Hydraulic burden and swimming behaviour of benthic fish in a vertical slot fish pass

The majority of the world’s rivers is fragmented and fish passes are frequently installed to enable fish passage across river obstacles. Beside native, also invasive species, such as round goby, can use fish passes to disperse upstream in uninvaded ecosystems. The present study aimed at evaluating a prototype hydraulic barrier in a vertical slot fish pass that impedes passage of the invasive round goby and enables passage of native, comparable species, such as gudgeon and bullhead. In addition, the study was designed to provide basic insight in the differences between the hydraulic forces experienced by the fish species in the flow and their individual behavioural response to different flow conditions.

The barrier was designed to homogenize the flow field over an extended distance to exceed the swimming capability of the round goby and impede resting by station holding due to the smooth bottom. The selective effect was created by the flow field that induced hydraulic burden, varying between species due to the individual body shapes of the fish.

The performance of the hydraulic barrier was assessed using a three step-approach: Flow measurements (Step 1), measurements of the hydraulic forces experienced by preserved fish (Step 2) and live fish swimming behaviour observations in the same flow field (Step 3). This approach was applied in a flow channel study under homogenised flow conditions and in a nearly full-scale vertical slot fish pass model. The purpose of this approach was to evaluate the prototype hydraulic barrier (i), create basic evidence about how flow affects the fish species individually and whether the hydraulic burden experienced in the flow field differ between species (ii), and to assess how the hydraulic burden experienced by benthic fish affect the passage behaviour across the prototype hydraulic barrier and at an unaffected vertical slot (iii).

The results showed: There was a species selective effect of the prototype hydraulic barrier. No round goby passed the barrier at the highest water discharge tested (130 L/s), while six gudgeon passed the barrier and four bullhead attempted passage but immediately returned. This passage behaviour agreed with the hydraulic forces measured as gudgeon experienced significantly lower forces at 130 L/s water discharge compared to the other species. Because the hydraulic forces differed between species and corresponded to the live fish swimming behaviour, and the live fish passage behaviour differed between water discharges, it is probable that the selective effect of the barrier was mediated by the created flow field. Beside this selective effect, the prototype barrier had a general impact on the fish migration behaviour. There were significantly less passages recorded at the prototype hydraulic barrier compared to the downstream untreated slot.

The fish indeed experience hydraulic burden differing between species and these hydraulic burden correspond to the individual swimming styles of the corresponding species. The fish respond to these hydraulic burden individually, depending on their own species biology. Nevertheless, there was a general adaptation of the passage behaviour to increased water discharge: All species swam faster with more speed variation and with straighter paths upstream the barrier.

With providing basic evidence about the individual hydraulic burden an invasive species experiences in comparison to two native, comparable species, this thesis is a further step towards species selective fragmentation of rivers for ecosystem conservation purposes. These findings open an avenue to fish pass design adapted to the fish community of specific ecosystems and will inform fish pass engineers, decision makers and researchers who work about the behavioural response of fish to flowing water.

Die Mehrzahl der Flüsse auf der Erde sind fragmentiert. Um Fischen die Wanderung stromauf zu ermöglichen, werden häufig Fischpässe an Wanderhindernissen eingesetzt. Neben einheimischen Arten wandern ebenfalls invasive Arten, wie etwa die Schwarzmundgrundel, diese Fischpässe hinauf und bedrohen stromauf gelegene Ökosysteme. Die vorliegende Arbeit beschreibt einen Protoypen einer hydraulischen, selektiven Sperre für den Einsatz in ‘Vertical slot’ Fischpässen. Diese Sperre soll den Aufstieg der invasiven Schwarzmundgrundel verhindern und gleichzeitig für einheimische und vergleichbare Arten, wie etwa Gründling oder Groppe, durchlässig sein. Zusätzlich soll die Studie Grundlagenwissen über die individuellen hydraulischen Hürden und die Art und Weise, wie die Fische mit Ihrem Schwimmverhalten auf die Hürden umgehen, schaffen.

Das Funktionsprinzip der Sperre ist, dass die Strömung über eine gewisse Distanz homogenisiert wird und die bodenlebenden Fische durch die glatte Oberfläche daran gehindert werden, sich am Boden festzusetzen und auszuruhen. Der selektive Effekt soll durch ein Strömungsfeld erzeugt werden, das aufgrund unterschiedlicher Körperform der Fische unterschiedliche hydraulische Kräfte auf die Fischkörper erzeugt. Damit erzeugt das Strömungsfeld aufgrund der individuellen Körperform der Fischarten unterschiedliche hydraulische Hürden für die Fische bei der Überwindung der Sperre.

