World Environmental and Water Resources Congress 2007: Restoring Our Natural Habitat © 2007 ASCE

Common Modeling Mistakes Using HEC-RAS

Presented by:
Iwan M. Thomas1
David T. Williams2, Ph.D., P.E., D.WRE
Downloaded from ascelibrary.org by OLD DOMINION UNIVERSITY on 06/27/14. Copyright ASCE. For personal use only; all rights reserved.

1, 2
Respectively, Project Engineer and National Technical Director for Water Resources,
PBS&J, 9275 Sky Park Court, suite 200, San Diego, CA 92123

Abstract

The U.S. Army Corps of Engineers River Analysis System (HEC-RAS), developed by
the Hydrologic Engineering Center, has long been the standard for one- dimensional
hydraulic modeling. Despite the wide spread use of the program, even seasoned
engineers can fall victim to some common modeling mistakes using HEC-RAS. Four
potential modeling mistakes and oversights will be examined herein:

(1.) Necessary adjustment of channel bottom elevations for internal bridge cross sections
(termed BU and BD in HEC-RAS verbiage), which are copied from the upstream and
downstream sections of bridges (Sections 2 and 3 of the bridge sequence) are often
ignored and can result in the incorrect computation of the water surface profile within the
structure and immediately upstream.
(2.) Frequently ignored expansion and contraction coefficients for super-critical flow
(most will use just the subcritical flow coefficients) and the drastic impact that the
incorrect application can have on the location of hydraulic jumps when in the mixed flow
mode.
(3.) The overlooked importance of proper cross section spacing through drop structures.
(4.) The need to trim cross sections to a reasonable elevation above the water surface
elevation so as not to impact the critical depth computation process is not widely
understood and rarely practiced.

The impacts on water surface elevation computation for these common errors are not
insignificant and can be easily avoided if the engineer is aware of the correct modeling
technique. In a field where calculations require accuracy to 0.01 feet, even minor errors
in modeling technique can result in unnecessary high cost solutions. The following paper
will document these situations through discussion of simple applications and the
mechanics of calculation performed by HEC-RAS as well as presentation of methods to
improve modeling accuracy. These are subtle errors that even experienced modelers may
fall victim to.

Impact of Internal Bridge Cross Section Elevation Adjustment

Hydraulic Modeling of flow through bridges is one of the common applications for HEC-
RAS. However, despite the frequency of this type of analysis, an error in the geometry of
the bridge cross sections can often be found that may result in computation of incorrect

Copyright ASCE 2007 World Environmental and Water Resources Congress 2007
World Environmental and Water Resources Congress 2007
 World Environmental and Water Resources Congress 2007: Restoring Our Natural Habitat © 2007 ASCE

water surface elevations in the vicinity of the bridge, usually in an area of particular
importance within the model. This becomes particularly important if running in mixed
mode and locating a hydraulic jump within the vicinity of the bridges. As is described by
the HEC-RAS Users Manual (HEC, 2002), the accepted method of analysis for bridges
utilizes a four cross-section sequence in order to adequately reflect contraction and
expansion zones upstream and downstream of the bridge, respectively. As is shown in
Figure 1, (reproduced from the HEC-RAS Users Manual) the four cross sections are
Downloaded from ascelibrary.org by OLD DOMINION UNIVERSITY on 06/27/14. Copyright ASCE. For personal use only; all rights reserved.

referred to by the numbers one through four from downstream to upstream.

Figure 1. Standard four bridge cross section layout.

The most downstream and upstream cross sections, numbers one and four, respectively,
reflect conditions where flow is no longer affected by the structure. Cross sections 2 and
3 are to be taken immediately downstream and upstream, respectively, at the toe of the
bridge embankment and reflect the natural ground. These two bounding cross sections
are copied by HEC-RAS to become the internal bridge cross sections (termed BU and BD
in HEC-RAS usage). When these cross sections are copied, no adjustment is made to the
ground points. A “stair stepping” of the ground profile results when horizontal flat
profiles are created upstream and downstream of the bridge. The ground slope through
the structure is artificially increased and can cause a dramatic impact in the computed
water surface elevation even in cases where the bridge has little or no obstruction of the
flow. HEC-RAS has a seldom used feature that allows the user to adjust the geometry of
the internal bridge cross sections to match the actual ground within the bridge and
eliminate errors that result from “stair stepping” of the ground profile. This feature can
be used to create an entirely new set of station and elevation points or adjust the
elevations of the copied cross sections (i.e., BU and BD cross sections).

