5th International Symposium on Hydraulic Structures Brisbane, Australia, 25-27 June 2014

Hydraulic Structures and Society: Engineering Challenges and Extremes

ISBN 9781742721156 - DOI: 10.14264/uql.2014.25

Flow Depth and Energy Coefficient Relationships for Stepped

Spillways

S.L. Hunt1 and K.C. Kadavy1

1
United States Department of Agriculture, Agricultural Research Service
Hydraulic Engineering Research Unit
Stillwater, OK 74075
UNITED STATES
E-mail: Sherry.Hunt@ars.usda.gov

Abstract: A multi-year, large-scale physical model study of stepped chutes was conducted over a
broad range of design parameters (i.e. step heights, slopes, and unit discharges). Air entrainment
developed naturally as the flow descended the chute. Air entrainment began to develop downstream
of the surface inception point at L/Li = 1.0, and it became fully developed at approximately L/Li = 2.0.
Depending on the chute slope, energy loss in the model ranged from 20% near L/Li = 1.0 to as much
as 85% near L/Li = 6.0. The contributions of this study are the development of flow depth, clear-water
flow depth, and energy coefficient relationships for stepped chutes with slopes ranging from 10 to 30.
Although stepped chutes provide significant energy dissipation, large steps may have an adverse
effect on the resulting Froude number; consequently, a longer stilling basin may be required then one
might expect. The results of this research are expected to make a significant impact of the design of
these structures.
Keywords: stepped chute, energy dissipation, inception point, stilling basin, and air entrainment.

1. Introduction
The United States Department of Agriculture (USDA) Natural Resources Conservation Service
(NRCS) has provided financial and technical assistance for the construction of more than 11,000 flood
control dams. The majority of these earthen embankments were originally constructed as low hazard
dams for providing flood control of agricultural land, but hazard creep (i.e. a change in hazard
classification from low to significant or high hazard) has occurred due to changing demographics in the
vicinity of some of these dams. Infrastructure has limited options for dam rehabilitation.

Spillway capacity is a deficiency commonly identified for these aging embankment dams. Stepped
chutes can provide increased spillway capacity. Other design advantages of stepped chutes are ease
of construction, application to the existing embankment slope, and the significant energy dissipation
provided as noted by Chanson (1994), Boes (2000), Ohtsu et al. (2004), Meireles and Matos (2009),
and Hunt and Kadavy (2009). The USDA-NRCS estimates approximately 10% or 1,100 dams will
require the use of stepped spillways in their rehabilitation design. Due to aging infrastructure and the
need of stepped spillways applied to flatter slopes ( 30), research on stepped spillways has
significantly increased in recent years worldwide. The USDA-Agricultural Research Service (ARS),
the principal in-house research agency within the USDA, provides research for finding solutions to
agricultural problems. The USDA-ARS Hydraulic Engineering Research Unit (HERU) provides
research for the support of the USDA Small Watershed Program. The research program at the HERU
is two-fold: 1) developing design guidelines for the use of hydraulic structures for flood control and
embankment overtopping protection and 2) developing computational software for predicting
embankment breach by overtopping and internal erosion.

The purpose of this paper is to describe the research program at the HERU as it relates to stepped
chute design and the design relationships for flow depth, y or ycw; and energy coefficient, developed
from this research. Relative energy loss (H/Ho) for stepped spillways is dependent on slope, y or
ycw, velocity, and the energy coefficient, . Most researchers (i.e. Chanson 1994, Boes 2000, Boes
and Hager 2003; Ohtsu et al. 2004) have rearranged energy loss relationships in terms of friction
factor due to the lack of supporting data in the development of relationships for y, ycw, and
2. EXPERIMENTAL SET-UP
A multi-year series of experiments on stepped chutes has been conducted at the USDA-ARS HERU
(Hunt and Kadavy 2009 and Hunt et al. 2013). Model design and construction of the stepped chutes
were based on recommendations set forth by Chanson (2002), Boes and Hager (2003), and Pfister
and Hager (2011) to minimize the occurrence of scale effects. Table 1 summarizes the major design
attributes of the stepped chutes. More detailed descriptions of the models are outlined by Hunt and
Kadavy (2009) and Hunt et al. (2013).