Die Sperre wurde in einem dreistufigen Verfahren getestet: Strömungsmessungen (Stufe 1), Messungen der hydraulischen Kräfte, die die Fische in der Strömung erfahren (Stufe 2) und Beobachtungen des Schwimmverhaltens lebender Fische im selben Strömungsfeld (Stufe 3). Dieses Verfahren wurde in Vorversuchen in einem Schwimmkanal unter standardisierten Strömungsverhältnissen und in einem physischen, fast maẞstabsgetreuen ‘Vertical slot’ Fischpass angewendet. Dabei sollte die Funktionalität der hydraulischen Sperre untersucht werden (i), Grundlagenwissen darüber geschaffen werden, wie Strömung die Fischarten unterschiedlich beeinflusst und ob die Fischarten unterschiedliche hydraulische Hürden bei der Passage der Sperre überwinden müssen (ii) und untersucht werden, wie diese hydraulischen Hürden das Schwimmverhalten der Fische unterschiedliche beeinflussen.

Die Ergebnisse zeigen: Die hydraulische Sperre hat eine artselektive Wirkung. Beim höchsten getesteten Wasserabfluss (130 L/s) passierten keine Schwarzmundgrundeln die Barriere, während sechs Gründlinge die Barriere passierten und vier Groppen eine Passage versuchten, jedoch sofort zurückkehrten. Dieses Durchgangsverhalten stimmt mit den gemessenen hydraulischen Kräften insofern überein, dass Gründlinge im Vergleich zu den anderen Arten signifikant geringere Kräfte bei 130 L/s Wasserabfluss erfahren. Da die hydraulischen Kräfte zwischen den Arten unterschiedlich sind, dem Schwimmverhalten der lebenden Fische entsprechen und das Verhalten der lebenden Fische zwischen den Wassereinleitungen unterschiedlich ist, ist eine selektive Wirkung der Barriere durch das erzeugte Strömungsfeld sehr wahrscheinlich.

Neben diesem selektiven Effekt hatte die Barriere einen generellen Einfluss auf das Fischaufstiegsverhalten. An der hydraulischen Sperre wurden im Vergleich zum stromabwärts gelegenen, unbehandelten Schlitz, deutlich weniger Passagen der Fische aufgezeichnet. Tatsächlich erfahren die Fische je nach Art unterschiedliche hydraulische Belastungen in der Strömung. Auf diese hydraulische Belastung reagieren die Fische individuell, entsprechend ihrer eigenen Artbiologie. Es wurde allerdings auch eine generelle Anpassung des Schwimmverhaltens an stärkere Strömung bei allen Arten beobachtet: Alle Arten schwammen schneller mit mehr Geschwindigkeitsvariation und mit geraderen Wegen die hydraulische Sperre hinauf.

Diese grundlagenwissenschaftlichen Erkenntnisse über die hydraulischen Hürden in Fischpässen einer invasiven Art im Vergleich zu zwei einheimischen, vergleichbaren Arten, sind ein wichtiger Schritt Richtung artselektive Fragmentierung von Flüssen zum Schutz des Ökosystems. Diese Thesis bietet Grundlagen für die Gestaltung von Fischaufstiegshilfen, die an die Anforderungen des Ökosystems angepasst sind. Damit hat sie Relevanz für Fischpass-Entwickler, Entscheidungsträger und Wissenschaftler, die über das Verhalten von Fischen in der Strömung forschen.

At the beginning was a fish in my hand. It was a cloudy day at the time when I went to school. I was a passionate angler and loved to fish some great barbels and eels in the lower Rhine near Cologne. I spent numerous hours there, rather than learning for my school exams and it was just a great time. Then came the day when I held this fish in my hand. I caught it in the Rhine some seconds ago and I just wondered. It was a small fish, about 10 cm long, which looked a bit curious. It was hard to identify the species but with my unexperienced view, I identified this fish as a European bullhead. Knowing that bullhead prefer rivers with good water quality, I was very happy, solved the fish from the hook carefully and released it into the river. Indeed, happiness did not stay long. It were only seconds and the small bell on the top of my rod rang again and I caught the next ‘bullhead’. This continued until the evening and I packed my stuff and went home. The fishing day was very disappointing because I caught numerous small ‘bullheads’ but not a single big and strong barbel. At this time, I did not guess that this day would have important impact on my future.

The time of good fishing at the river Rhine was over and I never caught a great barbel again. I still spent years fishing for barbels at the Rhine, but the only fish I caught was this small fish that looked like a bullhead. However, I knew from fishing magazines that it was an invasive species, called round goby. There was hope it might disappear someday, but the round goby stayed.

I finished school, studied Biology in Düsseldorf, and decided to continue my studies in a Masters Program at the University of Rostock. Looking forward to many fishing opportunities in Rostock, I thought to leave the round goby behind –not knowing that round gobies had distributed in the Western Baltic Sea already and became even bigger in brackish environments. I just felt a bit tracked by the round goby but accepted the opportunity to investigate predation of the round goby on Atlantic herring eggs in my Master thesis, which I performed at the Thünen-Institute for Baltic Sea fisheries. Fortunately, this thesis opened the door to Basel for investigating methods to impede the upstream migration of the round goby via active swimming.

This story reports my inner motivation for the following thesis. This story and my personal experience taught me one thing: Every catastrophe provides a chance. It is up to us what we make out of it.