Copyright ASCE 2007 World Environmental and Water Resources Congress 2007
World Environmental and Water Resources Congress 2007
 World Environmental and Water Resources Congress 2007: Restoring Our Natural Habitat © 2007 ASCE

To illustrate the potential impact of adjusting the elevations of the internal bridge cross
sections in a HEC-RAS model to match the ground profile, a sample reach was created.
The sample reach is a simple prismatic channel with a bottom width of 100 feet, side
slopes of 2:1 (H:V), a channel slope of 0.0025, and Manning n value of 0.017. Cross
sections were spaced at 100 foot intervals. In order to isolate the effect of the “stair
stepping” of the channel, it was determined that any bridge added would be input such
that it did not obstruct the flow. Three different geometries were used for comparison:
Downloaded from ascelibrary.org by OLD DOMINION UNIVERSITY on 06/27/14. Copyright ASCE. For personal use only; all rights reserved.

(1) No bridge was added to the channel.

(2) Bridge was added and internal cross sections (BU and BD) were adjusted according to
the channel slope.
(3) Bridge was added with no adjustment to internal bridge cross sections.

The bridge added to the model was 20 feet wide and located such that the upstream and
downstream faces were 40 feet from their respective bounding cross sections. For each
of the models water surface elevations were computed for a discharge of 5,000 cfs.
Figure 2 shows the water surface profiles in the vicinity of the bridge for each of the three
geometries (A) no bridge, (B) bridge with adjustment to internal cross sections, and (C)
bridge without adjustment to internal cross sections.

Figure 2. Water surface profile comparison, (A) no bridge modeled, (B) bridge with adjustment to
internal cross sections, (C) bridge without adjustment to internal cross sections.

Theoretically, there should be no impact from a bridge for which the low chord of the
deck is above the water surface elevation and which contains no abutments or piers that
obstruct the flow; however, as a result of the bridge routine computation, a maximum
increase of 0.03 feet is recorded at the upstream bounding cross section when the bridge
is added despite adjustment of internal cross sections. If the bridge is added and there is
no adjustment to the internal cross sections, an additional increase of 0.45 feet is
computed (0.48 feet total) at the upstream bounding cross section. This additional

Copyright ASCE 2007 World Environmental and Water Resources Congress 2007
World Environmental and Water Resources Congress 2007
 World Environmental and Water Resources Congress 2007: Restoring Our Natural Habitat © 2007 ASCE

increase results in increased water surface elevations to at least 500 feet upstream of the
bridge.

As can be seen by the increase in water surface elevation reported above and the impact
on water surface profiles shown in Figure 2, the impact of adjustment of internal cross
sections is non-trivial even in simple applications.
Downloaded from ascelibrary.org by OLD DOMINION UNIVERSITY on 06/27/14. Copyright ASCE. For personal use only; all rights reserved.

In real world applications, there is a large degree of variability in the width of bridge
embankments which are a function of the width of the road, height of the embankment
above natural ground and the side slope of the embankment. Thus upstream and
downstream bounding cross sections may be a considerable distance apart, which,
depending on the slope of the river, may result in a more pronounced “stair stepping” of
the ground profile through the structure resulting in even more dramatic effects on
computed water surface elevation than those exhibited. Adjustment of the internal bridge
cross sections is a simple operation if the engineer is aware of the feature and should be
considered proper modeling technique and practiced whenever possible.

Impact of Expansion and Contraction Coefficients on Water Surface Profiles Under

Super Critical Flow Conditions

The majority of analysis conducted using HEC-RAS is done in the sub-critical regime.
As a result, engineers generate a level of familiarity with the features and inputs that
correspond to that regime. This familiarity can result in potential misapplication of some
model inputs, particularly contraction and expansion coefficients used in the supercritical
regime. Most engineers have the contraction and expansion coefficients of 0.3 and 0.5 or
other values to be used in most areas with constriction (e.g., bridges and/or culverts) or
expansion of flow committed to memory. These coefficients however are generally
applicable to the subcritical regime. In the supercritical regime, coefficients for
contraction and expansion should be much lower: “around 0.05 and 0.1 respectively”
according to the HEC-RAS Hydraulic Reference Manual (HEC, 2002) for typical
bridges. These recommendations for supercritical flow are much less documented and
not well known and, as a result, they are often not utilized.