Table 1. Summary of Model Parameters.

Model Parameters Model A Model B Model C
Chute slope (degree) 14.0 18.4 26.6
Chute width (m) 1.83 1.83 1.83
Total drop (m) 1.52 5.49 3.05
Unit discharge, q (m3/s/m) 0.105 0.824 0.150 1.83 0.146 1.25
Step height, h (mm) 40, 76, 152 smooth (0.46), 19, 40, 76, 152
40, 76, 152, 305
F* 1.15 75.3 smooth, 9,600 116,000 1.40 70.4
steps, 0.6 428
h/dc 0.093 1.47 0.027 - 2.27 0.086 1.14

Ohtsu et al. (2001, 2004) and Chanson (2002) summarize acceptable methods for determining
skimming flow conditions. The flow conditions for these studies ranged from nappe to skimming, such
that the step height (h) to critical depth (dc) ratio was 0.027 h/dc 2.27. For the purposes of this
paper, skimming flow data, 0.027 h/dc 1.1, are used in the evaluation and development of
relationships.

Flow depths and velocity profiles were measured using a back-flushing Pitot tube coupled with a
differential pressure transducer or a two-tipped fiber optic probe attached to a manually-operated point
gauge and coupled with a data acquisition system. Centerline velocity profiles typically consisted of
10 measurements beginning at the step edge and ascending up through the flow profile at each
station. Approximately 10 stations were recorded with the same stationing used for all step heights.
The Pitot tube coupled with the pressure transducer was used to measure velocities in both the non-
aerated flow region, from the crest to Li and the aerated flow region from the Li to the end of the chute.
Although traditionally used to measure non-air entrained flows, a back-flushing Pitot tube, according to
Matos et al. (2002), is capable of measuring velocities in air entrained flows if the air concentrations
are less than 70%. To achieve a more accurate velocity measurement in air entrained flows, the Pitot
tube was back-flushed before each measurement to avoid air bubbles in the tube and the lines to the
pressure transducer. The two-tip fiber optic probe was used to measure velocities and air
concentrations in the flow region at and downstream of the inception point. The use of the Pitot tube
and the fiber optic probe in the aerated flow region allowed for a direct comparison between the two
methods for measuring velocities in air entrained flows.

3. DISCUSSION
The series of experiments conducted at the USDA-ARS HERU provides a comprehensive database
for broad-crested stepped chutes conducted at a large-scale for slopes ranging between 10 30.
This research has provided substantial data for the development of relationships related to the
inception point (Li), flow depth (y), clear-water flow depth (ycw), air concentration (C), and energy
coefficient (). The impact these relationships have on the design of the stilling basin will also be
discussed. Chansons (1994, 2002) research covers a broad range of data as well, but his work is
concentrated on ogee-crested weir stepped chutes, with steeper slopes, > 30. Boes (2000), Ohtsu
et al. (2001, 2004), Boes and Hager (2003), Meireles and Matos (2009), Frizell and Svodboda (2012),
Meireles et al. (2012) and others have conducted research that is more focused (i.e. smaller range of
, h, h/dc, q, drop height, etc.) and limited to some extent due to the size and/or scale of models. Yet,
despite these boundaries, the researchers have contributed significantly to development of design
guidelines (i.e. inception point, energy dissipation, flow depth, etc.) for these structures. The following
sub-sections provide the relationships developed from this research and a comparison to other
available relationships and data from literature (Chanson 1994, 2002; Ohtsu, et al. 2001, 2004;
Chanson and Toombes 2002; Ward 2002; Felder and Chanson 2007; Meireles and Matos 2009; Bung
2011; Meireles et al. 2012).

3.1. Inception point

The inception point, Li, is most commonly defined as the surface location where the turbulent
boundary layer reaches the free surface and noted as Li as shown in Fig. 1 (Chanson 2002, Takahashi
et al. 2006, Meireles and Matos, 2009, Pfister and Hager 2011, Meireles et al. 2012, Hunt and
Kadavy, 2013). It is often referred to as the surface inception point because it is noted as the location
were aeration is first visually observed on the water surface. Air entrainment begins to develop
downstream of the surface inception point at L/Li = 1.0, and it became fully developed at
approximately L/Li = 2.0 (Bung 2011, Hunt et al 2013).
Growing
Boundary Non-aerated
Layer Flow Region Aerated Flow
Region
Inception
Point
Flow

Li

Figure 1. Inception point for broad-crested weir stepped spillways.

