.
To assess the utility of the methods of Graf and
Calibration of HEC-1 (U.S. Army Corps of Engineers,
others (1982b) and Singh (1981), TC and R are esti-
1990) reduces the importance of the second character-
mated for each watershed studied by the method of
istic because the multiple periods of effective precipita-
Graf and others (1982b), and tps is estimated by the
tion are adequately deconvoluted in the calibration
method of Singh (1981). These estimated values are
process if the direct-runoff hydrograph is well defined
compared to the values obtained from calibration of
with a distinct peak (characteristic 1). Further,
HEC-1 for rainfall and runoff data for each watershed.
Viessman and others (1989, p. 186) recommend that
the direct runoff for the selected storm should range
from 0.5 to 1.75 in. The design storms to be simulated
DETERMINATION AND EVALUATION OF with the synthetic unit hydrographs will typically result
SYNTHETIC UNIT-HYDROGRAPH in direct runoff values in this range. Further, Laurenson
PARAMETERS and Mein (1985, p. 87) stated that small storms, result-
ing in less than about 0.4 in. of runoff, are often more
Determination of parameters for synthetic unit
difficult to fit than large storms because of extreme
hydrographs involved storm selection on the basis of a
areal variability of runoff, partial-area runoff, and large
detailed analysis of available rainfall and runoff data,
differences in the time distribution of effective precipi-
manual computations of the selected storms to deter-
tation resulting from small errors in the applied abstrac-
mine hydrograph-time lag and an approximate peak
tion model. Therefore, selection of storms resulting in
factor for the SCS dimensionless unit hydrograph,
at least 0.4 in. of direct runoff could reduce problems
and calibration of the selected storms with HEC-1
resulting from nonuniform spatial distributions of
(U.S. Army Corps of Engineers, 1990). Calibration
effective precipitation (characteristic 3). Thus, only
was done to obtain optimal values of TC and R for the
storms resulting in at least 0.4 in. of direct runoff were
Clark unit hydrograph and UL for the SCS dimension-
considered in this study (with the exception of one
less unit hydrograph. Evaluation of the parameters
storm with 0.36 in. of direct runoff on Green Lake
determined in this study included consideration of the
Ditch). Hydrographs that were affected by snowmelt
fit quality of the calibrated hydrographs and compari-
were not considered in this study.
son of the parameters to the results of previously
developed relations for the estimation of parameters Determination of direct-runoff hydrographs
for synthetic unit hydrographs. requires separation of (subtraction of) base flow from
the total-runoff hydrograph. Base flow was estimated
by extending the trend in flow prior to the start of the
Storm Selection storm (linearly or with the application of the standard
recession curve described below) to the time of peak
A unit hydrograph is the discharge-time graph discharge. After the time of peak discharge the base
(hydrograph) of a unit volume of direct runoff resulting flow increased linearly to the time when the total-
from a spatially evenly distributed effective precipita- runoff hydrograph consisted of only base flow. This
tion (approximately equal to precipitation excess if time was defined as the point on a semilogarithmic plot
interflow is small) with a uniform intensity over a given of the total-runoff hydrograph (with discharge on the
duration. Thus, storms for determination of parameters logarithmic scale), where the recession limb becomes
for synthetic unit hydrographs should be selected to approximately linear as described in Chow (1964,
conform to some extent with the definition of a unit p. 14-10).
hydrograph. Ideally, Viessman and others (1989, Storms may be spaced in time such that well-
p. 186) recommend that the storms utilized to defined rises in the hydrograph with distinct peaks

Determination and Evaluation of Synthetic Unit-Hydrograph Parameters 13

result, but the second rise begins in the latter part of the discharge (qP) was determined by dividing the direct-
recession curve of the first rise. In this case, rises in the runoff peak discharge by the depth of direct runoff.
hydrograph must be separated so that the direct-runoff Determination of the time-to-peak was more difficult.
hydrographs from each storm may be evaluated. Storm For storms with a single, distinct period of effective
separation was done on the basis of a standard reces- precipitation (1 to 5 hours of more than 0.1 in/h) or with
sion curve. The standard recession curve was devel- an extended period of light (less than 0.09 in/h) varied
oped by averaging the recession curves for the storms effective precipitation, the time from beginning of
on the given watershed that were not affected by addi- effective precipitation to direct-runoff peak discharge
tional rainfall during the recession period. Typically, was taken as TP. For storms with several hours of more
the agreement among these recession curves was close. than 0.1 in. of effective precipitation separated by a few
In storm separation, the standard recession curve was hours, these multiple hours of “heavy rain” (multiple
matched to the recession curve of the first rise and uti- 5-minute periods of heavy rain were considered for
lized to extend the normal recession under the second Terre Faire Ditch and Green Lake Ditch) result in
rise. In some cases, the second rise began at discharges incomplete rises in the rising limb of the direct-runoff
above those utilized in the standard recession curve and hydrograph and (or) prolonged high discharges in the
the direct-runoff hydrographs resulting from the two vicinity of the peak discharge. Thus, the time from
storms could not be reliably separated. The computed beginning of effective precipitation to direct-runoff
depth of direct runoff, direct-runoff peak discharge, peak discharge may be considerably longer than the
and time-to-peak discharge for the direct runoff for time-to-peak for the separate periods of heavy rain. For
each storm on each watershed are listed in table 10 (at these storms, the time-to-peak corresponding to the
the end of the report) for the storms utilized to develop period of the largest effective precipitation is taken as
and verify the hydrograph-parameter estimation equa- TP. The peak factor was then computed as the product
tions. of the estimated TP and qP divided by the watershed
area. For the second type of storm, the computed peak
factor is less than the actual peak factor because the
Hydrograph-Parameter Determination by measured direct-runoff peak discharge is primarily the
Manual Computation result of a fraction of the total depth of direct runoff;
therefore, qP is underestimated. The appropriate frac-
The hydrograph-time lag is the time difference tion of the depth of direct runoff can only be assessed
between the center of mass of the direct-runoff by deconvoluting the direct-runoff hydrograph. The
hydrograph and the effective precipitation hyetograph. computed peak factors for each storm on each water-
The direct-runoff hydrograph is obtained as previously shed are listed in table 10 for the storms utilized to
described. The effective precipitation hyetograph was develop and verify the hydrograph-parameter estima-
estimated by applying a uniform loss rate to the hyeto- tion equations.
graph of basin average precipitation (determined with
Thiessen polygons). The uniform loss rate was adjusted
such that effective precipitation equalled the depth of Hydrograph-Parameter Determination by
direct runoff. The centers of mass of the direct-runoff Rainfall-Runoff Model Calibration
hydrograph and effective precipitation hyetograph
were then computed on the basis of 1-hour data for the The time of concentration and watershed-storage
watersheds larger than 5 mi2 and 5-minute data for the coefficient for the Clark (1945) unit hydrograph were
watersheds smaller than 1 mi2. The computed determined by calibrating HEC-1 (U.S. Army Corps
hydrograph-time lags for each storm on each watershed of Engineers, 1990) for hyetographs of basin average
are listed in table 10 for the storms utilized to develop precipitation (determined with Thiessen polygons) and
and verify the hydrograph-parameter estimation equa- direct-runoff hydrographs for a total of 66 storms on
tions. the nine watersheds utilized to develop and 11 storms
The peak factor (qPTP/A) for the SCS dimen- utilized to verify the hydrograph-parameter estimation
sionless unit hydrograph was approximated utilizing equations. Optimal values of the initial loss and con-
the assumption that the direct-runoff hydrograph tinuing-loss rate also were determined in HEC-1 cali-
resulted from a single period of uniform effective bration, primarily to match the depths of effective
precipitation. Therefore, the unit-hydrograph peak precipitation and direct runoff, and were not used