Furthermore, in cases of true mixed flow, it is unclear which set of coefficients should be
used since they vary by flow regime. The application of these different sets of
coefficients can result in dramatic changes in the water surface profile within a reach.
Consider this example reach: 5000 cfs peak discharge, a rectangular concrete channel
(Manning n of 0.013) with a bottom width of 100 feet constricting over 100 feet in the
downstream direction to a similar channel with a bottom width of 80 feet. Cross sections
were modeled at 25 foot intervals. As a result of the constriction, a hydraulic jump is
induced.

Figure 3 shows the comparison of computed water surface profiles using each of the three
likely sets of coefficients (contraction/ expansion): supercritical (0.05/ 0.1), subcritical
(0.3/ 0.5), and unchanged (0.1/ 0.3 for natural channels).

Copyright ASCE 2007 World Environmental and Water Resources Congress 2007
World Environmental and Water Resources Congress 2007
 World Environmental and Water Resources Congress 2007: Restoring Our Natural Habitat © 2007 ASCE
Downloaded from ascelibrary.org by OLD DOMINION UNIVERSITY on 06/27/14. Copyright ASCE. For personal use only; all rights reserved.

Figure 3. Water surface profile comparison.

It is clear from the profiles in Figure 3 that the effects of modification of coefficients are
significant. These effects are summarized in Table 1.

Table 1. Summary of effect of varying coefficients on water surface profile

Coefficient of Location of Hydraulic Jump Maximum Depth in

Contraction/ Expansion (River Station) Hydraulic Jump (feet)
0.05/ 0.1 2250 5.98
0.3/ 0.5 2375 6.54
0.1/ 0.3 2275 6.09
Note: contraction begins at river station 2200 and is complete at river station 2100

From Table 1, it should be noted that the change in coefficients can result in the
movement of the computed location of a hydraulic jump 125 feet and cause an increase in
the height of the hydraulic jump of 0.56 feet at a discharge of 5000 cfs.

As a result of the uncertainty in selection of coefficients exemplified by Figure 3, in order

to determine the true regime of the study area, it may be necessary to create two separate
models; one for super critical flow and one for subcritical flow. Each of these models
can be run separately with the appropriate coefficients; if the regime is maintained
throughout the model, then results of that model should be acceptable. If however, both
models tend to the mixed flow regime, then engineering judgment must be exercised in
the application of coefficients and the interpretation of results.

Copyright ASCE 2007 World Environmental and Water Resources Congress 2007
World Environmental and Water Resources Congress 2007
 World Environmental and Water Resources Congress 2007: Restoring Our Natural Habitat © 2007 ASCE

Modeling Drop Structures in HEC-RAS

When modeling drop structures in HEC-RAS it is common to use a series of cross

sections to capture the features of the proposed or existing structure and approximate the
water surface elevation profile through the structure. Water flowing through a drop
structure inherently can be seen to have momentum characteristics, because HEC-RAS
performs calculations based on energy, not momentum there are limitations in
Downloaded from ascelibrary.org by OLD DOMINION UNIVERSITY on 06/27/14. Copyright ASCE. For personal use only; all rights reserved.

applicability of the computed profile.

Upstream and downstream of a typical drop structure water is flowing in the subcritical
regime. As water flows through the structure it will pass through critical depth, enter the
supercritical flow regime, and then before exiting will go through a hydraulic jump and
re-enter the subcritical flow regime. Thus, hydraulic modeling in HEC-RAS must be run
under the mixed flow option.

Additionally through the drop structure the water surface elevation and velocity are
changing rapidly thus, cross section spacing is critical for producing an accurate profile.
In the approach section for a drop structure similar to the one shown in Figure 4 with a
vertical drop into the stilling basin a cross section should be located at the upstream side
of the drop, the next cross section upstream should be very close, only a few feet.
Subsequent cross section spacing moving upstream should slowly increase until regular
spacing is reached. Within the structure itself cross sections should be located at and
slightly upstream of features of the structure such as baffle blocks and the downstream
end of the stilling basin so as to bound them. Downstream of the structure cross section
spacing can return to normal intervals quickly. For drop structures that use an ogee
spillway into the stilling basin cross section spacing is very similar to that used when a
vertical drop is present however many additional cross sections should be added along
the spillway to accurately represent the smooth bottom profile.