Chanson (1994, 2002) provided the framework for developing a Li for ogee-crested weir stepped
chutes for primarily steep slopes. Boes and Hager (2003) likewise developed an ogee-crested weir
stepped chute relationship for Li not for the surface like most others but for the location where the air
concentration near the pseudo-bottom is approximately 0.1. Meireles and Matos (2009), developed a
Li relationship similar to Chansons (1994, 2002), but their relationship predicts Li for broad-crested
stepped chutes with a much narrower range of and Froude surface roughness, F*. When Meireles
and Matos (2009) relationship is extended beyond its bounds, the relationship over-predicts Li for F* >
10 according to Hunt and Kadavy (2013). Hunt and Kadavy (2013) optimized Li for broad-crested
stepped chutes for a much broader range of and F* as outlined in Eqs. 1 and 2. For the purposes of
this paper, Hunt and Kadavys (2013) relationships will be used for determining Li.

Li
5.19F*
0.89
0.1 F* 28 (1)
ks

7.48F*
Li 0.78
28 F* 105 (2)
ks

where F* = Froude number defined in terms of roughness height: F* = q/[g(sin)ks3]0.5, g =

gravitational constant, and ks = the surface roughness = h cos().

3.2. Flow Depth

With this extensive database, interest lies in comparing the data to already existing relationships
available in literature and potentially increasing the applicability of these relationships, so to not re-
invent the wheel per se. Bung (2011) provides an independent relationship for clear-water flow depth
'
downstream of Li as a function of F*, H d , and h. When compared to data obtained from models A, B,
and C as shown in Fig. 2, Bungs (2011) relationship fits the data well for lower values of F*, but it
begins to skew slightly from the data for larger values of F*. According to Hunt and Kadavy (2009) and
Hunt et al. (2013), the flow depth normalized by the critical depth, y/dc, decreases rapidly from the
crest to L/Li = 1.0 for models A and B, and then for L/Li > 1.0, ycw/dc becomes relatively constant. A
similar observation was made for Model C. Based on these observations, Chansons (1994, 2002)
relationship and Meireles and Matos (2009) relationship for depth at the inception point, di, may be
relevant if one was to assume ycw remains relatively constant such that di = ycw as Hunt and Kadavy
(2009) and Hunt et al. (2013) suggests. Fig. 2 shows that Chansons relationship over-predicts ycw
data obtained from Models A, B, and C. Chansons relationship was applied outside its bounds (i.e.
36.4), so this may explain some of the over-prediction of the data. However, Meireles and Matos
(2009) relationship predicts the data well for its applicable range. Extending the relationship beyond
those bounds, their relationship appears to over-predict at higher values of F*.

Given relationships from Bung (2011), Chanson (1994, 2002), and Meireles and Matos (2009) tend to
over-predict ycw from Models A, B, and C, a best fit of the data, Eq. 3, was obtained using a similar
multi-variant, non-linear regression statistical analysis as Chanson (1994, 2002).

d i y cw
0.34(sin ) 0.13 F*0.62 (3)
ks ks

This best fit provides a coefficient of determination, R2, near 1.0 and a standard error of 0.07. A
comparison of Eq. 3 to data from Models A, B, and C along with data obtained from Chanson and
Toombes (2002), Felder and Chanson (2007), and Ward (2002) proves to be a reasonable method for
determining ycw/ks as shown in Fig. 3. To compare data obtained by Chanson and Toombes (2002),
Felder and Chanson (2007), and Ward (2002), assumptions were made regarding the data in relation
to L/Li, which could lead to the discrepancies in their observed data as compared to data obtained from
models A, B, and C.