During the work at my dissertation, I experienced two more catastrophes with such a force, that years were required to recognize the chance provided. Shortly after the beginning of my research project, my wife and I said goodbye to our first born son, who died late in pregnancy before giving birth for unknown reasons. About two years later, my younger brother passed away by suicide. After these events, my research project became least important of everything and it took a lot of mental strength to just keep going. However, much power was provided by my wonderful wife and my fantastic kids. Both living kids were born during my dissertation project. Now, when I am sitting here at my desk and write down my thesis, I can hear them playing in the living room and the chance from the experienced becomes clear to me: It is our responsibility to take care of the younger members of our community and we must provide them with all the tools and advice available to develop themselves, the society, the globe, and science.

Rivers have undergone significant changes in the two past centuries (Habersack & Piégay, 2007). Today, the majority of the world’s rivers are fragmented by anthropogenic barriers (Belletti et al., 2020). Indeed, river connectivity has been reported to be important for fish diversity (Shao et al., 2019), which is why diverse passage constructions have been developed (Katopodis & Williams, 2012). Numerous fish pass types were developed following a trial-and-error approach, focusing predominantly on the passage of economically relevant species, such as salmonids (Katopodis & Williams, 2012; Roscoe & Hinch, 2010). Following the application of fish passes that were actually designed for salmonid species for the whole fish community, Mallen-Cooper & Brand (2007) observed a poor performance of these fish passes for native species and even an increase of non-native species. The Convention on Biological Diversity raised the need to adapt fish passes to the requirements of species of relevance for the ecosystems (United Nations, 1992). To develop fish pass structures accounting for the individual differences in swimming performance between species, a combination of research approaches from fluid dynamics, engineering and behavioural ecology is promising (Castro-Santos et al., 2012).

A modern, frequently investigated fish pass type is the vertical slot fish pass, which is commonly installed at transversal structures, such as weirs (Wu et al., 1999). Vertical slot fish passes consist of a rectangular channel with partition walls. These partition walls contain vertical openings which separate the channel into pools. The flow through these slots creates a jet with accelerated flow velocities that has to be passed by the fish migrating upstream (Wu et al., 1999). Due to its construction, vertical slot fish passes are robust against variation in water discharge and maintain steady hydraulic conditions at slots and pools (Katopodis & Williams, 2012).

The threat of invasive species

Invasive species have become a world-wide threat to aquatic ecosystems (Havel et al., 2015). Designing fish passes that support specific species of importance for the ecosystem and hold back undesired species can support species selective fragmentation of the river (Rahel & McLaughlin, 2018). Selective fragmentation is a possible management option to impede the spreading of invasive species (Rahel & McLaughlin, 2018). Ensuring the passage of native species and impeding the passage of invasive species over river obstacles is a major challenge for decision makers and requires advances in integrated interdisciplinary research (Rahel and McLaughlin, 2018).

One of the 100 worst aquatic invasive species has been reported the round goby (Hirsch et al., 2016). The round goby arrived in Switzerland 2012 (Kalchhauser et al., 2013) and continued to disperse upstream the River Rhine (Lutz et al., 2020). Beside diffusion across short distance, also long distance dispersal via active swimming (approximately 500 m per year) have been reported for round goby (Bronnenhuber et al., 2011). Because of the individual swimming style of the round goby (Tierney et al., 2011), the idea of selective hydraulic barrier seems feasible (Hoover et al., 2003), which is why the round goby swimming performance has been assessed by Tierney et al. (2011) and Hoover et al. (2003). Indeed, to create a hydraulic selective barrier for the round goby that is passable by native species, it must be understood how the round goby swimming response to the flow conditions in the barrier differ between comparable, native species.

Swimming in fish

Fish are generally well adapted to move in their fluid environment (Lauder and Madden 2007). The main properties of water, especially its high density and incompressibility, have played an important role in fish evolution (Sfakiotakis et al., 1999). There are numerous ways fish create vortices in the water body to move in flowing water (Fish & Lauder, 2006; Lauder, 2015; Triantafyllou et al., 2000), which have evolved with time (Sfakiotakis et al., 1999). One specialized swimming style has been reported for benthic fish, such as the round goby: the ability to hold station at the bottom to save energy in flowing water (Gilbert et al., 2016; Tierney et al., 2011). Small body size (7-8 cm length, 1-2 cm height) facilitates station holding by escaping the flow at cobble substrates. In addition, larger pectoral fins enabled a benthic fish, darter (Etheostoma tetrazonum Hubbs & Black, 1940), to produce negative lift forces pressing the fish to the bottom (Carlson & Lauder, 2011).

Model species

Such behaviour, to produce downwards directed forces using pectoral fins for station holding and thereby increasing the swimming capacity, has also been reported for the invasive round goby (Neogobius melanostomus Pallas, 1814) (Gilbert et al., 2016; Tierney et al., 2011). In addition, Tierney et al., (2011) reported a general tendency to swim upstream of the round goby at moderate velocities, while it has been reported that the water discharge must be greater than 125 cm/s to prevent round goby upstream migration. A similar bottom-dwelling fish species is European bullhead (Cottus gobio L.), which occupies a similar ecological niche as the round goby (Roje et al., 2021). Bullhead prefer creeks with higher velocities and coarse substrate (Davey et al., 2005; Van Liefferinge et al., 2005), while low swimming capacities have been reported (Tudorache et al., 2008). Indeed, bullheads profit from a similar station holding behaviour based on usage of pectoral fins to ‘anchor’ on the bottom (Tudorache et al., 2008).

