highway safety publications journals road maintenance engineering injury litigation traffic safety facts accident collision investigation consulting court liability transportation organizations departments of transportation safety organizations road construction legal cases traffic accidents collisions highway safety.  See our highway safety expert services and publications.
Road Engineering Journal
Copyright © 1997 by TranSafety, Inc.
November 1, 1997
TranSafety, Inc.
(360) 683-6276
Fax: (360) 335-6402

Designing Highway Culverts That Do Not Impede the Movements of Resident Fish Species

State agencies have been pressuring the Virginia Department of Transportation (VDOT) to install bridges instead of culverts across trout streams. Various factors associated with culverts, including shallow water and greater turbulence, can cause difficulties for migrating fish and affect their genetic diversity and long-term survival. Bridges usually do not create these difficulties; however, they are considerably more costly to install and maintain. As such, culverts are often more feasible; and, if installed properly, "can reduce the adverse effects on fish while maintaining hydraulic efficiency."

G. Michael Fitch studied the effects of highway culverts on the migratory abilities of nonanadromous (resident) fish in Virginia. He reported his findings in "Avoidance of Nonanadromous Fish Passage Impedance Caused by Highway Culverts" (Transportation Research Record 1559).


In contrast to anadromous fish (saltwater fish that migrate to spawn in fresh water), little research had addressed the effects of culverts on nonanadromous fish in high-gradient streams. In addition, VDOT had no guidelines on using culverts rather than bridges for streams providing habitat to nonanadromous fish. The study at hand detailed the problems fish have in passing through existing culverts. The intent was to provide guidelines for installing culverts that eliminate fish passage problems and criteria for avoiding the use of culverts where they would cause passage problems. The result could be dramatic reductions in the need for corrective action and significant decreases in installation, maintenance, and retrofitting costs.


The study focused on a literature review and field survey. The literature review was intended to determine "the maximum swimming speeds various species could maintain for given distances, the minimum water depth in the culvert barrel required for passage, and the maximum outfall height that could be maneuvered." It also sought to learn what other states had done for similar nonanadromous fish-passage problems. The field survey was designed to validate conclusions from the literature by studying the specific culvert characteristics that both allow and obstruct passage for nonanadromous species. Because of their numerous trout streams, the field survey took place in the Salem and Staunton Districts of Virginia. Survey findings were supplemented with information from the Virginia Department of Game and Inland Fisheries. The researchers sampled wild (as opposed to hatchery-bred) brook, rainbow, and brown trout.

In the field survey, researchers initially examined twenty-five culverts; of those, they chose six for further study. Figure 1 is a photograph of one of those six culverts.

Culvert study site in Bath County, Virginia

Information gathered for these six culverts "included depth of flow and stream flow velocity, which was measured at 0.6 of the flow depth using a Teledyne Gurley Model 622 flow meter." The survey crew took these measurements at the inflow, middle, and outflow sections of the culverts. Researchers recorded each culvert's length and slope and also the streambed slope upstream and downstream--to compare streambed slope to the slope of the culvert itself (i.e., not as steep, steeper, or the same). They examined the outlet pool of each culvert for factors that would affect flow, such as depth and debris, and measured the outfall height (distance from the culvert's bottom at the outlet to the top of the outlet pool). The field survey also looked for hydraulic jumps, which are sudden rises in the turbulent water surface depth. Figure 2 shows slopes, stream velocities, and water depths for the six culverts studied.

Culvert slopes and resultant velocities and depths

The fish were detected by electrofishing with a fish shocker. Researchers collected fish from immediate areas upstream and downstream, and from culvert barrels. They identified, clipped, and logged the fish, then released them downstream. Someone visited the sites about four weeks later to find if fish had moved upstream through the culverts. The author mentioned that future researchers might automate this process with radio telemetry tracking devices, which are becoming less expensive and small enough to be used on nonanadromous trout.


The literature review revealed "substantial research" on fish-passage problems through culverts, much of it from western states. Research on the maximum swimming speed of trout was old and showed a variety of values, but it was deemed still accurate. The literature recommended not using baffles to control flow speeds under normal circumstances.

Research findings on minimum water depths that would allow most trout to pass through a culvert ranged from a low value of 8 cm (3 inches) to a high of 30 cm (12 inches). Some researchers concluded minimum depths of 12 cm (5 inches) and 15 cm (6 inches) were necessary; however, to some degree minimum depth depends on fish size--larger fish require more water. The literature revealed several other factors that affect the passage of fish through culverts. Most related to flow velocity or water depth.

Two researchers "recommended that all culverts be countersunk at least 15 cm [6 inches]. . . ." Maryland has required that all new culverts be countersunk a minimum of 30 cm (12 inches). One researcher reported that adult trout could maneuver an outfall height of about 15 cm (6 inches), but most researchers agreed the bottom of the culvert at the outfall should be below the outlet pool to prevent a barrier. Research also confirmed that concrete aprons, often used with corrugated metal pipe arch culverts, should not be used.

The literature revealed significantly conflicting views on the slope at which culverts should be placed; Fitch found it to be "the most highly debated issue for fish passage through culverts." A number of researchers said the slope should always be 0 percent; others found a 0.5 percent slope acceptable. Some suggested that slopes of up to 5 percent would work, as long as they used a baffling system.