100 1.0E+3 ModelsA,B,andC

Model A Eq.3,3.4deg
Model B Eq.3,14deg
Eq.3,18.4deg
Model C Eq.3,26.6deg
F-C (2007), 21.8 deg. 1.0E+2 Eq.3,52deg
Ward(2002),26.6deg,L/Liars>=1.0
10 Bung (2011), 26.6 deg.
Chanson (1994, 2002), 52 deg
CT(2002),21.8&15.9deg
FC(2007),21.8deg
ycw/ks

CT(2002),4.0deg,smoothchute
ycw/ks

Chanson (1994, 2002), 18.4 deg.

1.0E+1 CT(2002),3.4deg
M-M (2009)

1
1.0E+0

1.0E1
0
1.0E+0 1.0E+1 1.0E+2 1.0E+3 1.0E+4 1.0E+5
1 10 100 1000 F*
F*
Figure 2. Experimental ycw/ks data under skimming Figure 3. Experimental ycw/ks data under
flow conditions and L/Li > 1.0 compared with skimming flow conditions and L/Li > 1.0 compared
relationships from Bung (2011), Chanson (1994, with data from Chanson and Toombes (C-T)
2002), and Meireles and Matos (M-M) (2009),and (2002), Ward (2002), and Felder and Chanson
data from Felder and Chanson (F-C) (2007). (2007) and to Eq. 3.

Contradicting previous findings by Meireles and Matos (2009), Bung (2011), and Meireles et al.
2/3
q
(2012), Hunt et al. (2013) indicate that ycw/dc is a function of h/dc. Given d c 1 / 2 , then Eq. 3
g
can be rearranged into Eq. 4.

0.063
y cw h
0.34 cos 0.063 sin 0.18 for L/Li 1.0 and 0.027 h/dc 1.1 (4)
dc dc
The predicted and measured ycw/dc downstream of Li fall within 10% error. Eq. 4 provides a direct,
simplistic method for determining ycw that often relies on the air concentration, C, and the bulked flow
depth, y90.

For L/Li 1.0, Hunt et al. (2013) recognized that flow depth, y, is a function of not only h/dc and , but
also of L/Li. By assuming no air entrainment at Li, one can assume implicitly that y/dc = ycw/dc at L/Li =
1.0. To provide continuity between ycw/dc from Eq. 4 and y/dc for L/Li 1.0, the intercept for y/dc was
set equal to Eq. 4. From a multi-variant, non-linear statistical analysis, the refined relationship for y/dc
is as follows:

y L
0.22
h
0.063

0.34 cos 0.063 sin 0.18 (5)
d c Li dc

for 0.1 (L/Li) 1.0 and 0.027 h/dc 1.1

Eq. 5 yields a R2 = 0.96 and a standard error of 0.03. The predicted and measured y/dc agree within
10%. It should be noted that Eq. 5 was not developed to predict the nappe profiles near the crest
and therefore is not valid for L/Li < 0.1.

3.3. Energy Coefficient

Researchers like Chanson (2002), Boes (2000), Matos (2003); Boes and Hager (2003); Meireles and
Matos (2009), Hunt and Kadavy (2009), and Frizell and Svoboda (2012) have long noted that stepped
chutes provide significant energy dissipation, resulting in a reduction in length for the stilling basin.
The total energy loss to a given step location is

H H o H (6)

Vo2
where H o y o H crest 1.5d c , Hcrest = vertical height from the chute crest to the stilling basin
2g
V2
floor, H y cw cos , V = mean velocity, Vo = approach velocity, yo = approach depth above
2g
datum, ycw = equivalent clear water flow depth, = chute slope, g = gravitational acceleration, and =
energy coefficient. For locations upstream of the inception point, L/Li < 1.0, ycw = measured flow
depth, y. Considering mass conservation, V is equal to q/ycw.

To determine H or the relative energy loss, H/Ho, at the toe of the chute all the parameters are fairly
straight-forward and easily determined through assumptions and a sequence of relationships, Eqs. 1,
2, 4, and 5. The only unknown is . Because of the lack of supporting data and the development of
relationships for determining y, ycw, and , researchers (i.e. Chanson 2002; Boes 2000; Matos 2003;
Ohtsu et al. 2001, 2004.) have relied on the friction factor to determine energy loss. The data
provided from models A, B, and C provide the framework for developing relationships for .

