Skip to contentUnited States Department of Transportation - Federal Highway AdministrationFHWA HomeFeedback

Hydraulics Engineering

<< Previous Contents Next >>

Design For Fish Passage at Roadway - Stream Crossings: Synthesis Report

6 Design Considerations, Methods and Tools

How to use this chapter

  • Understand important biological, geomorphic and hydraulic considerations when providing for fish passage
  • Appreciate the constraints to providing for fish passage at culverts
  • Learn of the three main categories of design for fish passage: Geomorphic Simulation, Hydraulic Simulation and Hydraulic Design
  • Refer to assembled computational procedures for determining flow depth and the stability or mobility of streambed material
  • Learn what computer software is readily available to assist in assessment and design

6.1 Objectives

A culvert designed for fish passage must also meet applicable flood conveyance requirements, such as allowable headwater elevations during the design flood and local Federal Emergency Management Agency (FEMA) regulations.

Many factors will determine the suitability of any particular set of fish passage design criteria for a culvert installation, replacement or retrofit. With fish passage as the overall goal, economics, site logistics, regulatory requirements and roadway characteristics may dictate a particular design procedure. The following categories have been developed to aid in the classification and selection of design approaches based on project goals. These goals are based on biological, geomorphic and hydraulic considerations.

Category 1: Geomorphic Simulation, also known as stream simulation, approaches recreate or maintain natural stream reach geomorphic elements including slope, channel-bed width, bed materials and bedform by using the reference reach. The approaches are based on the assumption that crossings approximately matching natural conditions will readily pass fish that are moving in the natural channel. This approach has five benefits: (1) it will provide passage for fish more readily than for much narrower spans; (2) it may provide passage for other aquatic and some terrestrial organisms; (3) for discharges less than approximate bankfull values, it will not increase downstream channel velocities; (4) for discharges exceeding bankfull values, it will increase downstream velocity less than for culverts with narrower spans, and (5) maintenance requirements (debris clearing, streambed manipulation) should be less than for narrower spans.

Category 2: Hydraulic Simulation techniques utilize embedded culverts, natural or synthetic bed mixes and natural roughness elements such as oversized rock to provide hydraulic conditions conducive to fish passage. These techniques operate on the assumption that providing hydraulic diversity similar, but not identical, to that found in natural channels will create a fish passable structure without checks for excessive velocity or turbulence. Hydraulic Simulation will generally have the benefit of creating smaller spanning structures that have a reduced cost when compared to Geomorphic Simulation.

Category 3: Hydraulic Design techniques create water depths and velocities that meet the swimming abilities of target fish populations and life stages during specific periods of fish movement. Hydraulic Design is most often used in retrofit projects. General considerations include the effect of culvert slope, size, material and length. Flow control structures such as baffles, weirs, or oversized substrate are commonly utilized to create acceptable hydraulic conditions. This technique generates a smaller diameter culvert that keeps cost of materials to a minimum. Installation costs, however, are highly variable due to unique designs of baffles, weirs, steps or other controls. Hydraulic Design produces a less conservative design for fish passage than Geomorphic or Hydraulic Simulation.

All of the fish passage design methods seek to allow passage when fish are believed to be moving in the natural stream system. Note that none of the methods provide for unaltered flow hydraulics during flood events used for design by state DOTs.

Infrastructure Safety and Service Life
Culverts must also be built with consideration of safety and service life. Larger span culverts will have a greater cross-sectional area for passing flood events. Hydraulically designed culverts will have a smaller initial cost, but require additional maintenance and monitoring to avoid debris accumulation (Bates et al. 2003).

6.1.1 Biological Considerations Fish Passage Requirements

Crossing designs create different levels of stream reach connectivity. In general, Geomorphic Simulation creates the greatest connectivity, followed by Hydraulic Simulation and Hydraulic Design. In all cases, passage is presumed not to occur for discharges exceeding approximately bankfull conditions. The emphasis for fish passage design, therefore, is for discharges much smaller than those used in flood conveyance checks. A few pertinent questions can significantly narrow design option selection based on project goals.

What are the species of fish and life stages for which passage should be provided?

This question requires consultation with the fisheries biologist team member and likely consultation with natural resource agencies. The answer may depend upon, for example, regulations (e.g., Endangered Species Act), the desire to exclude invasive species, or economics (sport fishing considerations).

Hydraulic Designs can be completed to cater to a particular fish species and life stage; however, such a structure may provide a barrier to weaker swimming fishes at some or all flows.

What is the weakest swimming fish species and life stage for which passage is required?

Example: Adult Salmon; Juvenile Salmon; resident trout; benthic fish; all species and life stages present.

All techniques are designed to ensure fish passage; however, Geomorphic and Hydraulic Simulation approaches will allow passage for a wider variety of fish species and other aquatic organisms.

Do we know the swimming abilities, behaviors and timing of these species?

See Section 2.2.

At what flows, and time periods are these fish migrating? What is the allowable delay?

Design may depend on the timing of fish migration and relative flows. Delay impacts may be less crucial for resident fish than a spawning salmon. This problem can be compounded, for example, by several culverts in series or culverts that provide passage only after short delays. Ecological Significance

Further consideration should be paid to the ecological significance of the roadway-stream crossing. The only way to truly preserve habitat at a crossing is to use a bridge or open bottom structure. Figure 6.1 shows a representation of the range of ecological solutions available at a roadway-stream crossing. Extreme ends of the spectrum include traditional design for flood capacity, and bridges or road removals that will permit valley and floodplain processes.

Diagram of the range of ecological solutions. It is a double-headed arrow with a minus sign on the left side and a plus sign on the right one. It indicates the order of ecological solutions for culvert installations. From left to right: Flood capacity; Hydraulic design: pass target species/ life stage; Hydraulic simulation: pass all fish, sediment, and debris; Geomorphic simulation: pass all fish, aquatic organism, sediment and debris; Floodplain continuity; and Permit valley and floodplain processes. As you move towards the right there is increased ecological connectivity.
Figure 6.1 Range of ecological solutions at culvert installations (adapted from Gubernick 2006)

It is recommended that new culvert designs incorporate auxiliary barrels on the floodplains at the roadway crossing. Adding barrels at an elevation higher than bankfull will provide a flowpath for waters when discharge exceeds bankfull values. Such relief will decrease the main channel velocity and scour potential. This procedure is described in more detail in sections 7.2.4, 7.3.7 and 8.2.1.

6.1.2 Geomorphic Considerations

Site geomorphology is another important consideration in design for fish passage. Slope, channel location, channel stability and bed material are all examples of geomorphic elements that affect design selection. For example, installations located at slope breaks or in sediment sensitive areas may have a high propensity to degrade, aggrade or elicit a change in channel conditions, eventually creating another barrier or destroying valuable habitat (Bates et al. 2006). Form and Key Features

Channel form and key features can aid in understanding channel processes including sediment transport, channel stability, and channel migration (Bunte and Abt 2001). Key features describe stream elements such as large woody debris (LWD), rock, vegetation, or channel confinement, all of which can play a large part in channel form and stability (Montgomery and Buffington 1998). While features such as LWD may be prominent in some channels, exact placement and development of such influences, and associated features, may be fairly unpredictable (Montgomery and Buffington 1993), and an understanding of overall influence and importance will be essential. Stability

Channel stability refers to the likelihood that a channel will retain its current placement, gradation, shape and form over time. Channels in highly entrenched mountain streams will be less likely to show lateral or vertical changes over time, while meandering valley streams may show great variation both laterally and vertically in response to minimal inputs (Montgomery and Buffington 1993; Rosgen 1994). Morphological Adjustment Potential/Ability

A crossing can be built to buffer for slight lateral and vertical channel adjustments. Although this can increase the size and initial cost of a structure, benefits can include decreased maintenance requirements and increased design life. Rigid Structure in Dynamic Environment

Bates et al. provide a detailed discussion about managing the stream profile at crossing. All culverts act as constraints in a dynamic environment, remaining at a specific location and elevation, and prevent channels from maintaining their natural processes (2006). By attempting to understand the possible impacts of a crossing on the channel, it is possible to select design options that provide optimum fish passage while ensuring acceptable design life and maintenance requirements.

