lacey, jay; university of sherbrooke, civil engineering dept. … · 2016. 2. 2. · draft...

Draft

Numerical Study of an Innovative Fish Ladder Design for

Perched Culverts

Journal: Canadian Journal of Civil Engineering

Manuscript ID cjce-2014-0436.R2

Manuscript Type: Article

Date Submitted by the Author: 02-Sep-2015

Complete List of Authors: Duguay, Jason; Université de Sherbrooke, Civil Engineering Lacey, Jay; University of Sherbrooke, Civil Engineering Dept.

Keyword: environmental < MANUSCRIPT CLASSIFICATION, water supply-irrig;drain < Hydrotechnical Eng., fluid mech & hydrodyn < Hydrotechnical Eng., highways < Transportation, transportation structures < Construction

https://mc06.manuscriptcentral.com/cjce-pubs

Canadian Journal of Civil Engineering

Draft

1

Numerical study of an Innovative Fish Ladder Design for Perched Culverts

Jason M. Duguay1 R.W. Jay Lacey2

Corresponding Author:

Jason M. Duguay, Phd. Candidate

Departement de génie civil, Université de Sherbrooke 2500 boul. de l'Université, Sherbrooke, QC J1K 2R1 [email protected] R.W. Jay Lacey, Associate Professor Departement de génie civil, Université de Sherbrooke 2500 boul. de l'Université, Sherbrooke, QC J1K 2R1 [email protected] Telephone: 819 821-7110

Page 1 of 35



Draft

2

ABSTRACT

A fish ladder designed to facilitate fish passage at the outlet end of perched culverts is

investigated with a 3D computational fluid dynamics model. The fish ladder consists of a

series of alternating arch baffles with geometries providing options for fish passage over

varying flow and debris conditions within the ladder. At high flows, the baffle’s protruding

center arch increases pool depth, reducing the volumetric bulk turbulence of the pools and

improving fish habitat. The arch baffle is compared to a standard baffle design currently in

use and demonstrates potential advantages for fish passage. A recirculation zone of low

velocity occupies a large volume of the pool believed to provide appropriate hydraulic habitat

for resting and staging jump attempts upstream. This numerical study provides an

acceptable design for future physical prototype testing in the laboratory or field to verify

hydraulics and evaluate fish passage effectiveness.

Keywords: Perched Culverts, Fish Passage, Fish Ladder, Numerical Modeling.

Page 2 of 35



Draft

3

INTRODUCTION

Barriers to fish migration caused by perched culverts are an unfortunate reality of the

extensive road and rail networks that span North America. Perched culverts are the result of

a combination of incorrect culvert installation and sediment scour. Through erosive

processes, an excessively large scour hole can develop over the immediate tailwater reach of

the culvert resulting in a vertical drop between the culvert invert and the downstream water

surface varying anywhere between a few centimeters to several meters. Perched culverts are

known to have an important influence on the distribution of fish species (Norman et al. 2009;

Davis and Davis 2011; Doehring et al. 2011; Harvey and Railsback 2012) in fluvial habitats.

The construction of a fish ladder or rock ramp at the downstream end of the culvert is a

common approach to improve connectivity between the upstream and downstream reaches.

However, in remote regions, many fish ladder designs may prove impractical due to a lack of

available resources (i.e., concrete, large boulders). Furthermore, the difficult terrain on

which many perched culverts are found restricts the use of large construction equipment.

Replacing a perched culvert with one designed with fish passage in mind, though arguably the

ideal practice, may not always be feasible due to budgetary constraints or the need to avoid

inconvenient road blockages. In light of these practical considerations, many perched culverts

go unaddressed. Hence, a transportable and relatively easily installed fish ladder would be

useful to improve fish passage at many perched culverts located in remote regions or on

difficult to access terrain.

Structural Plate Corrugated Steel Pipe (SPCSP), due to its lightweight and segmented

design, can be transported to site and assembled by small construction teams. Recent

development of a thermoplastic co-polymer coating for SPCSP has been demonstrated to

increase service life of SPCSP in corrosive and abrasive conditions beyond expectations of

Page 3 of 35



Draft

4

galvanized SPCSP (Penny and Villeneuve 2003). The aforementioned qualities of SPCSP

demonstrate its applicability for use in fish ladder design. The present study employs a

computational fluid dynamics (CFD) model to investigate the flow field developed by an

innovative fish baffle (arch baffle) placed within a half round polymer coated SPCSP culvert.

The arch baffle could equally be fitted within other common smooth surface culvert materials

(e.g., high density polyethelyne, concrete) without substantial influence on the flow field,

however, the practical advantages of the lightweight segmented lengths of SPCSP and the

potential benefits of the low velocity resting zones between the corrugations make SPCSP an

interesting choice of construction material.

BACKGROUND

Fish Passage Stressors

The Department of Fisheries and Oceans (DFO) presents a list of stressors known to

inhibit fish passage at in-stream structures (Savoie and Hache 2002). Among these stressors, a

few are of particular importance for fish ladder design for application at perched culverts.

The flow field of the arch fish ladder will be evaluated with the following four stressors as

principle design criteria; (1) inadequate water depths, (2) excessive velocities, (3) substantial

vertical drops and (4) elevated turbulence levels. Numerous research efforts have investigated

the effects of these four stressors on fish passage performance on North American fish species

and are briefly discussed in the context of the work in the following paragraphs.