According to Chow (1959), can be assumed equal to unity if the channel is of regular cross-section
with a fairly straight alignment such that the effect of nonuniform velocity distribution on the computed
velocity head and momentum is small. Meireles et al. (2012) indicate that = 1.0 is typically adopted
for design purposes. Boes (2000) and Matos (2003) reports up to 1.21 for stepped spillways. To
examine the variation of and to determine the effect of the nonuniform velocity distribution on the
computed velocity head and momentum, Eq. 7 as outlined by Chow (1959) was used.

v dA v A
3 3

(7)
V 3A V 3A
where v is the velocity for an incremental area, A, in the velocity profile, V is the mean velocity of the
equivalent clear water depth-velocity profiles, and A is the total area of the profiles. Assuming two-
dimensional flow that is uniform across the width, A may be replaced by ycw.

According to Meireles et al. (2012), is nearly independent of dc/h for rather steep slopes (45
53); thus, Meireles et al. (2012) proposed a relationship for as a function of L/Li only for L/Li < 1.0.
Fig. 4 illustrates as it relates to L/Li and h/dc for model B. peaks a L/Li 1.0, but it varies between
1.0 and 1.20 for all models tested. Downstream of L/Li = 1.0, decreases rapidly and then trends
towards a constant value for a given h/dc. From the data obtained from models A, B, and C,
appears to be a function of L/Li, , and h/dc.

1.25 h/dc=.035-.10 .11-.20

.21-.40 .41-85
1.20 us.035-.10
us .21-.40
us .11-.20
us .41-.85

1.15

1.10
1.05
1.00
0.0 1.0 2.0 3.0 4.0 5.0 6.0
L/Li
Figure 4. Example of as it relates to the normalized length, L/Li for 0.027 h/dc 1.1 for model B.

Using a multi-variant, non-linear analysis for the data between 0.1 L/Li 1.0 and 0.027 h/dc 1.1,
Eq. 8 was developed, yielding a R2 = 0.82 and a standard error of 0.018.

0.128 sin
h L
0.632

1.0 1.025 1.0 (8)

dc Li

for 0.1 (L/Li) 1.0, 0.027 h/dc 1.1, and 10 30

The resulting error between the predicted and measured is approximately 5%. Another multi-
variant analysis was used to develop a relationship for for L/Li > 1.0 and 0.027 h/dc 1.1. A
constraint was set such that Eq. 8 would equal Eq. 9 at L/Li = 1.0.

0.128 sin
h L
2.37

1.0 (1.025 1.0) 0.718 0.723 (9)
dc Li

for L/Li > 1.0 and 0.027 h/dc 1.1

This analysis yielded a R2 = 0.78 and a standard error of 0.022. Again, the error between the
predicted and measured was within 5% for the majority of the data. The majority of the data
outside of the confidence intervals occurred between 1.0 L/Li 2.0 where the flow is more erratic in
behaviour and thus may have attributed to the error. Additionally, air concentrations and velocities
were not measurable with the fiber optic probe in the lower portion of the depth-velocity profiles
between 1.0 L/Li 2.0 either. Boes and Hager (1998), Chanson (2000), and Hunt and Kadavy
(2010) indicate velocity profiles trend toward a one-sixth power law distribution near the inception
point. As a result, these missing points in the velocity profiles were assumed to follow a one-sixth
power-law distribution. Consequently, this assumption may have contributed to the observed error.
Another observation that should be noted is if Eq. 9 were simplified to Eq. 10, such that is assumed
constant for L/Li > 1.0, then one would expect no more than an additional 2% error.

0.128 sin
h
1.025 for L/Li > 1.0 and 0.027 h/dc 1.1 (10)
dc

With all of the parameters either selected (i.e. h, , etc.) or calculated through a series of relationships,
Eqs. 1, 2, 4, 5, 6, 8, 9, and/or 10, the energy loss along any point within the stepped chute may be
determined.