6.1.3 Hydraulic Considerations Flood Flow Conveyance

Flood flow capacity must be considered at all roadway-stream crossings. Each state has established flood flow requirements for culverts as a function of roadway category. Typical values of required flood capacity range from the 4% chance flood (25-yr) to the 1% chance flood (100-yr). If designing for fish passage, culvert size is often larger than that required for flood conveyance; however, hydraulic capacity must still be checked to ensure adequate flood flow conveyance. Culvert Flow Characteristics

Slope and span will have a large impact on culvert flow characteristics. Crossings that are designed to create passage for specific fish and lifestages may require additional hydraulic considerations such as low and high fish passage flows and induced turbulence. Targeted Fish Passage at Design Flows

Hydraulic Design options require detailed hydrologic information in order to ensure fish passage at specific periods of fish migration, while Geomorphic and Hydraulic Simulation methods attempt to match (or closely mimic) natural stream reach characteristics, and require little to no additional hydrologic information. Passage for All Fish

Geomorphic and Hydraulic Simulation techniques are intended to provide passage for all fish species within the reach through any period during which they are moving. It may be difficult, or very costly, to provide passage for all fish by designing for specific hydraulic conditions. Sediment Transport

Culverts that maintain a natural bed will be sized to retain natural reach sediment transport properties (Bates et al. 2003; Bates et al. 2006; National Marine Fisheries Service Southwest Region 2001). If crossings constrict flow, there will likely be associated impacts on sediment transport including aggradation upstream and increased velocities, scour and degradation downstream from the structure (Castro 2003). Outlet Control

For fish passage velocities and depths to be met, it is recommended that flow remain subcritical through the culvert and at the outlet, requiring that culverts be designed to maintain outlet control (Alaska Department of Fish and Game and Alaska Department of Transportation 2001; Behlke et al. 1991; Bates et al. 2003). Characteristics governing outlet control include culvert inlet area and shape, barrel area and shape, barrel slope, barrel length, barrel roughness, and water surface elevation at the culvert outlet (Norman et al. 2005). Depressed inverts, or artificial roughness created by weir baffles, and deep corrugations can also be used to slow velocities within the culvert barrel (Behlke et al. 1991). Figure 6.2 from Hydraulic Design Series 5 depicts a culvert under outlet control with an unsubmerged entrance. Outlet control may also exhibit a submerged entrance with part of the barrel or the entire barrel flowing full.

Elevation view of a culvert with outlet control. In this case the inlet is not submerged and the culvert does not flow full.
Figure 6.2 Culverts under outlet control (Norman et al. 2005)

6.2 Constraints

Other than biological, geomorphic and hydraulic considerations, a number of project and site constraints will help determine the appropriateness of a particular design technique. These include, but are not limited to, funding, cost, right-of-way, and physical, environmental, and regulatory issues. Most of these issues apply to all roadway-stream crossings but have unique aspects when applied to fish passage.

6.2.1 Costs

Culvert barrel cost increases with span. If this were the only parameter used to compare the cost of design methods, Geomorphic Simulation techniques would be more costly than Hydraulic Simulation. Span cost is even less applicable to Hydraulic Design, where baffle, weir and other appurtenances can cost more than the culvert barrel.

Total roadway-stream crossing cost includes several other capital and recurring items such as installation and long-term maintenance. To date there are insufficient case histories to conclusively state that Geomorphic Simulation techniques incur greater life cycle costs than Hydraulic Simulation or vice versa. Limited experience in Alaskan rivers indicates that there is about a 20-30% capital cost reduction for Hydraulic Simulation installations when stream slopes exceed 3%, and little difference for stream slopes less than 3% (Gubernick, Personal Communication). These costs do not include long-term maintenance, posited to be inversely related to culvert span.

The cost of replacement may be prohibitively high due to deep fill or location (Interstate highway, for example). These cases may favor Hydraulic Design techniques.

Total life cycle cost for fish passage culverts is therefore difficult to compare for different design methods. It should not be assumed, for example, that one method is more costly than another based on culvert span. A complete cost analysis must be made for each crossing.

Many design techniques are still considered experimental, and long-term monitoring is still required to understand the true impacts and implications of a selected method (Chapter 10). Careful consideration of goals and requirements should be taken before selecting design criteria.

6.2.2 Right of Way

Right of way will affect the ability of designers to modify the channel outside of the culvert structure. Some design situations will require hydraulic control structures to ensure adequate backwatering, or to control channel slope, scour, and incision. Right of way costs may limit options on small retrofit projects with limited budgets. Clear communication with local landowners will provide an understanding of right of way, and innovative agreements and easements may extend access beyond existing rights of way.

6.2.3 Physical Constraints

In addition to right of way, a number of physical barriers or obstacles could force the designer to consider the costs of moving those obstacles vs. a change in design direction. Examples include utility crossings, extreme gradient changes, and incised or degrading channels. A roughened channel (Hydraulic Design, Section 7.3.4), for example, may be required instead of a Geomorphic Simulation procedure when protecting a utility. Correcting a perched culvert may also dictate a method that protects the streambed from incision.

6.2.4 Environmental Constraints

Environmentally sensitive areas will require a high degree of design consideration. For example, at a new crossing in a salmon spawning area, it may be pertinent to design an open bottom structure that allows natural substrate to remain relatively undisturbed through the crossing. To illustrate, a culvert barrier replacement in northern California utilized natural substrate and experienced salmon spawning within the structure only two years after installation (Furniss, Personal Communication).

6.2.5 Regulatory Constraints

Regulatory requirements, like those discussed in Chapter 1 may reduce design options. For example, the presence of endangered or threatened fish species will require specific and immediate consideration, and if passage for weak swimming fish is required, Hydraulic or Geomorphic Simulation may be the best option.

6.2.6 Alignment

Proper culvert alignment requires consideration of channel shape, morphology, and culvert length. Installations that run perpendicular to the road will allow the shortest installations. In addition, flow exiting a culvert at an angle is likely to induce scour (Baker and Votapka 1990; White 1997), requiring wider culverts or channel treatments to protect against stream movement (Bates et al. 2006). Highway alignment should avoid sharp stream bends, severe meanders, confluences or other areas of converging and diverging flow (Maryland State Highway Administration 2005). When situations require installation at a skew, Figure 6.3 depicts a series of alignment options. Following the current channel form will require a longer culvert. Straightening the channel will shorten the crossing but require channel protection. Creating a wider crossing will provide a slight buffer for channel migration but may also significantly increase material and construction costs (Bates et al. 2006). Also, the aligned culvert might result in the greatest direct habitat loss, and the perpendicular culvert might result in the greatest overall disturbance and the greatest risk due to the skew of the culvert to the stream. For locations at skews or bends, all three options should be considered, and the final design is often a combination of the three.

Graphical representation of three different alignment options for culverts. From left to right: (1) Skewed culvert; (2) Culvert perpendicular to road requiring channel realignment; (3) Wider crossing and headwalls eliminates skew.
Figure 6.3 Alignment options for a skewed roadway-stream crossing (Bates et al. 2006)

Treatments recommended for minimizing culvert length include adding headwalls, steepening embankments, and narrowing and lowering the road (Bates et al. 2003; Maryland State Highway Administration 2005). Specifications for such options are included in HDS-5 (Norman et al. 2005).

6.3 Design Approaches

6.3.1 No Impedance

DEFINED - No Impedance - Crossing design produces no impedance to aquatic organism passage by spanning both the channel and floodplain.

Aside from road removal or relocation, bridges provide optimum biological, geomorphic and hydraulic connectivity (Robison et al. 1999). Often bridges will be more expensive to install and have shorter effective lives than culverts (Venner Consulting and Parsons Brinkerhoff 2004). The No Impedance procedure will not be described further

6.3.2 Geomorphic Simulation (Category 1)

DEFINED - Geomorphic Simulation approaches are based on recreating or maintaining natural stream reach geomorphic elements including slope, channel-bed width, bed materials, and bedform.

The basis of these methods is the presumption that crossings matching natural conditions will readily pass fish that are moving in the natural channel. For this reason, analysis of fish passage flows is not required. Design methods are based on a reference reach (see Glossary). Geomorphic Simulation is also known as Stream Simulation (Bates 2006; WDFW 2000). This method has expedited regional permitting in some regions of the country (Bates, Personal Communication). Biological Characteristics

Successful installations should pass fish, debris, and sediment at rates very closely resembling the natural stream reach. Geomorphic Simulation assumes passage is provided for all fish species and life stages moving through the natural channel for all flows at which they are moving. Culverts spans wider than the bankfull width can provide dry bank margins that can serve to provide passage for aquatic and terrestrial organisms. Geomorphic Characteristics

To allow natural processes to occur within the culvert, the crossing slope must remain close to that of the natural channel. A review of such culverts in Washington State found that installations remaining within 25% of natural channel slope successfully replicated natural channel conditions (Barnard 2003). New open bottomed and embedded installations can be placed to minimize disturbance of bed material, or laid below grade and backfilled with natural material to maintain natural channel grade.

Geomorphic Simulation creates wide spanning culverts that exceed channel bed width. For example, in Washington, Barnard found that these structures should be 1.3 times the channel bankfull width in order to replicate stream processes (2003). In new installations, wide spanning culverts allow crossings to maintain natural bed material.

The wide-spanning culverts and open bottom structures needed to meet such requirements will allow a slight buffer against lateral and vertical stream adjustments (Bates et al. 2006). Although success has been achieved in high gradient situations, methods simulating the natural stream have been limited to gravel and cobble beds with only a few applications in sand bed streams (Bates et al. 2006). Hydraulic Characteristics

Geomorphic Simulation avoids the need for consideration of target species/life-stage, timing of fish migration, or fish passage hydrology. Since crossings are generally much larger than culverts designed for hydraulic capacity alone, Geomorphic Simulation will typically control design (hydraulic capacity must still be checked to meet the required headwater-flood policy). Data Requirements
  • Channel and valley type (Section 6.5.2)
  • Channel longitudinal profile and control points such as rock outcroppings, ledges and immobile bed features
  • Channel and floodplain cross sections
  • Reference reach characteristics
    • Channel geomorphic characteristics
    • Bedforms
    • Bed and bank material
  • Adjustment potential (vertical and horizontal) and alignment
  • Peak flow for culvert flow capacity
  • Sediment size distribution in upstream channel
  • Flood design flow

Geomorphic Simulation is illustrated in Section 7.1.