Barrier velocities

Fish demonstrate three swim speed modes. The fastest of the three, the burst speed

Page 4 of 35



Draft

5

(sometimes called darting speed), is normally used to evade predators and navigate rapid

reaches of a river such as a choke or low level cascade. A fish’s burst-speed can only be

maintained for a brief period of time (generally less than 10 seconds) and recuperation is needed

after its use. In contrast, a fish’s sustained speed or prolonged speed can be held for a duration

on the order of minutes. Fish generally use their sustained speed to traverse swift reaches of

rivers and streams. Finally, the cruising speed is maintainable for an indefinite period of time

and is used for feeding or migrating over slower reaches of water (Clay 1995). The velocity

thresholds that distinguish these three swimming modes depend on a number of factors

including the species of fish, body length and size, maturity and ambient water temperatures

(Clay 1995; Katopodis and Gervais 2012). The burst-speed of the target species is a critical

design parameter to consider when selecting the geometry of the baffle. Excessive pool

velocities can also diminish the availability of suitable hydraulic refuge zones. Over the last half

a century many studies have been performed with the intent of defining critical swimming

speeds for a variety of fish species. Bell (1990) compiled swim speed thresholds for fishes of

social and economic importance from a variety of sources. Some of Bell’s data is displayed in

Fig. 1 in an adapted form for fish species common to North America. Maintaining velocities

within a fish ladder in the range of the sustained swim speed of a target fish species ensures

a velocity barrier will not hinder its movements. Flow zones characterized by velocities in the

range of the species’ critical swim speed should be minimized and efforts to provide adequate

hydraulic refuge in high velocity should be taken.

Figure 1. Adult fish cruising, sustained and burst swimming speeds.

Turbulence

Page 5 of 35



Draft

6

Turbulence is generally understood as the random fluctuations of the three velocity

components (u', v', w') around their time averaged means (��, �, ��). Many studies have

investigated the influence of turbulence on fish passage, with the majority focusing on how

turbulence affects preferred holding positions and swimming energetics. Fish have been found

to prefer zones of low turbulence (Smith et al. 2005; Silva et al. 2012a) and excessive turbulence

has been shown to significantly reduce fish passage success rates in pool and weir type fishway

(Fouche and Heath 2013). A number of metrics are commonly used to characterize turbulence

levels within a flow for the purpose of fish passage. The volumetric dissipated power (VDP) is

a common metric used to measure the bulk turbulence within pool type fishways (Rajaratnam

et al. 1986; Chorda et al. 2010; Fouche and Heath 2013). VDP, with units of W/m3, gives a

global evaluation of the turbulence within a region of flow. Recommended values vary between

150 W/m3 and 200 W/m3, depending on the size and species of the target fish (150 W/m3 for

trout and 200 W/m3 for salmon; Larinier et al. (1994)). Equation 1 is used to calculate VDP

(�� ) with as density (kg/m3), Q flow rate (m3/s), g as the gravitational constant (m/s2), V as

the volume of water in the pool (m3) and ∆h being the elevation drop between the pools of the

fish ladder.

�� =

� ∆ℎ

� Eq. 1

Turbulent kinetic energy (TKE or k (J/kg)), provides a local evaluation of turbulence

and is used to identify zones of elevated turbulence, which are thought to be undesirable for

fish to hold station (Lacey et al. 2012). TKE is defined as the sum of the variance of the three

velocity components at a point in the flow, � = 0.5�� +��

� + ��, where σi is the

standard deviation of the respective velocity component. Few studies have quantified

Page 6 of 35



Draft

7

maximum k thresholds above which a flow volume becomes an undesirable as a holding

station for specific target species. A study on Iberian barbel (Luciobarius bocagei) in a pool

and weir type fishway demonstrated the affinity of individuals, especially smaller adults, for

areas of low k (Silva et al. 2011). The study also noted the use of low k zones as resting areas

prior to attempts to traverse high velocity orifices, suggesting the importance of areas of low k

in fish ladder design.

Jump height, pool depth and passageway width

The jumping capabilities of numerous trout and salmon species have been investigated by

a number of researchers (Lauritzen et al. 2005; Kondratieff and Myrick 2005; Brandt et al.

2005; Kondratieff and Myrick 2006; Lauritzen et al. 2010). Their efforts have demonstrated

that passage success over a vertical barrier depends not only on the barrier’s height but also the

depth of the plunge pool and the width of the passageway. Adequate plunge pool depth is

necessary to ensure fish can accelerate to speeds required to propel themselves over the

obstacle. The DFO suggests values of ∆h between pools varying from 100 mm for streams

on small watersheds (< 2.5km2) up to 200 mm for larger watersheds for the design of baffles

for use in road culverts (Savoie and Hache 2002). These are general guidelines which help

ensure the vertical drop between water surface levels is not a vertical barrier to the passage of

weaker juvenile fish.

Brandt et al. (2005) concluded that fish ladders designed for juvenile brook trout

(Salvelinus fontinalis) should have a maximum drop height of 0.1 m, pool depth of 0.1 m (yet,

should be as deep as possible) and a passageway width as wide as possible. Brandt et al.

(2005) determined that increases in passageway width improved jumping success rates

significantly. Kondratieff and Myrick (2006) demonstrated that brook trout with body lengths

Page 7 of 35



Draft

8

between 0.10 - 0.15 m could jump 0.64 m and larger brook trout could attain heights of 0.74 m.

For a given species, maximum jump height is highly dependent on the complex interaction of

body size, maturity and other parameters such as water temperature, depth and flow regime

(i.e., plunging or skimming flow; Branco et al. (2013)). However, reducing the maximum ∆h

observed within a fish ladder is important to ensure that the ascension of the target species

with the weakest jumping capacity is not hindered.

THE ARCH BAFFLE

The present study employs a 3D computational fluid dynamics model (CFD) to

investigate the hydraulic characteristics of the fish baffle form (arch baffle) presented in Fig. 2.