3.4. Energy Loss

Stepped spillways have been long noted in studies to provide significant energy loss. The significance
of the energy loss is attributed to many (i.e. Chanson 1994, 2002; Boes 2000; Matos 2003; Boes and
Hager 2003; Meireles and Matos 2009; Hunt and Kadavy 2009; Meireles et al. 2012; and Frizell and
Svoboda 2012) including the authors to believe that stepped spillways can lead to a reduction in the
length of the downstream stilling basin when compared to traditional smooth chute spillways. The
USBR (1973) outlines the design of stilling basins based on the incoming Froude number, Fr. Fr is a
function of velocity and flow depth as shown in Eq. 11, but may be rewritten in terms of ycw and dc
2/3
q
given d c 1 / 2 :
g
1 .5
V y
Fr 0 .5
cw (11)
(gy ) dc

Frizell and Svoboda (2012) indicate fairly high Fr for stepped chutes modelled in their research. A
Type III stilling basin was identified as an appropriate stilling basin for containing the hydraulic jump for
these higher velocities and Fr. It should be noted that Frizell and Svoboda (2012) introduced high
velocities down the stepped chute using a jet box. Because of the implementation of the jet box, Li
was not allowed to developed naturally as one would expect. From models A, B, and C, the flow
developed naturally as it descended the chute, and the resulting velocities measured no more than 7.5
m/s. These velocities were considerably less than those reported by Frizell and Svoboda (2012).
Consequently, models A, B, and C yielded lower Fr than those reported by Frizell and Svoboda
(2012). As shown by Fig. 5, the Fr from models A, B, and C, ranged from 3.3 to 5.4, indicating either a
Type IV (i.e. 2.5 Fr 4.5) or Type III (i.e. Fr > 4.5) stilling basin is adequate for containing the
hydraulic jump in the basin according to USBR (1973) criteria. Type IV stilling basins are much longer
than Type III stilling basins for a given Fr (USBR 1973). Frizell and Svoboda (2012) indicate that a
Type III has a little more applicability (i.e. Fr > 4.0) than originally presented by the USBR (1973).
Given Fig. 5 indicates that Fr may be as low as 3.3 for large step heights, further investigation on the
applicability of the Fr for Type III stilling basins is warranted. A Type IV stilling basin may prove costly,
so if a Type III stilling basin does not satisfy the design requirements, a new economically feasible
stilling basin design might be developed.
 7.0 Model A Model B
Model C Eq. 11: 14.0 deg
6.0 Eq. 11: 18.4 deg Eq. 11: 26.6 deg

5.0
r
F
4.0
3.0
2.0
0.0 0.2 0.4 0.6 0.8 1.0
h/dc
Figure 5. Fr as it relates to h/dc and for L/Li 1.0.

4. CONCLUSIONS
A multi-year, large-scale physical model study of stepped chutes was conducted over a range of h, ,
and q at the USDA-ARS HERU in Stillwater, OK, USA. Relationships for y, ycw, and were
developed from the data obtained from this present study. y and ycw were found to be dependent on
h/dc, , and/or L/Li, and they were independently verified with data from literature. The data for
yielded a maximum value of approximately 1.2 near L/Li =1.0. For L/Li > 1.0, decreased rapidly
before trending towards a constant, with a minimum value of approximately 1.0. Data from this study
indicate is dependent on L/Li, h/dc, and contradicting previous research by Meireles et al. (2012)
that is independent of h/dc. Additionally, this study indicates that large step heights can have an
impact on the resulting Froude number and subsequent stilling basin design. Additional research is
recommended to further investigate the applicability of traditional stilling basins for stepped chutes.