6.3.3 Hydraulic Simulation (Category 2)

DEFINED - Hydraulic Simulation techniques utilize embedded culverts, natural or synthetic bed mixes, and natural roughness elements such as oversized rock, to provide hydraulic conditions conducive to fish passage. These techniques operate on the assumption that providing hydraulic diversity similar, but not identical, to that found in natural channels will create a fish passable structure without checks for excessive velocity or turbulence. Many techniques are based on regional design experience.

Regardless of specific criteria, Hydraulic Simulation will generally create smaller spanning structures that have a reduced capital cost but higher maintenance requirements (debris removal) when compared to Geomorphic Simulation. Biological Characteristics

By creating a crossing that resembles natural stream slope and substrate, passage is assumed adequate for fish in the stream reach. This assumption is often based on regional experience and project monitoring (Alaska Department of Fish and Game and Alaska Department of Transportation 2001; Maryland State Highway Administration 2005; Robison et al. 1999; Miles, Personal Communication; Browning 1990). In Alaska, experience has found that culverts following Hydraulic Simulation, they call it "Stream Simulation," criteria adequately pass fish, and permitting has been expedited (Alaska Department of Fish and Game and Alaska Department of Transportation 2001). Techniques developed by Maryland State Highway Administration (2005) and Browning (1990) check channel velocities for compliance with local stream flows. Although structures aren't specifically oversized to provide stream bank margins, low flows may provide dry bank areas that will allow terrestrial organisms to pass (Miles, Personal Communication). Geomorphic Characteristics

Hydraulic Simulation creates hydraulic roughness, low flow paths, and resting areas conducive to fish passage by utilizing natural or artificial bed material (Robison et al. 1999; Browning 1990), or oversized substrate that remains stable during design floods (Alaska Department of Fish and Game and Alaska Department of Transportation 2001). Bed structures and key pieces are used to create flow diversity and resting areas, ideally matching bed characteristics of the natural channel.

Culvert span is generally close to or slightly less than bankfull (Alaska Department of Fish and Game and Alaska Department of Transportation 2001; Browning 1990; Robison et al. 1999; Maryland State Highway Administration 2005), allowing sediment and debris flow to continue through the crossing at flows up to bankfull. Substrate does not necessarily mimic stream reach substrate and form as in Geomorphic Simulation.

Some Hydraulic Simulation approaches create a stable channel within the culvert (i.e. Alaska Department of Fish and Game and Alaska Department of Transportation 2001; Bates et al. 2003). In such a case, bed load and suspended load still move through the culvert, but foundation bed material is not scoured out at high flows (i.e. a 50-yr event). This requires less flow area within the culvert barrel, as higher flows can pass through the culvert without scouring the bed material (Miles, Personal Communication).

In situations where a mobile bed is created, or allowed to develop within the crossing, sediment and debris movement is similar up to bankfull flows. Bed material can be washed out during a flood event, leaving a bare culvert, and, without upstream grade control, lead to upstream progressing channel incision. Recruitment may replace material that is scoured out, but it cannot be relied upon to do so. Regardless of bed stability, fines must be part of the bed material mixture to seal voids and avoid flows going subsurface, which would create a low flow barrier. Hydraulic Characteristics

Culvert spans designed for Hydraulic Simulation are generally very close to, or slightly less than, bankfull width. Methods that call for increased bed sizing and roughness will decrease flow velocity but increase turbulence.

Hydraulic capacity must be checked to ensure adequacy. Data Requirements

Channel and valley type (Section 6.5.2)
Channel longitudinal profile and control points such as rock outcroppings, ledges and immobile bed features
Channel and floodplain cross sections
Sediment size distribution in upstream channel
Adjustment potential (vertical and horizontal) and alignment
Flood design flow

Hydraulic Simulation is illustrated in Section 7.2.

6.3.4 Hydraulic Design (Category 3)

DEFINED - Hydraulic Design techniques create water depths and velocities that meet the swimming abilities of target fish populations during specific periods of fish movement. General considerations include the effect of culvert slope, size, material, and length. Flow control structures such as baffles, weirs, formal fishways or oversized substrate are commonly utilized to create adequate hydraulic conditions.

Hydraulic Design is most applicable to retrofits, but can be used for new and replacement culverts. This technique generates a smaller diameter culvert, while still meeting fish passage criteria including leap height, average cross-sectional velocity, flow depth, and drop height. Hydraulic Design is specifically tailored to meet target fish species requirements, but produces a less connected design than Geomorphic or Hydraulic Simulation. These designs are applicable for slopes up to 5% (Robison et al. 1999; Bates et al. 2003; Katopodis 1992). Biological Characteristics

Hydraulic Designs have been shown to aid in upstream migration by providing resting pools, low velocities, and deep flow (Gregory et al. 2004). These techniques utilize the swimming abilities of target fish populations in order to develop hydraulic criteria necessary to ensure fish passage. The target fish species and lifestage should be determined through consultation with fisheries biologists, and will generally focus on the weakest swimming fish known to require passage during specific periods of fish movement. Designs to meet specific hydraulic criteria are likely to constrict flow, disrupt ecosystem connectivity, and require a more rigorous design and permitting process than geomorphic or Hydraulic Simulation (i.e. Alaska Department of Fish and Game and Alaska Department of Transportation 2001; Bates et al. 2003). Hydraulic Design does not account for ecosystem requirements or the movement of non-target species. Geomorphic Characteristics

Hydraulic Design is applicable over a range of slopes. Installations on mild slopes may create fish passable conditions without grade control structures, while moderately sloped (1-3.5%) installations and retrofits may require weirs or baffles to attain fish passable conditions (Bates et al. 2003; Alaska Department of Fish and Game and Alaska Department of Transportation 2001).

The structures created by Hydraulic Design are more likely to affect flow through and around the structure than those designed by Geomorphic or Hydraulic Simulation. Localized aggradation and degradation due to channel constriction may have to be addressed (Castro 2003), and regular debris maintenance is generally required for Hydraulic Design culverts. This can be especially important in retrofit situations where structure modifications, such as baffles or weirs, have the propensity to catch and hold debris, increasing the risk of debris clogging (Bates et al. 2003). Hydraulic Characteristics

Low and high fish-passage flows must be determined to ensure that hydraulic criteria are met during periods of fish movement (Chapter 5). This requires knowledge of the times of the year and flow regimes at which fish move within the natural channel. In new installations, fish passage considerations will generally control structure size, but flood conveyance must still be checked. Smaller diameters, especially when combined with the effects of baffles, or other roughness elements, can restrict passage of water and debris through the culvert, decreasing the flood flow capacity while increasing the likelihood of plugging and culvert failure. Data Requirements
  • Channel longitudinal profile
  • Target fish species and requirements such as swimming and leaping ability, depth requirements and time of year for movement
  • Channel cross sections
    • Channel geomorphic characteristics
    • Bed and bank material
  • Adjustment potential (vertical and horizontal) and alignment
  • Low fish passage flow
  • High fish passage flow
  • Flood design flow

Hydraulic Design is illustrated in Section 7.3. Further Considerations

This design approach is often recommended as a last alternative, when other possibilities are found to be unfavorable (Alaska Department of Fish and Game and Alaska Department of Transportation 2001; Bates et al. 2003; Flosi et al. 1998; Robison et al 1999; Maine Department of Transportation 2004). In Washington for example, design guidelines recommend that use of Hydraulic Design be limited to culvert retrofits, producing inexpensive, short-term, benefits until the crossing can be replaced (Bates et al. 2003).

Baffles have a much larger failure rate than other techniques. They are prone to clogging, and are difficult to prefabricate as settling may cause the baffles to pop out leading to damage to the culvert itself and to culvert failure (Robison et al. 1999; Gardner 2006). Hydraulically designed structures will have a shorter design life, increased maintenance needs, and a more intensive permitting process than Geomorphic or Hydraulic Simulation culverts.

6.4 Design Selection

The selection of an appropriate design technique will be the result of project goals and the design techniques applicable to a particular situation or region of the country. In Chapter 7, design techniques from across the country are explained within the context of the design categories listed above. Design examples are included in the Chapter 8 to further clarify the design process.

A first step in the decision process is to verify the necessity of a road crossing. Abandonment or removal of a crossing may be a plausible and desirable solution for fish passage problems, especially on forest land where road use is intermittent or logging and fire traffic can be rerouted with little consequence (Robison et al. 1999). Temporary structures and fords might also be considered.

It is recommended that State DOTs meet with State natural resource agencies and appropriate federal agencies (for example, U.S. Army Corps of Engineers) to discuss these methods for general applicability for a region.

Agreements between State Departments of Transportation and Resource agencies can greatly expedite the design and permitting process, ensuring that the requirements of all parties are met satisfactorily through a common vision. For example, Alaska and Oregon currently have agreements between their respective resource agencies aimed at more timely approval of permit applications for culvert installations, and recognizing the priority of replacement/repair of current fish passage barriers (Venner Consulting and Parsons Brinkerhoff 2004).

A comparison table of the design categories is presented in Table 6.1.

