Characteristics of Ephemeral Hydrographs in the
Southwestern United States
Downloaded from ascelibrary.org by UNIV OF NEVADA-RENO on 12/17/13. Copyright ASCE. For personal use only; all rights reserved.
Rina Schumer 1; Anna Knust 2; and Douglas P. Boyle 3
Abstract: A collection of archived hydrographs for 28 ephemeral streams provided a unique opportunity to quantify hydrograph characteristics
for channels of the arid/semiarid Southwestern United States. Recurrence of flow and volume calculated using annual maxima converge with
those of the complete series by 7 years, suggesting that records of annual maxima are sufficient for estimates of extreme event recurrence. The
relationship between peak discharge and volume for a flow event cannot be tightly constrained without consideration of event duration. As a
result, volumes VðTÞ, for recurrence intervals T, cannot be estimated from corresponding peaks Qp ðTÞ. Instead, volume recurrence should be
estimated based on the joint distribution of peak discharge and duration. Although the density governing event duration in these semiarid study
regions decays as a power law with time, the authors could not identify a probability density that provides a good fit to volume magnitude.
Deterministic relationships exist between peak discharge and volume for an event of a given magnitude for the Las Vegas, Nevada; Phoenix,
Arizona; and Albuquerque, New Mexico regions. DOI: 10.1061/(ASCE)HE.1943-5584.0000643. © 2014 American Society of Civil Engineers.
Author keywords: Frequency analysis; Flow duration; Regional analysis; Arid lands; Flood volume.
Introduction
Recurrence of ephemeral flows is a unique subset of frequency
analyses because of the extreme temporal and spatial variability
of flash floods (Mason et al. 1999a, b, 2000; Reis and Crouse
2002). However, flow and volume recurrence methods prescribed
by flood control districts and other agencies rarely distinguish
between perennial and ephemeral channels. Because ephemeral
channels are often characterized by many years of zero flow and
have record lengths that are insufficient for meaningful statistical
analyses (Hjalmarson and Thomas 1992; IACWD 1982), prescribed methods for all streams are typically based on perennial
flow data. Nevertheless, ephemeral hydrograph characteristics,
including peak, duration and volume, are key variables for engineering studies aimed at mitigating flooding and land use change,
particularly in the growing urban areas of the arid and semiarid
Southwestern United States. However, long-term, high temporalresolution ephemeral flow data are rare in the southwestern United
States and the world, leading to surprisingly little work on regional
flow and volume frequency from statistics instead of modeling.
Here, a dataset is described that contains high temporal-resolution
ephemeral flow data for 28 channels from urban regions in the
Southwestern United States. These data and the analyses in this
1
Associate Research Professor, Division of Hydrologic Sciences, Desert
Research Institute, 2215 Raggio Parkway, Reno, NV 89512 (corresponding
author). E-mail: [email protected]
2
Assistant Research Hydrologist, Division of Hydrologic Sciences,
Desert Research Institute, 2215 Raggio Parkway, Reno, NV 89512. E-mail:
[email protected]
3
Associate Professor, Dept. of Geography, Univ. of Nevada, Reno,
104B Mackay Science Hall MS0154, Reno, NV 89557. E-mail:
[email protected]
Note. This manuscript was submitted on October 26, 2011; approved on
July 17, 2012; published online on August 6, 2012. Discussion period open
until June 1, 2014; separate discussions must be submitted for individual
papers. This paper is part of the Journal of Hydrologic Engineering,
Vol. 19, No. 1, January 1, 2014. © ASCE, ISSN 1084-0699/2014/1-1017/$25.00.
work can be used to ground truth rainfall-runoff models and flood
mitigation design studies.
Until recently, only annual maxima and average daily flows
were published in the U.S. Geological Survey’s (USGS) National
Water Information System (NWIS) database for gaged streams. To
compensate, the USGS has published methods for estimating peak
flow recurrence in both gaged (IACWD 1982) and ungaged (Mason
et al. 1999a, b, 2000; Reis and Crouse 2002) channels. Volume
frequency is a critical variable in sizing detention basins, estimating
dam storage (Sherif et al. 2011), and performing sediment transport
analyses in semiarid watersheds. Because streamflow databases
typically do not provide sufficient data for flash-flood volume
estimation, volume-frequency relationships for gaged streams are
typically developed for flood mitigation studies using inflow
hydrographs that are predicted by physically based or statistical
rainfall-runoff models. Rainfall-runoff models rely on design
storms to simulate precipitation when data are not available. Model
results from gaged streams are used to inform similar models of
ungaged channels in the same climatic region.