Gudgeon (Gobio gobio L.) is another bottom-dwelling fish occurring in similar habitats like round goby and bullhead. Contrary to the round goby and bullhead, gudgeon has more a bentho-pelagic life style and no specialized station holding behaviour has been reported for gudgeon.

Because bullhead and gudgeon are native species of the local fish community in the River Rhine and its tributaries near Basel and the mentioned similarities in their species biology and swimming behaviour, they were selected as representatives for native fish in the present study.

How do fish experience the flow?

To create a hydraulic barrier with a flow field that affects the passage behaviour of fish species differently, it is important to understand how the fish experience the flow differently. Several studies aimed at perceiving the flow field from a fish perspective –for example via artificial lateral line systems (Chambers et al., 2014; Fuentes-Pérez et al., 2015; Venturelli et al., 2012). There are robots mimicking swimming fish to understand basics in fish kinematics (Thandiackal et al., 2021). Indeed, these studies are commonly based on simplified fish shapes and neglect the effect of the individual body shape of the fish on their hydraulic burden experienced. Fish morphology is directly related to swimming costs (Ohlberger, Staaks, & Hölker, 2006) and several studies showed morphological adaptation in fish from riverine habitats (Dashinov et al., 2020; Franssen et al., 2013; Imre, 2002; Meyers & Belk, 2014). It is possible that these adaptations in body shape in habitats of increased flow occurred to reduce the hydraulic forces experienced.

Mathematical models are commonly used to estimate the hydraulic forces fish bodies experience in flowing water, such as Drucker & Lauder (1999), Sällström & Ukeiley (2014), Van Wassenbergh et al. (2015). Indeed, more direct measurements at real fish would probably provide a more realistic picture of the hydraulic burden. Therefore, Sagnes et al., (2000) implemented measurements of drag experienced by dead fish, when describing the change in hydraulic drag during ontogenesis of grayling larvae (Thymallus thymallus L.). Further studies applying physical force measurements focus on single species and artificial fish (Barrett et al., 1999; Quicazan-Rubio et al., 2019).

Measuring the direct forces experienced and comparing these forces between species might increase the understanding of how flow affects the fish species differently. This can be the key for understanding how flow fields can be changed to induce selective passage of specific fish species corresponding to the ecosystem requirements.

Objectives and hypotheses

One purpose of this study was to provide the fundamentals for a hydraulic barrier that impedes the further upstream migration of the round goby, while maintaining passage of comparable native species, such as bullhead or gudgeon. The prototype barrier was designed and tested based on the idea of a selective, hydraulic barrier that impedes station holding by a smooth surface and homogenous flow conditions over an extended distance, exceeding the swimming performance of the round goby while native species can ascend (Hoover et al., 2003). Nevertheless, to create species selective hydraulic conditions in a fish pass, it is important to understand how the fish species experience the flow differently and how the flow affects the individual swimming behaviour of the fish.

To fill this knowledge gap, a three step-approach is implemented in the present thesis: This approach consisted of flow measurements (1. Step), force measurements on preserved fish (2. Step), and live fish swimming behaviour observations (3. Step). The direct forces experienced by preserved fish in the flow were measured as a proxy for the hydraulic burden live fish encounter when swimming against flow. The hydraulic burden fish experience was defined as the hydraulic force the fish body experiences from the flow field while passage. These hydraulic burden were then compared to live fish observations to describe how hydraulic burden mediate the live fish swimming behaviour. This was done to create evidence about how benthic fish respond behaviourally to flow and how this response differs between species.

The aim of the present thesis was to create basic evidence about how the flow affects the fish species individually and whether the hydraulic burden experienced in the flow differ between species (Paper I, III, IV):

Hypothesis 1: The hydraulic burden the fish experience while passage differ between species.

Further, the present thesis aimed at evaluating a prototype, hydraulic, and selective barrier that was designed to impede the passage of the invasive round goby (Paper II, III):

Hypothesis 2: The hydrodynamics in the barrier prevent the passage of the invasive round goby while native, comparable species can ascend.

Finally, to assess how the hydraulic burden experienced by the fish affect passage behaviour, the hydraulic burden experienced by the fish were compared with the live fish swimming behaviour (Paper I, III, IV):

Hypothesis 3: Species with smaller hydraulic burden have increased passage capability.