6.4.1 Summary Matrix of Design Approaches
Table 6.1 Summary of Geomorphic, Biologic and Hydraulic Characteristics of Various Crossing Options
Category Description Relative Width Characteristics
Biological Geomorphic Hydraulic
NA No Impedance ≥100-yr floodplain Pass all fish and aquatic organisms Unchanged Q100 unconstricted
1 Geomorphic Simulation ≥bankfull Pass all fish and aquatic organisms Natural Substrate; Mobile bed; Stability of substrate usually not checked Unaltered for Q slightly above bankfull; Check Q100
2 Hydraulic Simulation ≤bankfull Reported to pass all fish and aquatic organisms Oversized substrate; Stationary bed; Stability of bed usually checked Similar for Q slightly less than bankfull; Check Q100
3 Hydraulic Design variable; usually <bankfull Pass target species at target life stage Artificial channel Must meet target species and life stage requirements; Check for Q100

6.5 Analysis Tools and Software

Analysis tools and computer software can be useful in the design of fish passable structures. The programs/websites are recommended or specified for use by many design/assessment documents. In addition, reference to Table 1.2 will be helpful for accessing other pertinent FHWA publications.

6.5.1 Channel Geometry Channel Width

The correct determination of channel width is an important prerequisite for many of the design techniques described in this report. Width measurements should describe stable, straight channel conditions between bends and outside the influence of a culvert or an artificial or unique constriction (Bates et al. 2003). Two common design parameters include bankfull width and active channel width. Bankfull width is the result of geomorphic processes, while active channel width is more related to an ordinary discharge. In entrenched and non-adjustable systems bankfull and active channel width may be very similar, while evaluation in other areas, such as meandering valley streams, might show great discrepancies (Bates et al. 2006).

Active Channel Width:
The "active channel" describes the stream width at current and recent discharges, beyond which permanent features such as terrestrial vegetation begin to dominate (Hedman and W.M.Kastner 1977). For engineering purposes, the active channel can be distinguished by the ordinary high water (OHW) mark - the elevation delineating the highest water level that has been maintained for a sufficient period of time to leave evidence on the landscape (Taylor and Love 2003). Representations may also include erosion, shelving or terracing, change in soil characteristics, a break or destruction of terrestrial vegetation, moss growth on rocks along stream margins, vegetation changes from predominantly aquatic to predominantly terrestrial, or the presence of organic litter or debris (Taylor and Love 2003; Bates et al. 2003).

Bankfull Width:

Bankfull width describes stream characteristics during channel forming events. Bankfull flow is thought to mark the condition of incipient motion, with impacts on long-term form, function and stability of the channel (Williams 1978). This is typically recognized as a 1- to 2-year event, when flow within the channel just begins to spill over into the active floodplain (Leopold et al. 1964). When floodplains are absent or difficult to ascertain, as in entrenched mountain streams, markers used to determine bankfull and active channel show little variation (Bates et al. 2003). Difficulty in determining bankfull flow in the field prompts some to provide guidelines for estimation of bankfull width based on surveyed cross sections and return period flow (i.e. Maine Department of Transportation 2004). This type of estimation may show great disparity when compared with field observations of channel-bed width (Mussetter 1989).

Cross section of a channel showing the bankfull width above the active channel width. In certain systems bankfull and active channel can be very similar, and active channel indicators are often to describe bankfull flow when a floodplain is not present as in an entrenched system.
Figure 6.4 Depiction of bankfull channel width compared to active channel width (Taylor and Love 2003) (Note that in certain systems bankfull and active channel can be very similar, and active channel indicators are often used to describe bankfull flow when a floodplain is not present as in entrenched systems) Channel Profile and Condition

It is extremely important to understand structure impacts on the channel over time including incision, scour, headcut and regrade (Bates et al. 2006). This requires an accurate survey of the longitudinal profile (River and Stream Continuity Partnership 2004). A longitudinal profile should include the culvert site and 20 channel widths or a minimum of 61-91 m (200-300 ft) up- and downstream of the structure (Castro 2003; Bates et al. 2003). Recent experience shows this distance to be a minimum. Maryland DOT surveys 152 m (500 ft) on each side of the culvert (Kosicki, Personal Communication 2007). This will allow an understanding of the final channel bottom elevation resulting in the vicinity of the replacement structure, ensuring proper invert elevations, embedment, and slope. A good survey is also useful in assessing the potential for downstream flooding, alteration of upstream and downstream habitat, potential for erosion and headcutting, and stream stability in general (River and Stream Continuity Partnership 2004).

Channel Evolution:
Although a crossing may seem stable, there are various levels of stability, natural or anthropogenic, and it is important to examine upstream and downstream channel conditions to understand the current channel condition. Figure 6.5 depicts channel evolution after an initial channel incision moved the stream from a stable state.

Diagram of a Channel Evolution Model. From top to bottom: (1) "Stable", the bank height is less than the critical bank height; (2) "Incision", the bank height is greater than the critical bank height; (3) "Widening", the bank retreats; (4) "Stabilizing", the bank height is equal to the critical bank height; (5) "Stable", the bank height is less than the critical bank height.
Figure 6.5 Critical bank height is inherently unstable and will result in bank failure and stream widening (Castro 2003)

Channel Incision, Headcut and Regrade:
As channels continually evolve and migrate, channel adjustment can lead to structure failure. Installations that fail to recognize channel processes may compromise fish passage and alter the quantity and quality of stream corridor habitat (Castro 2003).

In situations where a current culvert installation is acting as a control point, removal, replacement with a larger structure, or lowering may allow channel incision to progress upstream uncontrollably, or until another control point is reached. Regrade will be more immediate and pronounced in sand bed streams (Bates et al. 2003). Stream reaches actively aggrading or incising will cause Geomorphic Simulation culverts to be ineffective, and Hydraulic Design or Simulation incorporating channel grade controls (Section may be more suitable.

6.5.2 Stream Classification

Classifying a stream containing a crossing or the site of a crossing impels the design team to collect meaningful data and discuss the stream dynamics before choosing a design procedure.

Systems for stream classification are useful tools in building awareness of stream form and function. Methods describe the channel in terms of cross-sectional shapes, morphological parts of the stream, and interactions between flow and sedimentation (Bunte and Abt 2001). The following section is intended to introduce the user to popular methods in stream classification and geomorphology, but is not sufficient for structure design. Coordination with a local geotechnical engineer and geomorphologist is necessary for ensuring structure performance. For more information it will be useful to examine references included below and Hydraulic Engineering Circular No. 20 (Lagasse et al. 2001). Montgomery and Buffington

Montgomery and Buffington created a stream classification system based on channel systems in the Pacific Northwest that is applicable to similar regions elsewhere. Their methodology follows changes in channel morphology as steep headwater streams run through steep valleys and hillslopes, gentle valleys, and eventually low gradient valleys (Bunte and Abt 2001). As water flows to the ocean, channel types generally transition from cascade, step-pool, plane bed, pool-riffle and dune-ripple, as shown in Table 6.2. Channel bedform is described by the type and size of sediment, sediment transport capabilities, and hydraulic conditions within a stream reach. Table 6.2 from Bunte and Abt summarizes this classification system with respect to channel geomorphic and hydraulic conditions.

Table 6.2 Stream Classification by Montgomery and Buffington (from Bunte and Abt 2001)
Stream gradient, range and mode
Stream Typical bed material Dominant Sediment Dominant sediment storage Typical pool spacing*
0.03 - 0.20
(0.08 - 0.20)
Cascades Cobble-boulder Fluvial, hillslopes, debris flows Around flow obstructions < 1
0.02 - 0.09
(0.04 - 0.08)
Step-pool Cobble-boulder Fluvial, hillslopes, debris flows Bedforms 1 - 4
<0.02 - 0.05
(0.02 - 0.04)
Plane-bed, forced pools Gravel-cobble Fluvial, bank failure, debris flows Overbank None
<0.001 - 0.03
Pool-riffle Gravel Fluvial, bank failure Overbank, bedforms 5 - 7
< 0.001 Dune-ripple Sand Fluvial, bank failure Overbank, bedforms 5 - 7

Values in parentheses are the modes of the observed stream gradient distribution; * in terms of channel widths

A reach-scale categorization allows streams to be categorized based on relative positions within the watershed and sediment transport characteristics. This type of analysis is useful in understanding the potential response of a channel reach to a crossing installation. Montgomery and Buffington define reach level morphologies as source, transport and response reaches (Montgomery and Buffington 1993).

Source reaches contain as much or more sediment than the stream can transport. Transport reaches are high gradient supply-limited channels, which are unlikely to respond quickly or severely to disturbance. This includes bedrock, cascade and step-pool channels. Response reaches are lower gradient transport-limited channels with a high potential for morphological adjustment in response to sediment input. This general classification covers plane-bed, pool-riffle and braided channels. The transition from transport to response reach is where the impacts of increased sediment supply will have the largest impact, as sediment supplied by the transport reach will readily settle out at the first reach that cannot maintain sediment transport capacity (Montgomery and Buffington 1993).