Current procedures for estimating flash flow and volume
recurrence in arid regions have great uncertainty for a variety of
reasons. First, rainfall and runoff data for an ephemeral stream
rarely exist, limiting the potential for model calibration. Second,
standard rainfall-runoff models are based on the assumption that
runoff recurrence mirrors rainfall recurrence. This assumption
may not be valid in many ephemeral streams (Farquharson et al.
1992). Furthermore, although rainfall-runoff models assume that
precipitation is spatially uniform over the entire basin, storm cells
in the arid Southwest have spatial variability at scales smaller than
basin size (Fernandez et al. 1999; Hawkins and Pole 1989). In ungaged basins, regional envelope curves relating basin area to maximum observed discharge are used to verify model results. Lack
of data generally precludes the evaluation of models for volume
estimation.
Archived streamflow records are evaluated in this study for
28 ephemeral watersheds in Nevada, Arizona, and New Mexico.
In addition to documenting hydrograph characteristics, data
with regional flow and volume envelope curves are compared,
and the following questions are addressed for streams in the
10 / JOURNAL OF HYDROLOGIC ENGINEERING © ASCE / JANUARY 2014
J. Hydrol. Eng. 2014.19:10-17.
Downloaded from ascelibrary.org by UNIV OF NEVADA-RENO on 12/17/13. Copyright ASCE. For personal use only; all rights reserved.
Fig. 1. Site locations relative to the Southwestern U.S. arid/semiarid urban study areas (see site descriptions in Table 2)
Southwestern United States: (1) Are annual maxima sufficient for
extreme event recurrence estimates for ephemeral streams? and
(2) What is the relationship between peak flow and volume for
ephemeral flow events?
Study Locations
Las Vegas, Nevada; Phoenix, Arizona; and Albuquerque, New
Mexico are located in arid and semiarid regions of the mountainous
Southwestern United States, where evapotranspiration rates typically exceed precipitation (Fig. 1). Climate differs in the three
regions, most notably for this study, in terms of precipitation.
The average annual precipitation is approximately 203, 152, and
356 mm in the Phoenix, Las Vegas, and Albuquerque regions,
respectively. Extreme variations in temperature and precipitation
occur in the study areas. The flow in many channels is the direct
result of precipitation, whose seasonal occurrence varies from
region to region (Table 1). Convective summer thunderstorms lead
to most of the flow events in Las Vegas and Albuquerque, whereas
winter storms lead to most of the moisture in Phoenix. Temperature
and snowfall are largely dependent on elevation in all study regions.
and Google Earth to identify flashy streams with fair or better records and no upstream controls, diversions, or signs of significant
changes in morphology that would affect runoff. Four sites within
Las Vegas and no sites within the Albuquerque city limits met these
criteria. The sites to research were thus extended outside of the urban areas but within flood regions defined by Thomas et al. (1997),
where flow occurrence was statistically similar to those inside the
urban areas (Fig. 1). Specifically, published average daily flow
records were used to determine the fraction of events that occur
in each season at the prospective sites, where seasons were defined
as spring (April, May, June), summer (July, August, September),
and fall/winter (October, November, December, January, February,
March). Sites were grouped geographically and the Wilcoxon
rank-sum test (Hosking and Wallis 1997) was used to determine
whether the average fraction of events occurring each season were
statistically similar to those in the corresponding urban area. This
analysis identified sites responding to climatic patterns similar to
those in the urban study areas.
Table 1. Fraction of Events Occurring in Each Season
Region
Data Collection
Candidate streams were chosen from the USGS NWIS database,
USGS Water Books, land use records, historic aerial photographs,
Phoenix
Las Vegas
Albuquerque
Spring (%)
Summer (%)
Fall/winter (%)
6
12
18
34
48
65
60
40
17
JOURNAL OF HYDROLOGIC ENGINEERING © ASCE / JANUARY 2014 / 11
J. Hydrol. Eng. 2014.19:10-17.
Downloaded from ascelibrary.org by UNIV OF NEVADA-RENO on 12/17/13. Copyright ASCE. For personal use only; all rights reserved.