A typical drop structure utilizing a vertical drop into the stilling basin is shown in Figure
4. Approximate cross section locations are also shown.

Figure 4. Typical drop structure with recommended cross section locations

Copyright ASCE 2007 World Environmental and Water Resources Congress 2007
World Environmental and Water Resources Congress 2007
 World Environmental and Water Resources Congress 2007: Restoring Our Natural Habitat © 2007 ASCE

Using HEC-RAS with recommended cross section spacing through a drop structure will
provide a generally accurate water surface profile upstream and downstream of the
structure; however, but it should not be considered a substitute for a detailed external
analysis. As previously mentioned, because of the necessary momentum considerations
not determined by HEC-RAS, the water surface profile through the structure should be
considered approximate and external analysis using equations specifically developed for
Downloaded from ascelibrary.org by OLD DOMINION UNIVERSITY on 06/27/14. Copyright ASCE. For personal use only; all rights reserved.

the particular structure is required for determination of profile information for use in
detailed design. HEC-RAS can be a very useful tool in preliminary design because of its
ease of use and where multiple simulations can be conducted rapidly such as the
sensitivity to tailwater conditions.

Modeling of drop structures in HEC-RAS requires great care in proper cross section
spacing and application of results. Figure 4 gives the designer a good idea of the required
locations and spacing of cross sections.

Impact of Maximum Cross Section Height on Critical Depth Computation

HEC-RAS has two methods for determining critical depth for a given cross section. The
first (parabolic) method computes the specific energy at various elevations, fits a
parabolic curve to them, and reports the local minimum that it finds as critical depth.
This method is generally sufficiently accurate, particularly for cross sections that have a
regular geometry (a defined channel with small overbank areas). However, if the cross
section has wide flat overbanks or ineffective flow areas or levees that could be
overtopped, then it is possible that there will be multiple local minimums and the
parabolic method may find (possibly erroneously) only one local minimum rather than
the global minimum. For cross sections of this nature, the second (secant) method will
provide a more reliable critical depth computation. The secant method selects 30
different elevations at which to calculate the specific energy based on the total height of
the cross section and then searches the results for all local minimums.

Figure 5 shows the relation of depth versus specific energy for two different types of
cross sections. A simplified specific energy curve for a cross section that is well behaved,
with only a single minimum energy, is shown in Figure 5-A. Figure 5-B shows a
generalized example of specific energy curve for a cross section with multiple local
minimums, potentially the result of very wide overbank areas or ineffective flow areas.

Copyright ASCE 2007 World Environmental and Water Resources Congress 2007
World Environmental and Water Resources Congress 2007
 World Environmental and Water Resources Congress 2007: Restoring Our Natural Habitat © 2007 ASCE
Downloaded from ascelibrary.org by OLD DOMINION UNIVERSITY on 06/27/14. Copyright ASCE. For personal use only; all rights reserved.

Figure 5. Relation of depth versus specific energy, (A) single local minimum, (B) multiple local
minimums.

The cross section displayed in Figure 6 has both an ineffective flow area and levee that
may be overtopped resulting in a specific energy curve that will contain multiple local
minimums. The cross section in Figure 6 also displays the measurement of the maximum
overall cross section height (hmax) and the maximum channel height (hch), important
parameters in determination of the specific energy curve for that cross section.

Ineffective
Levee Flow Limit

hmax

hch

Figure 6. HEC-RAS cross section plot with overall maximum height (hmax) and maximum channel
height (hch).

For cross sections with a maximum overall height less than 1.5 times the maximum
channel height, the 30 elevations selected for specific energy computation are evenly
distributed. However, if the maximum overall cross section height is greater than 1.5
times the maximum channel height, then 25 of the 30 elevations are evenly distributed
between the minimum cross section elevation and the maximum channel elevation. The

Copyright ASCE 2007 World Environmental and Water Resources Congress 2007
World Environmental and Water Resources Congress 2007
 World Environmental and Water Resources Congress 2007: Restoring Our Natural Habitat © 2007 ASCE

remaining five elevations used to generate the specific energy curve are evenly
distributed between the maximum channel elevation and the maximum cross section
elevation. Thus, for a cross section that has been cut with a very high maximum height
and a relatively shallow channel, in high flow situations, where critical depth is above the
channel, the critical depth determination is performed using very few data points.