14 Equations for Estimating Synthetic Unit-Hydrograph Parameter Values for Small Watersheds in Lake County, Illinois
further in the development of the estimation equations. Model-fit efficiency values greater than 0.9,
The quality of the calibration was assessed on the basis generally, indicate a close match between measured
of the coefficient of model-fit efficiency (Nash and and simulated direct-runoff hydrographs. The model-
Sutcliffe, 1970): fit efficiency for 15 of the 66 storms utilized to develop
and 2 of the 11 storms utilized to verify the
n n hydrograph-parameter estimation equations was less
than 0.9. Four of the 17 storms were on Bull Creek, but
∑ ( Qmi – Qm ) 2 – ∑ ( Qmi – Qsi ) 2 model-fit efficiencies greater than 0.875 resulted for 3
=1 i–1
EFF = i------------------------------------------------------------------------------------
- , (11)
n of those storms. Three of the 17 storms were on Green
Lake Ditch. Green Lake Ditch is the smallest and most
∑ ( Qmi – Qm ) 2 impervious watershed considered, and the hydrographs
i 1
indicate a rapid response to rainfall that is difficult to
where simulate. The remainder of the storms with model-fit
Qmi is the measured direct runoff at time i, efficiencies less than 0.9 were distributed among the
Qm is the average measured direct runoff for the other watersheds. The average model-fit efficiency and
storm, percentage error in simulation of direct-runoff peak
discharge are listed in table 3 for each watershed. The
Qsi is the simulated direct runoff at time i, and
average percentage errors listed in table 3 indicate that
n is the number of simulated hydrograph ordi- under optimal (calibrated) conditions application of the
nates. Clark unit-hydrograph method results in average over-
Multiple starting points were utilized, as necessary, in estimations of the measured peak discharge from 5.9 to
the nonlinear optimization applied in HEC-1 to ensure 19.1 percent for watersheds in Lake County.
that a close match between the measured and simulated The unit-graph lag for the SCS dimensionless
direct-runoff hydrographs was obtained. The percent- unit hydrograph (Snider, 1971) was determined by cal-
age error between the measured and simulated direct- ibrating HEC-1 (U.S. Army Corps of Engineers, 1990)
runoff peak discharges was computed as a measure of for hyetographs of basin average precipitation (deter-
the reliability of applying the Clark unit-hydrograph mined with Thiessen polygons) and direct-runoff
method in Lake County. The time of concentration and hydrographs for a total of 66 storms on the nine water-
watershed-storage coefficient for the Clark unit sheds utilized to develop and 11 storms utilized to
hydrograph, the time-to-peak for the Snyder synthetic verify the hydrograph-parameter estimation equations.
unit hydrograph, the model-fit efficiency, and the per- The application of the fixed peak factor (qPTP/A = 484)
centage error in the simulated direct-runoff peak dis- in the SCS dimensionless unit-hydrograph method
charge are listed in table 11 (at the end of the report) for precludes obtaining close fits between the measured
the storms utilized to develop and verify the and simulated direct-runoff hydrographs. Therefore,
hydrograph-parameter estimation equations. the model-fit efficiency was not computed for the

Table 3. Average values of measures of calibration quality for the calibrated Clark unit-hydrograph method and the calibrated
Soil Conservation Service dimensionless unit-hydrograph method for all storms on selected watersheds in Lake County, Ill.,
utilized to develop and verify the equations for estimation of synthetic unit-hydrograph parameters
[SCS, Soil Conservation Service]
SCS dimensionless
Clark unit hydrograph unit hydrograph
Model-fit Error in simulated Error in simulated
Number of efficiency peak discharge peak discharge
Watershed storms (percent) (percent)

Bull Creek 11 0.913 19.1 120

Terre Faire Ditch 4 .933 16.8 56.3
Indian Creek 11 .954 9.5 46.2
Green Lake Ditch 4 .855 17.4 74.2
North Branch Chicago River 10 .961 5.9 71.6
Skokie River at Lake Forest 10 .938 16.4 99.3
Skokie River near Highland Park 7 .952 9.3 59.3
Squaw Creek 8 .885 15.0 75.1
Flint Creek 12 .942 7.4 38.9

Determination and Evaluation of Synthetic Unit-Hydrograph Parameters 15

 Table 4. Range and mean of unit-hydrograph peak factors of the type applied in the Soil Conservation
Service dimensionless unit-hydrograph method determined by manual hydrograph analysis for all storms
on selected watersheds in Lake County, Ill., utilized to develop and verify the equations for estimation of
synthetic unit-hydrograph parameters
Watershed Minimum Mean Maximum

Bull Creek 112 236 355

Terre Faire Ditch 306 415 494
Indian Creek 333 397 447
Green Lake Ditch 121 276 449
North Branch Chicago River 155 320 474
Skokie River at Lake Forest 189 319 411
Skokie River near Highland Park 190 363 478
Squaw Creek 143 267 422
Flint Creek 305 390 487

calibrated hydrographs for the SCS dimensionless unit- Lake Forest and the North Branch Chicago River were
hydrograph method. The unit-graph lag for the SCS considered in the development of the method of Singh.
dimensionless unit hydrograph and the percentage The results of applying the methods of Graf
error in the simulated direct-runoff peak discharge are and others (1982b) and Singh (1981) to the gaged
listed in table 12 (at the end of the report) for the storms watersheds in Lake County and the results of HEC-1
utilized to develop and verify the hydrograph-parame- (U.S. Army Corps of Engineers, 1990) calibration for
ter estimation equations. The average percentage error these watersheds are shown in tables 5-7. The value of
in simulation of direct-runoff peak discharge is listed TC, estimated with the method of Graf and others, is
for each watershed in table 3. within the range of values determined in this study for
The average percentage errors listed in table 3 Terre Faire Ditch, Indian Creek, North Branch Chicago
indicate that under optimal (calibrated) conditions River, and Squaw Creek (table 5), and for each of these
application of the SCS dimensionless unit-hydrograph watersheds the difference between the estimated and
method results in average overestimations of the mean values is greater than 40 percent. The value of R,
measured peak discharge of 38.9 to 120 percent for estimated with the method of Graf and others, is
watersheds in Lake County. These results are supported within the range of values determined in this study for
by the range and mean of the computed peak factors Bull Creek, North Branch Chicago River, Skokie River
listed in table 4. For each of the nine watersheds in at Lake Forest, and Skokie River near Highland Park
Lake County studied, the mean of the estimated peak (table 6). As expected, the agreement between the
factors is considerably less than the fixed value of 484 estimated and mean values for both Skokie River
applied in the SCS dimensionless unit-hydrograph watersheds is within 20 percent. The value of tps
method. estimated with the method of Singh is within the
range of values determined in this study for Terre Faire
Ditch, Indian Creek, North Branch Chicago River, and
Comparison of Calibrated Hydrograph- Squaw Creek (table 7). These results indicate that the
Parameter Values with Results of previous studies may yield reliable estimates for some
Previous Relations watersheds in Lake County, but that development of
relations for estimation of hydrograph parameters
For estimation of unit hydrographs for water- specifically for application in Lake County will provide
sheds in Illinois, Graf and others (1982b) developed valuable information for engineers and planners.
relations between TC and R for the Clark (1945) unit
hydrograph and watershed characteristics, and Singh
(1981) developed relations between tps of a modified EQUATIONS FOR ESTIMATING
Snyder (1938) unit hydrograph and watershed charac- SYNTHETIC UNIT-HYDROGRAPH
teristics. Data from the Skokie River streamflow gages PARAMETERS
at Lake Forest and near Highland Park were considered
in the development of the method of Graf and others. The equations for estimating four synthetic unit-
Data from the streamflow gages on the Skokie River at hydrograph parameters (time of concentration and

16 Equations for Estimating Synthetic Unit-Hydrograph Parameter Values for Small Watersheds in Lake County, Illinois
Table 5. Time of concentration for the Clark unit-hydrograph method estimated with the method of Graf and others (1982b)
compared with the minimum, mean, and maximum values determined from calibration for all storms on selected watersheds
in Lake County, Ill., utilized to develop and verify the equations for estimation of synthetic unit-hydrograph parameters
Estimated from Determined from storm calibration
Graf and others Minimum Mean Maximum
Watershed (hours) (hours) (hours) (hours)

Bull Creek 8.94 1.03 1.81 4.62

Terre Faire Ditch .30 .12 .53 1.17
Indian Creek 3.58 3.57 12.1 23.8
Green Lake Ditch 1.11 .09 .098 .11
North Branch Chicago River 11.6 1.17 6.38 18.8
Skokie River at Lake Forest 6.99 1.04 2.09 5.34
Skokie River near Highland Park 8.61 1.42 4.75 7.05
Squaw Creek 6.93 4.03 12.1 27.5
Flint Creek 5.66 13.1 30.7 57.7

Table 6. Watershed-storage coefficient for the Clark unit-hydrograph method estimated with the method of Graf and others
(1982b) compared with the minimum, mean, and maximum values determined from calibration for all storms on selected
watersheds in Lake County, Ill., utilized to develop and verify the equations for estimation of synthetic unit-hydrograph
parameters
Estimated from Determined from storm calibration
Graf and others Minimum Mean Maximum
Watershed (hours) (hours) (hours) (hours)

Bull Creek 20.9 19.6 29.0 38.1

Terre Faire Ditch .70 .93 1.47 2.67
Indian Creek 8.35 16.7 28.7 45.9
Green Lake Ditch 2.58 .83 .89 .99
North Branch Chicago River 27.2 17.7 24.3 28.7
Skokie River at Lake Forest 16.3 13.1 19.8 26.9
Skokie River near Highland Park 20.1 13.2 22.9 31.8
Squaw Creek 16.2 34.6 53.2 73.1
Flint Creek 13.2 36.0 52.1 78.3

Table 7. Time-to-peak for the Snyder unit-hydrograph method estimated with the method of Singh (1981) compared
with the minimum, mean, and maximum values determined from calibration for all storms on selected watersheds in
Lake County, Ill., utilized to develop and verify the equations for estimation of synthetic unit-hydrograph parameters
Estimated Determined from storm calibration
from Singh Minimum Mean Maximum
Watershed (hours) (hours) (hours) (hours)