The arch baffle was conceived through numerous simulations and baffle design modifications

performed by the authors and is based on a conceptual idea by Ken Hannaford, biologist for

the Government of Newfoundland and Labrador, Canada. The arch baffle consists of a lower

principal passageway and a higher secondary passageway. The two are separated by a convex

arch protruding through the water surface under all but the highest flow rates. In Fig. 2, the

radii of the principal and secondary passageway are given in terms of the radius, R, of the

fish ladder material (in this case SPCSP where R = 1.22 m), y is the depth between the bottom

of the baffle and the lowest point of the principal passageway, y’ is the distance from the

lowest point of the principal passageway and the highest point of the baffle, and Y is the

distance from the bottom of the baffle to the lowest point of the secondary passageway. The arch

baffle dimensions could be modified to accommodate the use of a square or rectangular cross-

section fish ladder design and is not limited to use in corrugated steel pipe. The front view of

the downstream facing ramp is also shown in Figure 2. The ramp respects a 2:1 slope

Page 8 of 35



Draft

9

beginning at the lowest point of the trough of the principal passageway and extends outwards

by a distance of 0.2 m irrespective of the radius of the material used in accordance to Savoie

and Haché (2002).

For the present study, the dimensions of the arch baffle were chosen to match those

required to construct a planned physical prototype ladder on a perched culvert located in

Newfoundland, Canada. This culvert shows promise as a potential site for future field testing

and the physical dimensions of the modeled baffle are presented in Fig. 3. The radius (R) of

the culvert is 1.22 m, giving dimensions of the arch baffle depicted in Fig. 2 of Y = 0.73 m, y’

= 0.30 m and y = 0.50 m.The protruding convex arch serves to direct flow through the

principal and secondary passageways, and to retain water for increased pool depths at higher

flow rates. The wide principal passageway is thought to provide fish with a large surface area

for improved jumping success as demonstrated in Brandt et al. (2005). Furthermore, the

curved form of the baffle is intended to reduce blockages by eliminating abrupt edges for

debris to catch upon. The principal passageway of the baffles are offset over the entire length of

the fish ladder as observed in Fig. 4. A recent study (Silva et al. 2012b), found that for orifice

pool and weir type fish ladders, offsetting the orifices markedly improved ascension rates

compared to an inline orifice arrangement. Silva et al. (2012)b observed velocities in the

recirculation region of the pool to be considerably reduced for the offset orifice

arrangement. In light of these findings, it was decided to offset the baffles of the arch fish

ladder in the hope of achieving similar reductions in velocity and development of low velocity

resting pools.

The baffle spacing of the numerical model was chosen to respect the maximum ∆h of 200

mm prescribed by the DFO for watersheds > 6 km2. Baffle spacing is a function of slope

Page 9 of 35



Draft

10

and can be determined by dividing the chosen value of ∆h by the slope in decimal form

(Savoie and Haché 2002). This is 100 mm higher than the maximum 100 mm ∆h suggested

for juvenile brook trout by Brandt et al. (2005). Yet the present study intends to be as

general as possible by designing a fish ladder that can be used for commonly larger

watersheds and therefore a ∆h of 200 mm was retained. However, baffle spacing can be

shortened to reduce ∆h for applications on smaller watersheds or where ∆h is of great concern.

THE DFO BAFFLE

In the present study, the hydraulics of the arch baffle were compared to those of the

DFO baffle design (Fig. 5) developed under identical flow conditions. The DFO baffle design

was used for comparative purposes since it is a widely used in fish passage applications in

Canada. The dimensions and geometry of the DFO baffle followed the recommendations

outlined in Savoie and Hache (2002) for a watershed size of 6 to 10 km2. The DFO

recommendations stipulate a notch depth of 0.25 m, a baffle height of 0.5 m, a drop between

baffles of 0.2 m and a notch width of 0.3 m (Fig. 5). Unlike the arch baffle, whose dimensions

vary with culvert radius, the DFO baffle dimensions depend solely on watershed size. Though

a watershed size was not known for this numerical study, a 6 to 10 km2 catchment basin was

assumed. This catchment area is representative of a significant number of small to medium

sized streams where perched culverts are commonly found.

NUMERICAL MODEL

The numerical code employed (Flow-3D, Flow Science Inc.) solves the Reynolds

averaged Navier-Stokes (RANS) equations with a coupled two equation k-� turbulence closure

model. The k-� model was applied for its ability to give reasonable approximations of free

Page 10 of 35



Draft

11

surface flows (Rodi 1980) and also for the k spatial output data. A 3D computer assisted

design (CAD) model of the fish ladder (see Fig. 4) was first drawn and then imported into

Flow-3D as a stereolithography file (.stl) with the following physical characteristics; a length

of 10 m, a diameter of 2.44 m, corrugations with a pitch of 0.230 mm, a depth of 0.064 mm

and a radius of 0.057 mm. Stereolithography files accurately represent the surfaces of 3D

objects as an array of interconnected triangular facets. The numerical solver used herein has an

integrated function (FAVORize) which allows accurate resolution of complex geometrical

surfaces such as the corrugations and the curved form of the arch baffle with minimal effort in

mesh construction. The fish ladder was set on a slope of 8.5% and the baffles were separated

by a distance of 2.37 m starting at 0.75 m from the upstream end of the fish ladder to maintain

the ∆ℎ =200 mm between pools.