Based on the above mentioned hypotheses, four publications were created with individual research questions (Fig. 1):

Two experiments (Fig. 1) were performed to fill the knowledge gap. The first experiment focused on the hydraulic burden and the fish behaviour of the three test species (round goby, gudgeon and bullhead) under standardized and homogenous flow conditions in a flow channel in the laboratory at the University of Basel, Switzerland. In the second experiment, the prototype hydraulic barrier, as well as the hydraulic forces, and live fish swimming behaviour of the same species, were assessed in a realistic, full-scale fish pass model (test rig) at the Karlsruher Institute for Technology. To keep both experiments as comparable as possible, the three step-approach was applied in both experiments always with the same test species (round goby, gudgeon, and bullhead). To describe the dependency of the hydraulic forces experienced by the fish and the swimming behaviour from the flow velocity, both experiments were performed for different water discharges.

After the live fish observations in both experiments, the fish were euthanized with an overdose of MS-222 and preserved in formalin. Formalin increased the stiffness of the fish and preserved their body postures in a standardized way. After three days in 4% formalin with straight body postures, the fish were transferred stepwise via 40% ethanol (24 h) and 60% ethanol (24 h) to 75% ethanol solution to reduce toxicity of the fish (Paper I). The preserved fish were then punctuated vertically at the assumed centre of gravity with a needle and connected to the fixation stick. This fixation stick held the fish at position in the flow field and transduced the force experienced to the force sensor. All animal experiments were approved by the Swiss cantonal authorities (permits Nr. 2934 and 2846) and the German regional authorities (permit Nr. G217_17- IWG).

1. Flow channel study (Paper I and II)

We used a Loligo ® flow channel (swim Tunnel Respirometer, 185 L volume, 88 * 25 * 25 cm measuring chamber) to describe the flow, the forces and the swimming behaviour of our test species in a rectangular flume that had a shape being representative for our idea of a hydraulic barrier (homogenized flow over a smooth bottom for a distance of 88 cm) for a first quantification of flow conditions, the hydraulic forces and the swimming behaviour of live fish in a rectangular flume.

Flow channel study: Flow measurements

The flow velocity was measured using an acoustic Doppler velocimeter (ADV, Nortek Vectrino®) at 36 positions in the measurement chamber and at three adjusted velocities (0.25, 0.55, and 0.85 m/s) that were considered relevant for swimming in benthic fish (Tierney et al., 2011; Tudorache et al., 2008).

Flow channel study: Hydraulic forces

The hydraulic forces experienced by preserved fish were measured by a Vernier ® Go Direct Force and Acceleration Sensor (GDX-100609). The near ground-drag force was measured at 45 measurement points in the measurement chamber and at 0.25, 0.55, and 0.85 m/s flow velocity. This measurement was performed for single individuals of round goby, gudgeon, and bullhead with attached fins to display the spatial distribution of the forces above the bottom in the measurement chamber. In the next step, the hydraulic forces were measured at the central position in the flow channel at different velocity steps to measure the dependency of the hydraulic force from the flow velocity. To determine the effect of the fins on the forces experienced, three fish per species with attached and three fish with completely spread fins were tested in this measurement.

Flow channel study: Live fish swimming behaviour

First, prolonged swimming trials (Ucrit) were performed with 18 round goby, 12 gudgeon, and 12 bullhead. The single fish were released into the measurement chamber and left for 20 min in stagnant water for acclimatisation before the water flow was set to 0.15 m/s for 10 min. In the following, the flow speed was increased by 0.10 m/s every ten minutes until the fish could not hold position or showed signs of fatigue.

Afterwards, sprint speeds (Usprint) of the three species were evaluated for 18 round goby, 11 gudgeon and 12 bullhead. After 5 min acclimatisation at 0.05 m/s flow velocity, the flow was increased by 0.05 m/s every 10 s until the fish reached fatigue. Following a 10 min break in the measurement chamber with no flowing water, the experiment was repeated until the fish refused to swim or were unduly stressed. The fish were recorded during both assessments from vertical perspective by Gopro ® cameras (Hero 4).

2. Fish pass experiment at Karlsruher Institute for Technology (Paper II, III, and IV)

A nearly full scale test rig (1:1.6 scaled physical model of the fish pass in Koblenz, Germany) at the Theodor-Rehbock Hydraulic Laboratory at Karlsruher Institute for Technology (KIT), Germany (Fig. 1B) was used to assess the prototype hydraulic barrier and to describe the hydraulic burden experienced by preserved fish (round goby, gudgeon, and bullhead) and passage behaviour of live fish under realistic conditions. Water discharges can vary in real vertical slot fish passes, which is why three different water discharges (80, 105 and 130 L/s) were tested in the experiments. The two lower discharge rates were chosen to increase the probability of recording migration behaviour of the live fish, because they did not pose a challenge to the swimming capacities of the tested species as revealed from the flow channel study (Paper II). The 130 L/s water discharge was included in the experiment because we found this discharge led to the most representative flow velocities compared to actual best-practice vertical slot fish passes (Bombač et al., 2017). Flow measurement and live fish observations were performed at 80, 105, and 130 L/s water discharge. The force measurements were performed at 80 and 130 L/s and not at 105 L/s water discharge to increase the sample size of tested fish in the time available.

