Wetted Ramps Selectively Block Upstream Passage of Adult Sea Lampreys

Abstract

Dams fragment stream habitats and fishways around dams typically serve few species that are strong swimmers or jumpers. We tested a prototype wetted ramp designed to allow upstream passage of small-bodied fishes while blocking upstream movement of invasive sea lampreys in the Laurentian Great Lakes. We tested short, smooth ramps with 5–10 mm water depth in various combinations of ramp angle, water flow, and swim channel width with the aim to selectively block adult migrating sea lampreys (Petromyzon marinus) while passing sub-adult white suckers (Catostomus commersonii) and creek chubs (Semotilus atromaculatus). Sea lampreys easily passed a 0.75 ramp at a 5° angle, but very few individuals passed a similar ramp at a 10° angle, and none passed a longer ramp at a 5° angle. Limiting the amplitude of tailbeats in a narrow channel did not hamper lampreys or the other species. Greater water flow, and thereby greater immersion depth on the ramp, fostered passage for all species. Smaller-bodied individuals of creek chubs and white suckers performed best on the ramp. We showed that wetted ramps could be incorporated into fishways at low-head dams to aid the passage of smaller-bodied fishes while also blocking the spawning migration of adult sea lampreys.

1. Introduction

Stream fragmentation threatens fish diversity, with millions of instream barriers in the US alone hindering fish movement [1,2]. Although many barriers have fishways, their effectiveness varies, and the number of species that can pass them is limited [3]. Improving stream connectivity for fishes needs to balance with other ecological and economic needs [4,5]. For example, in the Laurentian Great Lakes, re-connecting fragmented river habitat competes with reducing access of invasive sea lampreys Petromyzon marinus L. to their spawning grounds [6,7]. Native to the Atlantic Ocean, the anadromous sea lamprey is a parasitic fish that feeds on the blood and body fluids of large-bodied fishes, and it is causing substantial damage to the Great Lakes ecosystem [5]. Currently, lampricide treatments targeting juveniles in rivers and instream migration barriers are used to suppress sea lampreys to ecologically and economically acceptable levels [8,9]. Of the native and desirable species residing in the Great Lakes, approximately 65% are estimated to be negatively impacted by the barriers used for sea lamprey control [10]. Low-head dams, which are widely installed into streams of that region, block upstream movement of sea lampreys and non-jumping finfishes. Modifications to these dams, including adjustable-crest barriers, trap-and-sort fishways, and the use of odorants and other signals to divert lampreys, are being explored [11]. These methods are still under development or have other shortcomings [10,11,12], prompting the ongoing search for further solutions.

This study evaluates the potential for wetted ramps, also called “low-head ramped weirs” [13], to selectively prevent the upstream migration of adult sea lampreys. Previous research on wetted ramps has demonstrated their potential to pass fish of various shapes and swimming abilities [14,15,16] despite the challenge of a shallow water depth [3]. However, studies have produced mixed results regarding their effectiveness in selectively blocking adult sea lampreys [15,17]. In tests with sea lampreys, smooth wetted ramps inclined at 10° blocked 85% of them, with steeper inclines proving impassable. Yet, when finfish species were tested under similar conditions, passage rates were generally low (0–40%) and varied widely among species [14,15,17]. This study aimed to refine the design of wetted ramps to fully block adult sea lamprey upstream migration while allowing greater passage success of finfish species.

Based on previous findings [15], we tested the effect of two ramp inclinations and lengths [13,16], water discharge over the ramps, and ramp width on upstream passage of adult sea lampreys and two common riverine Great Lakes fishes, creek chub Semotilus atromaculatus and white sucker Catostomus commersonii. We predicted that steeper and longer ramps would reduce the passage success of lampreys while greater discharge (and thereby greater water depth on the ramp) would foster passage of all three test species. Sea lampreys are anguilliform swimmers with a wide tail beat amplitude [18,19]. We hypothesized that narrowing the swim channel would hamper sea lampreys, but not finfishes, passage by limiting their lateral tail movements and thereby reducing their swimming efficiency.