Bull Creek 5.98 1.79 2.78 5.17

Terre Faire Ditch .75 .23 .54 1.10
Indian Creek 11.1 3.91 11.9 22.5
Green Lake Ditch .75 .14 .145 .15
North Branch Chicago River 9.63 1.82 6.67 17.7
Skokie River at Lake Forest 7.76 1.80 2.87 5.52
Skokie River near Highland Park 9.56 2.24 5.19 7.13
Squaw Creek 8.84 4.87 12.6 27.9
Flint Creek 11.7 13.6 29.9 54.0

watershed-storage coefficient for the Clark unit oped utilizing the results of calibration and manual
hydrograph, unit-graph lag for the SCS dimensionless computations for 66 storms on nine small watersheds in
unit hydrograph, and hydrograph-time lag) were devel- Lake County, Ill. Eleven storms on eight of the small

Equations for Estimating Synthetic Unit-Hydrograph Parameters 17

watersheds were utilized to verify the estimation equa- separate equations for estimating the hydrograph
tions. The procedure applied to develop the estimation parameters were developed with watershed area and
equations, the resulting equations, the results of equa- main channel length as the primary watershed charac-
tion verification, and the range of applicability of the teristics. The equations for estimating the synthetic
equations are presented below. unit-hydrograph parameters in hours as a function of
watershed and storm characteristics with watershed
area as the primary watershed characteristic are
Equation Development

Equations for estimating synthetic unit- TC = 39.1 A0.577(I+1)-1.146D0.781, (13)

hydrograph parameters were developed utilizing multi-
ple linear regression to relate the logarithms of
hydrograph parameters to logarithms of watershed R = 123 A0.390(I+1)-0.722S-0.303, (14)
area; main channel length and slope; percentages plus
one of impervious, forest, and wetland cover; depth, UL = 44.5 A0.483(I+1)-0.805D0.336, (15)
duration, and maximum 1-hour depth of effective pre-
cipitation; and depth, duration, and maximum 1-hour and
depth of precipitation. The multiple linear regression in
logarithms resulted in estimation equations of the form
TL = 119 A0.345(I+1)-0.690S-0.182D0.187. (16)

hpi = a W1b1 W2b2 ... S1c1 S2c2 ... , (12) where I is the percentage of impervious cover, and
D is the depth of effective precipitation in inches. The
where hpi is hydrograph parameter i, Wj are watershed equations for estimating the synthetic unit-hydrograph
characteristics j, Sk are storm characteristics k, a is a parameters in hours as a function of watershed and
coefficient, bj are exponents corresponding to water- storm characteristics with main channel length as the
shed characteristics j, and ck are exponents correspond- primary watershed characteristic are
ing to storm characteristics k. Nonlinear equations,
such as equation 12, between hydrograph parameters,
and watershed and storm characteristics have been TC = 87.5 L0.868(I+1)-1.563D0.780, (17)
determined theoretically from the kinematic wave
approximation (Ragan and Duru, 1972), experimen-
tally in the laboratory (Shen, 1974), and empirically R = 81.1 L0.759(I+1)-0.994, (18)
from field data (Snyder, 1938; Rao and others, 1972;
and others). The multiple linear regression was applied UL = 74.9 L0.777(I+1)-1.133D0.371, (19)
in a stepwise approach. Watershed and storm character-
istics were added to the regression model (eq. 12) one and
at a time, and characteristics were retained in the
regression model only if the corresponding exponents
were statistically significant (the corresponding TL = 105 L0.637(I+1)-0.930D0.214. (20)
95-percent confidence interval for the parameter did
not include zero) and the sign of the exponent was The multiple correlation coefficient and
correct from a physical viewpoint. For example, standard error in the logarithms for these equations
hydrograph-timing parameters should increase with are listed in table 8. With the exception of the Clark
increasing area, main channel length, and percentage of time of concentration, more than 90 percent of the
open water, and decrease with increasing slope and per- variance in the logarithms of each of the hydrograph
centage of impervious cover (which indicates a parameters is explained by equations 13-20. The
decrease in overland flow roughness and, typically, an hydrograph parameters estimated with equations
increase in channelized drainage). utilizing area (equations 13-16) and the hydrograph
Watershed area and main channel length are parameters determined through calibration or manual
both scale parameters and strongly correlated. Thus, computations are shown in scattergrams in figures 4-7,

18 Equations for Estimating Synthetic Unit-Hydrograph Parameter Values for Small Watersheds in Lake County, Illinois
 Table 8. Standard error and multiple correlation coefficient for logarithmic data in the equations for estimating synthetic
unit-hydrograph parameters for Lake County, Ill.
Synthetic Multiple correlation coefficient Standard error
unit- Equation Equation Equation Equation
hydrograph utilizing utilizing utilizing utilizing
parameter area length area length

Clark time
of concentration 0.873 0.845 0.735 0.808
Clark storage
coefficient .961 .957 .296 .308
Soil Conservation Service
unit-graph lag .968 .961 .272 .286
hydrograph-time lag .963 .963 .247 .242

respectively. The results for the equations utilizing comparison of the results from these equations with
main channel length (equations 17-20) are similar to the values determined from calibration indicated high
those shown in figures 4-7 and are not included here. scatter in the estimates. Therefore, the equations for
The watershed and storm characteristics estimating TC and R separately are more reliable for
included in equations 13-20 are consistent with physi- application in Lake County.
cal reasoning and the results of controlled experiments
(Shen, 1974), which supports application of these
equations on ungaged watersheds. For example, the Equation Verification
storage in the stream system upstream from a point
During the 1995 Water Year (October 1, 1994-
(streamflow gage) should be a function of the physical
September 30, 1995), 11 storms on eight of the nine
characteristics of the watershed independent of storm
watersheds, which met the criteria for storm selection
characteristics. Shen (1974) found that for laboratory
discussed previously, were identified for verification of
experiments of the rainfall-runoff process on impervi-
the equations for estimation of synthetic unit-
ous surfaces, the hydrograph-timing parameters were
hydrograph parameters. These verification storms were
mildly affected by watershed slope. Thus, it is reason-
analyzed manually and through HEC-1 calibration to
able that main channel slope is included in only one determine storm and hydrograph characteristics, and
of six equations for estimating hydrograph-timing synthetic unit-hydrograph parameters in the same
parameters. Further, inclusion of the main channel manner as the 66 storms utilized to develop the estima-
slope could improve the equations for estimating TC; tion equations. The direct-runoff depth and peak dis-
however, the exponent on slope would be positive charge, time-to-peak discharge, hydrograph-time lag,
indicating TC increases with slope. This is not reason- and peak factor for the verification storms are listed in
able from a physical viewpoint, and this result further table 10. The time of concentration, watershed-storage
indicates the relative insensitivity of hydrograph- coefficient, Snyder unit-hydrograph time-to-peak,
timing parameters to slope. Finally, Shen (1974) found model-fit efficiency, and percentage error in the peak
that hydrograph-timing parameters increased with the discharge for the HEC-1 calibration of the Clark unit
product of storm intensity and duration (storm depth). hydrograph for the verification storms are listed in table
Thus, it is reasonable that each synthetic unit- 11. The unit-graph lag and percentage error in peak dis-
hydrograph time parameter increases with the depth of charge for the HEC-1 calibration of the SCS dimen-
effective precipitation. sionless unit hydrograph for the verification storms are
Equations for estimating TC+R and R/(TC+R) as listed in table 12.
a function of watershed and storm characteristics, as Equation 13 was utilized to estimate TC values
done by Graf and others (1982b), also were evaluated. for the Clark unit-hydrograph method for the verifica-
Equations for estimating TC+R were obtained with tion storms as a function of the watershed area, percent-
multiple correlation coefficients greater than 0.96. age of impervious cover, and depth of effective
However, equations for estimating R/(TC+R) had precipitation. The estimated TC values and the
multiple correlation coefficients less than 0.7, and measured TC values for the verification storms are