All available archived strip charts were obtained from regional
USGS offices and digitized. Strip charts were not available for all
candidate streams; data were obtained for 9 Las Vegas sites, 10
Phoenix sites, and 9 Albuquerque sites (Table 2). Strip charts
contain stage hydrographs for each flow event recorded at the gage
site. Rating curves used to convert stage to discharge are not stored
within strip charts and were often unavailable. However, it was
standard practice for USGS field technicians to hand-tabulate
discretized discharge for each flow event on ephemeral stream records. These tables were compared and cross-checked with stage
hydrographs, notes found in the archives, and published average
daily flows and peak flows so that discharge recorded at the highest
possible temporal resolution. Catchment area ranges from 2 to
1,256 km2 for the 28 study sites. The longest site record obtained
from the USGS archives is 22 years. The number of flow events per
year of complete record ranges from 1 to 50. In the annual maxima
series, records extend from the 1950s to present. The longest
record of annual maxima is 51 years. Streamflow records in the
partial-duration series represent time periods from the 1960s to
present. The average annual precipitation at these sites ranges from
Table 2. Characteristics of Study Sites and Period of Flow Records
Region
Site
Phoenix
aguafriahumboldt
boulder
humbug
newriverbell
Agua Fria River
near Humboldt, AZ
Boulder Creek near
Rock Springs, AZ
Humbug Creek near
Castle Hot Springs, AZ
New River at Bell
Road near Peoria, AZ
Gage
number
Basin
area (km2 )
Elevation
(m)
AMS period
of recorda
PDS period
of recorda
09512450
1,256
1,342
2000–2006
2000–2006
09512830
98
576
1984–1993
1988–1993
09512860
155
546
1984–1994
1988–1994
09513835
479
367
1963–1984;
1990–1993
1969; 1974;
1976–1979;
1981; 1984; 1993
1961–1962; 1967;
1969; 1991–1998
newriverglendale
New River near
Glendale, AZ
09513910
837
323
salttrib
Salt River Tributary in
South Mountain Park,
Phoenix, AZ
09512200
5
430
1943; 1955;
1960–1979;
1990–1998
1961–2006
skunk
Skunk Creek near
Phoenix, AZ.
09513860
168
449
1960–2006
sycamore
Sycamore Creek near
Fort McDowell, AZ.
Turkey Creek near
Cleator, AZ
Wet Bottom Creek
near Childs, AZ
Amargosa River at
Beatty, NV
C-1 Channel near Warm
Springs Road at
Henderson, NV
Caruthers Canyon near
Ivanpah, CA
09510200
425
536
1960–2006
1962–1963; 1969;
1971–1972; 1976;
1979–1981;
1986; 1988
1961; 1969;
1972–1976;
1981; 1984;
1992–2006
1988–2006
09512600
232
958
1980–1992
1988–1992
09508300
94
708
1968–2006
1986–2006
10251217
1,186
1,007
1995–2005
1994–2005
09419740
10
570
1991–2006
1991; 1993–2006
09423350
2
1,734
1964–2007
10250800
448
841
1963–1989
09419674
260
681
1983–2005
1965; 1967;
1971–1972;
1974; 1976; 1978;
1981–1984;
1986–1989;
1999–2007
1965–1967;
1981; 1985
1994–2005
09419610
24
2,377
1961–1994
1968–1969;
1971–1981;
1983; 1985; 1987;
1990; 1994
1966; 1970–1974;
1976–1977;
1979–1981
1989–1999
1990–2006
turkey
wetbottom
Las Vegas
USGS
site name
amargosa
c1channel
Caruthers
Darwin
flamingo
leecanyon
Darwin Canyon near
Darwin, CA
Flamingo Wash at Decatur
Blvd. at Las Vegas, NV
Lee Canyon near Charleston
Park, NV
lovell
Lovell Wash Near Blue
Diamond, NV
10251980
137
1,168
pittman
Pittman Wash at Wigwam
Parkway near Henderson, NV
Sloan Channel Tributary at
Las Vegas Blvd. near North
Las Vegas
09419695
177
628
1965–1981;
1987; 1999;
2003–2007
1988–2000
09419659
45
566
1989–2007
sloan
12 / JOURNAL OF HYDROLOGIC ENGINEERING © ASCE / JANUARY 2014
J. Hydrol. Eng. 2014.19:10-17.
Table 2. (Continued.)
Region
Albuquerque
aboarroyo
fourmiledraw
Downloaded from ascelibrary.org by UNIV OF NEVADA-RENO on 12/17/13. Copyright ASCE. For personal use only; all rights reserved.
grants
lascruces
losesteros
losesterostrib
saltcreek
sanvicente
southsevenrivers
a
USGS
site name
Site
Abo Arroyo near Blue
Springs, NM
Fourmile Draw near
Lakewood, NM
Grants Canyon at Grants, NM
Las Cruces ARR near Las
Cruces, NM
Los Esteros Creek above
Santa Rosa Lake, NM
Los Esteros Creek Tributary
above Santa Rosa Lake, NM
Salt Creek near Tularosa, NM
San Vicente Arroyo at
Silver City, NM
South Seven Rivers
near Lakewood, NM
Gage
number
Basin
area (km2 )
Elevation
(m)
AMS period
of recorda
PDS period
of recorda
08331660
616
1,672
1997–2000
1997–2000
08400000
686
1,016
1952–2006
08343100
08363600
34
35
1,968
1,235
1962–1995
1961–1966
1976; 1983; 1985;
1987; 1990–1992;
1995; 1997;
2001–2006
1991–1995
1961–1694
08382730
170
1,456
1974–1990
1974; 1976–1997
08382760
686
1,454
1973–1997
08480595
08477600
35
69
1,235
1,792
1996–2005
1938; 1954–1965
1974–1980;
1982–1990
1996–2005
1961–1965
08401200
570
1,014
1964–2006
1989–1990;
2002–2006
Period of record in complete water years.