2. Materials and Methods

2.1. Study Location and Animals

The experiments were carried out at Hammond Bay Biological Station (HBBS) near Cheboygan, Michigan, and at the Saline Fisheries Research Station (SFRS) in Saline, Michigan. The experiments with sea lampreys were conducted during Year 1 at HBBS, while finfishes (white sucker and creek chub) were tested in both Year 1 (SFRS) and Year 2 (HBBS and SFRS). Adult migrating sea lampreys were collected from traps in the Cheboygan River, the Ocqueoc River, and the Carp Lake Outlet. They were stored in two 2000-L holding tanks until used in the experiments. The suckers used at HBBS were collected from Black Mallard Creek via seine nets and held in a separate holding tank. Sea lampreys were held for 2–20 days (longer for the test of the extended ramp) prior to testing, following the protocols used at the USGS Hammond Bay research facility. All finfishes were tested within 72 h of capture. Individuals were used only once. White suckers and creek chubs for Year 2 were caught in the Saline River near the research station using a backpack electroshocker. Adult sea lampreys were tested during the sea lamprey migratory period (late April to early June). Tests with chubs and suckers were carried out after the experiments with lampreys in Year 1 and in the fall of Year 2. The water temperature in holding and experimental tanks averaged 13.5 °C (range 12–15 °C) for lamprey and 20 °C (range 16–22.5 °C) for the other species.

2.2. Experimental Setup and Procedures

2.2.1. Sea Lamprey

For sea lamprey trials, plastic ramps were constructed, measuring 1 m Length × 0.50 m Width with an additional 0.25 m submerged onramp at the downstream end to guide fish into the channel. Treatment 1 had a rectangular wetted channel 90 mm wide and 50 mm deep. Treatment 2 had a square channel of 50 mm × 50 mm and was designed to limit the tailbeat amplitude of sea lampreys attempting to swim up. Ramps were marked at 50 mm increments to the maximum height of 0.75 m. A bundle of 5 mm diameter tubes at the top of the ramps laminarized the ramp flow. A successful passage was defined as any time a fish touched these flow straighteners and then fell back. For sea lamprey tests, ramps were placed in a 700 L rectangular tank (1.5 m × 0.75 m × 1 m) that was divided into three sections: downstream staging pen, ramp plus supports, and 20 L head tank plus supports and submersible recirculation pump(s). Water was recirculated for the duration of each test with a small input of colder lake water to offset heat created by the pumps. After three tests, the water was fully exchanged. Figure 1 shows the arrangement used at SFRS. The setup was kept dark during observations to simulate nighttime migratory conditions. An infrared camera and IR lights were used to record animal behavior. All videos were recorded at 30 frames/s using a digital video recorder (DVR).

Each test run involved four individual sea lampreys observed for four hours. Three fish in each test run were individually marked by small reflective tags attached to two locations along the dorsal fin by monofilament line, and one fish was left untagged. Tagging was needed because most of the lampreys were too close in length to distinguish them by size in the video recordings. A total of 129 lampreys were tested in 33 experimental runs (4 fish × 33 runs = 132 fish − 3 died/escaped fish). Fish size was measured after each test run. The sea lampreys averaged 480 mm in total length (45 mm SD, range 190–570 mm) and 233 g (52.5 g SD) in weight.

Three experimental variables were manipulated during sea lamprey tests: ramp inclination, channel width, and flow rate. Angles of 5° and 10° from the horizontal were chosen based on previous research with similar designs [15,17]. Two levels of discharge were tested, 0.3 L/s (“low flow”) and 0.6 L/s (”high flow”). An additional treatment (using 32 lampreys in eight runs) was tested in late June of Year 1: an extended ramp of 1.75 m length with 50 mm width wetted channel and 0.6 L/s flow.

Table 1. Treatment combinations used in the wetted-ramp-style experiment and resulting water velocities and water depths.

Species	Ramp Angle	Channel Width	Ramp Length (m)	Flow Rate (L/s)	Mean Flow (m/s)	Mean Water Depth (mm)
Lamprey	5°	50 mm	1	0.3	0.44	8.3
Lamprey	5°	50 mm	1	0.6	0.94	16.7
All species	10°	50 mm	1	0.3	1.05	6.7
All species	10°	50 mm	1	0.6	1.17	11.7
Lamprey	5°	50 mm	1	0.3	0.82	5
Lamprey	5°	50 mm	1	0.6	1.05	16.7
All species	10°	50 mm	1	0.3	1.00	4
All species	10°	50 mm	1	0.6	1.15	8.4
Lamprey	5°	50 mm	1.75	0.6	0.91	15

lists the treatment combinations, including water velocity and depth in the wetted channels.

2.2.2. Finfish

The observations of chubs and suckers were carried out in a similar manner to the sea lamprey test but using only one angle (10°) because the 5° ramp was easy to pass for lampreys (see Results section). The experimental apparatus at SFRS was similar to that at HBBS, with the main difference being a slight reduction (~15%) in overall tank size. Three untagged individuals of the same species but distinctly different sizes were picked for each test run, and their lengths were measured after concluding the 4-h recordings. These fish were left untagged to minimize handling stress, while their greater variability in length (compared to the lamprey) allowed distinguishing up to three untagged individuals in the video from one test run. A total of 129 finfishes (69 chubs, 60 suckers) were tested in 43 experimental runs. Creek chubs averaged 16.4 cm (3 cm SD, range 9–25 cm) in total length, while white suckers averaged 22.8 cm (6.4 cm SD, range 8–38 cm).