A crossing location within a particular reach, as well as the proximity of other reaches will help a designer ascertain the potential impacts and geomorphic response of the stream. Crossings that fall at the intersection of two different channel types, for example, could indicate channel incision, or that the crossing is located at a point of geomorphic transition (Bates et al. 2006). Crossings placed in a response reach typically will require extra consideration of channel processes and morphological impacts (channel aggradation and lateral movement). Rosgen Stream Classification

Rosgen channel classification is based on five morphometric parameters of the channel and its floodplain: entrenchment ratio, width-depth ratio at bankfull flow, sinuosity, stream gradient and mean bed particle size (Rosgen 1996). These characteristics are used to distinguish seven stream types, represented by capital letters A to G. Table 6.3 lists the morphological characteristics of Rosgen's stream types.

Table 6.3 Morphological Characteristics of the Major Rosgen Stream Types (Bunte and Abt 2001)
Stream Type Morphological characteristics
A Step-pool, or cascading: plunge and scour pools, high energy, low sediment storage, stable;
B Riffles and rapids: some scour pools, bars rare, stable;
C Pool-riffle sequences: meandering, point bars, well developed floodplain, banks stable or unstable;
D Braided: multiple-channels, shifting bars, scour, deposition, high sediment supply, eroding banks;
DA Anastomosing: multiple channels, pool-riffle, vegetated floodplain, adjent, wetlands, stable banks;
E Meadow meanders: well-developed floodplain, riffle-pool, relative high sediment conveyance;
F Valley meanders: incised into valleys, poor floodplain, pool-riffle, banks stable or unstable;
G Gullies: incised into hillslopes and meadows, high sediment supply, unstable banks, step-pool.

Channels can be further distinguished using numbers to represent bed material and particle size, and lowercase letters to represent deviation from expected channel slopes. For example, a stream classified as C4b is a C-type stream with a gravel bed and gradient within the range of 0.02-0.039, which is more typical of a B-type stream (Rosgen 1994). Accurate classification requires a longitudinal and cross-sectional channel survey and sediment sample analysis. Summary of Channel Classification

All stream classification systems can be useful in understanding basic channel reach geometry and dominant geomorphic processes. This can be valuable in predicting channel response to modification or culvert replacement. Certain channel types can carry specific design challenges. For example, risk of floodplain constriction and/or lateral adjustment is associated with Rosgen C, D and E channels (Bates et al. 2006). As mentioned above, plane bed, pool-riffle, and dune-ripple channels are associated with response reaches, and are likely to show the most dramatic response to disturbance (Montgomery and Buffington 1993).

For further discussion of stream classification and applicability to channel crossing design, it is useful to review the original documents by Rosgen (1994; 1996), Montgomery and Buffington (1993; 1998), Bunte and Abt (2001), and Bates(2006). It is important to note that these design techniques or classification systems are not well tested outside the regions for which they were created. Installations in low gradient, highly mobile sand bed streams may require special consideration.

6.5.3 Stream Morphology

As a rigid structure in a dynamic environment, culverts must be designed with channel processes in mind. Effective designs consider the channel and watershed context of the crossing location. Channels are continually evolving, and an understanding of stream adjustment potential must be addressed. Without consideration, well intended plans could have detrimental or completely ineffective results/impacts on the stream system and related habitat (Castro 2003; Furniss 2006). Gradient

Past channel degradation can require channel modification, or considerations of the impact of increased slope on channel stability, substrate and future conditions (Robison et al. 1999; Bates et al. 2006; Bates et al. 2003). A true Geomorphic Simulation can only be completed when culvert bed slopes very closely match the slopes of the adjacent stream channel. Oversized sediment utilized in Hydraulic Simulations provides more leeway with regards to stream slope, but also require that crossing slopes be close to the adjacent channel. Bed Material and Embedded Culverts

The benefits of natural streambeds and embedded culverts are widely recognized in fish passage applications (e.g. Venner Consulting and Parsons Brinkerhoff 2004; Bates et al. 2003; Taylor and Love 2003; Clarkin et al. 2003). Bed material provides barrel roughness, which provides areas of low velocity that may be conducive to fish passage, mimics natural hydraulics, and is self sustaining when designed properly (White 1997). Key Roughness Elements

In order to provide fish migration paths and resting areas many design techniques utilize key roughness elements to create diversity in flow velocity, depth, and energy dissipation (Robison et al. 1999; Bates et al. 2006; Browning 1990). Key roughness elements describe any number of materials that can be used to provide hydraulic roughness and diversity to a crossing including oversized substrate, constructed channel features including banks, stone sills, boulder clusters, log sills, and baffles. Such features are intended to increase bed stability and provide resting areas and hydraulic diversity conducive to fish passage. Subsurface Flows

Crossings that are filled with a coarse simulated bed mix may allow low flows to seep between rocks - and move solely in the subsurface - until interstitial spaces have been sealed with fine particles. To limit streambed permeability, an appropriate proportion of fine material must be included in the bed mix (5-10%) (United States Forest Service 2006a; Bates et al. 2006). During channel construction, placement of a sediment barrier fabric, mud or straw wattles (Browning 1990; Gubernick, Personal Communication), or washing fines into the streambed during construction can effectively seal the voids (Bates et al. 2006).

6.5.4 Estimating Roughness with the Manning Equation

Estimating roughness with the Manning equation (Equation 6.1, Chow, 1959) is most often used to estimate uniform flow depth given a design discharge:

Equation 6.1
Q =KAS1/2R2/3 = VA


K = 1.0 for SI and 1.486 for CU units, respectively
Q = Channel discharge, m3/s (ft3/s)
S = Channel slope, m/m (ft/ft)
R = Hydraulic radius (cross-sectional area/wetted perimeter), m (ft)
A = Cross-sectional area, m2 (ft2)
V = Average channel velocity, m/s (ft/s)
n = Manning's "n" (channel roughness coefficient)

Of primary importance is to determine the Manning's "n" value, or channel roughness coefficient, for low and high values of discharge and for flood discharges. This will govern sediment stability and the hydraulic properties within the culvert barrel. Estimates for Manning's "n" may be found in HEC-20 (Lagasse et al. 2001), Chow (1959) or as determined at U.S. Geological Survey stream gage sites (Barnes 1967). For coarse streambed material, the procedures described in Chapter 7 use the following specialized equations. Limerinos Equation


  • Limerinos 1970 (as discussed in Bates et al. 2003)


  • Experience shows a more accurate prediction in higher-velocity situations.


  • Equation is based on data where 0.9<R/D84 <6.9 and 0.02<n<0.107.
  • The error range for n/R1/6is +42.9% to -33.7%.
Equation 6.2
1.16 + 2log(R/D84)


K = 0.1129 for SI and 0.0926 for CU units, respectively
R = hydraulic radius, m (ft)
D84 = the dimension of the intermediate axis of the 84th percentile particle, m (ft) Jarrett's Equation


  • Jarrett 1984 (as discussed in Bates et al. 2003)


  • Average velocity is less than 0.9 m/s (3 ft/s)
  • Based on data where slope is between 0.2% and 4%
  • May be applicable up to an 8.25% slope where 0.4<R/D84<11 and 0.03<n<0.142


  • Error range of n on the test data is wide, +44% to +123%
  • It is implied that, as slope increases, sediment size increases and so does roughness.
Equation 6.3

N= KSf0.38R-0.16


K = 0.32 for SI and 0.39 for CU units, respectively
Sf = the friction slope of the channel
R = hydraulic radius of the channel, m (ft) Mussetter's Equation


  • Mussetter 1989 (as discussed in Bates et al. 2003)


  • Derived from data in Colorado mountain streams, with sediment distributions similar to those recommended by WDFW guidelines.
  • Fish passage velocity calculations


  • Derived from data where slope is between 0.54% and 16.8%, 0.25<R/D84<3.72, and 0.001<f<7.06 (0.036<n<4.2)
  • Error range is between +3.8% to +12%.
  • Accuracy decreases when velocity is greater than 0.9 m/s (3 ft/s).
Equation 6.4

1.49R0.17/(n)(g)0.5 = (8/f)1/2 = 1.11(y/D84)0.46 (D84/D50)-0.85 Sf-0.39


y is the mean depth, ft

Note: If Equation 6.4 is used in metric units, the constant 1.49 will equal 1.0.

6.5.5 Bed Mobility

The design engineer must understand the basic concepts of particle sizing and stream stability in order to specify a sediment mixture for the proposed culvert that is appropriate for the selected design approach. For example, Geomorphic Simulation seeks to mimic the natural streambed sediment mixture, while Hydraulic Design usually uses a coarser mix of sediment in the culvert barrel than found in the adjacent stream channel. Nine methods are here presented to assist the engineer in determining the stability of the streambed within the culvert barrel. These are taken from the USFS Stream Simulation design, the WDFW Stream Simulation design and the Roughened Channel design, as illustrated in sections 7.1.1, 7.1.2 and 7.3.4, respectively. Consultation with team members familiar with these procedures will be beneficial. These tools may be used to test the sensitivity of substrate mixes to entrainment. Simulated Streambed Design


  • USFS Stream Simulation - DRAFT Manual (Bates et al. 2006)

When natural bed material cannot be used, a well-graded mix of materials should be created to closely approximate the particle size distribution of the reference reach. The most important elements of a constructed bed are large particles to provide bed structure, and fines to limit bed permeability and bind the bed mix together. Analysis of bed material can be done through a sieve analysis, but is most commonly done through a pebble count. When distribution is calculated by a pebble count, D100, D84, D50 of the reference reach are taken directly from the surface pebble count, and smaller grain sizes are determined through use of the Fuller-Thompson equation (6.5).

