Test of a Screw-Style Fish Lift for Introducing Migratory Fish into a Selective Fish Passage Device

Abstract

Barriers are an effective mechanism for managing invasive species, such as sea lamprey in the Laurentian Great Lakes but are detrimental because they limit the migration of desirable, native species. Fish passage technologies that selectively pass desirable species while blocking undesirable species are needed. Optical sorting tools, combined with newly developed computer learning algorithms, could be used to identify invasive species from high-resolution imagery and potentially isolate them from an assortment of the Great Lakes fishes. Many existing barriers lack fishways, and optical sorting may require fish to be dewatered for image capture. The Archimedes screw, a device originating from 234 BCE, offers the potential to continuously lift fish and water over low-head barriers or into an optical sorting device. To test the efficacy of an Archimedes screw and fish lifting to capture and pass Great Lakes fishes, we built a field-scale prototype and installed it at the Cheboygan Dam, Michigan in the USA in 2021. The fish lift safely transported 704 fish (688 of which were suckers (Catostomidae)) in 11 days. The passage of the suckers through the fish lift increased with the water temperature and attraction flow. There were no observed injuries in the transported fish or mortalities in a subset of suckers held post-transport.

1. Introduction

The Great Lakes’ restoration action plans have called for the removal of barriers or the development of bypass methods to reconnect the Laurentian Great Lakes (hereafter the Great Lakes) fish to tributaries [1]. At the same time, these action plans call for the prevention of the spread of invasive species [1]. One of the benefits of removing dams is the increased passage of migratory species, but this action may also allow the passage of invasive species [2], diseases, and even contaminants into otherwise protected watersheds. The conflicting management actions surrounding the barriers lead to the connectivity conundrum [3]. Sea lamprey (Petromyzon marinus (Linnaeus, 1758)) control in the Great Lakes is just one invasive species whose control is dependent on the barriers to migration. In total, 494 lowermost barriers, that is the first barrier between a tributary and its receiving water, prevent adult sea lamprey from accessing spawning habitats in tributaries to all five of the Great Lakes and eliminate the need for a lampricide treatment upstream of these dams [4]. However, one of the impacts of these barriers is the blockage of most native fish species migrating from the Great Lakes, which makes the barriers a major impediment to restoring the habitat and migratory of fish communities [5,6]. Fish passage technologies that selectively allow the passage of desirable species while blocking undesirable species are needed to address the connectivity conundrum surrounding the barriers [3,7,8].

Currently, the only proven method to selectively pass and block fishes is through manual trap-and-sort. While effective at preventing the movement of undesirable fishes, such as the sea lamprey, a number of desirable fish can pass through the manual trap-and-sort and vary widely (e.g., 7–88%) [9], and the technique can be time-consuming, and it can be detrimental to fish health. Thus, there is a need to develop automated sorting methods that minimize human intervention and the handling of fish. Recently, an image library of the Great Lakes fish (6246 images) was collected using an optical scanner (Whooshh FishL^TM Recognition scanner, Whooshh Innovations Inc., Seattle, WA, USA) and used to train a convolutional neural network [10]. The algorithm identified individual fish species with ~95% accuracy and correctly identified sea lamprey from all other species with >99% accuracy [10]. Although the sorting of fish based on the species’ level identification with an optical scanner has not yet been implemented, the same technology is currently used to determine the size of the fish, which then drives the operation of the sorting chute doors, guiding the fish to appropriately sized transport tubes [11]. Therefore, a step towards species-level sorting using an optical scanner is not out of reach. The scanner required the fish to be removed from the water and passed through the scanner, single file, on a wetted ramp to collect consistent and high-resolution images. While other imaging tools exist that do not require fish dewatering, accurate species identification can be hindered by high turbidity [12] or the presence of multiple fish [13]. The volitional entry of highly motivated and mobile fish, such as Pacific salmonids into optical scanners that require fish to be dewatered, can be achieved using short technical fishway sections (e.g., Alaska Steeppass) with a false weir [11]. However, technical fishways are not always designed for a non-salmonid passage and can be much less effective at attracting and passing non-salmonids [14,15,16]. Therefore, a solution for collecting and lifting fish from the water is necessary to further develop an optical sorting tool for non-salmonids.