Equations for Estimating Synthetic Unit-Hydrograph Parameters 19

shown in figure 4. Equation 14 was utilized to estimate accurately simulating design hydrographs may be eval-
R values for the Clark unit-hydrograph method for the uated by comparison of hydrographs simulated utiliz-
verification storms as a function of the watershed area, ing the hyetograph of effective precipitation
percentage of impervious cover, and main channel determined in hydrograph time-lag computations and
slope. The estimated R values and the measured R TC and R estimated with equations 13 and 14 or 17 and
values for the verification storms are shown in figure 5. 18, respectively. This comparison provides a more
Equation 15 was utilized to estimate UL values for the stringent test of the utility of the estimation equations
SCS dimensionless unit hydrograph for the verification than the comparison of measured and estimated TC and
storms as a function of the watershed area, percentage R values in figures 4 and 5. The percentage errors in the
of impervious cover, and depth of effective precipita- estimated peak discharge and time-to-peak discharge
tion. The estimated UL values and the measured UL val- for the verification storms simulated with the Clark
ues for the verification storms are shown in figure 6. unit-hydrograph method utilizing TC and R estimated
Equation 16 was utilized to estimate TL values for the with equations 13 and 14 or 17 and 18, respectively, are
verification storms as a function of the watershed area, listed in table 9. For 8 of the 11 verification storms,
percentage of impervious cover, main channel slope, the error in the peak discharge is less than 25 percent
and depth of effective precipitation. The estimated TL with either set of estimation equations. For 10 of the
values and the measured TL values for the verification 11 verification storms, the error in the time-of-peak
storms are shown in figure 7. The synthetic unit- discharge is less than 20 percent with either set of esti-
hydrograph-parameter values estimated for the verifi- mation equations. The average percentage error in the
cation storms with equations 13-16 are scattered peak discharge and time-to-peak is less than 10 percent
around the line of perfect agreement between estimated with either set of estimation equations.
and measured values (1:1 line on figs. 4-7) in a similar The simulated and measured hydrographs are
manner as the synthetic unit-hydrograph parameter val- shown in figures 8-18 (at the end of the report). The
ues estimated for the storms utilized in the develop- agreement between the measured and simulated hydro-
ment of the equations. Therefore, the accuracy of graphs is excellent for the storm of April 26, 1995, on
equations 13-16 when estimating synthetic unit- Indian Creek (fig. 9), the North Branch Chicago River
hydrograph parameters for independent storms is simi- (fig. 12), and Flint Creek (fig. 17); and the storm of
lar to the fit accuracy of equations 13-16 listed in table November 5, 1994, on Green Lake Ditch (fig. 10). The
8. The verification results for the equations utilizing agreement between the measured and simulated hydro-
main channel length (equations 17-20) are similar to graphs is good for the storm of November 27, 1994, on
those shown in figures 4-7 and are not included here. Green Lake Ditch (fig. 11), the storm of May 23, 1995,
The time of concentration tends to be underesti- on the North Branch Chicago River (fig. 13), and the
mated with equations 13 (fig. 4) or 17. The effects storm of April 26, 1995, on Squaw Creek (fig. 16).
of the underestimation of TC on the ultimate goal of The agreement between the measured and simulated

Table 9. Percentage error in the estimated peak discharge and time-to-peak discharge for the verification storms on selected
watersheds in Lake County, Ill., simulated with the Clark unit-hydrograph method utilizing estimated values of time of
concentration and watershed-storage coefficient
Error in peak discharge Error in time-to-peak
Equations Equations Equations Equations
Storm based on based on based on based on
Watershed date length area length area

Bull Creek 4/26/95 52.1 41.5 16.0 12.0

Indian Creek 4/26/95 .17 3.25 -13.2 -5.26
Green Lake Ditch 11/05/94 -4.16 .38 .00 .00
Green Lake Ditch 11/27/94 -27.9 -23.7 .00 .00
North Branch Chicago River 4/26/95 -2.94 -8.46 7.14 7.14
North Branch Chicago River 5/23/95 4.51 -2.26 -20.0 -20.0
Skokie River at Lake Forest 4/26/95 47.7 45.7 -6.67 -6.67
Skokie River near Highland Park 4/26/95 22.8 28.0 -11.1 -11.1
Squaw Creek 4/26/95 22.3 16.5 -17.1 -19.5
Flint Creek 4/26/95 -12.2 -3.54 -9.62 -9.62
Flint Creek 5/23/95 -12.0 -2.29 -39.5 -39.5

20 Equations for Estimating Synthetic Unit-Hydrograph Parameter Values for Small Watersheds in Lake County, Illinois
 60
MEASURED TIME OF CONCENTRATION (TC ), IN HOURS

55

50
Line of perfect agreement

45

40

35

30

25

20

15

10

Calibration storms
5
Verification storms

0
0 5 10 15 20 25 30 35 40 45 50 55 60

COMPUTED TIME OF CONCENTRATION (TC ), IN HOURS

Figure 4. Time of concentration for storms on nine watersheds in Lake County, Ill., for the Clark unit-hydrograph method
measured and computed as a function of watershed area, percentage of impervious cover, and depth of effective precipitation.

Equations for Estimating Synthetic Unit-Hydrograph Parameters 21

 60
MEASURED WATERSHED STORAGE COEFFICIENT (R), IN HOURS

55

50

45

40

35

30 Line of perfect agreement

25

20

15

10

5 Calibration storms
Verification storms

0
0 5 10 15 20 25 30 35 40 45 50 55 60

COMPUTED WATERSHED STORAGE COEFFICIENT (L R), IN HOURS

Figure 5. Watershed-storage coefficient for storms on nine watersheds in Lake County, Ill., for the Clark unit hydrograph
measured and computed as a function of watershed area, percentage of impervious cover, and main channel slope.

22 Equations for Estimating Synthetic Unit-Hydrograph Parameter Values for Small Watersheds in Lake County, Illinois
 60

55
MEASURED UNIT-GRAPH LAG (UL ), IN HOURS

50

45

40

Line of perfect agreement

35

30

25

20

15

10

5 Calibration storms
Verification storms
0
0 5 10 15 20 25 30 35 40 45 50 55 60

COMPUTED UNIT-GRAPH LAG (U L ), IN HOURS

Figure 6. Unit-graph lag for storms on nine watersheds in Lake County, Ill., for the Soil Conservation Service dimensionless
unit hydrograph measured and computed as a function of watershed area, percentage of impervious cover, and depth of
effective precipitation.

Equations for Estimating Synthetic Unit-Hydrograph Parameters 23

 60
MEASURED HYDROGRAPH TIME LAG (T L ), IN HOURS

55

50

45

40

35

30

25
Line of perfect agreement

20

15

10

5 Calibration storms
Verification storms

0
0 5 10 15 20 25 30 35 40 45 50 55 60

COMPUTED HYDROGRAPH TIME LAG (T L ), IN HOURS

Figure 7. Hydrograph-time lag for storms on nine watersheds in Lake County, Ill., measured and computed as a function of
watershed area, percentage of impervious cover, main channel slope, and depth of effective precipitation.

24 Equations for Estimating Synthetic Unit-Hydrograph Parameter Values for Small Watersheds in Lake County, Illinois
hydrographs is poor for the April 26, 1995, storm on runoff response to rainfall on this watershed is very
the Skokie River near Highland Park, but this may be rapid, and it is difficult to separate the runoff peaks
the result of an unusually shaped measured hydrograph resulting from closely spaced rainfall bursts. A reliable
near the peak discharge (fig. 15). Thus, only 3 of the 11 calibration or determination of the synthetic unit-
verification storms—storm of April 26, 1995, on Bull hydrograph parameters for the June 23, 1994, storm on
Creek (fig. 8) and the Skokie River at Lake Forest this watershed was not possible. Thus, no independent
(fig. 14), and the storm of May 23, 1995, on Flint Creek verification of the estimation equations was done.
(fig. 18)—were poorly simulated utilizing the estima- The objective of developing the estimation equa-
tion equations. tions is reliable simulation of design hydrographs for
small (less than 25 mi2), urban watersheds in Lake
County. The range of values for area (0.06-37 mi2),
Application Limits for the Estimation main channel length (0.33-16.6 mi), and main channel
Equations slope (3.13-55.3 ft/mi) for the watersheds utilized to
develop the estimation equations is representative of
The verification results presented in table 9 and
the conditions for small watersheds in Lake County.
figures 8-18 indicate that application of the estimation
However, the range of the percentage of impervious
equations may result in very accurate simulation of
cover (7.32-40.6 percent) may not include the full
hydrographs from actual storms without calibration.
range of conditions of interest for small watersheds in
Thus, application of the estimation equations to deter-
Lake County. For example, it may be necessary to com-
mine synthetic unit-hydrograph parameters for design-
pare peak discharges before and after development, and
storm simulation may result in reliable design hydro-
the conditions before development may include low
graphs. The application limits for the estimation equa-
percentages of impervious cover. In such cases, the
tions must be considered for proper utilization of the
computed hydrographs should be checked carefully to
estimation equations. These application limits are dis-
determine if the results are reasonable.
cussed below.
In the development of the estimation equations The example of the Southwest Fork of the South
by stepwise multiple linear regression, the watershed Branch of Ravine 10 illustrates the limitation of the
characteristics (area, length, percentage of impervious application of the estimation equations with respect to
cover, and, to a lesser extent, main channel slope) were the percentage of impervious cover. This watershed
found to be the primary predictor variables for the syn- includes 48 percent impervious cover, whereas the esti-
thetic unit-hydrograph parameters. The depth of effec- mation equations were developed for watersheds with
tive precipitation provided a small adjustment factor percentages of impervious cover from 7.32 to 40.6 per-
among storms on the same watershed. Thus, the good cent. The rainfall and runoff data from the Southwest
verification results for the simulated hydrographs Fork of the South Branch of Ravine 10 indicate that
presented in table 9 and figures 8-18 were expected unit hydrographs may be difficult to derive and apply
because the verification storms were measured on the on watersheds with high percentages of impervious
watersheds utilized to develop the estimation equa- cover. Thus, if the estimation equations are applied to
tions. Rainfall and streamflow gages were installed on areas with greater than 40.6 percent impervious cover,
December 7, 1993, and April 6, 1994, respectively, on the computed hydrographs should be checked carefully
the Southwest Fork of the South Branch of Ravine 10 to determine if the results are reasonable.
at Highland Park, Ill., to provide data for independent The maximum measured depth of effective pre-
verification of the estimation equations. During the cipitation for the storms utilized to develop the estima-
period from April 6, 1994, to September 30, 1995, tion equations was 2.16 in. from Terre Faire Ditch
only the storm of June 23, 1994, resulted in runoff that for the storm of June 30, 1993. Computed depths of
met the criteria for unit-hydrograph derivation in this effective precipitation for long-duration (greater than
study. The Southwest Fork of the South Branch of 6 hours), high-return period (50- or 100-year) storms
Ravine 10 drains 0.218 mi2 with 48 percent impervious on areas with high percentages of impervious cover
cover near downtown Highland Park, Ill. The final may be more than twice the maximum measured value
drainage channel in the watershed is a 54-in. storm utilized in the development of the estimation equations.
sewer. Because of the small size, high percentage of Thus, application of the estimation equations could
impervious cover, and the storm sewer drainage, the result in estimated hydrograph-timing parameters, TC,