Fish pass experiment: Flow measurements

The acoustic Doppler (a similar device from the flow channel study) was mounted at an electronic, programmable, carriage (Isel ®) that was first adjusted over the prototype barrier (Fig 1). This carriage positioned the flow measurement probe at predefined measurement points automatically. Flow was measured at 14 measurement points, which were equally distributed over the barrier and ca. 2 cm above the bottom. Subsequently, the electronic carriage was adjusted above the untreated slot and the flow was measured at 19 measurement points which were equally distributed in the vicinity of the slot, ca. 2 cm above the bottom.

Fish pass experiment: Hydraulic forces

The hydraulic forces experienced by seven fish per species at 130 L/s water discharge and five fish per species at 80 L/s water discharge were measured using a water resistant (IP 68) force-torque sensor (Nano 17, ATI ®), which was integrated in a PVC-hull to protect the sensor from disturbing flow. The force measurement probe was adjusted at the same electronic carriage and measurements were performed at the same measurement positions from the flow measurements. Four measurement points at the untreated slot were excluded from the force measurement because the probe design impeded positioning of the probe above these points.

Similar to the flow channel study, the fish were punctuated and connected to the fixation stick prior the carriage positioned the fish over the first measurement point. At the measurement point, the hydraulic forces in X-, Y-, and Z-direction were recorded for 60 s with a data collection frequency of 1000 Hz. After the period of 60 s, the probe with the fish was positioned over the next measurement point and the data recording was started again.

Fish pass experiment: Live fish swimming behaviour

The fish were released downstream the untreated vertical slot and were able to swim freely in the test rig (Fig. 1B) unaffected by human presence for two hours to ensure the fish displayed only voluntary swimming behaviour. Because it was not clear how the fish might behave at the beginning of the experiments, the experimental period of two hours was chosen to increase the probability to observe a meaningful amount of untreated slot and barrier passages. Three cameras (Security-Center IR CCTV-Camera 380 TV-lines, IP 68, Abus ®, Germany) were installed (one above the vertical slot, two above the prototype hydraulic barrier) and recorded the fish while passage. In addition, metal grids were used to block the vertical slot and barrier after the period of 2 h to enable counting the fish that had migrated finally after this period of 2h.

The experiment was performed for three water discharges (80, 105, and 130 L/s). Because 130 L/s was assumed most representative for the conditions in real fish passes (Bombač et al., 2017), the experiment was performed for three times at this water discharge. To increase the number of fish included in the experiments and thereby the probability to record migration behaviour with the cameras, all fish (43 ±4 individuals per species) available were tested in every run. To discriminate between species in the video analysis, species were tested separately.

Following the live fish experiment, the video footage was screened by eye and the time and number of defined events (Appendix 1) was recorded in the program Blender (v. 2.79). Afterwards, the paths of the fish that displayed these events were tracked in X-, Y- coordinates. These trajectories were used to compare the swimming patterns between the species at the different water discharges and to compare the swimming behaviour with the hydraulic burden experienced.

Figure 1: The experiments were performed in a flow channel at the University of Basel (A) and in a nearly full-sized, physical vertical slot fish pass model at the Karlsruhe Institute for Technology (B). The flow direction is represented by grey arrows (fd). The water (w) in the flow channel was accelerated by a rotor (r), which was powered by an electric engine (e). Flow straighteners (fs) ensured homogenised flow in the measurement chamber (m), where the measurements were performed. At the test rig, the observations focused on an untreated vertical slot (Sl) and the prototype hydraulic barrier (Ba). Three cameras (c1, c2, c3) where installed above slot and barrier for live fish observations. All experiments were performed for the three test species and at different water discharges to describe how the flow affected the fish species differently.

Click here to download the high resolution image as PDF

The prototype hydraulic barrier

The experiments revealed a species selective effect of the prototype hydraulic barrier. No invasive round goby passed the barrier at the highest water discharge tested (130 L/s), while six native gudgeon passed the barrier and four native bullhead attempted passage but immediately returned. That the selective effect was hydraulically mediated is probable because round goby has strongly changed its passage behaviour across the barrier between water discharges. Round goby was the species that passed the barrier most frequently of all tested species (18 times) at the lowest water discharge (80 L/s). In addition, round goby showed a general upstream dispersal tendency at the untreated slot downstream the barrier as indicated by 65 round gobies that passed at 130 L/s water discharge. One further observation supporting that the selective effect was hydraulically mediated was: gudgeon, the species with most passages across the barrier at 130 L/s water discharge, experienced the smallest hydraulic forces on the barrier and might therefore have experienced the lowest hydraulic burden while passage.