Archimedean screws, as devices first conceived for raising and transferring water, offer a means to continuously transport fish out of the water and into an optical sorting tool. They use a rotating helical blade to lift water continuously (i.e., pump) or extract energy from the water (i.e., turbine) [17]. Other mechanical fish lifts (e.g., elevators, locks, or Hidrostal pumps) have the means to transport fish out of the water [18] but can delay or prevent fish passage, as they only move fish at discrete times. In addition to providing continuous transport, Archimedes screw turbines are commonly used to provide a generally safer alternative to conventional turbines for downstream passage [19,20], although fish mortality has been reported to be 17–19% for downstream migrating eel [21]. Alternatively, high survival has been demonstrated during upstream passage trials for a variety of fishes. McNabb et al. [22] found that 93–98% of juvenile chinook salmon (Oncorhynchus tshawytscha (Walbaum, 1972)) survived their upstream passage through a 3 m diameter by a 11.5 m long Archimedes screw lift. Another study by Vriese [23] found no mortalities passing nine European freshwater species (N = 99) upstream in a pilot test of a 70 cm diameter screw. Although these findings of low mortality support the further advancement of Archimedes screws to lift fish, this application of Archimedes screws has not been widely investigated. One key knowledge gap is an understanding of the downstream entrance hydraulic conditions of this style of fish lift and how they impact fish capture. Further investigations into these phenomena would help clarify the potential utility of Archimedes screw fish lifts as a tool for advancing selective passage objectives and contending with the connectivity conundrum surrounding the barriers in the Great Lakes and elsewhere.

We designed a field-scale Archimedes screw fish lift (ASFL) and deployed it in a sea lamprey crowding structure in a Great Lakes tributary that is frequently inundated with large numbers of native migratory fish and invasive sea lamprey. The objectives of this pilot study were to (1) characterize the hydraulic conditions at the downstream entrance of an ASFL, (2) investigate their relationship with fish capture and eventual transport, and (3) document survival of native fishes transported by the ASFL. We hypothesized that the hydraulic conditions at the downstream entrance of an ASFL influences fish capture. We predicted that fish capture in a field-scale prototype ASFL would be increased during treatments with an added small jet of attraction flow, compared to treatments without added attraction flow.

2. Materials and Methods

2.1. Study Site

The ASFL was tested at the Cheboygan Lock and Dam on the Cheboygan River, MI, in the USA during the period of 25 April–6 May 2021. The mean spring river discharge is 19 m³/s. The Cheboygan Lock and Dam, located approximately 2.5 km upstream of the mouth in Lake Huron, is comprised of a large earthen embankment, wood, and steel sheet pile spillway with six gates, a non-functional fishway, and an integral sea lamprey trap. The trap complex features four pot-style sea lamprey traps that can be removed via steel slots and are connected to a large pre-trap area that crowds fish near the trap entrances (Figure 1).

The Archimedes screw is comprised of one or more helical surfaces or blades around a central shaft and surrounded by a fixed or rotating outer cylinder. While the screw rotates, water and fish are entrained in the bucket, the volume formed between adjacent flights of the helical surface and constrained by the outer cylinder [24], and transported upstream. The prototype ASFL used in the study was custom fabricated out of stainless steel (Moran Ironworks, Onaway, MI, USA). The ASFL has a single helical blade and a rotating outer cylinder (the blade and outer cylinder were continuously welded together). The ASFL was rotated by an AC gearmotor (Dayton 4ZJ51, Lake Forest, IL, USA), and the rotation speed was controlled using a variable frequency drive (Schneider Electric ATV12H037F1, Rueil-Malmaison, FR), drawing 0.28 kW/h. The geometry of the screw and critical dimensions are provided in Figure 2 and

Table 1. Dimensions and characteristics of the ASFL depicted in Figure 2.