In comparing the literature review with field survey results, Fitch found that the literature on nonanadromous species included "some reasonably specific maximum flow velocities and outfall heights for fish passage." However, across the board "these values were significantly less than those published for most anadromous species." For the most part, field tests yielded results that corroborated the literature and matched the expected results. Some of the results from the field survey were limited because the numbers of trout collected were low. But researchers observed fish passing through three of the culverts during sampling, which showed "that the flow velocities, water depths, and outfall heights measured on those particular instances were maneuverable." The numbers of fish caught and studied were low "from a statistical perspective," yet the low numbers echoed the numbers from the literature review and verified that nonanadromous species are more difficult to recapture for study than are anadromous species, which swim in large numbers during migration.

Flow depths and large outfall heights were the two factors that most impeded fish from passing through the culverts. Only one culvert showed a flow velocity above the lowest maximum value recommended by the literature, though this did not mean that flow velocities never rose to more acceptable levels.

One culvert experienced problems with depth of flow, although all six culverts had depths of about 15 cm (6 inches) in at least part of the barrel, which is above minimum levels recommended in the literature. Figure 3 is a photograph showing flow depth thinning as the water enters a culvert.

Critical flow near culvert inlet

Moreover, water depths are not constant throughout the year. The author explained:

. . . stream flow volumes decrease as summer progresses, decreasing the depth of flow. Coupled with the fact that culverts are normally oversized because of conservative watershed computer models, shallow depth of flow is one of the major impedances to fish passage.

Outfall height also proved a problem for passage, especially for younger fish. Two of the culverts had outfall heights that would pose problems for juvenile trout. Figure 4 shows such an outfall.

Perched outfall restricting passage

The field survey and literature review revealed several critical factors in avoiding fish- passage problems. Slope is perhaps the most significant factor when installing culverts, because it affects many of the other hydraulic characteristics of culverts. The author concluded:

. . . because flow velocity, depth of flow, and outfall heights are affected by slope in combination with many other factors, no specific maximum slope is recommended unless all other variables are held constant as well. Since these variables will obviously change from site to site, the results of this study indicate that culverts should be installed at the same gradient as the streambed, provided maximum flow velocity values are not exceeded. . . .

A culvert installed steeper than the streambed can increase flow and decrease depth of flow below minimum values. A culvert installed at a gradient less than the streambed can cause a hydraulic jump at the barrel's inlet, which in turn can cause turbulence and reduce a fish's buoyancy. In streambeds with slopes of 2 to 3 percent, stream flow velocities will begin to impede the passage of nonanadromous trout, unless a mechanism is installed to slow velocity in the culvert's barrel. Figure 5 illustrates problems created by a culvert placed at less than the streambed slope.

Problems when culvert barrel is placed at slope less than that of streambed

Baffles are not recommended, but increasing the roughness coefficient of the culvert's bottom is a solution. This will reduce stream flow velocity and thus increase depth of flow in the culvert's barrel. There has been little research on increasing roughness in the bottom of the culvert. Fitch described one study that:

experimented with peripheral rings to increase the roughness of circular and square culverts. Reductions in flow velocity of 50 to 70 percent were obtained, depending on slope, flow rate, and ring size. On the basis of the research . . . and the field observations of this study, it is postulated that if the bottom of the culvert could be cast to raise the roughness coefficient to a value equal to the streambed bottom, . . . no increase in flow velocity would occur.

This approach has the potential to solve a number of the problems of fish passage through culverts. In fact, "if the cost of developing and installing culverts with such a bottom design is less than the cost of bridge construction and maintenance, culverts could feasibly be used at more stream crossings without impeding the passage of resident trout." Figure 6 is a drawing of such an artificially rough culvert bottom.

Proposed artificial bottom


Based on the literature review, the study concluded that the maximum flow velocities for anadromous fish are much greater than those of Virginia's nonanadromous trout. Virginia's trout species can maneuver a maximum flow velocity of 1.2 meters (4 feet) per second, a minimum depth of flow of 9 cm (3.5 inches), and maximum outfall height of 10 cm (4 inches). These factors all depend on culvert slope. Flow velocity was not the factor most responsible for impeded fish passage; instead, outfall heights and shallow flow depths contributed most to impeding fish passage in the culverts studied. One solution to the fish passage problem would be to increase the roughness coefficient of culvert bottoms, which in turn would allow installing culverts at steeper slopes without affecting fish passage.

Culverts can be the best way to cross trout streams in Virginia, provided certain actions are taken.

  1. The culvert should be on the same slope as the streambed.
  2. The slope of the stream should be less than 3 percent.
  3. The flow velocity should not exceed 1.2 meters (4 feet) per sec under normal flow conditions.
  4. The culvert barrel should be properly countersunk at the outlet.

In addition, newly installed culverts should not use baffles to control stream flow, and concrete aprons should not be used at culvert outlets.

If these actions are not possible or feasible, then bridges should be constructed.

Finally, future researchers should use a different method of tracking the movements of target species. For example, fish might have internal anchor tags with directions for anglers who might catch the fish. The author recommended additional research to determine if creating culverts that simulate stream bottoms is feasible. This would mean creating bottoms with roughness coefficients greater than that of smooth concrete.

Copyright © 1997 by TranSafety, Inc.

Back to Index       Top of Page