90 to 430 mm. Site elevations range from 323 m above mean sea
level (amsl) to 2,377 m amsl.
Results
Envelope curves of maximum observed discharge versus catchment
area play an important role in validating predictions of flow and
delineating geographic flood regions (Crippen and Bue 1977;
Gaume et al. 2009; Jarvis 1926; Meirovich et al. 1998; Mimikou
1984; Rybski and Bunde 2009). A discharge envelope curve for
both perennial and ephemeral streams in the Southwestern United
States was developed by Thomas et al. (1997) using more than
1,300 gaging stations. The maximum observed peak for a given
basin area for all study sites fell within that regional envelope
(Fig. 2), though well below the upper limit of flows observed in
Fig. 2. Maximum observed flow study sites fall within the Southwestern regional envelope using 1,300 ephemeral and perennial streams
perennial streams. This is presumed to be the result of a lack of
base flow in ephemeral channels, as well as transmission losses
through the channel during runoff events. In addition, although
the Phoenix and Albuquerque area streams appear to follow a
similar trend of increasing maximum discharge with basin area,
for the nine Las Vegas area streams, the maximum discharge appears to fluctuate between 11 and 100 m3 =s for basin area between
2 and 1,000 km2 without a monotonic trend. This is also likely
explained by transmission loss variability in the Las Vegas area.
Envelope curves demonstrating historic bounds for ephemeral
stream flood volumes of the Southwestern United States will
have great value for truthing rainfall runoff models used for the
estimation of detention basin storage. A single example of regional
ephemeral flow and volume statistics and envelope curves exists
for Israeli streams (Meirovich et al. 1998), which is used as a
benchmark for comparing southwestern U.S. data. The maximum
observed flood volume for each study site is bounded by the areavolume envelope curve developed for ephemeral streams in the arid
Negev desert (Meirovich et al. 1998). However, the U.S. ephemeral
stream historic maximum observed volumes have more scatter than
those of Israeli streams [Fig. 3 and Meirovich et al. (1998), p. 210].
Historic observations also do not appear to follow the general trend
of increasing volume with area, as found in Israel. This led to the
investigation of additional controls on flood event volume.
Flow volume VðTÞ is commonly related to peak flow Qp ðTÞ
using V ¼ aQbp , where a and b are coefficients often related to basin and climatic characteristics. For example, the volume-discharge
relationship V ¼ 14.5Q−0.914 fit 105 single-peak hydrographs from
summer thunderstorm flood events from 18 ephemeral and perennial streams in southern Utah (Eychaner 1976). Peak flow–volume
relationships for all ephemeral flow events in the study regions fell
into a window with well-defined maxima [Fig. 4(a)]. The authors
were able to further clarify the peak discharge-event volume relationship by accounting for hydrograph duration. Specifically, a
deterministic power-law relationship between peak and volume
exists for a given flow event duration [e.g., Table 3, Fig. 4(b)].
Volumes VðTÞ for recurrence intervals T can be estimated from
corresponding peaks Qp ðTÞ if recurrence intervals of the volumes
are the same as those of the peaks (Devulapalli and Valdes 1996;
Eychaner 1976). This would be a useful relationship to exploit
JOURNAL OF HYDROLOGIC ENGINEERING © ASCE / JANUARY 2014 / 13
J. Hydrol. Eng. 2014.19:10-17.
Downloaded from ascelibrary.org by UNIV OF NEVADA-RENO on 12/17/13. Copyright ASCE. For personal use only; all rights reserved.
Fig. 3. Maximum observed volume per area at the study sites are
bounded by the envelope curve developed for streams in the Negev
Desert, Israel (adapted from Meirovich et al. 1998)
because the USGS National Flood Frequency (NFF) program has
published regional regression equations to estimate flood peak
discharges for various recurrence intervals (Jennings et al. 1994;
Reis and Crouse 2002). However, these relations cannot be used
in this semiarid study area where the discharge recurrence and
volume recurrence for most individual flow events do not coincide
(Fig. 5). In the extreme example of Wet Bottom Creek, Arizona, a
December 4, 1992 flood event resulted in the 21-year peak but only
the 1-year volume. As in many areas, variation in timing, intensity,
and areal distribution of rainfall, as well as direction of the storm in
relation to the basin, leads to problems in relating peak discharge to
volume (Lennartz and Bunde 2009).