Dimension	Prototype ASFL	Variable
Outside diameter (mm)	762	O.D.
Inner diameter (mm)	101	I.D.
Length (m)	3.05	L
Vertical lift (m)	1.50	H
Angle of inclination (degrees)	27	β
Bucket length (mm)	762	S
Bucket volume (liters)	50
Motor torque (N·m)	115
Rotation (rpm)	12

. An attraction flow of 200 L/min was provided by a submersible pump (Grundfos Unilift AP12 1HP, Brookshire, TX, USA) and two symmetric Schedule 80 PVC pipes (average inner diameter of 3.75 cm) attached to the support beam at the ASFL entrance and aimed at a 45-degree angle towards the centerline of the ASFL (Figure 3). The steel frame supporting the entrance of the ASFL was set within a stoplog slot at the northwest corner of the sea lamprey trap (in place of Trap D, Figure 1c). Steel wire mesh on the supporting frame was used to ensure fish could not bypass the ASFL. Fish transported by the ASFL were deposited into a floating enclosure immediately upstream of the ASFL.

2.2. Hydraulics

The velocity fields at the entrance of the ASFL were measured using an acoustic Doppler velocimeter (ADV) (Vectrino Plus—NotrekUSA, Annapolis, MD, USA). Velocity measurements were taken in front of the ASFL entrance with and without attraction flow along horizontal planes at 30 cm (top of the ASFL) and 80 cm (bottom of the ASFL). For each horizontal plane, measurements were taken at 16.5, 33, and 49.5 cm along the x-axis and 12.7, 19.0, 38.1, 57.1, and 76.2 cm along the y-axis. A total of 60 points were sampled. Velocity measurements were taken at a frequency of 50 Hz over a 120 s duration. Velocity data were filtered and despiked following the recommendations of Wahl [25] using a custom MATLAB (R2021b, Mathworks, Natick, MA, USA) script. These measurements were used to quantify the time-averaged velocities, flow patterns, fluctuating velocities, and turbulent kinetic energy. Velocity fluctuations in each direction $\left({\mathit{u}}^{\prime },{\mathit{v}}^{\prime },{\mathit{w}}^{\prime }\right)$ are the difference between the instantaneous velocity $\left(\mathit{u},\mathit{v},\mathit{w}\right)$ and time-averaged velocity ($\overline{\mathit{u}},\overline{\mathit{v}},\overline{\mathit{w}})$.

$$\begin{array}{c}{\mathit{u}}^{\prime }=\mathit{u}-\overline{\mathit{u}}\\ {\mathit{v}}^{\prime }=\mathit{v}-\overline{\mathit{v}}\\ {\mathit{w}}^{\prime }=\mathit{w}-\overline{\mathit{w}}\end{array},$$

(1)

Turbulent kinetic energy (TKE) is the mean kinetic energy per unit mass associated with turbulent structures in the flow and is calculated by:

$$\mathrm{TKE}=\frac{\mathbf{1}}{\mathbf{2}}\left(\overline{\mathit{u}{\prime }^{\mathbf{2}}}+\overline{\mathit{v}{\prime }^{\mathbf{2}}}+\overline{\mathit{w}{\prime }^{\mathbf{2}}}\right),$$

(2)

where the overbar indicates the time-averaged value.

2.3. Experimental Procedure

The ASFL was operated for consecutive 24 h trials between 25 April and 6 May 2021, with attraction flow turned on or off at 12:00 p.m. on alternating days. Fish caught in the upstream floating enclosure at the end of each trial were counted, identified, and inspected for wounds. All fish were measured for total length. When more than 20 fish of the same species were captured in the trap, only 20 individuals of that species were randomly measured. Water temperature was measured daily at the time of trap check. As the most prevalent species transported by the ASFL, 10 suckers (Catostomidae) were held for mortality monitoring at the U.S. Geological Survey Hammond Bay Biological Station for 3 weeks post transport through the ASFL. Suckers were visually inspected, daily via a side viewing window on the aquaria and again during handling at the end of the holding period. Finally, fish captured in the pot-style sea lamprey traps during the experimental period were counted and identified to species or higher taxonomic levels to further characterize the fish community available for capture by the ASFL during the experimental trials.