2.3. Data Extraction and Analysis

Each ramp was divided into three sections: entrance (0–100 mm from the waterline), mid ramp (100–400 mm), and top (400–700 mm). Apart from determining “successful passage” (reaching 0.75 m on the ramp plane/1.75 m on the extended ramp), we extracted from the video for each “attempt” the mean height on the ramp plane (MH) to the closest cm, swimming speed over the ramp (ground speed), and tail beat frequency (for lamprey only). An “attempt” to climb was analyzed anytime a fish reached at least 100 mm on the ramp plane. Mean height for an individual was calculated by averaging the greatest heights achieved during each attempt, up to a maximum of the first 10 observed attempts for that individual. For calculating mean height, each successful passage event was included as a value of 0.75 m (1.75 m for the long ramp). Due to the ramp not leading to an upstream tank, any fish could have repeated successful passage attempts included in the calculation of mean height. Swim speed was determined by counting the number of frames it took a fish to travel a specified distance and dividing this number by the frame rate. Swim speeds (m/s) were standardized to fish body lengths (BL/s) for most analyses. Tail beat frequency (in Hz) was defined as the number of full undulations of the tail in one recorded second (30 video frames) in the mid-ramp section. This number was counted to the nearest 1/4 tailbeat. Where a full 30 frames could not be counted, no tailbeat frequency was recorded for that attempt.

The means of sea lamprey and finfish entry velocity (swim speed over ground, measured at 0–150 mm from entry point onto the ramp), mid-ramp swim speed (between 200 and 400 mm from entry point), and mean height achieved on the ramp plane were analyzed using a 2-way factorial ANOVA after pooling data from the narrow and wide wetted channels (see Results). A t-test was used to compare mean tailbeat frequency between angles of inclination. We transformed data using square root (for swim speeds) and inverse square root transformations (−1/√x for mean heights) to achieve a normal distribution. Mean heights achieved on the ramp by the suckers and chubs were analyzed with a Kruskal–Wallace test because data transformation did not achieve normality.

3. Results

3.1. Attempt Rates, Passage Success Rates, and Effect of Channel Width

In total, 80% of tested lampreys, 60% of chubs, and 50% of suckers attempted to swim up the ramp within the 4-h observation window. The rate of successful passage over the 1 m long wetted ramp was much greater for the two finfish species (97% of attempting individuals, or 39/42 chubs and 29/30 suckers) than for sea lampreys (39% of 102 attempting individuals). Passage success of lampreys was much higher at the 5° angle than at 10° (74% vs. 2.4%). The two sea lampreys that passed the 10° ramp did so in the treatment with a narrow channel and high flow.

The performance of lampreys was very similar on the narrow (50 mm) vs. wider (90 mm) wetted channels (t-tests on the mean height on ramp, entry, and mid-ramp swim speed: all NS). Therefore, the data from the two widths were pooled for the analysis of the other treatment effects. White suckers, but not creek chubs, achieved greater mean height (MH) on the narrow channel (suckers: 73 cm vs. 60 cm, U = 191, p_(α) = 0.001; chubs: 71 cm vs. 68 cm, U = 270, NS). Ramp width did not have any significant effect on the observed swim speeds of suckers or chubs (entry speed: F_1.69 = 0.7, NS; for mid-ramp speed: F_1.69 = 0.04, NS) and, therefore, the data from two widths were pooled for subsequent analyses.

3.2. Effect of Ramp Angle and Water Flow

Ramp angle did not affect the swim speed of lampreys near the middle of the ramp (Figure 2a, F_1.90 = 0.01, NS), but at the 5° ramp inclination, the mean entry speed onto the ramp and mean height (MH) achieved on the ramp were significantly greater (MH: F_1.90 = 102, p_(α) < 0.001; entry velocity: F_1.90 = 8.16, p_(α) = 0.005) (Figure 2b,c). Analysis of lamprey tail beat frequency and swim speed on the ramp suggested that greater effort was required at the 10° inclination to reach the same average ground speed as in 5° treatments (0.38 BL/s). At the 10° angle, sea lampreys showed much greater average tail beat frequencies while on the ramp (Figure 3, F_1.90 = 107, p_(α) < 0.0001).