Fuller-Thompson equation:

Equation 6.5



d = particle size of interest, mm (ft)
P= percentage of the mixture smaller than d
D100= largest size material in the mix, mm (ft)
n = parameter that determines how fine the resulting mix will be. A value of 0.5 produces a maximum density mix when particles are round

This equation can be rearranged to find any particle size, for example:

D16 = 0.321/nD50

D5 = 0.101/nD50

When distribution is calculated by a pebble count, D100, D84, D50 of the reference reach are taken directly from the surface pebble count, and smaller grain sizes are determined through use of the Fuller-Thompson equation (6.5). This is based on D50, and creates a simulated bed mix. (This application has not been field tested, and professional judgment is recommended). Modified Shield's Equation


  • USFS Stream Simulation - DRAFT Manual (Bates et al. 2006)


  • Riffles and plane-bed channels with channel-bed gradients less than 5%
  • Sand and gravel bed streams with low relative roughness (flow depth considerably greater than streambed particle size)
  • Poorly graded streambed (particles represent a narrow range of class sizes)


  • D84 between 10 and 250 mm (2.5 to 10 inches)
  • Particle size of interest ≤ 20-30 times D50.

The modified Shield's equation is used to determine particle stability based on critical shear stress. Particle stability is compromised when boundary shear stress in the channel is greater than a critical stress threshold. Boundary shear stress is calculated using Equation 6.6.

Equation 6.6

τ = γRS


τ = average boundary shear stress, Pa (lb/ft2)
γ= specific weight of water, N/m3 (lb/ft3)
R= hydraulic radius (Cross-Sectional Area of Flow divided by Wetted Perimeter - calculated at design flow), m (ft)
S= slope, m/m (ft/ft)

Once boundary shear stress has been calculated, a critical stress threshold is calculated using Equation 6.7.

Equation 6.7

τci = τ*D50s - γ)Di0.3D500.7


τci = critical shear stress at which the sediment particle of interest begins to move, Pa (lb/ft2)
τ*D50 = dimensionless Shield's parameter for D50 particle size (this has been experimentally derived for a number of particle sizes, see Table 6.4)
D50 = diameter of the median or 50th percentile particle size of the channel bed, m (ft)
Di = diameter of the particle size of interest (typically D84 or D95 for stream simulation), m (ft)

Table 6.4 Angle of Repose, Shield's Parameter and Critical Shear Stress Values for Gravel-, Cobble- and Boulder-Sized Particles (Bates et al. 2006)
Particle size classification Particle size, Di
Angle of repose (Φ), f
Shield's parameter a, τ* Critical shear stress, τc
very large boulders > 2048 42 0.054 37.37
large boulders 1024-2048 42 0.054 18.68
medium boulders 512-1024 42 0.054 9.34
small boulders 256-512 42 0.054 4.67
large cobbles 128-256 42 0.054 2.34
small cobbles 64-128 41 0.052 1.13
very coarse gravels 32-64 40 0.05 0.54
coarse gravels 16-32 38 0.047 0.25
medium gravels 8-16 36 0.044 0.12
fine gravels 4-8 35 0.042 0.057
very fine gravels 2-4 33 0.039 0.026

a equation used to determine Shield's parameter for gravel-, cobble-, and boulder-sized particles: τ* = 0.06 tanΦ Critical Unit Discharge Approach


  • USFS Stream Simulation - DRAFT Manual (Bates et al. 2006)


  • Channels with gradients exceeding 10%
  • Flow depth is shallow with respect to channel-bed particle diameter (situations where discharge is much easier to determine than depth).

This approach is based on unit discharge, and a value of critical unit discharge will be compared to channel unit discharge to determine particle entrainment (particle lifting into flow).

Equation 6.8 is used to calculate channel unit discharge.

Equation 6.8

Q= Q/w


q= Unit discharge, m3/s/m (ft3/s/ft)
Q= Discharge, m3/s (ft3/s)
w= the width of the channel at a given cross section, defined by active channel width, m (ft)

Equation 6.9 is used to predict the entrainment of the particle size of interest.

Equation 6.9
qcD50 =0.15g0.5D501.5


qc-D50= the critical unit discharge to entrain the D50 particle size, m2/s (ft2/s)
D50 = the median or 50th percentile particle size, m (ft)
g= gravitational acceleration, m/s2 (ft/s2)
S= slope, m/m (ft/ft)

More generally,

Equation 6.10
qci = qcD50   Di   b


qci= the critical unit discharge to entrain the particle size of interest, m2/s (ft2/s)
Di= the particle size of interest, m (ft)
D50= the median or 50th percentile particle size, m (ft)
b= measure of the range of particle sizes that make up the channel bed (quantifies the effects on particle entrainment of smaller particles being hidden and of larger particles being exposed to flow)

Equation 6.11
b = 1.5   D84   -1


D84= the 84th percentile particle size, mm (ft)
D16= the 16th percentile particle size, mm (ft)


  1. Equation 6.8 is used to calculate the unit discharge for bankfull flow
  2. Equation 6.9 is used to find the critical unit discharge (qc-D50) needed to entrain the D50 particle size at the given cross section.
  3. Equation 6.11 is used to calculate the sorting of the channel bed (b).
  4. Equation 6.10 is used to calculate the critical discharge (qci) needed to entrain the particle of interest at any given cross section.
  5. Compare the critical unit discharge (qci) to the unit discharge (q) in the channel at the specified flow. If the unit discharge is less than the critical discharge the particle size of interest will not be entrained (particle will remain immobile). If unit discharge is greater than critical discharge the particle size of interest will be entrained. Boundary Shear Threshold Analysis


  • Williams 1983 (as discussed in USFS Stream Simulation - DRAFT Manual, Bates et al. 2006)


  • Williams equations indicate the upper and lower thresholds in boundary shear stress required to initialize movement of a given particle size.


  • Equation 6.12 was developed from particles between 15 to 900 mm (0.05 to 2.73 ft).
  • Equation 6.13 was developed from particles between 10 to 3300 mm (0.03 to 10 ft).
  • Both equations express shear stress in customary units (lb/ft2).
Equation 6.12

τci-u = 0.0814Di

Equation 6.13

τci-i = 0.00355Di


τc-u = is the upper critical shear stress value (lb/ft2) for determining particle mobility and immobility for the particle size of interest.
Τc-I = is the lower critical shear stress value (lb/ft2) for determining particle mobility and immobility for the particle size of interest
Di = is the particle size of interest, mm (ft)


  1. Calculate the average boundary shear stress using Equation 6.6 for the flow of interest (e.g. bankfull).
  2. Using Equations 6.12 and 6.13, calculate the upper and lower critical shear stress values for the particle size of interest at any given cross sections (e.g. D84).
  3. To determine if the particle will be immobile, mobile, or potentially mobile, compare the average boundary shear stress for a particular flow to the upper and lower critical shear stress values for the particle size of interest.

If the average shear stress (τ) is greater than the upper critical shear stress (τc-u), the particle will be mobile at this flow. If the average boundary shear stress (τ) is less than the lower critical shear stress (τc-i), then the particle will be immobile for these flow conditions. If the average boundary shear stress is between the upper and lower critical shear stress values, then the particle has potential to move at these flow conditions. U.S. Army Corps of Engineers Riprap


  • USFS Stream Simulation - DRAFT Manual (Bates et al. 2006)
    Roughened Channel Design, WDFW (Bates et al. 2003)


  • D84/D15 ratio typically less than 3-7 in practice
  • Sizing immobile key pieces


  • Considers angular rock (not specifically applicable to round rock)
  • Rock may move as smaller rocks surrounding key pieces move. Similar-sized rock should be used to support key pieces.

The U.S. Army Corps of Engineers has developed two riprap models for designing riprap bank protection. These were developed through laboratory and analytical work, and consider angular rock, which is resistant to sliding and rolling. Note that round rock may have to be significantly larger than angular rock to achieve similar levels of stability (Abt, 1988).

Manuals are available at

For slopes from 2-20%

Equation 6.14
D30 = 1.95S0.555(1.25q)2/3


D30= dimension of the intermediate axis of the 30th percentile particle, m (ft)
S= the bed slope, m/m (ft/ft)
q= the unit discharge, m2/s (ft2/s)
g= acceleration due to gravity, m/s2 (ft/s2)

1.25 is a safety factor that may be increased, and designers are cautioned against using this method for rock sizes greater than 0.15 m (6 in).

The U.S. Army Corp of Engineers recommends angular rock with a uniform gradation (D85/D15 = 2). This is not preferred for fish passage situations due to porosity issues. An approximate factor for scaling D30 of a uniform riprap gradation to one that is appropriate for stream channels is 1.5, so that:

Equation 6.15

D84 =1.5D30


D84 = dimension of the intermediate axis of the 84th percentile particle, m (ft) Reference Reach Approach


  • WDFW Stream Simulation (Bates et al. 2003)

Maximum particle size and appropriate sediment size distribution can be determined by examining reaches directly upstream from the culvert, or nearby reaches with similar characteristics to the design channel (e.g. unit discharge, slope, geometry and relative stability). Unit-Discharge Bed Design


  • WDFW Stream Simulation (Bates et al. 2003)

J.C. Bathurst developed the following equation to predict the size of D84 particles that would be on the threshold of motion for a given critical discharge in high gradient streams with heterogeneous beds (1987).