2.4. Statistical Analysis

The analysis of the number of fishes transported by the ASFL was analyzed using a Generalized Linear Mixed Model (GLMM). GLMM analysis was only completed for suckers as no other species had more than four captures. The number of suckers transported by the ASFL was analyzed using the mean daily water temperature and attraction flow (on/off) as the fixed effects and the day as a random effect. The linear regression and normality assumptions were explored using Pearson model residuals. The histogram of the model residuals was normally distributed at about zero, and a plot of normalized residuals against the fitted values did not indicate a systematic dependence of the variance on the fitted values. Examining the residual lag-plots did not reveal signs of the possible non-independence of observations from the unmodeled autocorrelated processes [26]. We further considered reduced models with a single fixed effect, either by the attraction flow or water temperature. To determine which model was most well supported by the data, an information theoretic approach was performed using Akaike’s information criterion (AIC), where the model with the lowest AIC value was the best model. The data were analyzed using the fitglme function in MATLAB.

3. Results

3.1. Hydraulics

The flow topology differed between conditions with and without attraction flow (Figure 4). Without attraction flow, the flow pattern on the right side of the ASLF near the surface (30 cm depth) is predominately towards the ASFL, whereas with the attraction flow present, flow is directed away from the ASFL. Flow near the bottom of the ASFL (80 cm depth) is less impacted by the attraction flow, and areas of flow towards the ASFL entrance are indicative of water rushing into the empty ASFL bucket. Similar patterns in turbulence were also observed relative to attraction flow presence (Figure 5). The only notable difference in TKE values is near the surface with attraction flow on, where TKE increases three-fold compared to the no attraction flow case.

3.2. Fish Passage

A total of 704 fish, 688 of which were suckers, were transported through the ASFL (

Table 2. Total count and average and maximum length of fish transported by the ASFL during the entire test period. Note, only 169 (25%) of the suckers caught were measured for total length.

Species	Scientific Name	No. Passed	Ave. Length (mm)	Max Length (mm)
northern pike	Essox lucius (Linnaeus, 1758)	3	577	635
rock bass	Ambloplites rupestris (Rafinesque, 1817)	4	178	218
rainbow trout	Oncorhynchus mykiss (Walbaum, 1792)	2	257	410
sea lamprey	Petromyzon marinus (Linnaeus, 1758)	3	475	482
smallmouth bass	Micropterus dolomieu (Lacépède, 1802)	1	352	352
yellow perch	Perca flavescens (Mitchill, 1814)	3	208	224
suckers	Catostomidae	688	451	534

). The temperature steadily increased throughout the test period (Figure 6). While the number of suckers transported by the ASFL increased throughout the test period, more were transported when the attraction flow was on (101 ± 66 [mean ± SD]) versus off (30 ± 40). The model most supported by the data contained the temperature and attraction flow as the covariates and day as a random factor (

Table 3. List of model and AIC scores for Generalized Linear Mixed Model of suckers lifted by the ASFL. Model variables with fixed effects include average water temperature (Temp °C) and attraction flow treatment (Attract), while day was modelled as a random effect.

Model	AIC	ΔAIC
Temp + Attract + (Day)	120.76	-
Attract + (Day)	124.37	3.61
Temp + (Day)	125.45	4.69

). Both the attraction flow and temperature were positively associated with the daily number of suckers transported by the ASFL (

Table 4. Fixed effects from the top Generalized Linear Mixed Model to explain the number of suckers transported by the ASFL.

Model Term	Coefficient	SE	df	t	p-Value
Intercept	−101.50	51.10	8	−1.99	0.08
Water temp.	16.46	6.09	8	2.70	0.03
Attraction flow	68.37	22.53	8	3.04	0.02

). All 10 fish held for visual monitoring appeared in good condition 3-weeks post transport with no bruising, abrasions, hemorrhaging, or mortalities observed.