Equations for Estimating Synthetic Unit-Hydrograph Parameters 25

UL, and TL substantially higher than observed. For R = 123 (1.03)0.390 (20 + 1)-0.722 (38.34)-0.303 =
example, 5 in. of effective precipitation results in TC, 4.58 hours;
UL, and TL values of 1.93, 1.33, and 1.17, respectively,
times the values for 2.16 in. of effective precipitation and estimated with equations 17 and 18, respectively,
(applying equations 13, 15, and 16, respectively). as
These large increases in the hydrograph-timing param-
eters may be physically justified. For large floods, sub-
stantial overbank flow and subsequent flood-wave TC = 87.5 (2.87)0.868 (20 + 1)-1.563 (3.46)0.780 =
attenuation result. Thus, large values of TC, UL, and TL 4.94 hours, and
may reflect flood-wave attenuation. Nevertheless, the
reasonableness of TC, UL, and TL values estimated for R = 81.1 (2.87)0.759 (20 + 1)-0.994 = 8.76 hours.
design storms with large values of precipitation excess
should be carefully considered before these values are These values of TC and R could then be input to
utilized for design-hydrograph computation. HEC-1 (U.S. Army Corps of Engineers, 1990) along
with the design hyetograph, SCS curve number, and
watershed characteristics to compute the design
Application Example hydrograph. The design hydrographs obtained using
the area-based and length-based equations should be
The Buffalo Creek at Lake Zurich, Ill., watershed compared, and the most reasonable equation applied.
forms part of the southern boundary of the Indian The differences in TC and R estimated with the area-
Creek watershed and the eastern boundary of the Flint based and length-based equations appear substantial;
Creek watershed (fig. 1). This watershed drains however, in simulation of multiperiod storms the differ-
1.03 mi2 with a main channel length and slope of ences in the final computed hydrographs may be small
2.87 mi and 38.34 ft/mi, respectively. The dominant as illustrated in figures 8-18.
soils in this watershed are Markham silt loam, Beecher The value of the unit-graph lag for the SCS
silt loam, and Morley silt loam, which are SCS hydro- dimensionless unit hydrograph for this design storm on
logic-soil type C (Soil Conservation Service, 1969). the Buffalo Creek at Lake Zurich, Ill., watershed is esti-
The land use in this watershed is assumed (for this mated with equation 15 as
example) to be 1 acre residential lots with 20 percent
impervious cover and an SCS curve number of 79 (Soil
UL = 44.5 (1.03)0.483 (20 + 1)-0.805 (3.46)0.336 =
Conservation Service, 1986). To illustrate the applica-
tion of the estimation equations, it is assumed that the 5.91 hours,
50-year, 24-hour storm is the critical storm for hydro-
logic design for this watershed, and the synthetic unit- and estimated with equation 19 as
hydrograph parameters are sought to simulate the
design hydrograph for this watershed. On the basis of UL = 74.9 (2.87)0.777 (20 + 1)-1.133 (3.46)0.371 =
the isohyetal maps in Bulletin 70 (Huff and Angel, 8.55 hours.
1989, p. 57), the depth of the 50-year, 24-hour storm for
this watershed is 5.75 in. Applying the SCS curve num- These values of UL could then be input to HEC-1 or
ber, the depth of effective precipitation for the TR20 (Soil Conservation Service, 1982) with the
50-year, 24-hour storm is 3.46 in. Thus, the values of design hyetograph, SCS curve number, and watershed
the time of concentration and storage coefficient for the characteristics to compute the design hydrograph.
Clark unit-hydrograph method for this design storm on Again, the design hydrographs obtained using the area-
the Buffalo Creek at Lake Zurich, Ill., watershed are based and length-based equations should be compared,
estimated with equations 13 and 14, respectively, as and the most reasonable equation applied. The design
hydrograph obtained on the basis of the SCS dimen-
sionless unit-hydrograph method should be expected
TC = 39.1 (1.03)0.577 (20 + 1)-1.144 (3.46)0.781 = to result in substantial overestimates of the peak
3.22 hours, and discharge similar to the values reported in table 3.

26 Equations for Estimating Synthetic Unit-Hydrograph Parameter Values for Small Watersheds in Lake County, Illinois
 The value of the hydrograph-time lag for this utilized to verify (test) the estimation equations. The
design storm on the Buffalo Creek at Lake Zurich, Ill., synthetic unit-hydrograph parameters were determined
watershed is estimated with equation 16 as by calibration applying the U.S. Army Corps of Engi-
neers (1990) HEC-1 flood hydrograph package (TC, R,
and UL) or by manual analysis of the rainfall and runoff
TL = 119 (1.03)0.345 (20 + 1)-0.690 (38.34)-0.182 data (TL). Relations between synthetic unit-hydrograph
(3.46)0.187 = 9.56 hours, parameters and watershed and storm characteristics
were determined by multiple linear regression of the
and estimated with equation 20 as logarithms of the parameters and characteristics.
The watershed characteristics considered
included area; length and slope of the main channel;
TL = 105 (2.87)0.637 (20 + 1)-0.930 (3.46)0.214
and percentages of impervious, forest, and wetland
= 15.8 hours. cover. The storm characteristics considered were dura-
tion, depth, and maximum 1-hour depth for total and
Again, the analyst should select the most reasonable effective precipitation. Area and main channel length
value of TL between these estimates for further hydro- are watershed-scale parameters and highly correlated.
logic analysis of Buffalo Creek at Lake Zurich, Ill. Thus, separate sets of equations were developed with
area and main channel length as the starting parame-
ters. Percentage of impervious cover, main channel
SUMMARY AND CONCLUSIONS slope, and depth of effective precipitation also were
identified as important characteristics for the estima-
Design hydrographs computed from design
tion of synthetic unit-hydrograph parameters. The
storms, simple models of abstractions (interception,
estimation equations utilizing area had multiple corre-
depression storage, and infiltration), and synthetic unit
lation coefficients of 0.873, 0.961, 0.968, and 0.963 for
hydrographs provide vital information for stormwater,
TC, R, UL, and TL, respectively. The estimation equa-
flood-plain, and water-resources management through-
tions utilizing main channel length had multiple corre-
out the United States. The U.S. Geological Survey, in
lation coefficients of 0.845, 0.957, 0.961, and 0.963 for
cooperation with the Lake County Stormwater Man-
TC, R, UL, and TL, respectively.
agement Commission, studied rainfall and runoff data
for small watersheds in Lake County collected between Simulation of the measured hydrographs for
1990 and 1995 to develop equations for estimation of the verification storms utilizing TC and R, obtained
synthetic unit-hydrograph parameters on the basis of from the estimation equations, yielded good results
watershed and storm characteristics. The synthetic without calibration. The peak discharge for 8 of the
unit-hydrograph parameters of interest were the time of 11 storms was estimated within 25 percent, and the
concentration (TC) and watershed-storage coefficient time-to-peak discharge for 10 of the 11 storms was
(R) for the Clark unit-hydrograph method, the unit- estimated within 20 percent. Thus, application of
graph lag (UL) for the Soil Conservation Service the estimation equations to determine synthetic unit-
dimensionless unit hydrograph, and the hydrograph- hydrograph parameters for design-storm simulation
time lag (TL) for the linear-reservoir method for unit- may result in reliable design hydrographs as long as the
hydrograph estimation. Data from nine small water- physical characteristics of the watersheds under con-
sheds ranging in area from 0.06 to 37 mi2 were utilized sideration are within the range of those for the water-
in this study including Bull Creek near Libertyville, sheds considered in this study (area: 0.06-37 mi2, main
Terre Faire Ditch at Libertyville, Indian Creek at Praire channel length: 0.33-16.6 mi, main channel slope:
View, Green Lake Ditch at Buffalo Grove, North 3.13-55.3 ft/mi, and percentage of impervious cover:
Branch Chicago River at Deerfield, Skokie River at 7.32-40.6 percent). The estimation equations are most
Lake Forest, Skokie River near Highland Park, Squaw reliable when applied to watersheds with areas less
Creek at Round Lake, and Flint Creek near Fox River than 25 mi2. In the applications of the estimation equa-
Grove. Data from 66 storms with effective-precipita- tions, the percentage of impervious cover is most likely
tion depths greater than 0.4 in. on these watersheds to be outside of the range of conditions in the study
were utilized to develop the estimation equations and watersheds, and in such cases the computed synthetic
data from 11 storms on 8 of these watersheds were unit-hydrograph parameters and design hydrographs