Envelope curves summarize the extent of historic floods, but
generally cannot inform flow or volume frequency (Castellarin
et al. 2005). Plotting the positions for each site demonstrates that
the annual maximum series (AMS) and the partial duration series
(PDS) for peak flows converge by the 7-year event (Fig. 6).
Although the PDS is a complete record of flow events, for almost
all study sites, the largest flow event in any given year far exceeds
the second largest flow event, leaving annual maximum a useful
statistic for predicting extreme (>7-year) events. Using AMS resulted in an increase in the period of record for most sites.
Regional frequency relationships are often developed by
relating recurrence data to physical or climatic characteristics using
multiple-regression techniques. If recurrence predictions are required for levels outside the period of record, a theoretical probability distribution can be fit to observed data for extrapolation of
recurrence estimates from a regional curve [e.g., Stedinger and
Tasker (1985)]. There are no specific recommendations for fitting
extreme volumes with probability densities, although theoretical
and empirical justification has been given for fitting a variety of
densities to flow recurrence. Bulletin 17B (IACWD 1982) recommends the use of the Log-Pearson III distribution for all stream
types. The extreme value (EV)-1 or Gumbel distribution provided
the best fit to recurrence of ephemeral flows in Saudi Arabia
(Al-Turbak and Quraishi 1987; Sorman and Abdulrazzak 1987),
whereas the lognormal distribution and gamma (log-Pearson Type
III) distributions provided good fits for streams in Israel (Ben-Zvi
and Meirovich 1997). Generalized extreme value densities were developed for each region in a study of 162 stations from northwest
Africa, Iran, Jordan, Saudi Arabia, Botswana, and South Africa
(Farquharson et al. 1992). Visual assessment and the AndersonDarling statistic were used in this study for goodness of fit to
determine the best-fit distribution for the AMS of peak and volume
for the study sites. The commonly used Weibull, gamma, logPearson Type III, lognormal, and extreme-value distributions were
tested for fit to recurrence intervals (Table 4). No single probability
density was acceptable for all sites in a given region for annual
maxima of volumes. Furthermore, the p-values for goodness of
fit tests tended to be high, indicating poor distributional fits. Lack
of fit of probability densities to flow recurrence in the arid/semiarid
Southwest has been noted in other studies (Hjalmarson and
Thomas 1992).
Fig. 4. (a) Peak-volume data clouds are similar for three study regions. A single envelope curve bounds the maximum volume corresponding to a
3
3
given peak, V ¼ 90293Q0.99
p , where V is in m and Qp is in m =s; (b) upper-boundary models for 24-, 12-, and 6-h flow event lengths for Phoenix
(dashed lines) (see Table 3 for best-fit lines for x-hour event maxima and best fit for all regions)
14 / JOURNAL OF HYDROLOGIC ENGINEERING © ASCE / JANUARY 2014
J. Hydrol. Eng. 2014.19:10-17.
Table 3. Summary of Volume-Peak Models for the Three Study
Regions; Best-Fit and Upper Boundary Equations Are Shown, Where x
Is the Peak Streamflow for the Runoff Event (m3 =s) and yðxÞ Is the
Estimated Volume (m3 )
Region
Event
duration
(h)
Best-fit
R2 a
Upper
boundary
24
12
6
24
12
6
24
12
6
yðxÞ ¼ 25700 x0.60
yðxÞ ¼ 13569 x0.66
yðxÞ ¼ 6962 x0.69
yðxÞ ¼ 26665 x0.71
yðxÞ ¼ 13477 x0.69
yðxÞ ¼ 8822 x0.77
yðxÞ ¼ 15694 x0.89
yðxÞ ¼ 11689 x0.66
yðxÞ ¼ 5340 x0.74
0.96
0.97
0.98
0.90
0.96
0.97
0.96
0.98
0.94
yðxÞ ¼ 53302 x0.89
yðxÞ ¼ 23535 x0.75
yðxÞ ¼ 12240 x0.78
yðxÞ ¼ 54601 x0.91
yðxÞ ¼ 38210 x0.94
yðxÞ ¼ 18814 x0.91
yðxÞ ¼ 44830 x0.82
yðxÞ ¼ 16221 x0.70
yðxÞ ¼ 9807 x0.75
Phoenix
Downloaded from ascelibrary.org by UNIV OF NEVADA-RENO on 12/17/13. Copyright ASCE. For personal use only; all rights reserved.
Las Vegas
Albuquerque
a 2
R is presented for the best-fit models.