Mesh Generation and Flux surfaces

Structured mesh blocks were used to discretize the computational domain. A group of six

mesh blocks with cell sizes of 25 mm were created to define the computational domain around

the ladder. Visual verification of Flow-3D’s representation of the baffle and corrugation

geometries suggested that a 25 mm cell size produced sufficient resolution. Given the

minimum cell size, the thickness of the baffles was set at 25 mm instead of the thickness of

polymer coated steel (6 mm) to be used in the fabrication of the actual baffles. The use of a

thicker baffle in the numerical model allows the geometries to be resolved without resorting to

excessively small cell sizes necessitating long simulation times (in the order of days or

weeks). The additional thickness is believed to have only minor effects on the simulated flow

field. Porous flux surfaces were defined at each of the baffles to record the flow rates at

given periods in time during the simulation. The flux surface data output was used in

Page 11 of 35



Draft

12

conjunction with the time dependent mass averaged mean kinetic energy to determine when

steady-state was achieved. Furthermore, a volumetric sampling volume was used to obtain the

pool volume needed for the calculation of VDP (Eq. 1).

Boundary and Initial Conditions

The majority of mesh planes constituting the mesh blocks of the computational

domain were given symmetry boundary conditions. The code assumes a no-slip wall

boundary condition for cells intersecting geometric forms. Velocities in these cells are

estimated using a log-law wall function. The entrance of the upstream mesh block was defined

as a volumetric flow rate boundary and the exit of the lower mesh block was defined as a

pressure boundary with a fixed water level below the ramp. The chosen boundary conditions

allow flow to move into the domain, over each of the baffles and exit the computational

domain at the downstream end.

Numerical Trials

Flows through the arch and DFO baffles were simulated for 2 flow rates (0.0615 and

0.150 m3/s) at a slope of 8.5% matching the slope of the potential field test culvert in

Newfoundland. A number of preliminary simulations were performed to determine the flow

rate required to cause the DFO notch to run full (i.e., water level of the first pool just about

to cascade over the weir portion of the DFO baffle). A flow rate of 0.0615 m3/s was

determined. The velocity and turbulent kinetic energy fields at the passageways of both the

arch design and the DFO design, respectively, were compared at this flow rate and is referred

to hereafter as the low flow rate. Both baffles were then further tested at a high flow rate of

0.150 m3/s, which was found adequate for an appreciable flow to pass through the secondary

Page 12 of 35



Draft

13

passageway. The length of the fish ladder is a determinant factor of its cost, thus fish ladders

are not uncommonly installed on relatively high slopes (> 8.5%) to minimize length. Two

final simulations (0.150 m3/s and 0.0615 m3/s) were performed of only the arch baffle at a

slightly higher slope of 10% to evaluate the flow field at a steeper gradient. For these last two

simulations the baffle spacing was shortened from 2.37 m to 2.04 m to respect the 200 mm drop

between pools.

The findings presented in the results and discussion section of this article only pertain to

the specific geometries and assumptions used in the development of the model as outlined in

the previous sections. The influence of adapting the ladder to accommodate variations in

culvert diameter, flow rate or slope on velocity and turbulence distributions should be

evaluated with either a numerical or physical model.

RESULTS AND DISCUSSION

Low flow rate

Velocities

Figures 6a and 6c present the 3D velocity magnitude colored streamlines of a

characteristic pool at the 0.0615 m3/s flow rate for the arch and DFO baffles, respectively.

The general flow pattern in the arch baffle (Fig. 6a) shows flow moving through the

principal passageway, along the corrugations, across the bottom of the fish ladder and then

channeled through the principal passageway of the downstream baffle. A large recirculation

region is developed in the region of flow labeled 1 in Fig. 6a. Also, large low velocity zones

(labeled 2 in Fig. 6a) are located immediately upstream of the principal passageways. The

flow pattern of the DFO baffle (Fig. 6c) is observed to move through the notch, expand

Page 13 of 35



Draft

14

symmetrically on both sides before contracting through the next downstream notch. The flow

pattern of the DFO baffle is also characterized by two upstream low velocity resting zones on

either side of the notch as indicated by the labels 3 in Fig. 6c.

The velocity magnitudes in Figs. 6a and 6c are within the same order of magnitude,

with the highest velocities being approximately 2.5 m/s where the jet enters the downstream

pool surface of the DFO and arch baffles. The velocities developed at the passageway for

both the arch and DFO baffles are below or near the lower burst swimming threshold for adult

individuals from the salmonid species presented in Fig. 1. However, due to their weaker

swimming capacity, juvenile brown and cutthroat trout as well as Artic char may encounter

some difficulty at both the arch and the DFO baffle passageways over the tested conditions.

Field testing of a full scale arch fish ladder is necessary to verify velocities over a range of

common discharges.

Turbulence

The spatial distribution of k near the passageway of both the arch and DFO baffles at the

0.0615 m3/s flow rate was also evaluated. Figures 6b and 6d depict the k distributions for the

arch baffle and the DFO baffle, respectively. Both baffles demonstrate elevated levels of k

concentrated in the flow volume where the jet impinges on the corrugations of the downstream

pool. The DFO baffles, however, exhibit higher k values (> 0.050 J/kg) in proximity of the

corrugations compared to that of the arch baffle where k values are < 0.040 J/kg. For the DFO

baffle, the impinging jet is characterized by a substantial vertical momentum component. This

momentum is abruptly dissipated over a shallow depth on the bottom center of the fish ladder. In

contrast, for the arch baffle, the vertical component of momentum comprised in the jet issuing

from the principal passageway is less abruptly dissipated over the curved corrugations as it is

Page 14 of 35



Draft

15

entrained downwards towards the next principal passageway. Furthermore, the arch baffle pool

depths are higher (approximately 0.5 m at 0.150 m3/s compared to 0.3 m for the DFO baffle),

meaning energy is allowed to dissipate over a greater depth in the arch baffle, permitting further

reductions in the magnitudes of k. Turbulent kinetic energy levels for the DFO baffle may be

lowered by increasing the depth of the pools (i.e., baffle height) and by increasing the notch

width to provide energy dissipation over a larger surface area of the bottom of the pools.