Summary and Conclusions 27

must be checked carefully to determine if the results are Kuichling, E., 1889, The relation between the rainfall and the
reasonable. discharge of sewers in populous districts: Transactions,
American Society of Civil Engineers, vol. 20, p. 1-60.
Laurenson, E.M., and Mein, R.G., 1985, RORB-Version 3
runoff routing program user manual: Department of
REFERENCES CITED Civil Engineering, Monash University, Clayton,
Chow, V.-T., ed., 1964, Runoff, section 14: Handbook of Victoria, Australia, 155 p.
applied hydrology, New York, McGraw-Hill Book Leighton, M.M., Ekblaw, G.E., and Horberg, L., 1948, Phys-
Company, variable pagination. iographic divisions of Illinois, Illinois State Geologic
Chow, V.-T., Maidment, D.R., and Mays, L.W., 1988, Survey Report of Investigations 129, Champaign, Ill.,
Applied hydrology: New York, McGraw-Hill Book 33 p.
Company, 572 p. Mitchell, W.D., 1954, Floods in Illinois, magnitude and
Clark, C.O., 1945, Storage and the unit hydrograph: Trans- frequency: Illinois Department of Public Works and
actions, American Society of Civil Engineers, vol. 110, Buildings, Division of Waterways, 386 p.
p. 1419-1488. Nash, J.E. and Sutcliffe, J.V., 1970, River flow forecasting
Duncker, J.D., Vail, T.J., and Melching, C.S., 1995, Regional through conceptual models, Part 1--A discussion of
rainfall-runoff relations for simulation of streamflow principles: Journal of Hydrology, v. 10, p. 282-290.
for watersheds in Lake County, Illinois: U.S. Geologi- O’Kelly, J.J., 1955, The employment of unit hydrographs to
cal Survey Water-Resources Investigations Report determine the flows of Irish arterial drainage channels:
95-4023, 71 p. Proceedings, Institution of Civil Engineers (Ireland),
Ford, D.T., Morris, E.C., and Feldman, A.D., 1980, Corps of vol. 4, no. 3, p. 365-412.
Engineers’ experience with automatic calibration of Ragan, R.M., and Duru, J.O., 1972, Kinematic wave nomo-
precipitation-runoff model: Water and Related Land graph for times of concentration: Journal of the Hydrau-
Resource Systems, Y. Haimes and J. Kindler, eds., New lics Division, American Society of Civil Engineers,
York, Pergamon Press, New York, p. 467-476. vol. 98, no. HY10, p. 1765-1771.
Franz, D.D., and Melching, C.S., in press, Full equations Rao, A.R., Delleur, J.W., and Sarma, B.S.P., 1972, Concep-
(FEQ) model for the solution of the full, dynamic tual hydrologic models for urbanizing basins: Journal
equations of motion for one-dimensional unsteady of the Hydraulics Division, American Society of Civil
flow in open channels and through control structures: Engineers, vol. 98, no. HY7, p. 1205-1220.
U.S. Geological Survey Water-Resources Investiga- Shen, Y.-Y., 1974, Experimental investigation of the effects
tions Report 96-4240 of rainstorm and watershed characteristics on surface
Graf, J.B., Garklavs, G., and Oberg, K.A., 1982a, Time of runoff: Ph.D. Thesis, Department of Civil Engineering,
concentration and storage coefficient values for Illinois University of Illinois at Urbana-Champaign, Urbana,
streams: U.S. Geological Suvey Water-Resources Ill., 197 p.
Investigations Report 82-13, 35 p. Singh, K.P., 1981, Derivation and regionalization of unit
Graf, J.B., Garklavs, G., and Oberg, K.A., 1982b, A tech- hydrograph parameters for Illinois (dam safety pro-
nique for estimating time of concentration and storage gram): Illinois State Water Survey Contract Report 258,
coefficient values for Illinois streams: U.S. Geological Champaign, Ill., 84 p.
Survey Water-Resources Investigations Report 82-22, Snider, D., 1971, Hydrographs, Chapter 16: Soil Conserva-
10 p. tion Service National Engineering Handbook,
Hershfield, D.M., 1961, Rainfall frequency atlas of the U.S. Department of Agriculture, Washington, D.C.,
United States for durations from 30 minutes to 24 hours variable pagination.
and return periods from 1 to 100 years: Weather Bureau Snyder, F.F., 1938, Synthetic unit-graphs: Transactions,
Technical Paper 40, U.S. Department of Commerce, American Geophysical Union, vol. 19, p. 447-454.
Washington, D.C., 115 p. Soil Conservation Service, 1969, General soil map, Lake
Huff, F.A., and Angel, J.R., 1989, Frequency distributions County, Illinois: Unpublished map on file in the
and hydroclimatic characteristics of heavy rainstorms Grayslake office of the Natural Resources Conservation
in Illinois: Illinois State Water Survey Bulletin 70, Service, scale 1:190,080.
Champaign, Ill., 177 p. Soil Conservation Service, 1982, TR-20 computer program
Johanson, R.C., Imhoff, J.C., Kittle, J.L., Jr., and for project formulation hydrology: U.S. Department of
Donigan, A.S., Jr., 1984, Hydrological simulation Agriculture, Washington, D.C., variable pagination.
program - FORTRAN (HSPF), user’s manual for Soil Conservation Service, 1985, National engineering hand-
release 9.0: U.S. Environmental Protection Agency book, Section 4. hydrology: U.S. Department of Agri-
Report EPA-600/3-84-066, 767 p. culture, Washington, D.C., variable pagination.

28 Equations for Estimating Synthetic Unit-Hydrograph Parameter Values for Small Watersheds in Lake County, Illinois
Soil Conservation Service, 1986, Urban hydrology for small Viessman, W., Jr., Lewis, G.L., and Knapp, J.W., 1989, Intro-
watersheds: Soil Conservation Service Technical duction of hydrology, third edition: New York, Harper
Release 55, U.S. Department of Agriculture, Washing- and Row, Publishers, 780 p.
ton, D.C., variable pagination. Woodward, D.E., Merkel, W., and Sheridan, J., 1995, NRCS
Turner, H.M. and Burdoin, A.J., 1941, The flood unit hydrographs background and future: Water
hydrograph: Journal of the Boston Society of Civil Resources Engineering, W.H. Espey, Jr. and
Engineers, vol. 28, no. 3, p. 232-281. P.G. Combs, eds., American Society of Civil Engineers,
U.S. Army Corps of Engineers, 1990, HEC-1 flood New York, v. 2, p. 1693-1697.
hydrograph package, user’s manual: U.S. Army Corps
of Engineers Hydrologic Engineering Center,
Davis, Calif., variable pagination.

Summary and Conclusions 29

TABLES 10-12
FIGURES 8-18
Table 10. Direct-runoff hydrograph characteristics and parameters determined from manual analysis of hydrographs on
watersheds in Lake County, Ill., for storms utilized to develop and verify the equations for estimation of synthetic unit-
hydrograph parameters
[in., inches; ft3/s, cubic feet per second; qP, unit-hydrograph peak discharge in cubic feet per second per inch of direct runoff; TP, time to peak in hours; A,
watershed area in square miles; *, indicates that the time-to-peak corresponding to the period of the largest effective precipitation is taken as TP]

Direct-
Direct- runoff
runoff peak Time-to-peak Hydrograph Peak
Storm depth discharge discharge time lag factor
Watershed date (in.) (ft3/s) (hours) (hours) (qPTP/A)

Bull Creek 8/17/90 0.66 83.1 15 20.0 300

11/27/90 .66 80.7 *13 28.7 252
3/27/91 .54 76.2 11 15.5 246
4/08/91 .53 53.9 *17 26.0 274
4/14/91 .59 66.5 11 26.7 197
5/25/91 .94 82.7 8 28.4 112
3/22/93 .55 72.3 17 33.3 355
3/31/93 1.08 72.3 17 38.9 181
7/08/93 .88 79.9 8 36.2 115
7/18/93 .40 61.8 9 24.8 221
14/26/95
1.11 94.2 25 30.1 342
Terre Faire Ditch 5/25/91 .72 9.7 *1.75 2.68 306
6/30/93 2.16 31.0 *2.00 2.21 373
7/08/93 1.14 24.4 *1.75 2.00 486
7/18/93 1.23 25.6 *1.83 2.35 494

Indian Creek 5/09/90 .53 364 23 18.4 442

11/04/90 .42 217 *25 46.2 362
4/08/91 .62 371 25 28.6 419
4/14/91 .56 420 21 23.8 441
5/25/91 .78 612 16 26.4 352
3/22/93 .71 391 29 35.3 447
3/31/93 1.13 517 26 42.1 333
4/14/93 .52 325 *22 27.9 385
6/08/93 .48 152 50 60.4 444
7/18/93 .46 218 30 31.4 398
14/26/95 1.00 584 *21 31.1 344

Green Lake Ditch 5/25/91 .36 7.3 1.33 2.07 449

6/30/93 .82 19.2 .75 1.97 293
111/05/ .63 9.1 *1.00 5.94 242
94
111/27/
.58 8.4 *.50 3.32 121
94