Equation 6.16

D84 = 3.54S0.747(1.25qc)2/3/g1/3


D84 = intermediate axis of the 84th percentile particle in the sediment distribution, m (ft)
S = energy slope of the proposed channel
qc= the critical unit discharge (total design discharge divided by the width of the bankfull channel) at which incipient motion of D84 occurs, m2/s (ft2/s)
g = the acceleration due to gravity, m/s2 (ft/s2)

This is recommended as a starting point for development of sediment mixes in high gradient streams. Two design categories are recommended based on slope.

  1. If channel slope is less than 4%, bed-changing flows may vary greatly. J.E. Costa's paleohydraulic analysis (described below) may be used to determine the magnitude of the bed changing flow for a given particle size.
  2. If channel slope is greater than 4%, 100-year flood is used for design flow. This will closely predict the same size particle as that found in natural channels with similar Q100 and Wch. This is the goal of stream simulation.

These methods generally agree, but should both be checked. These are mobile or nearly mobile particles at these flows. If it is advisable to create a bed that is more stable, particle sizes should be increased. If bed slope approaches or exceeds 1.25 times the natural reach slope, it may not be possible to simulate stream conditions, and a Hydraulic Simulation approach or a Hydraulic Design approach, such as Roughened Channel may be considered (Section 7.3.4). Paleohydraulic Analysis


  • WDFW Stream Simulation (Bates et al. 2003)

Paleohydraulic analysis uses the maximum particle size and flood depth to determine the discharge of flash floods. An equation developed by Costa (1983) to understand velocity based on particle size is useful in substrate sizing for stream channel design. Users should consult Costa (1983) to supplement their understanding of this procedure. This equation and the accompanying table (6.5) are in customary units.

For determining depth, velocity (ft/s) is given by:

Equation 6.17

V = 9.57(D84)0.487


D84 = is arrived at by an iterative procedure, ft


  • D84 is assumed, allowing velocity to be calculated by Equation 6.17.
  • Divide design flow by velocity to get cross-sectional area of flow.
  • Find depth from proposed channel cross section.
  • Use Table 6.5 to find the associated particle size.
  • When the resulting particle size agrees with the initial estimate, the particle design is considered suitable for design.
Table 6.5 Prediction of Water Depth for a Given Maximum Particle Size that Has Been Moved (Bates et al. 2003)
(Data has been converted to English Units; some values are log-interpolated)
Slope -> 0.005 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
Particle Size, ft Depth, ft
0.2 1.2 0.9 0.6 0.5 0.5 0.4 0.4 0.4 0.4 0.4 0.4
0.5 3 2.1 1.5 1.3 1.2 1 1 0.9 0.9 0.9 0.8
1 6 4.1 2.9 2.5 2.2 1.9 1.8 1.8 1.7 1.6 1.5
1.5 8.8 5.9 4.1 3.6 3.1 2.7 2.6 2.5 2.4 2.2 2.1
2 11.3 7.4 5.2 4.5 3.9 3.4 3.2 3.1 2.9 2.8 2.7
2.5 13.6 8.9 6.2 5.4 4.7 4.1 3.9 3.7 3.5 3.3 3.2
3 15.6 10.2 7.1 6.1 5.3 4.6 4.4 4.2 4 3.8 3.6
3.5 17.6 11.4 7.9 6.9 6 5.2 4.9 4.7 4.5 4.3 4.1
4 19.5 12.6 8.7 7.5 6.6 5.7 5.4 5.2 4.9 4.7 4.5
4.5 21.3 13.7 9.4 8.2 7.2 6.2 5.9 5.7 5.4 5.1 4.9
8.1 36.4 23.1 15.6 13.5 11.7 10.1 9.6 9.1 8.6 8.2 7.8
10.5 45.6 28.9 19.4 16.7 14.4 12.5 11.8 11.2 10.6 10 9.5

At higher slopes, the Costa equation predicts smaller particle sizes than the Bathurst equation, all other conditions being equal (Bates et al. 2003). Critical Shear Stress Method


  • Roughened Channel Design, WDFW (Bates et al. 2003)

Critical shear stress methods are used to estimate the initial movement of particles. Particles movement occurs when the maximum shear stress, τ0max, within the channel exceeds a calculated critical shear stress, τc. Critical shear stress is the shear stress required to cause movement of a given particle size (see Sections and The maximum shear stress is 1.5 times γRS, where γ is the unit weight of water, R is the hydraulic radius, and S is the slope. Data used to derive these equations are largely from low-gradient situations, although design charts show slopes up to 10% and particle sizes up to 0.58 m (1.9 ft) (Bates et al. 2003).

6.5.6 Countermeasures for Channel Instability

As a rigid structure in dynamic environment, culverts may require consideration of riprap and channel modification to address scour and channel degradation or incision (Bates et al. 2003; Robison et al. 1999; Maryland State Highway Administration 2005). An undersized culvert will destabilize the adjacent stream reach. A number of alternatives are available to protect the impacted channel. Modification of the channel both up- and downstream of the structure can decrease the slope required at the culvert installation, helping to meet velocity, gradient and embedment requirements. Scour Countermeasures

Riprap refers to oversized rock strategically placed within the channel to control scour and erosion. Application of riprap for energy dissipation is outlined in Hydraulic Engineering Circular 14 - Hydraulic Design of Energy Dissipators for Culvert and Channels (Thompson and Kilgore 2006). Figure 6.6 depicts improper use of riprap for a fish passage situation. When utilized, voids in riprap should be filled with fines to prevent flows from going subsurface (Maine Department of Transportation 2004).

Photo of perched culvert outlet with riprap placed downstream that blocks fish passage.
Figure 6.6 Downstream riprap will dissipate energy and reduce scour, but must be placed with fish utilization in mind (USFS 2005)
(Note - riprap at this culvert exit effectively blocks fish passage)

Energy Dissipation Pool:
The state of Maine requires an energy dissipation pool at culvert outlets (Maine Department of Transportation 2004). These pools allow fish to rest before attempting to enter a structure, ensuring proper culvert outlet hydraulics and backwatering. General requirements include a pool width greater than or equal to 2 times the culvert span, and a pool length greater than or equal to 3 times the culvert span. Weirs are used to maintain the appropriate flow elevation and flow capacity. If the pool does not backwater the culvert outlet during the design period, the Energy Dissipation Factor (Section 3.2.5) is checked to ensure that it is less than or equal to 4ft-lb/ft3/s (Maine Department of Transportation 2004). Channel Modifications

Downstream channel modifications may be necessary to ensure proper culvert backwatering or to control crossing slope. Upstream channel modification can include erosion or grade control structures (detailed below), or a tapering of channel banks to smooth out the impacts of an inlet constriction (Robison et al. 1999). Such grade controls are frequently an element of a cost-effective retrofit; they are also used on replacement projects. A number of techniques for channel modification are included in Table 6.6.

Table 6.6 Comparison of Channel Profile Design Structures Used to Control Grade either Upstream or Downstream of a Culvert (adapted from Bates et al. 2003 with additional comments from other sources)
Grade Control Advantages Disadvantages Limitations
Log Sills Downstream bed-elevation control Limited to <5% final gradient (affects length to catch channel grade) Minimum spacing of 15 ft. Limited to <5% gradient. Allowable drop depends upon fish requiring passage. No wet/dry cycles between.
Baffles Increase hydraulic roughness Turbulence, hydraulic profile raised, debris and structural problems. No small fish passage. Slope less than or equal to 3.5%.
Plank Sills Hand Labor Less durability Limited to <5% gradient streams, small streams.
Roughened Channel Natural appearance, flexible, can provide passage for all fish. Technical expertise required. Technical fish-passage analysis required. Limited to <3% gradient streams.
Boulder Controls Flexible, allowing channel to regrade slowly Should only be used for downstream use if culvert is sufficiently embedded. Maximum drop of
0.75 ft.
Fishway Can provide passage for most fish Expensive. Technical expertise and site-specific, flow-regime data required. Debris and bedload problems. Narrow range of operating flow. Difficult to provide passage for all fish, all of the time. Roughened Channel

Roughened channels can be constructed within the natural channel to control channel shape, slope and form. This may be especially pertinent in areas where past degradation causes a culvert installation to be placed at a severe slope. Methods and equations used in the design of roughened channels can be found in Section 7.3. Grade Control Structures

Grade control structures may be necessary upstream or downstream of a culvert to control longitudinal profile and water surface elevations. Downstream of a culvert these installations typically backwater the culvert and stabilize steepened reaches. Figure 6.7 depicts the placement of downstream grade control. Such structures have been shown to cause problems with fish passage (Browning 1990), and a clearance of 20 ft between the culvert outlet and the first downstream control is recommended (Bates et al. 2003; Robison et al. 1999). Upstream of a culvert, grade control is used to stabilize a reach and protect against current or future headcutting.

This type of structure, depicted in Figure 6.8, should end no closer than 35-50 ft from the culvert inlet (Bates et al. 2003).