However, the DFO baffle was designed for application inside roadway culverts (not

necessarily fish ladders) where baffle height has a significant influence on hydraulic capacity.

Therefore the height of the DFO baffle studied here may not be the most appropriate for use in a

fish ladder, where the influence of the baffle height on the hydraulic capacity of the culvert is

less of a concern.

The centerline offset of the principal passageway of the arch baffle allows the zone of

elevated k to be concentrated near the side walls of the fish ladder, leaving the center of the pool

and the region near the opposing wall with lower values of k (< 0.020 J/kg). This may benefit

fish by providing a large zone of low turbulence near the passageway for recuperation before

staging jump attempts. The DFO design demonstrates zones of low k immediately upstream to

the left and right (labeled 4 in Fig. 6d) of the zone of high k where the jet impinges on the

corrugations in the center of the pool. It can be seen in Fig. 6d that these zones are shallower

and smaller in comparison to the resting zone of the arch baffle (label 1 Fig. 6a), yet still provide

resting areas at low flow rates near the upstream passageway with k values of approximately

0.010 J/kg. These zones could be enlarged by heightening the DFO baffle.

Drop heights and Pool Depth

At the low flow rate of 0.0615 m3/s, the ∆ℎ between the downstream and upstream

Page 15 of 35



Draft

16

water levels for both baffles fluctuates close to 0.2 m. This is well below the maximum jump

height of 0.635 m determined by Kondratieff and Myrick (2006) for brook trout with body

lengths between 10 and 15 cm. The pool depth near the upstream passageway developed by

the arch baffle (approximately 0.495 m) and the DFO design (approximately 0.3 m) also

respect the minimum pool depth suggested by Kondratieff and Myrick (2006) of 10 cm and

follows the guidelines in Savoie and Hache (2002) for watersheds > 6 km2.

High flow rate

Figure 7a presents the streamline velocity distribution at the baffle and throughout the

pool of the arch baffle fish ladder for a high flow rate of Q = 0.150 m3/s. From Fig. 7a the

regions of high velocity (> 1 m/s) are shown to be relegated to the sides of the fish ladder in

the wake of the principal and secondary jets. The remainder of the pool is characterized by

velocity magnitudes < 1 m/s similar to that observed for the low flow rate. From Fig. 7a it

can be seen that the velocities at both the principal and secondary passageways are in the

range of 2 to 2.5 m/s, still respecting the lower threshold of burst swim speeds for the

majority of fish species presented in Fig. 1. The majority of the flow passes through the

principal passageway, dissipating energy over the side wall corrugations before impinging on

the downstream baffle wall, then diverts up and over the proceeding baffle. The jet issuing

from the principal passageway impinges along the wall of the fish ladder, thus relegating a

significant amount of energy dissipation to a small area of the pool. This design feature

reduces the magnitude k in the center of the pool compared to those observed in the center of

the DFO pool. Also, a large volume of water exhibiting low values of U and k is situated

close to the upstream baffle between the outfall of the principal and secondary passageways.

The location of this hydraulic refuge zone is likely suitable for fish to rest and stage upstream

Page 16 of 35



Draft

17

jump attempts depending on fish behavior and abilities. Flow from the secondary passageway

falls along the opposite wall and is washed towards the center of the pool along the bottom.

From here, a horizontal vortex develops into an upwelling vertical vortex near the main

downstream passageway.

At Q = 0.150 m3/s the secondary passageway begins to pass enough flow to provide for

fish passage. At the secondary passageway, the ∆h between the downstream water level and

the upstream water level is approximately 0.2 m, whereas the ∆h between the downstream

water surface elevation and the lowest elevation of the secondary baffle is approximately 0.1 m.

These ∆h values are below the maximum jump heights demonstrated for brook trout by

Kondratieff and Myrick (2006). It is interesting to note that the downstream water surface

level is higher than the lowest point of the primary passageway by 0.08 m (drowned weir

effect), suggesting that weaker fish may be able to swim directly between pools without the

need to jump, provided that fish are not hindered by the circulation pattern and eddies formed

(Silva et al. 2012a). Depth above the trough of the primary passageway will continue to

increase at higher flow rates due to the retaining function of the protruding convex arch of the

immediate downstream baffle. Furthermore, increases in flow rate would also result in a rise in

water surface level of the downstream pool which would approach the lowest elevation of the

secondary passageway. These features demonstrate the potential for the arch baffle design to

respect maximum values of ∆h as the flow rate increases through the ladder.

Turbulent kinetic energy distributions of the arch baffle at the high flow rate are

depicted along streamlines in Fig. 7b. At the high flow rate, k values in the pool increase

considerably and range from approximately 0.10 J/kg near the wall downstream of the

principal passageway to less than 0.020 J/kg in the recirculation zone near the

Page 17 of 35



Draft

18

passageways.

DFO baffle

The high flow rate through the DFO baffle is depicted in Figures 7c and 7d. A large

portion of the flow passes through the notch as the remaining flow spills over the weir

portions of the baffle. The majority of the flow is then entrained along the corrugations near

the centerline of the pool before contracting through the downstream notch, similar to the low

flow condition (Fig. 6c). The highest velocities occur along the centerline of the baffle near

the corrugations and occupy a large region of the central flow volume. A low velocity

recirculation zone is observed close to the water surface of the distal end of the pool (label 1,

Fig. 7c). Also, two recirculation zones flanking each side of the jet issuing from the notch can

be observed in Fig. 7c (label 2), which should be beneficial to fish. The cascading water over

the flat upper weir portion of the DFO baffle demonstrates velocities in excess of 3 m/s due to

the vertical acceleration of the flow as it free falls into the pool below. However, the majority

of fish passage should occur by fish swimming through the slot and not by jumping over the

flat upper weir portion, so these higher velocities are likely of little consequence. One possible

remedy to these high velocities would be decreasing ∆h by reducing the spacing of the DFO

baffle.