Fig. 6. AMS and PDS for Pittman Wash converge at 4 years; AMS and
PDS for all study sites converge by 7 years
the use of AMS reduces the labor involved in data collection and
production of distributions that describe flood statistics.
What Is the Relationship between Peak Flow and
Volume for Ephemeral Flow Events?
Fig. 5. Each data point represents a single-flow event; peak and volume
for individual flow events have different recurrence
Discussion
None of this study’s ephemeral study sites had rain gages. This is
true of most ephemeral streams in the Southwestern United States,
so it is generally not possible to relate storm statistics and resulting
flash flood characteristics. This makes flow and volume envelope
curves a useful check on rainfall-runoff models that are used for
developing flood mitigation structures.
Are Annual Maxima Sufficient for Extreme Event
Recurrence Estimates for Ephemeral Streams?
AMS and PDS series for all study sites converged by 7 years,
suggesting that the AMS is sufficient for describing recurrence
of flows on timescales greater than a decade. As described previously, although the PDS is a complete record of flow events, the
largest annual flows tend to far exceed the second-largest flow
events such that only the annual peaks contribute to the
>10-year recurrence statistics. This is a significant result because
Most (64%) of the single-peak flash flood hydrographs in the study
data set consist of a near-instantaneous peak with an exponentially
decaying falling limb. The bulk of hydrograph volume is a function
of the length of the falling limb, representing event duration.
Multiple-peak hydrographs act as overlapping single-peak events,
so the same argument applies. The short periods of extreme discharge do not contribute a large fraction of the total volume unless
event duration is small. This explains the difference between peak
and volume recurrence for flow events. Spatial and temporal variability of southwestern storm cells, combined with dependence of
event volume on both peak and duration, likely leads to scatter in
the basin area–maximum observed volume relationship. Even
though the Albuquerque, Phoenix, and Las Vegas regions experience different climates, the common shape of their ephemeral
hydrographs leads to similar peak-duration-volume relationships.
Lack of fit of commonly used probability densities to annual
maximum volumes may also be a function of the dependence of
volume on both peak flow and duration. Development of probabilistic volume–frequency relationships likely will require sufficient
data for a joint density for peak flow and event duration. Because
peak flow and event duration appear to be independent, a joint
density can be developed through convolution of the individual
densities (Ross 1994) or methods such as copulas (Zhang and
Singh 2006). The density describing the probability of encountering an event of length t appears easier to obtain than the density of
peaks. In the three study regions, short (<1 h) events were most
common and the probability of longer events decays as a power
law with time (Fig. 7).
What Variables Should Be Considered in Designing
Detention Basins for Ephemeral Channels?
Optimal design for a flood-mitigation structure specifies the smallest storage volume that will limit outflow to a maximum discharge.
It is discharge magnitude that causes damage, so it is necessary to
JOURNAL OF HYDROLOGIC ENGINEERING © ASCE / JANUARY 2014 / 15
J. Hydrol. Eng. 2014.19:10-17.
Table 4. Best-Fit Distributions for AMS Peaks and Volumes for Each Site with Associated p-values
Peak
Region
Downloaded from ascelibrary.org by UNIV OF NEVADA-RENO on 12/17/13. Copyright ASCE. For personal use only; all rights reserved.
Phoenix
Las Vegas
Albuquerquec
Site
aguafriahumboldt
boulder
humbug
newriverbell
newriverglendale
salttrib
skunk
sycamore
turkey
wetbottom
amargosa
c1channel
caruthers
darwin
flamingo
leecanyon
lovell
pittman
sloan
aboarroyo
fourmiledraw
lascruces
loesesterostrib
losesteros
saltcreek
sanvicente
southsevenrivers
Best-fit distribution
b
P-III
P-III
lognormal
Weibull
lognormal
Weibull
lognormal
P-III
LP3
Weibull
Weibull
Weibull
lognormal
Weibull
LP3
LP3
lognormal
Weibull
lognormal
lognormal
Weibull
LP3
LP3
Weibull
lognormal
LP3
lognormal
Volume
p-value
Na
Best-fit distribution
>0.25
0.21
0.21
>0.25
0.98
>0.25
0.39
>0.25
0.23
0.20
>0.25
>0.25
0.17
>0.25
>0.25
>0.25
0.68
>0.25
0.73
0.57
>0.25
>0.25
>0.25
0.09
0.49
>0.25
0.21
7
10
11
21
27
27
44
47
13
39
10
15
41
20
22
18
16
12
17
4
40
6
17
25
7
13
36
P-III
P-III
P-III
lognormal
lognormal
lognormal
Weibull
LP3
lognormal
lognormal
lognormal
lognormal
P-III
Weibull
LP3
Weibull
P-III
P-III
lognormal
lognormal
lognormal
lognormal
lognormal
LP3
NA
lognormal
lognormal
p-value
N
0.10
0.24
>0.25
0.11
0.56
0.52
>0.25
0.18
0.52
0.10
0.21
0.93
>0.25
>0.25
0.21
>0.25
>0.25
>0.25
0.87
0.87
0.54
0.53
0.87
>0.25
<0.05
0.51
0.11
5
5
7
8
3
6
18
17
4
18
12
12
22
5
11
8
5
10
14
4
4
4
22
16
10
5
5
N is the number of years of complete data.