The higher flow rate benefited the passage of lampreys. Their success rate was 51% at high flow versus 18% at low flow (pooled data from 5° and 10° treatments). Lampreys reached a significantly greater MH in high-flow treatments compared to low-flow (48 cm vs. 40 cm F_1.90 = 10.44, p_(α) = 0.002, Figure 2c). In the low flow treatment, lamprey mean entry speed onto the ramp was about 20% greater (0.45 BL/s for low flow vs. 0.37 BL/s for high flow, F_1.90 = 10.44, p_(α) = 0.002. Figure 2b) but the mean mid-ramp swim speed was similar (0.37 BL/s at low flow vs. 0.4 BL/s at high flow, F_1.90 = 0.67, NS. Figure 2a). The majority of observed lampreys (22/40) accelerated on the lower part of the ramp when the water flow was high, but only 28% (15/54) achieved, on average, positive acceleration at low flow (two-sample proportion test: Z = 2.65, p_(α) = 0.008). The flow rates we used did not affect the MH of creek chubs (U = 181, NS) or white suckers (U = 76.5, NS). Flow rate also did not affect the entry or mid-ramp swim speeds of chubs and suckers (entry speed: F_1.69 = 2.4, NS; mid-ramp speed: F_1.69 = 0.8, NS).

When comparing the performance of the two species, creek chubs achieved greater mid-ramp speeds (mean of 3.7 BL/s) than suckers (2.6 BL/s) (F_1.69 = 17.6, p_(α) < 0.001, Figure 4), but this was largely caused by the different size ranges of the two species in our experiment. There was no discernible effect of fish length on entry speeds, but on the ramp, larger individuals achieved lower absolute mean swim speeds (in m/s), with swim speed reduced by about 25% for every 100 mm increase in length (Figure 5). After controlling for the effect of body length, the two species swam at similar speeds on the ramp (ANCOVA of mid-ramp speeds in m/s with body length as a covariate, F_1.70 = 0.56, NS).

3.3. Effect of Ramp Length

Due to time constraints, we only tested sea lampreys on the 1.75 m/5° ramp. None of the 27 attempting lampreys reached the top of the extended ramp. The maximum achieved distance on the extended ramp was 57 cm, reached by two individuals. The MH achieved on this ramp was 46 ± 6.8 (SD) cm, which was less than the 60 cm observed on the standard ramp at the same angle/flow (Figure 2c, t = 9.4, p_(α) < 0.001).

4. Discussion

We report the successful use of a wetted ramp as a selective upstream migration barrier against an invasive fish species while maintaining adequate passage rates for other small-bodied fish species. In some configurations, the wetted ramps blocked 100% of the adult sea lampreys tested but enabled upstream passage of most of the finfishes we tested. As expected, a steeper ramp angle reduced the upstream distance fishes reached, and a greater ramp length eliminated the passage of lampreys. We found no support for our hypothesis that sea lamprey upstream movement could be inhibited by limiting their tail beat amplitude.

4.1. Attraction, Passage Success Rates, and Effect of Ramp Inclination

A passage device has to entice fish to attempt passage and allow the fish to pass without undue stress [20,21]. In our lab setup, most of the lampreys and about half the individuals of the other two species attempted passage in a short observation period. Fishway attraction varies widely under field conditions [21] and depends on many factors [22,23]. Successful passage can be strongly influenced by the angle of a passage device [3,15,24], which we confirmed for lampreys in this study. A ramp angle of 10° stopped all but two of the >50 sea lampreys we tested, while a 5° angle allowed 74% of tested lampreys to reach the top of the ramp. No lampreys passed the 1.75 m ramp at a 5° angle, but since those fish were tested later in the migratory season and did not swim up as far as individuals tested earlier on the shorter ramp, it is possible that a 1.75 m/5° ramp is passable by migrating sea lampreys in top condition. Further research is also needed to find out if that angle/length combination is suitable for small-bodied finfish species. The cause of the poorer performance of sea lampreys on the steeper ramp was probably a combination of greater water velocity, decreased water depth, and greater effect of gravity [3].

Water velocity is a limiting factor for fishway design and often dictates which species are capable of using a particular fishway, e.g., [3,25]. As water velocity increases, the number of species that can use a fishway decreases as fewer and fewer species are capable of sustained swimming long enough to traverse the device [3]. Adult sea lampreys are capable of using vertical slot fishways with similar water velocities to those found on the ramps tested here (~1.0–1.25 m/s vs. 0.4–1.2 m/s in our experiments) [5,26] and are estimated to have a maximum sprint swimming velocity of ~4 m/s [27,28], which means that water velocities alone did not prevent successful passage of sea lampreys in our study. When we add water velocity on the ramp to the ground speed we calculated from the video, our lampreys swam on the ramp at a maximum observed speed of about 1.5 m/s. We posit that the challenge of fast water to swim against was amplified by the limited water depth on the wetted ramp. When fish swim at or near the air–water interface, the potential thrust is reduced by surface waves that increase resistance [29]. Corniuk [19] found that adult sea lampreys swimming in 10 mm water depth used greater tail beat frequencies and larger tail beat amplitude to achieve the same swim velocity as fully-submerged individuals, which suggests lower swimming efficiency in very shallow water.