Graphical illustration of a downstream grade control where the flow is from right to left. Grade controls, consisting of a series of sills, are placed downstream from the culvert.
Figure 6.7 Downstream grade control (Bates et al. 2003)

Graphical illustration of upstream regrade channel-steepening options where flow is from right to left. Upstream oversteepening is mitigated by placing sills in series in the streambed.
Figure 6.8 Upstream regrade channel-steepening options (Bates et al. 2003) Tailwater Control

It may also be necessary to raise the tailwater elevation in order to backwater the culvert and provide minimum flow depths. Sometimes this is all that's required to retrofit a flat, short culvert. Many methods are available including:

  • Weirs
  • Sills
  • Constructed tailwater pools
  • Full or partial channel restoration
  • Riffle grade control structure/Roughened Channel

Flow over weirs can create velocity and depth barriers, and it may be necessary to design a series of weirs to provide fish passage and backwatering the culvert. Broad Crested Weirs

The Maine Department of Transportationdescribes the following method for the design of a rectangular notch weir-Broad Crested Weir (Maine Department of Transportation 2004). This is a channel-spanning structure at the culvert outlet, which can be used to ensure proper water surface elevation and backwatering. When the drop over a weir will create a barrier to fish passage, it will be necessary to include further control structures to create a series of manageable step pools while maintaining adequate culvert backwater. A series of notch weirs is depicted in Figure 6.9.

Photo of notch weirs downstream of a culvert installation acting to properly backwater the culvert, while maintaining manageable drops.
Figure 6.9 Notch weirs downstream of a culvert installation, acting to properly backwater the culvert, while maintaining manageable drops (United States Forest Service 2005)

Design Procedures are as follows:

At first pass, the weir height can be set at the desired water height (ignores the depth of flow over the weir).

Equation 6.18

Q = Cd(2/3)(2g/3)1/2bch13/2


Cd = discharge coefficient (0.9 assumed)
bc = channel width across the bar, m (ft)
h1 = water elevation upstream of the bar (referenced to bar elevation), m (ft)

Solving for h1

Equation 6.19

h1 = [Q/(Cd(2/3)(2g/3)1/2bc)]2/3

(Note the assumption 0.9 is in view of the uncertainty and variability in the weirs contemplated here.)

Flow over the weir will be critical, and velocity (vc) must be checked for fish swimming ability:

Equation 6.20

vc = (gh1)1/2


vc = critical velocity, m/s (ft/s)

This procedure uses constructed materials. Consultation with appropriate State and Federal agencies should occur to determine the acceptability of this design.

Channel regrade promoted by an undersized culvert installation can be a concern with culvert replacement or removal. Grade control structures can be used up and/or downstream of the structure to help protect against catastrophic channel regrade.

6.5.7 Culvert Shapes and Standards

A number of culvert shapes are available to meet the specific needs of a culvert site. Selection will be the result of site conditions including depth of cover, limited allowable headwater elevations, clogging potential, need for natural stream bottom, or structural and hydraulic requirements (Ballinger and Drake 1995). Common shapes for fish passage design include round and elliptical pipes, box culverts, and open-bottom arches. All types of culvert shapes have been used for fish passage, and selection is likely the result of site conditions and personal preference (Bates et al. 2003). Table 6.7 is a collection of noted advantages and disadvantages of culvert shapes and materials.

Table 6.7 Advantages and Disadvantages of Different Culvert Shapes for Fish Passage Installations (from White 1997; Norman et al. 2005; Bates et al. 2003; Robison et al. 1999)
Shape Advantages Disadvantages
Bridge Usually the best alternative for fish passage. Cost
Circular Structurally and hydraulically efficient. Greater depth of fill allowable for given span, and easier installation (in reference to Arch or Pipe Arch installations). More prone to clogging at high flows. Flexible walls in large culverts require special care during backfill construction.
Pipe-Arch and Elliptical Wider section available for low flows with less height. For buried culverts, installation can be difficult.
Arch Very good fish passage when sized adequately. Allow natural streambed material to be maintained in new installations. Expensive installation. Not practical when stable footings cannot be created.
Structural plate (Round or Arched) Can be placed on the bedding and partially backfilled with top plates left off. Distortion during compaction can lead to problems joining final pieces. Structural plate pipes should not be backfilled until all plates are completed and bolts tightened.
Box Easily adaptable to a variety of situations. Not as structurally and hydraulically efficient as other shapes due to angled corners.
Multi Cell Allow adequate capacity in low profile situations. Lower road bed elevation. Prone to clogging due to area between the barrels and smaller individual culvert size.

Corrugated metal culverts are commonly used in fish passage design. These structures provide boundary roughness that may be conducive to fish passage (Powers et al. 1997; Barber and Downs 1996; Behlke et al. 1989), as well as aiding in retention of bed materials (Bates et al. 2003). Culvert embedment is also commonly called for, with some exceptions in hydraulically designed culverts. When new installations utilize natural bed material, bottomless structures have the advantage of allowing natural substrate to remain in place.

Standards for bridges, culverts, foundations and backfill can be found in "Standard Specifications for Highway Bridges, 17th edition" (AASHTO HB-17, AASHTO, 01‑Sept, 2002).

6.5.8 Simulation Software

Several computer programs exist to assist the engineer in the design process. FishXing is most often used to assess culverts for fish passability and is often used in conjunction with FishBase. HY-8 v. 7.0 is used to analyze the detailed hydraulics of culvert flow and for design. HEC-RAS is used for design on larger rivers and for culverts in series where the water surface elevation of one culvert is affected by another. All of these programs predict average cross-sectional velocities within the culvert barrels. FishXing

FishXing (pronounced "fish crossing") is a fish passage analysis tool developed by the United States Forest Service. According to product description, FishXing provides the following features (United States Forest Service 2006a):

  • Allows for comparison of multiple culverts designs within a single project
  • Calculates hydraulic conditions within circular, box, pipe-arch, open-bottom arch, and embedded culverts
  • Contains default swimming abilities for numerous North American fish species
  • Contains three different options for defining tailwater elevations
  • Calculates water surface profiles through the culvert using gradually varied flow equations, including hydraulic jumps
  • Outputs tables and graphs summarizing the water velocities, water depths, outlet conditions, and lists the limiting fish passage conditions for each culvert

This software is free and available for download at

Noted limitations include:

  • Incomplete fish swimming ability data (although the program does provide the option for user input of swimming values)
  • Roughness coefficient selections limited and not always practical
  • Steep learning curve
  • Validation issues

This program has been recommended as a first cut analysis tool, but for concrete prioritization, design or analysis site visits and analysis should be completed (Cahoon et al. 2005). Analysis with field assessment and study has found FishXing to match results between 71-100% of the time (Rajput 2003; Cahoon et al. 2005). A powerful use for FishXing is in a culvert assessment of "indeterminate" designated crossings. The software may be able to move a designation to "passable" or "impassable." FishBase

FishBase is a searchable relational database catering to different professionals including research scientists, fisheries managers, zoologists and many more. It contains information on over 28,500 fish species, including pictures, data on swimming speeds, distribution, biology, and references. It is available on CD or on the web at HY-8 v. 7.0

The HY-8 v. 7.0 Culvert Analysis program was developed by FHWA in order to automate some of the information contained within HDS-5, "Hydraulic Design of Highway Culverts," HEC-14, "Hydraulic Design of Energy Dissipaters for Culverts and Channels," and HEC-19, "Hydrology." It is intended for hydraulic capacity design, but is useful in evaluating design flood stability, scour potential, and culvert barrel velocity. Maryland suggests the use of other programs for the calculation of tailwater rating curves (Maryland State Highway Administration 2005).

This software is free, and available for download at

Noted limitations include:

  • Cannot be used for horizontal or adversely sloped culverts
  • Cannot explicitly simulate embedded culverts (user must approximate modified shape) HEC-RAS

The Hydrologic Engineering Center River Analysis System (HEC-RAS) is a river modeling program developed by the U.S. Army Corps of Engineers. HEC-RAS can be used to perform hydraulic calculations for a full network of natural and constructed channels. Users have the ability to place culverts within channel context and perform analyses of one-dimensional steady and unsteady flow. The steady flow component is capable of modeling subcritical, supercritical and mixed flow regimes, while the unsteady flow component was developed primarily for subcritical flow calculations.

HEC-RAS is free and available for download at

Noted limitations include:

  • Very steep learning curve Commercial Programs

There are many commercial programs available for analysis and design of culverts, but their applicability has not been evaluated for this publication. A short discussion of many of these programs is available in Environmental Stewardship Practices, Procedures, and Policies for Highway Construction and Maintenance. Final Report for NCHRP Project 25-25, Task 4, National Cooperative Highways Research Program Transportation Research Board (Venner Consulting and Parsons Brinkerhoff 2004).

The FishPass Website is a sponsored project of the Bioengineering Section of the American Fisheries Society. FishPass is a forum for professional discussion of the biological and engineering science of upstream and downstream fish passage. Areas of discussion include fish passage technologies, projects, swimming capabilities and behavior and biological and engineering studies and events.

Subscription details are available at

<< Previous Contents Next >>

Updated: 09/22/2014

United States Department of Transportation - Federal Highway Administration