Turbulent kinetic energy distributions of the DFO baffle are presented in Fig. 7d. Again,

the zone of highest turbulence (0.10 J/kg) is found along the centerline of the pool in

proximity to the corrugations. Also, k values characterizing the recirculation zones on either

side of the issuing jet are elevated compared to the low flow rate. The cascading water

disrupts the low velocity recirculation zones observed at the low flow rate (label 3 Fig. 6c) by

creating distinctly turbulent zones on either side of the issuing jet (label 3, Fig. 7d). These

Page 18 of 35



Draft

19

turbulent zones may reduce access to appropriate flow refuge for resting and staging upstream

jump attempts. The extent to which the cascading water may affect fish passage should be

investigated in future field or laboratory studies. The high k values in the center of the pool

may be improved by constructing the ladder from a smooth material such as concrete or

HDPE, or alternatively, by increasing the pool depth in an effort to divert the flow from

impinging directly on the corrugations. The centrally located zone of high turbulence near the

upstream notch of the pool may hinder fish passage, however, the extents to which it does

should be investigated in a comprehensive field study.

Volumetric Dissipative Power

The volumetric dissipative power of the arch and the DFO baffles for the various tested

slopes and discharge are presented in Table 1. Both baffle designs respect the VDP suggestions

laid out by Larinier et al. (1994) for salmonids, with the arch baffle (60 W/m3) producing

roughly half the value of VDP as the DFO baffle (117 W/m3) at the low flow rate of 0.0615

m3/s. For the high flow rate of 0.150 m

3/s, the protruding convex arch of the arch baffle

increases pool depth by retaining flow. This has the advantage of increasing the retained

volume in the pool with subsequent reductions in VDP. A VDP value of 117 W/m3 is produced

in the arch baffle at the high flow rate compared to 227 W/m3 for the DFO baffle at the same

flow rate. At the high flow rate, the DFO baffle produces a VDP in excess of the recommended

value of 200 W/m3 (Larinier et al. 1994). This could be remedied by increasing the height of

the DFO baffle beyond that recommended for the use of the baffle in culverts as outlined in

Savoie et al. 2002.

Results of the 10% slope

Subsequent simulations were performed on the arch baffle at the slightly higher slope of

Page 19 of 35



Draft

20

10% (results not shown). Analysis of the simulation results demonstrated similar U, k and

general flow structure in the pool as observed for the 8.5% slope. The volumetric

dissipative power increased to 140 W/m3 compared to 110 W/m3 at 8.5% at the high flow

rate (see Table 1), this is due to the reduction in volume caused by the closer baffle spacing

necessary on higher slopes to maintain the maximum 0.2 m drop. Placement of the fish

ladder on slopes greater than 10% will require a closer baffle spacing if a maximum drop

height of 0.2 m is to be respected. The maximum possible slope will likely be limited to one

producing values of VDP < 200 W/m3 as suggested by Larinier et al. (1994) for salmonids

or another appropriate threshold for other fish species. Results from the present study,

however, suggest that the arch baffle is well suited to produce adequate hydraulic

conditions for fish passage at slopes up to 10%, however field studies are needed for

confirmation.

RECOMMENDATIONS AND PRACTICAL CONSIDERATIONS

Field or laboratory testing of the culvert is required on the arch baffle to verify spatial

velocity and turbulence distributions over a range of flow rates common to small to medium

streams at migratory periods of the year. A comprehensive study involving the passage of

live fish would be beneficial to evaluate the effectiveness of the design for a variety of fish

species, age groups and body sizes. Also the arch baffle should be evaluated for its ability to

pass various forms of debris and compared to the DFO baffle to highlight any possible design

advantages in this regard.

The aim of this study was to evaluate the flow field of the arch baffle fitted within a fish

ladder for installation on the outlet end of a perched culvert. This study does not address the

Page 20 of 35



Draft

21

use of the baffle within the interior of a culvert or investigate how the baffle may influence

hydraulic capacity. Interested readers should look to other sources discussing baffle

installation within culverts to understand potential effects on hydraulic capacity (Ead et al.

2002; Duguay and Lacey 2014). For its intended application, the arch baffle fish ladder, should be

installed so that the water level in the first pool (immediately downstream of the culvert outlet)

is equal to that of the original tailwater level used in the design of the culvert. Proceeding in

this fashion ensures minimal consequences on the hydraulic capacity of the culvert from the

addition of the fish ladder. The water level of the first pool can be manipulated by adjusting

the elevation of the first baffle (immediately downstream of the culvert outlet) in the fish

ladder. Such a practice may also benefit fish passage by providing sufficient water depth

along the invert of the culvert itself. Also, the lowest point of the principal passageway of the

last baffle (downstream end of the fish ladder) should be placed at an elevation below the

lowest downstream seasonal water level. Improper installation of the ladder’s downstream

invert may cause a perched condition with the first downstream baffle acting as a vertical drop

during seasonally low flows. The construction of a scour reducing bed design at the

downstream end of the fish ladder would reduce the likelihood of flow exiting the ladder

reproducing a perched condition.