P-III is the Pearson Type III or gamma distribution.
c
Insufficient data to characterize peak or volume distributions for Grant’s Canyon at Grants, NM.
a
b
Summary
Fig. 7. Flow events of less than 1 h are most common in the study area;
the probability of longer events decays as a power law
protect against the x-year discharge. In determining storage
volume, recurrence of event duration for the x-year discharge must
be considered in conjunction with the rate of detention basin
drainage. The consistent relationship among flow, duration, and
volume suggests that basin area and network morphology are less
significant in predicting flood volume.
High temporal resolution flash flood hydrographs were studied for
28 ephemeral streams in three climatic regions of the Southwestern
United States. From this data, the following conclusions can
be drawn:
1. Annual maxima series are sufficient for inferring flow recurrence on timescales greater than a decade.
2. The basin area–event volume data cloud and the envelope that
bounds it can be used to evaluate results of rainfall-runoff
models.
3. The relationship between peak discharge and volume for a
flow event cannot be tightly constrained without consideration
of event duration. As a result, volumes VðTÞ, for recurrence
intervals T, cannot be estimated from corresponding peaks
Qp ðTÞ. Instead, volume recurrence must be estimated based
on the joint distribution of peak discharge and duration.
4. The density governing event duration in these semiarid study
regions decays as a power law with time.
5. No single probability density provides a good fit to flow or
volume magnitude in the study streams.
6. The deterministic power-law relationship between peak discharge and volume for an event of a given magnitude over different climatic regions suggests that flow-event volumes are
more a function of ephemeral hydrograph shape (sharp front
with slow decay of falling limb) than basin area or geometry.
Acknowledgments
This work was supported by an Urban Flood Demonstration Program grant through the Army Corps. of Engineers. The authors
16 / JOURNAL OF HYDROLOGIC ENGINEERING © ASCE / JANUARY 2014
J. Hydrol. Eng. 2014.19:10-17.
wish to thank Thomas Halthom (Sacramento), Kerry Garcia
(Carson City), Phillip Bowman (Albuquerque), and Shirley
Francisco (Flagstaff) of the U.S. Geological Survey for locating
and providing access to archived data.
Downloaded from ascelibrary.org by UNIV OF NEVADA-RENO on 12/17/13. Copyright ASCE. For personal use only; all rights reserved.
References
Al-Turbak, A. S., and Quraishi, A. A. (1987). “Regional flood frequency
analysis for some selected basins in Saudi Arabia.” Regional flood
frequency analysis, V. P. Singh, ed., D. Reidel, Dordrecht, the
Netherlands, 27–34.
Ben-Zvi, A., and Meirovich, L. (1997). “Direct probabilistic description
of arrival times of runoff events within the year.” Stochastic Hydrol.
Hydraul., 11(6), 511–521.
Castellarin, A., Vogel, R. M., and Matalas, N. C. (2005). “Probabilistic
behavior of a regional envelope curve.” Water Resour. Res., 41(6).
Crippen, J. R., and Bue, C. D. (1977). “Maximum flood flows in the conterminous United States.” U.S. Geological Survey Water Supply Paper,
1887, U.S. Government Printing Office, Washington, DC, 52.
Devulapalli, R. S., and Valdes, J. B. (1996). “Volume-duration-frequencies
for ungaged catchments in Texas.” Texas A&M Univ., College Station,
TX.
Eychaner, J. H. (1976). “Estimating runoff volumes and flood hydrographs
in the Colorado River basin, Southern Utah.” Water resources investigations, U.S. Geological Survey, Salt Lake City, UT, 76–102.
Farquharson, F. A. K., Meigh, J. R., and Sutcliffe, J. V. (1992). “Regional
flood frequency analysis in arid and semi-arid areas.” J. Hydrol.,
138(3–4), 487–501.
Fernandez, P. C., Fattorelli, S., Rodriguez, S., and Fornero, L. (1999).
“Regional analysis of convective storms.” J. Hydrol. Eng., 10.1061/
(ASCE)1084-0699(1999)4:4(317), 317–325.