Non-sucker fish species transported by the ASFL included four game fishes native to the Great Lakes, rainbow trout, and invasive sea lamprey (Table 2). Several fish species known to be in the pre-trap crowding structure during the ASFL trials (based on fishes identified to species captured in the sea lamprey pot-style traps) were not transported by the ASFL. In order of most to least abundant, these included the emerald shiner (Notropis atherinoides (Rafinesque, 1818)), walleye (Sander vitreus (Mitchill, 1818)), round goby (Neogobius metanostomus (Pallas, 1814)), troutperch (Percopis omiscomaycus (Walbaum, 1792)), common shiner (Luxilis cornutus (Mitchill, 1817)), creek chub (Semotilus atromaculatus (Rafinesque, 1820)), brown trout (Salmo trutta (Linnaeus, 1758)), and rainbow smelt (Osmerus mordax (Mitchill, 1814)).

4. Discussion

The evaluation of the capture and transport of fishes through a prototype of the Archimedes Screw Fish Lift showed that a mixture of migratory Great Lakes fish could be successfully transported with no observed injury or mortality. The increase in sucker capture and transport when an additional attraction flow was provided also highlights the need for further investigations into how hydraulic conditions influence ASFL capture and encounter rates. Overall, this pilot-level study demonstrates that the ASFL has the potential to capture and transport Great Lakes migratory fish over small vertical differentials or into an optical sorting and fish passage tool.

The hydraulic conditions at the entrance of the prototype ASFL impacted the capture and subsequent transport of the sucker as the capture-increased attraction flow was present. While changes to the flow topology near the ASFL were minimal with the addition of 200 L/min attraction flow, they had a significant impact on fish capture. One possible explanation for the effect is the additional flow helped to reduce the extent and magnitude of reverse flow fields formed when water rushes into the empty bucket. Since fish migrating upstream exhibit strong rheotactic responses to flow [27], reverse flow fields may disorient or otherwise inhibit upstream movement. The fact the presence of attraction flow still increased the capture of suckers emphasizes that unmodified entrance conditions at the ASFL may not be sufficient to attract and capture significant portions of fish. Further investigations that employ fine-scale tracking of fish movements could help clarify hydraulic conditions favorable for fish capture in an ASFL. A key outstanding uncertainty in the future use of the prototype ASFL is whether fish would volitionally enter or would be captured with water filling the bucket.

Testing the efficacy of the ASFL as a standalone means for transporting Great Lakes migratory fishes over small vertical differentials would require manipulations of hydraulic conditions hypothesized to influence both ASFL capture and encounter. While our study provided useful insight toward the future deployment of an ASFL in a fish passage scenario with fishes congregated in a confined space, there is substantial uncertainty regarding how an ASFL could perform outside of a confined space and whether both encounter and capture rates can be enhanced with a more distinguishable attraction flow. In a more open configuration, large manipulations in attraction flow would likely be required to achieve a measurable effect on ASFL encounter and capture rate. Fishway entrance design guidelines suggest that flows used to attract fish toward fishway entrances should be approximately 5% of the overall river discharge and located and oriented to maximize the influence of the attraction flow hydraulic cue (parallel to flow when competing flows are low and perpendicular to flow when competing flows are high) [26]. Furthermore, turbulence enhanced attraction flow, which was found to guide sea lamprey movement in an open stream, [27] may be a useful technique for increasing ASFL encounter and capture rates and warrants further investigation.

Testing the ASFL with a balanced mixture of fishes representing diverse fish communities in the Great Lakes rivers and streams could provide useful insight into how efficacy may vary among species. Our current experimental design did not allow us to quantify the number of fish available to capture by the ASFL and, thereby, untangle the respective roles of the ASFL capture efficacy and the number of fish available for capture on the observed capture rate for each species. However, resource managers considering ASFL as a tool for contending with the connectivity conundrum could benefit from an understanding of how its efficacy may vary among species. Further work that quantifies capture and encounter rates directly by tracking the fate of tagged individuals and/or conducting trials with a known number of individuals in an enclosed setting could help build this understanding.

Our pilot-level study provides a valuable step towards understanding how the ASFL could be used in upstream passage scenarios, a promising yet understudied application of this technology. While more extensive testing is still required, the results of this study suggest that an ASFL with additional attraction flow could potentially be a viable solution to capture and transport fish into an optical sorting device that requires fish to be out of the water for inspection and image recognition.