Beside the selective effect, the prototype barrier had a general impact on the fish migration behaviour. There were significantly less passages recorded at the prototype hydraulic barrier compared to the downstream untreated slot across all water discharges. This might be explained by the need to overcome the untreated slot prior challenging the prototype barrier in our experiment. It is possible that the fish fatigued after passage of the untreated slot and were physiologically unable to pass the barrier accordingly. However, two hours experimental time may have enabled the fish to recover from the passage of the untreated slot. In addition, the same fish were used for the different experimental runs and the fish were therefore able to assess the flow fields and learn (Kieffer & Colgan, 1992) the energetically favourable routes across barrier and slot. Indeed, that there was no clear change of the fish behaviour between the three runs performed at 130 L/s water discharge (Paper IV), suggests that adaptation and learning across the experiments played rather a minor role. Another important point is that such flow field assessment effects of the fish can, beside in our experiments, also occur in real fish passes. At a barrier installed in a real fish pass, there would also be untreated slots that have to be overcome by the fish prior challenging the barrier. The experimental setup was as realistic as possible and the study was designed to exclusively record voluntary migration behaviour (the fish were completely unaffected by human presence for the experimental period of 2 h). Under these conditions the results showed: While having a species selective effect at 130 L/s water discharge, the barrier reduced the passage of all species at all water discharges compared to the untreated slot (Paper II).

Hoover et al. (2003) reported that a hydraulic barrier would have to maintain flow velocities greater than 0.75 m/s over an extended distance to exceed the physiological swimming endurance of the round goby. Tierney et al. (2011) mentioned flow speeds > 1.25 m/s without microhabitats, allowing the fish to recover, to be required for withholding the round goby. Our flow measurements reveal that the prototype barrier fulfilled the requirements mentioned by Hoover et al. (2003) but the flow speed was lower than the requirement suggested by (Tierney et al., 2011). The water flow velocity measured on the barrier at 130 L/s water discharge was 0.91 m/s ±0.08 SD (Standard deviation). Thereby, the water velocity above the barrier was higher and more homogenous than at the untreated slot (mean velocity of 0.74 m/s ±0.21 SD at 130 L/s water discharge). This reveals: The barrier created a near-homogenous jet stream (Liu, Rajaratnam, & Zhu, 2006) over the barrier length which affected the fish body differently, as indicated by force measurements. Thereby, we showed that the flow field created on the barrier primarily explains the differing live fish upstream migration behaviour, as revealed by the varied passage frequencies between untreated slot and barrier observed in the live fish observations (Paper II).

Although the prototype hydraulic barrier performed promisingly in the reported laboratory experiments, the barrier must be assessed in long term field experiments to describe its long term-ecosystem impacts. It is possible that vegetation or debris on the barrier could alter the hydraulics of the barrier over time. In addition, two hours experimental time can be too short when compared to the time available to round gobies challenging the barrier in a real fish pass. It is possible that the fish exercise and learn how to overcome the barrier over time (Kieffer & Colgan, 1992). If then only the most powerful swimmers of the goby population can pass the barrier, it seems possible that these fish found an upstream population with species of increased migration tendency and swimming capability. How these potentially negative effects need to be weighed against the barrier’s positive effects has to be assessed in long term-field experiments.

The hydraulic burden experienced by the fish

Based on the referenced literature, the present thesis provides the first experimental comparison of the hydraulic forces experienced between real, preserved benthic fish species, at homogenised flow and in vertical slot fish pass flow fields. The hydraulic burden experienced by the fish in flowing water were shown to depend on the flow velocity and on the fish body surface exposed to the flow (Paper I). Thereby, the pectoral fin size and position were observed to have important impact on the hydraulic burden experienced by the fish. Especially round goby and bullhead experienced stronger forces when pectoral fins were spread (Paper I), which supports their specialized swimming style. Benthic fish have been described to adjust pectoral fin postures in a way that the forces experienced press the fish to the bottom and support station holding (Coombs et al., 2007). The experiments overall revealed significant differences in the hydraulic burden between fish species with the smallest hydraulic burden experienced by gudgeon, largest vertical force directed to the ground experienced by bullhead and median forces experienced by round goby. Considering that gudgeon is described as a powerful semi-pelagic swimmer (Tudorache et al., 2008, Paper II), that bullhead is highly oriented to the ground and uses flow sheltered regions between rocks to withstand the flow (Gosselin et al., 2010; Tudorache et al., 2008; Van Liefferinge et al., 2005, Paper II), and that round goby has been described a versatile swimmer (Tierney et al., 2011), the results suggest that the hydraulic burden corresponded to the individual swimming styles of the fish.

These hydraulic burdens were recorded for different fish samples at two different positions with different flow fields in our experimental setup: At the barrier and the untreated slot. In both independent measurements, we recorded significantly smaller hydraulic burden for gudgeon at 130 L/s water discharge. This observation supports that the hydraulic burden variation between species remains stable between different flow fields of the untreated slot and the barrier.

The hydraulic forces varied between measurement positions especially at the untreated slot, where we also observed local variation in the flow field. Positive correlation between hydraulic forces and flow velocity suggest that the local flow determines the force the fish experience (Paper III and IV) and thereby facilitated passage, especially at the untreated slot, where the flow field was more diverse. This result was not expected because we observed a strong dependency of the hydraulic forces from the adjusted water flow velocity when measuring at one central point in the measurement chamber (Paper I). In the untreated slot however, we observed this local change in the hydraulic burden across measurement points although the water discharge was fixed to 130 L/s (Paper IV). This finding suggest that local flow conditions (e.g. flow direction, flow velocity or turbulence) determine the local force experienced. This means that fish might be able to reduce their hydraulic burden during passage by choosing swimming routes of reduced hydraulic forces. Thereby, fish passes of diverse flow conditions might favour fish passage and homogenous flow conditions can have selective effects.