CONCLUSIONS

The results from this CFD study, while requiring physical testing, demonstrate that the

arch baffle produces similar U and k to the DFO baffle, while potentially providing various

improvements in certain respects. The arch baffle develops lower values of VDP compared to

the DFO design, it also confines the zones of high k and U near the corrugations allowing a large

Page 21 of 35



Draft

22

and relatively tranquil recirculation zone to form just downstream of the passageways. Fish

may use this zone to rest and stage jump attempts. The ∆h values of the principal passageway

of the arch baffle were found to respect reasonable limits for salmonids species at the tested flow

rates. At high flow, it was determined that the principal passageway of the arch baffle will be

drowned by the water level of the adjacent downstream pool, possibly providing weaker fish

the opportunity to swim directly between pools without jumping. Additionally, the principal

passageway is drowned further at higher flows due to the protruding convex arch of the arch

baffle, dynamically increasing pool depths and aiding in the reduction of VDP with flow rate.

Furthermore, the increasing width of the passageways with increasing flow rate is thought to

improve jumping success rates as suggested by Brandt et al. (2005). At the high flow rate, the

secondary passageway is shown to develop sufficiently low velocities and ∆h values for the

passage of salmonid species. The secondary slot should provide an auxiliary passage option for

fish and debris in the situation where the principal passageway becomes blocked. Following

extensive physical testing in the laboratory or field, the arch baffle fish ladder should show

promise as a solution to fish habitat fragmentation issues at perched culverts throughout North

America.

ACKNOWLEDGMENTS

The authors would like to extend their gratitude to the Natural Science and Engineering

Research Council of Canada as well as the Corrugated Steel Pipe Institute for the financial

support for this study. Also, extended gratitude is given to Ken Hannaford for the original

concept of the arch baffle and its design features to improve fish passage. We would also like to

thank the two anonymous reviewers for their helpful comments and suggestions which have

Page 22 of 35



Draft

23

resulted in a much improved manuscript.

REFERENCES

Bell, M.C. 1990. Fisheries Handbook of Engineering Requirements and Biological Criteria.

U.S. Army Corps of Engineers, North Pacific Division. Portland, OR.

Branco, P., Santos, J.M., Katopodis, C., and Pinheiro, A., and Ferreira, M.T. 2013. Pool-Type

Fishways: Two Different Morpho-Ecological Cyprinid Species Facing Plunging and

Streaming Flows. PLoS ONE, 8(5): e65089. doi:10.1371/journal.pone.0065089.

Brandt, M.M., Holloway, J.P., Myrick, C.A., and Kondratieff, M.C. 2005. Effects of Waterfall

Dimensions and Light Intensity on Age-0 Brook Trout Jumping Performance. Transactions of

the American Fisheries Society, 134:2, 496-502. doi: 10.1577/T03-175.1.

Chorda, J., Maubourguet, M., Roux, H., Larinier, M., Tarrade, L., and David, L. 2010. Two-

dimensional free surface flow numerical model for vertical slot fishways. Journal of Hydraulic

Research, 48(2): 141-151. doi: 10.1080/00221681003703956.

Clay, C. H. 1995. Design of Fishways and Other Fish Facilities. CRC Press, Inc., Boca Raton,

FL. Second edition.

Davis, J. and Davis, G. 2011. The influence of stream-crossing structures on the distribution

of rearing juvenile pacific salmon. Journal of the North American Benthological Society,

30(4): 1117–1128. doi:10.1899/10-090.1.

Duguay, J. and Lacey, R. 2014. Effect of Fish Baffles on the Hydraulic Roughness of Slip-

Page 23 of 35



Draft

24

Lined Culverts. Journal of Hydraulic Engineering, 141(1): 04014065.

doi:10.1061/(ASCE)HY.1943-7900.0000942.

Doehring, K., Young, R., and McIntosh, A. 2011. Factors affecting juvenile galaxiid fish

passage at culverts. Marine and Freshwater Research, 62(1): 38–45. doi:10.1071/MF10101.

Ead, S., Rajaratnam, N., and Katopodis, C. 2002. Generalized study of hydraulics of culvert

fishways. Journal of Hydraulic Engineering, 128(11): 1018–1022. doi:10.1061/(ASCE)0733-

9429(2002)128:11(1018).

Fouché, P., and Heath, R. 2013. Functionality evaluation of the Xikundu fishway, Luvuvhu

River, South Africa. African Journal of Aquatic Science, 38(1):69-84.

doi:10.2989/16085914.2013.773418.

Harvey, B., and Railsback, S. 2012. Effects of passage barriers on demographics and stability

properties of a virtual trout population. River Research and Applications, 28(4): 479–489. doi:

10.1002/rra.1574.

Katopodis, C., and Gervais, R. 2012. Ecohydraulic analysis of fish fatigue data. River

Research and Applications, 28(4): 444–456. doi: 10.1002/rra.1566.

Kondratieff, M., and Myrick, C. 2005. Two adjustable waterfalls for evaluating fish jumping

performance. Transactions of the American Fisheries Society, 134(2): 503–508. doi:

10.1577/T03-174.1.

Kondratieff, M., and Myrick, C. 2006. How high can brook trout jump? A laboratory

evaluation of brook trout jumping performance. Transactions of the American Fisheries

Society, 135(2): 361–370. doi: 10.1577/T04-210.1.

Page 24 of 35



Draft

25

Lacey, R., Neary, V., Liao, J., Enders, E., and Tritico, H. 2012. The ipos framework: Linking

fish swimming performance in altered flows from laboratory experiments to rivers. River

Research and Applications, 28(4): 429–443. doi: 10.1002/rra.1584.

Larinier, M., Porcher, J., Travade, F., and Gosset, C. 1994. Passes à poissons -Expertises et

conception des ouvrages de franchissement. Conseil Supérieur de la Pêche, Paris (France).

Lauritzen, D., Hertel, F., and Gordon, M. 2005. A kinematic examination of wild sockeye

salmon jumping up natural waterfalls. Journal of Fish Biology, 67(4): 1010-1020. doi:

10.1111/j.0022-1112.2005.00799.x.