5. REFERENCES
Boes, R. M. (2000). Two-phase flow and energy dissipation on stepped spillways. Proc. of
Association of Dam Safety Officials Annual Meeting, ASDSO, Lexington, Kentucky.
Boes, R. M. and Hager, W. H. (1998). Fiber-optical experimentation in two-phase cascade flow. In
Proc. Intl. RCC Dams Seminar. (K. Hanson, ed.) Denver, Colo: Schnabel Engineering.
Boes, R. M. and Hager, W. H. (2003). Two-phase flow characteristics of stepped spillways. J. of
Hydr. Eng., 129(9), 661-670.
Bung. D. B. (2011). Developing flow in skimming flow regime on embankment stepped spillways. J.
Hydr. Res., IAHR, 49(5):639-648.
Chanson, H. (1994). Hydraulics of skimming flows over stepped channels and spillways. J. Hydr.
Res., IAHR, 32(3), 445-460.
Chanson, H. (2000). Discussion of characteristics of skimming flow over stepped spillways. J. of Hydr.
Eng., 126(11): 862-865.
Chanson, H. (2002). The Hydraulics of Stepped Chutes and Spillways. Steenwijk, The Netherlands:
A. A. Balkema Publishers.
Chanson, H. and Toombes, L. (2002). Energy dissipation and air entrainment in a stepped storm
waterway: an experimental study. J Irrigation and Drainage Engrg., ASCE, 128(5):305-315.
Chow, V. T. (1959). Open-channel hydraulics. McGraw-Hill, Book Company, Inc., Boston, MA.
Felder, S. and Chanson, H. (2007). Turbulence and turbulent length and time scales in skimming
flows on a stepped spillway. Dynamic similarity physical modeling and scale effects. Report No.
CH64/07, Hydraulic Model Report CH series, Division of Civil Engineering, Univ. of Queensland,
Brisbane, Australia.
Frizell, K. W. and Svoboda, C. D. (2012). Performance of type III stilling basins stepped spillway
studies do stepped spillways affect traditional design parameters? United States Dept. of the
Interior, Bureau of Reclamation, Hydraulic Laboratory Report HL-2012-02, Denver, CO.
Hunt, S. L. and Kadavy, K. C. (2009). Energy dissipation for flat-sloped stepped spillways using new
inception point relationship. Proc. of Association of Dam Safety Officials Annual Meeting, ASDSO,
Lexington, Kentucky.
Hunt, S. L. and Kadavy, K. C. (2013). Inception point for embankment dam stepped spillways. J.
Hydr. Eng., 138(9), 796-802.
Hunt, S. L., Kadavy, K. C., and Hanson, G. J. (2013). New flow depth relationships for embankment
dam stepped spillway design. Dam Eng.
Matos, J. (2003). Roller compacted concrete and stepped spillways. From new dams to dam
rehabilitation. Proc. International Congress on Conservation and Rehabilitation of Dams: Dam
Maintenance and Rehabilitation, 553-560.
Matos, J., Frizell, K. H., Andr, S., and Frizell, K. W. (2002). On performance of velocity
measurement techniques in air-water flows. Wahl, T. L., Pugh, C. A., Oberg, K. A., and Vermeyen, T.
B. (Eds.) Proc. Hydraulic Measurements and Experimental Methods, ASCE, Estes Park, CO.
Meireles, I. and Matos, J. (2009). Skimming flow in the nonaerated region of stepped spillways over
embankment dams. J. Hydr. Eng., 135(8), 685-689.
Meireles, I., Renna, F., Matos, J., and Bombardelli, F. (2012). Skimming, non-aerated flow on
stepped spillways over RCC dams. J. Hydr. Eng., 138(10), 870-877.
Ohtsu, I., Yasuda, Y., and Takahashi, M. (2001). Discussion of onset of skimming flow on stepped
spillways. J. Hydr. Eng., 127(6), 522-524.
Ohtsu, I., Yasuda, Y., and Takahashi, M. (2004). Flow characteristics of skimming flows in stepped
channels. J. Hydr. Eng., 130(9), 860-869.
Pfister, M., and Hager, W. H. (2011). Self-entrainment of air on stepped spillways. Inl J. of
Multiphase Flow, 37(2), 99-107.
Takahashi, M., Gonzalez, C. A., and Chanson, H. (2006). Self-aeration and turbulence in a stepped
channel: influence of cavity surface roughness. Intl. Jl.Multiphase Flow, 32:1370-1385.
Ward, J. (2002). Hydraulic design of stepped spillways. PhD Dissertation. Colorado State University,
Ft. Collins, Colorado.
United States of the Interior Bureau of Reclamation (USBR). (1973). Design of Small Dams. U. S.
Government Printing Office, Denver, CO.