Gaume, E., et al. (2009). “A compilation of data on European flash floods.”
J. Hydrol., 367(1–2), 70–78.
Hawkins, R. H., and Pole, R. A. (1989). “Standardization of peak-volume
transformations.” Water Resour. Bull., 25(2), 377–380.
Hjalmarson, H. W., and Thomas, B. E. (1992). “A new look at regional
flood-frequency relations for arid lands.” J. Hydraul. Eng., 10.1061/
(ASCE)0733-9429(1992)118:6(868), 868–886.
Hosking, J. R. M., and Wallis, J. R. (1997). Regional frequency analysis,
Cambridge University Press, Cambridge, U.K.
Interagency Advisory Committee on Water Data (IACWD). (1982). Interagency Advisory Committee on Water Data guidelines for determining
flood flow frequency: Bulletin 17B of the Hydrology Subcommittee,
USGS, Reston, VA.
Jarvis, C. S. (1926). “Flood flow characteristics.” Trans. Am. Soc. Civ.
Eng., 89(1), 985–1032.
Jennings, M. E., Thomas, W. O., and Riggs, H. C. (1994). “Nationwide
summary of U.S. Geological Survey regional regression equations
for estimating magnitude and frequency of floods for ungaged sites,
1993.” U.S. Geolgoical Survey Water-Resources Investigations Rep.
196, USGS, Reston, VA.
Lennartz, S., and Bunde, A. (2009). “Trend evaluation in records with longterm memory: Application to global warming.” Geophys. Res. Lett.,
36(16).
Mason, R. R., Jr., Fuste, L. A., King, J. N., and Thomas, W. O., Jr. (2000).
“The National Flood-Frequency Program—Methods for estimating
flood magnitude and frequency in rural areas in New Mexico.” USGS
Fact Sheet 055-00, USGS, Reston, VA.
Mason, R. R., Jr., King, J. N., and Thomas, W. O., Jr. (1999a). “The
National Flood-Frequency Program—Methods for estimating flood
magnitude and frequency in rural areas in Arizona.” USGS Fact Sheet
111-98, USGS, Reston, VA.
Mason, R. R., Jr., Ries, K. G., King, J. N., and Thomas, W. O., Jr. (1999b).
“The National Flood-Frequency Program—Methods for estimating
flood magnitude and frequency in rural areas in Nevada.” USGS Fact
Sheet 123-98, USGS, Reston, VA.
Meirovich, L., Ben-Zvi, A., Shentsis, I., and Yanovich, E. (1998).
“Frequency and magnitude of runoff events in the arid Negev of Israel.”
J. Hydrol., 207(3–4), 204–219.
Mimikou, M. (1984). “Envelope curves for extreme flood events in
northwestern and western Greece.” J. Hydrol., 67(1–4), 55–66.
Reis, K. G., III, and Crouse, M. Y. (2002). “The National Flood Frequency
Program, Version 3: A computer program for estimating magnitude and
frequency of floods for ungaged sites.” USGS Water-Resources
Investigations Rep., USGS, Reston, VA.
Ross, S. (1994). A first course in probability, MacMillan, Englewood
Cliffs, NJ.
Rybski, D., and Bunde, A. (2009). “On the detection of trends in long-term
correlated records.” Physica A, 388(8), 1687–1695.
Sherif, M. M., Mohamed, M. M., Shetty, A., and Almulla, M. (2011).
“Rainfall-runoff modeling of three wadis in the northern area of
UAE.” J. Hydrol. Eng., 10.1061/(ASCE)HE.1943-5584.0000280,
10–20.
Sorman, A. U., and Abdulrazzak, M. J. (1987). “Regional flood discharge
analysis southwest region of the kingdom of Saudi Arabia.” Regional
flood frequency analysis, V. P. Singh, ed., D. Reidel, Dordrecht, The
Netherlands, 11–25.
Stedinger, J. R., and Tasker, G. D. (1985). “Regional hydrological analysis
1: Ordinary, weighted, and generalized least squares compared.” Water
Resour. Res., 21(9), 1421–1432.
Thomas, B. E., Hjalmarson, H. W., and Waltermeyer, S. D. (1997).
“Methods for estimating magnitude and frequency of floods in the
southwestern United States.” USGS Water Supply Paper 2433, USGS,
Tucson, AZ.
Zhang, L., and Singh, V. P. (2006). “Bivariate flood frequency analysis using the copula method.” J. Hydrol. Eng., 10.1061/(ASCE)1084-0699
(2006)11:2(150), 150–164.
JOURNAL OF HYDROLOGIC ENGINEERING © ASCE / JANUARY 2014 / 17
J. Hydrol. Eng. 2014.19:10-17.