North Branch Chicago River 5/08/90 1.35 435 *11 32.6 180
8/17/90 .59 551 *10 19.8 474
11/27/90 1.32 449 *9 30.1 155
3/27/91 .61 195 19 30.4 308
4/08/91 1.09 395 *14 28.0 257
4/14/91 .86 363 22 24.2 471
6/18/93 .40 221 12 22.4 337
6/30/93 .67 271 10 24.5 205
14/26/95 1.69 544 *29 28.8 474
15/23/95 .40 133 20 27.6 338

Skokie River at Lake Forest 8/17/90 .78 270 12 17.2 320

11/05/90 .53 149 *15 18.6 324
11/27/90 .84 273 *14 18.5 350
3/27/91 .45 168 11 17.6 316
4/08/91 .52 186 11 13.0 303
4/14/91 .54 240 11 15.4 376
3/31/93 1.14 271 19 28.1 347
4/14/93 .78 180 14 26.0 249
7/08/93 .71 174 *10 27.5 189
14/26/95
1.18 300 *21 21.9 411

Table 10. 33
Table 10. Direct-runoff hydrograph characteristics and parameters determined from manual analysis of hydrographs on
watersheds in Lake County, Ill., for storms utilized to develop and verify the equations for estimation of synthetic unit-
hydrograph parameters—Continued
Direct-
Direct- runoff
runoff peak Time-to-peak Hydrograph Peak
Storm depth discharge discharge time lag factor
Watershed date (in.) (ft3/s) (hours) (hours) (qPTP/A)

Skokie River near Highland Park 11/05/90 0.70 335 *17 21.4 386
11/27/90 .94 465 *18 18.3 422
11/01/92 .41 224 *13 14.1 337
4/19/93 .40 175 15 23.7 311
6/30/93 .77 308 10 22.0 190
7/08/93 .41 151 *24 26.7 419
1
4/26/95 1.35 504 *27 25.4 478

Squaw Creek 4/08/91 .67 82.3 20 71.1 143

4/14/91 .83 97.7 *34 38.1 233
3/22/93 .72 110 33 57.0 293
3/31/93 .55 114 35 33.6 422
4/14/93 .70 87.7 33 58.3 240
4/19/93 .51 126 25 34.0 359
7/08/93 .44 76.7 22 47.8 223
14/26/95 .78 121 *25 42.1 225

Flint Creek 5/09/90 .57 182 41 56.9 354

8/19/90 1.84 408 56 77.9 336
11/27/90 .56 198 39 41.0 373
3/27/91 .56 162 55 62.0 429
4/14/91 .67 235 40 42.1 379
5/25/91 .41 157 41 46.9 424
4/08/93 .51 134 43 62.2 305
4/14/93 .66 184 42 64.4 316
4/19/93 .40 168 37 42.9 420
6/07/93 .51 130 64 69.4 441
14/26/95 1.14 395 52 57.7 487
15/23/95 .43 175 38 41.8 418

1
Storm utilized to verify the relations for estimating the synthetic unit-hydrograph parameters.

34 Equations for Estimating Synthetic Unit-Hydrograph Parameter Values for Small Watersheds in Lake County, Illinois
Table 11. Parameters for the Clark and Snyder unit-hydrograph methods determined from calibration of the U.S. Army Corps
of Engineers (1990) Flood Hydrograph Package HEC-1 and measures of calibration quality for data from watersheds in Lake
County, Ill., for storms utilized to develop and verify the equations for estimation of synthetic unit-hydrograph parameters
Clark unit hydrograph Snyder unit
Time of Watershed hydrograph Error
concen- storage time-to-peak Model- in peak
Storm tration coefficient discharge fit discharge
Watershed date (hours) (hours) (hours) efficiency (percent)

Bull Creek 8/17/90 1.50 24.0 2.68 0.930 16.9

11/27/90 1.03 24.1 1.79 .914 51.2
3/27/91 1.54 23.4 2.70 .897 22.4
4/08/91 1.03 32.3 1.81 .891 5.5
4/14/91 1.35 24.7 2.54 .879 28.8
5/25/91 2.78 34.2 3.75 .850 39.9
3/22/93 1.86 19.6 2.75 .957 22.7
3/31/93 1.03 38.1 1.81 .920 5.5
7/08/93 4.62 36.1 5.17 .914 16.3
7/18/93 1.56 27.0 2.73 .956 -8.0
1
4/26/95 1.62 35.9 2.79 .932 8.5
Terre Faire Ditch 5/25/91 .12 2.67 .23 .868 20.0
6/30/93 1.17 1.23 1.10 .960 9.7
7/08/93 .20 .93 .24 .968 20.8
7/18/93 .64 1.04 .60 .934 16.7

Indian Creek 5/09/90 3.57 16.7 3.91 .973 .6

11/04/90 9.05 42.0 9.35 .940 29.0
4/08/91 23.8 24.7 22.5 .926 4.3
4/14/91 4.77 22.5 5.08 .973 14.5
5/25/91 16.2 18.7 15.4 .977 8.3
3/22/93 19.7 24.1 18.7 .980 3.6
3/31/93 12.8 34.3 12.8 .985 3.5
4/14/93 17.2 23.5 16.5 .957 -2.2
6/08/93 3.66 45.9 4.66 .953 17.1
7/18/93 10.8 34.4 11.0 .853 18.8
14/26/95 11.1 28.5 11.1 .980 6.7

Green Lake Ditch 5/25/91 .10 .88 .15 .888 14.3

6/30/93 .09 .83 .14 .810 31.6
111/05/ .11 .99 .15 .820 11.1
94
111/27/ .09 .86 .14 .900 12.5
94

North Branch Chicago River 5/08/90 4.24 27.9 4.83 .971 11.9
8/17/90 1.30 20.6 2.25 .990 6.0
11/27/90 5.90 25.2 6.14 .967 16.5
3/27/91 4.69 27.6 5.10 .950 13.8
4/08/91 7.93 25.6 8.08 .985 -1.8
4/14/91 18.8 20.4 17.7 .935 -4.4
6/18/93 1.17 17.7 1.82 .900 5.9
6/30/93 2.39 28.7 3.12 .973 5.5
1
4/26/95 4.07 26.0 4.70 .970 2.2
1
5/23/95 13.3 22.9 13.0 .970 3.8

Skokie River at Lake Forest 8/17/90 2.47 16.8 2.95 .951 17.4
11/05/90 1.57 19.0 2.66 .934 20.8
11/27/90 5.34 16.4 5.52 .920 20.5
3/27/91 1.16 16.8 1.81 .939 17.6
4/08/91 1.35 16.4 2.24 .900 2.9
4/14/91 2.37 13.1 2.84 .948 26.8
3/31/93 2.70 22.1 3.54 .975 9.9
4/14/93 1.44 26.4 2.67 .966 16.7
7/08/93 1.45 26.9 2.68 .870 19.5
1
4/26/95 1.04 24.2 1.80 .973 11.7

Table 11. 35
Table 11. Parameters for the Clark and Snyder unit-hydrograph methods determined from calibration of the U.S. Army Corps
of Engineers (1990) Flood Hydrograph Package HEC-1 and measures of calibration quality for data from watersheds in Lake
County, Ill., for storms utilized to develop and verify the equations for estimation of synthetic unit-hydrograph
parameters—Continued
Clark unit hydrograph Snyder unit
Time of Watershed hydrograph Error
concen- storage time-to-peak Model- in peak
Storm tration coefficient discharge fit discharge
Watershed date (hours) (hours) (hours) efficiency (percent)

Skokie River near Highland Park 11/05/90 7.05 20.9 7.08 0.973 -0.6
11/27/90 6.85 16.5 6.90 .960 18.3
11/01/92 1.42 13.2 2.24 .939 12.5
4/19/93 3.72 19.1 4.10 .967 11.4
6/30/93 1.58 31.8 2.76 .935 13.9
7/08/93 6.85 31.3 7.13 .955 13.2
14/26/95 5.81 27.4 6.11 .935 -3.4

Squaw Creek 4/08/91 4.03 59.8 4.87 .890 5.5

4/14/91 27.5 73.1 27.9 .897 10.2
3/22/93 8.46 53.9 9.05 .868 19.1
3/31/93 17.0 45.2 17.1 .885 9.6
4/14/93 12.2 64.9 12.6 .920 12.5
4/19/93 11.9 34.6 12.1 .920 3.2
7/08/93 9.99 39.8 10.2 .868 40.3
14/26/95
5.94 54.2 6.75 .830 19.6

Flint Creek 5/09/90 36.7 55.3 36.2 .940 4.9

8/19/90 57.7 51.8 54.0 .965 13.3
11/27/90 17.3 54.0 17.5 .870 9.6
3/27/91 30.5 55.0 30.2 .980 6.8
4/14/91 17.5 46.0 17.6 .935 12.3
5/25/91 40.8 41.5 38.6 .900 11.5
4/08/93 13.1 78.3 13.6 .952 10.4
4/14/93 27.7 68.5 27.9 .980 1.6
4/19/93 33.1 36.0 31.4 .966 5.4
6/07/93 28.8 58.0 28.7 .930 5.4
14/26/95
34.5 44.9 33.4 .980 -1.0
15/23/95
30.9 36.3 29.7 .910 8.6

1Storm
utilized to verify the relations for estimating the synthetic unit-hydrograph parameters.