The wetted-ramp configurations we tested blocked very few of the white suckers and creek chubs. Similar success rates have been seen in galaxid species swimming on wetted ramps with comparable angles and water depths [16]. The finfish species we studied likely succeeded where lampreys failed due to their greater maximum swimming speeds [18,27]. White suckers are fairly strong swimmers [30,31] and can pass fishways against swift currents when fully submerged [25,32]. Much less is known about the swimming performance of creek chubs [33,34]. Kivari [14] tested the same two finfish species as we did on similar wetted ramps and found a much lower passage rate, which was probably due to the comparatively greater water velocity and lower water depth on the ramps in Kivari’s study [14].

4.2. Effect of Channel Width, Water Flow, and Fish Size

Our expectations that reduced wetted channel width would impede sea lampreys in their passage over the wetted ramp were not met. Corniuk [19] measured an average tail beat amplitude for adult sea lampreys in shallow (10 mm) water of 0.11 BL, or about 50 mm, which was the width of the narrow ramp in our experiment. Nearly all of the sea lampreys tested on the narrow ramp made contact with the sidewalls of the device as they swam up, and in a few cases, lampreys wedged themselves between the walls of the water channel to hold their position. Although sea lampreys are not capable of using their oral disc to aid in climbing, like the Pacific lamprey (Entosphenus tridentatus) [35], their ability to use surface structures as climbing aids has been well documented [36,37]. We conclude that curtailing the tailbeat amplitude of lampreys is not a promising method for impeding their upstream passage. Creek chubs and white suckers were not hampered by the narrow channel width we tested either; indeed, suckers seemed to gain a slight benefit from the narrower ramp, possibly aided by the greater water depth that resulted from constricting the width of water flow.

Increasing the pump discharge resulted in greater water depths on the ramps, which increased the relative submersion depth of fish using those devices. This was expected to increase the performance of both fish groups as swimming ability is positively correlated with propulsive surface area [29], and this surface area is diminished in very shallow water. At the low flow rate, lampreys entered the ramp at greater velocity but then lost speed, while at high flow, a majority of observed individuals managed to accelerate and thereby swim farther up the ramp. The suckers and chubs we tested seemed largely unaffected by the flow conditions we tested. Both species passed the 10° ramp with relative ease. However, we observed that larger individuals of the two species achieved lower ground speeds on the ramp. Previous research showed that wetted ramps are more challenging for larger-bodied individuals [14,17] and fishes with a less-fusiform body shape [15]. Kivari [14] used high-speed video recordings to show how larger creek chubs and white suckers apparently struggle more than smaller individuals when swimming on wetted ramps.

5. Conclusions

Wetted ramps could provide a suitable alternative or addition to traditional fish passage designs, especially where small-bodied fishes are a management concern [3,38,39]. No single fishway design is likely to pass all fish species [3,5,40], but a combination of ramps with varying configurations may allow a fairly wide range of target species to use wetted-ramp fishways [13,41]. More research will be needed, especially under field conditions, to determine what additional small-bodied species can pass wetted-ramp-style fishways of various lengths and angles, how water temperature impacts passage over wetted ramps (Reinhardt unpublished data), and what engineering solutions will be required to (a) maintain wetted ramps within the design parameters of water depth and velocity, (b) attract fish to ramps despite a low attraction flow over the ramp, and (c) avoid fouling under field conditions. One of the limitations of wetted ramps will be the height of obstacles, such as low-head dams, that can be made passable. At a 10° angle, a 1 m long ramp has a 0.18 m elevation gain, i.e., 5.6 m of ramps plus resting areas would be needed for a 1 m rise. While this large space requirement still compares favorably with nature-like fishways that have much gentler slopes [42] and, therefore, need longer runs for the same rise, it seems likely that wetted ramps will be most useful for fishways over smaller obstacles, such as culverts [39]. Wetted-ramp-style fishways also have the potential to meet the need for selective upstream fish passage of native species at the low-head barriers currently used for sea lamprey control throughout the Great Lakes Basin. If our findings hold up under field conditions, a short section of the wetted ramp could be used as a sorting component within existing fishways, possibly alongside other methods for separating migrating adult sea lampreys from native fishes [43].