Lauritzen, D., Hertel, F., Jordan, L., and Gordon, M. 2010. Salmon jumping: Behavior,

kinematics and optimal conditions, with possible implications for fish passageway design.

Bioinspiration and Biomimetics, 5(3): 035006. doi: 10.1088/1748-3182/5/3/035006.

Norman, J., Hagler, M., Freeman, M., and Freeman, B. 2009. Application of a multistate

model to estimate culvert effects on movement of small fishes. Transactions of the American

Fisheries Society, 138(4): 826–838. doi: 10.1577/T08-156.1.

Penny, D. J. and Villeneuve, D. 2003. Adapting European pavement tests to prove coating

performance on steel culverts. Ministère des transports du Québec. Available from

http://www.cspi.ca/sites/default/files/download/tacBrochure_103.pdf.

Rajaratnam, N., Van der Vinne, G., and Katopodis, C. 1986. Hydraulics of vertical slot

fishways. Journal of Hydraulic Engineering, 112(10): 909–927. doi:10.1061/(ASCE)0733-

9429(1986)112:10(909).

Rodi, W. 1980. Turbulence models and their application in hydraulics: a state of the art

Page 25 of 35



Draft

26

review. International Association for Hydraulic Research (IAHR), Delft, Netherlands.

Savoie, R. and Haché, D. 2002. Design Criteria for Fish Passage in New or Retrofit Culverts

in the Maritime Provinces. Fisheries and Oceans Canada, Moncton, NB.

Silva, A., Katopodis, C., Santos, J., Ferreira, M., and Pinheiro, A. 2012a. Cyprinid swimming

behaviour in response to turbulent flow. Ecological Engineering, 44: 314-328.

doi:10.1016/j.ecoleng.2012.04.015.

Silva, A., Santos, J., Ferreira, M., Pinheiro, A., and Katopodis, C. 2011. Effects of water

velocity and turbulence on the behaviour of iberian barbel (Luciobarbus bocagei,

Steindachner 1864) in an experimental pool-type fishway. River Research and Applications,

27(3): 360–373. doi: 10.1002/rra.1363.

Silva, A., Santos, J., Ferreira, M., Pinheiro, A., and Katopodis, C. 2012b. Passage efficiency

of offset and straight orifices for upstream movements of iberian barbel in a pool-type

fishway. River Research and Applications, 28(5): 529–542. doi: 10.1002/rra.1465.

Smith, D., Brannon, E., and Odeh, M. 2005. Response of juvenile rainbow trout to turbulence

produced by prismatoidal shapes. Transactions of the American Fisheries Society, 134(3):

741–753. doi: 10.1577/T04-069.1.

Page 26 of 35



Draft

27

TABLE 1: Volumetric dissipative power (VDP) obtained for the various fish ladder

configurations.

Baffle Slope (%) Flow Rate (m3/s) VDP (W/m3)

DFO 8.5 0.0615 125 DFO 8.5 0.150 227

Arch 8.5 0.0615 60

Arch 8.5 0.150 117

Arch 10 0.0615 70

Arch 10 0.150 140

Page 27 of 35



Draft

28

List of Figure Captions

Figure 1. Adult fish cruising, sustained and burst swimming speeds.

Figure 2. Approximate dimensions of the Arch baffle relative to the radius of the material

used for the fish ladder.

Figure 3. Modeled dimensions of the arch Baffle in millimeters.

Figure 4. Segment of the modeled arch baffle fish ladder showing the alternating pattern

of the primary and secondary passageways with the downstream ramp visible.

Figure 5. Modeled dimensions of the DFO baffle as outlined in Savoie and Haché (2002).

Figure 6. a) and b) U (m/s) and c) and d) k (J/kg) distributions throughout the flow field

for the arch (a, b) and DFO designs (c, d) as depicted along U colored streamlines at the

low flow rate (Q=0.0615 m3/s).

Figure 7. a) and b) U (m/s) and c) and d) k (J/kg) distributions throughout the flow field

of the arch baffle (a, b) and DFO baffle designs (c, d) at the high flow rate (Q = 0.150

m3/s).

Page 28 of 35



Draft

Figure 1. Adult fish cruising, sustained and burst swimming speeds. 52x15mm (600 x 600 DPI)

Page 29 of 35



Draft

Figure 2. Approximate dimensions of the Arch baffle relative to the radius of the material used for the fish ladder.

33x13mm (600 x 600 DPI)

Page 30 of 35



Draft

Figure 3. Modeled dimensions of the arch Baffle in millimeters. 54x34mm (600 x 600 DPI)

Page 31 of 35



Draft

Figure 4. Segment of the modeled arch baffle fish ladder showing the alternating pattern of the primary and secondary passageways with the downstream ramp visible.

85x83mm (232 x 232 DPI)

Page 32 of 35



Draft

Figure 5. Modeled dimensions of the DFO baffle as outlined in Savoie and Haché (2002). 49x29mm (600 x 600 DPI)

Page 33 of 35



Draft

Figure 6. a) and b) U (m/s) and c) and d) k (J/kg) distributions throughout the flow field for the arch (a, b) and DFO designs (c, d) as depicted along U colored streamlines at the low flow rate (Q=0.0615 m3/s).

120x85mm (300 x 300 DPI)

Page 34 of 35



Draft

Figure 7. a) and b) U (m/s) and c) and d) k (J/kg) distributions throughout the flow field of the arch baffle (a, b) and DFO baffle designs (c, d) at the high flow rate (Q = 0.150 m3/s).

119x83mm (300 x 300 DPI)

Page 35 of 35



lacey, jay; university of sherbrooke, civil engineering dept. … · 2016. 2. 2. · draft...

Documents