36 Equations for Estimating Synthetic Unit-Hydrograph Parameter Values for Small Watersheds in Lake County, Illinois
Table 12. Unit-graph lag for the Soil Conservation Service dimensionless unit hydrograph
determined from calibration of the U.S. Army Corps of Engineers (1990) Flood Hydrograph
Package HEC-1 and percentage error in calibrated-hydrograph peak discharge for data
from watersheds in Lake County, Ill., for storms utilized to develop and verify the equations
for estimation of synthetic unit-hydrograph parameters
[SCS, Soil Conservation Service]

Unit-graph lag for Error

the SCS dimension- in peak
Storm less unit hydrograph discharge
Watershed date (hours) (percent)

Bull Creek 8/17/90 11.2 96.4

11/27/90 8.42 138
3/27/91 7.26 179
4/08/91 16.0 66.7
4/14/91 11.0 130
5/25/91 15.2 119
3/22/93 9.52 93.3
3/31/93 26.5 62.5
7/08/93 10.1 216
7/18/93 6.43 182
1
4/26/95 25.0 34.0

Terre Faire Ditch 5/25/91 .64 110

6/30/93 1.28 29.0
7/08/93 .80 70.8
7/18/93 .96 15.4

Indian Creek 5/09/90 12.3 3.9

11/04/90 23.8 92.0
4/08/91 23.3 21.6
4/14/91 10.9 98.8
5/25/91 16.2 32.5
3/22/93 21.6 25.6
3/31/93 21.2 54.0
4/14/93 19.6 21.5
6/08/93 30.8 11.2
7/18/93 18.6 94.0
14/26/95
17.9 53.6

Green Lake Ditch 5/25/91 .53 71.4

6/30/93 .39 121
111/05/
.48 66.7
94
111/27/
.54 37.5
94
North Branch Chicago River 5/08/90 13.6 82.7
8/17/90 7.21 132
11/27/90 13.8 72.0
3/27/91 13.5 100
4/08/91 14.5 58.8
4/14/91 18.6 15.4
6/18/93 7.11 76.9
6/30/93 10.1 120
14/26/95 17.9 30.7
15/23/95 19.4 27.8

Skokie River at Lake Forest 8/17/90 8.16 96.0

11/05/90 9.15 98.6
11/27/90 9.84 59.3
3/27/91 6.19 139
4/08/91 8.19 67.3
4/14/91 6.39 107
3/31/93 11.5 69.7
4/14/93 11.4 120
7/08/93 9.62 153
14/26/95
11.6 83

Table 12. 37
 Table 12. Unit-graph lag for the Soil Conservation Service dimensionless unit hydrograph
determined from calibration of the U.S. Army Corps of Engineers (1990) Flood Hydrograph
Package HEC-1 and percentage error in calibrated-hydrograph peak discharge for data
from watersheds in Lake County, Ill., for storms utilized to develop and verify the equations
for estimation of synthetic unit-hydrograph parameters—Continued

Unit-graph lag for Error

the SCS dimension- in peak
Storm less unit hydrograph discharge
Watershed date (hours) (percent)

Skokie River near Highland Park 11/04/90 11.6 58.2

11/27/90 11.5 48.0
11/01/92 8.45 29.9
4/19/93 10.9 69.7
6/30/93 12.2 101
7/08/93 15.3 78.8
1
4/26/95 17.9 29.6

Squaw Creek 4/08/91 25.5 151

4/14/91 39.7 74.5
3/22/93 28.1 85.5
3/31/93 25.8 54.4
4/14/93 37.2 67.0
4/19/93 22.9 39.7
7/08/93 24.2 92.2
14/26/95
37.7 36.4

Flint Creek 5/09/90 42.2 30.2

8/19/90 55.3 27.8
11/27/90 28.7 69.7
3/27/91 35.6 44.7
4/14/91 29.8 40.8
5/25/91 41.6 17.8
4/08/93 27.4 132
4/14/93 39.4 56.5
4/19/93 33.3 22.6
6/07/93 38.2 47.0
14/26/95
26.6 6.8
15/23/95
33.1 -29.1

1
Storm utilized to verify the relations for estimating the synthetic unit-hydrograph parameters.

38 Equations for Estimating Synthetic Unit-Hydrograph Parameter Values for Small Watersheds in Lake County, Illinois
Figure 8.

Figure 8. Measured direct-runoff hydrograph and direct-runoff hydrographs simulated with the Clark unit-hydrograph method in the U.S. Army Corps of Engineers
Flood Hydrograph Package HEC-1 with the time of concentration and storage coefficient determined from the estimation equations for Lake County, Ill., for the storm of
April 26, 1995, on Bull Creek.
39
40
Equations for Estimating Synthetic Unit-Hydrograph Parameter Values for Small Watersheds in Lake County, Illinois

Figure 9. Measured direct-runoff hydrograph and direct-runoff hydrographs simulated with the Clark unit-hydrograph method in the U.S. Army Corps of Engineers
Flood Hydrograph Package HEC-1 with the time of concentration and storage coefficient determined from the estimation equations for Lake County, Ill., for the storm of
April 26, 1995, on Indian Creek
Figure 10.

Figure 10. Measured direct-runoff hydrograph and direct-runoff hydrographs simulated with the Clark unit-hydrograph method in the U.S. Army Corps of Engineers
Flood Hydrograph Package HEC-1 with the time of concentration and storage coefficient determined from the estimation equations for Lake County, Ill., for the storm of
November 5, 1994, on Green Lake Ditch.
41
42
Equations for Estimating Synthetic Unit-Hydrograph Parameter Values for Small Watersheds in Lake County, Illinois

Figure 11. Measured direct-runoff hydrograph and direct-runoff hydrographs simulated with the Clark unit-hydrograph method in the U.S. Army Corps of Engineers
Flood Hydrograph Package HEC-1 with the time of concentration and storage coefficient determined from the estimation equations for Lake County, Ill., for the storm of
November 27, 1994, on Green Lake Ditch.
Figure 12.

Figure 12. Measured direct-runoff hydrograph and direct-runoff hydrographs simulated with the Clark unit-hydrograph method in the U.S. Army Corps of Engineers
Flood Hydrograph Package HEC-1 with the time of concentration and storage coefficient determined from the estimation equations for Lake County, Ill., for the storm of
April 26, 1995, on North Branch of the Chicago River.
43
44
Equations for Estimating Synthetic Unit-Hydrograph Parameter Values for Small Watersheds in Lake County, Illinois

Figure 13. Measured direct-runoff hydrograph and direct-runoff hydrographs simulated with the Clark unit-hydrograph method in the U.S. Army Corps of Engineers
Flood Hydrograph Package HEC-1 with the time of concentration and storage coefficient determined from the estimation equations for Lake County, Ill., for the storm of
May 23, 1995, on North Branch of the Chicago River.
Figure 14.

Figure 14. Measured direct-runoff hydrograph and direct-runoff hydrographs simulated with the Clark unit-hydrograph method in the U.S. Army Corps of Engineers
Flood Hydrograph Package HEC-1 with the time of concentration and storage coefficient determined from the estimation equations for Lake County, Ill., for the storm of
April 26, 1995, on Skokie River at Lake Forest.
45
46
Equations for Estimating Synthetic Unit-Hydrograph Parameter Values for Small Watersheds in Lake County, Illinois

Figure 15. Measured direct-runoff hydrograph and direct-runoff hydrographs simulated with the Clark unit-hydrograph method in the U.S. Army Corps of Engineers
Flood Hydrograph Package HEC-1 with the time of concentration and storage coefficient determined from the estimation equations for Lake County, Ill., for the storm of
April 26, 1995, on Skokie River at Highland Park.
Figure 16.

Figure 16. Measured direct-runoff hydrograph and direct-runoff hydrographs simulated with the Clark unit-hydrograph method in the U.S. Army Corps of Engineers
Flood Hydrograph Package HEC-1 with the time of concentration and storage coefficient determined from the estimation equations for Lake County, Ill., for the storm of
April 26, 1995, on Squaw Creek.
47
48
Equations for Estimating Synthetic Unit-Hydrograph Parameter Values for Small Watersheds in Lake County, Illinois

Figure 17. Measured direct-runoff hydrograph and direct-runoff hydrographs simulated with the Clark unit-hydrograph method in the U.S. Army Corps of Engineers
Flood Hydrograph Package HEC-1 with the time of concentration and storage coefficient determined from the estimation equations for Lake County, Ill., for the storm of
April 26, 1995, on Flint Creek.
Figure 18.

Figure 18. Measured direct-runoff hydrograph and direct-runoff hydrographs simulated with the Clark unit-hydrograph method in the U.S. Army Corps of Engineers
Flood Hydrograph Package HEC-1 with the time of concentration and storage coefficient determined from the estimation equations for Lake County, Ill., for the storm of
May 23, 1995, on Flint Creek
49