As a consequence of the low resolution used to calculate the specific energy curve, at
Downloaded from ascelibrary.org by OLD DOMINION UNIVERSITY on 06/27/14. Copyright ASCE. For personal use only; all rights reserved.

high flows, there is a significant increase in uncertainty surrounding critical depth

computation results if they are indeed in the overbank areas. In well behaved cross
sections with specific energy curves similar to those seen in Figure 5-A, the uncertainty
will be minimized. In cross sections that have wide flat overbanks and overtopping
ineffective flow areas or levees, they may have specific energy curves similar to those
seen in Figure 5-B. If this is the case, the accuracy of determination of critical depths
will decrease dramatically. With so few computation points, it is unlikely that the
multiple local minimums and general variation in the specific energy curve will be
captured.

For such cases, utilizing the secant method of specific energy computation introduces the
possibility that critical depth computation may be influenced by the overall height of the
cross section, even in cases where the top of the cross section is well above the water
surface elevation. Due to ease of the use of GIS to generate cross sections, the possible
computational error introduced with cross sections that extend vertically well beyond the
water surface elevations further emphasizes the need for the application of engineering
judgment in all phases of modeling, particularly cross section layout and modification.

In order to avoid potential errors in critical depth computations, it is considered good

modeling practice to vertically trim cross sections to elevations that are reasonably low
but still above the design discharge water surface elevation.

Summary

Hydraulic modeling conducted using HEC-RAS produces water surface elevations that
are reported and interpreted with an accuracy of 0.01 feet. On such a scale even minor
changes to modeling methods can produce noticeable changes in computed water surface
elevations.

When modeling bridges an often overlooked feature allows the user to adjust internal
bridge cross sections to more accurately represent the ground profile. Stair stepping of
the ground profile that results from the default method of simply copying internal bridge
cross sections from upstream and downstream can cause significant errors to computed
water surface elevations in the vicinity of the bridge.

Appropriate expansion and contraction coefficients used in HEC-RAS in the supercritical

flow regime are far less documented and well known that their counterparts for
subcritical flow. The selection of these coefficients can have a profound impact on water
surface elevation computations and the determination of the location of hydraulic jumps

Copyright ASCE 2007 World Environmental and Water Resources Congress 2007
World Environmental and Water Resources Congress 2007
 World Environmental and Water Resources Congress 2007: Restoring Our Natural Habitat © 2007 ASCE

when they are induced. Special care must also be taken in selecting coefficients and
application of results in cases of true mixed flow.

When modeling drop structures in HEC-RAS using a series of cross sections, spacing and
location of those cross sections is of great importance to accurately represent the water
surface elevations that will be observed upstream and downstream of the actual structure.
It is also important to recognize that because of the momentum considerations not
Downloaded from ascelibrary.org by OLD DOMINION UNIVERSITY on 06/27/14. Copyright ASCE. For personal use only; all rights reserved.

determined by HEC-RAS the water surface profile through the structure should be
considered approximate only and not used as the basis for design.

The secant method of critical depth computation in HEC-RAS is considered the more
accurate of the two available methods. However, critical depth determination can be
susceptible to error because the calculation method is dependant on geometric parameters
(channel and overall height) of the cross section that should have no bearing on the actual
determination. Simply trimming the cross section to a reasonably low elevation but still
above the design discharge water surface elevation can minimize the potential error in
critical depth computation.

The common errors in HEC-RAS modeling discussed above all have the potential to
cause inconsistencies in the computed water surface elevations of hydraulic models.
These inconsistencies can be minor or, in some cases, extreme. However, in many cases,
fairly simple solutions can be applied to eliminate or minimize these potential
computational errors as long as the engineer is aware of these potential mistakes and the
features that best combat them.

Reference

Roberson, John A., Cassidy, John J., Chaudhry, M. Hanif, Hydraulic Engineering, John
Wiley and Sons, Inc, New York, 1998.

U.S. Army Corps of Engineers, Hydrologic Engineering Center, HEC-RAS User’s

Manual v 3.1, November 2002.

U.S. Army Corps of Engineers, Hydrologic Engineering Center, HEC-RAS Hydraulic

Reference Manual v 3.1, November 2002.

U.S. Department of Agriculture, National Resources Conservation Service, Introduction

to HEC-RAS Course, “Modeling Drop Structures in HEC-RAS Version 3.1,” March
2003.

Copyright ASCE 2007 World Environmental and Water Resources Congress 2007
World Environmental and Water Resources Congress 2007