The effect of the hydraulic burden on the fish swimming behaviour

The fish varied in their upstream migration behaviour between species at the barrier as revealed by the random forest machine learning approach (Paper III). It was possible to identify round goby and gudgeon with high accuracy, only with information about water discharge and swimming patterns. When predicting the species, the most important predictor variable was the water discharge. This agrees with our observation that round goby showed most passages at lower water discharge, while gudgeon passed most frequently at higher water discharge at the untreated slot (Paper IV) and barrier (Paper III). Because the water discharge was the only varied parameter between the experiments, these findings support, together with the observed reduced hydraulic burden measured for gudgeon at higher water discharge, that hydraulic burden have an important effect on the individual swimming behaviour of the fish. Beside this species specific effect, there was a general adaptation observed in the swimming behaviour of all fish at higher water discharge. All fish increased their mean swimming speed, swam with more variable speed and used straighter paths on their way upstream the barrier; probably to reduce the hydraulic burden when swimming upstream.

The hydraulic forces measured cannot reflect the general swimming behaviour of live fish in a flow field because this depends on the individual species biology (Blake, 2004; Coombs et al., 2007; Sfakiotakis et al., 1999) and individual personality traits (Hirsch et al., 2017; Lothian & Lucas, 2021). However, hydraulic forces can provide evidence of one piece of a puzzle of all determinants of fish swimming behaviour: The physical burden the fish experience in the flow due to their body exposed to the flow. It is known that flow conditions can support swimming or make swimming more difficult. For example, turbulence is reported to have important impact on the swimming performance in fish (Lupandin, 2005) and can have positive, as well as negative impact on fish swimming performance. Fish can make use of areas of reduced vorticity to save energy (Facey & Grossman, 1992) and also experience passive propulsion under specific flow conditions (Beal et al., 2006). Although smaller forces at measurements with fish compared to a reference run without fish were detected at 130 L/s at the untreated slot (Paper IV) and in the flow channel experiment (Paper I), no forces directed against flow direction were revealed by the measurements with the 3D-sensor when a fish was connected to the fixation stick. We do therefore not assume that the fish experienced passive propulsion in our experiments.

Measuring hydraulic forces experienced by the fish can be a refinement of the common approach, which is to assess fish passes by flow characterisation and live fish observations (e.g. Drucker & Lauder, 1999; Porreca et al., 2017; Sagnes & Statzner, 2009). The force measurements provide a new, direct and standardized way to describe how the created flow field, which is determined by the fish pass design, affects the fish physically. This approach does not require the use of live fish and thereby improves fish welfare by avoiding animal experiments. However, there is more research needed to understand how the species respond to the hydraulic burden individually.

Challenges

The direct forces applied to surrounding water of swimming fish are not directly measurable, which complicates the quantification of their locomotor forces (Drucker & Lauder, 1999). The applied approach to measure the forces by real, preserved fish was an approximation to live fish swimming several millimetres above ground but it did not account for the impact of movement and different body postures on the forces experienced.

The live fish observations of the first experiment revealed that especially round goby and bullhead used flow channel corners and friction forces to withstand the flow (Paper I and II). Because friction was not standardisable with the measuring device applied, it was not possible to account for the usage of friction forces to withstand the flow in benthic fish (Carlson & Lauder, 2011) in the present study. This important question, how benthic fish resist the flow by using friction forces, remains a future challenge.

Prior to testing preserved fish, there was the idea to 3D-scan real fish, print them by a 3D-printer, and measure the forces experienced by 3D-printed fish as a proxy for real fish. Using computer manipulation, it would have been possible to create multiple 3D-models of the same fish with different fin and body postures. This would have enabled to test the effect of fin and body posture on the forces experienced in a more standardized way than it was possible with the preserved fish. Another advantage would have been that the measurements would have been highly replicable because the model could be used several times, the model shape data could be stored and published in a repository and replicates of the model could be printed with a 3D-printerby researchers everywhere in the world. Indeed, it turned out at the beginning of this research project, that the 3D-scan and 3D-print represent two methodological steps adding important inaccuracy to the model due to technical limitations. Because the shape of the 3D-printed models varied obviously from real fish shapes, preserved fish turned out to be best representative for real fish for the experiments. Nevertheless, further development of the 3D-printing technique might enable its application in such studies in future.

Conclusion

The present work provides a description of the performance of a prototype hydraulic barrier which was designed to impede the upstream migration of the invasive round goby, while enabling passage of native, comparable species. With application of the three step-approach, it was possible to describe the barrier’s mechanism, which was based on the creation of flow conditions affecting the swimming behaviour of the fish species differently.

From an applied research perspective, it became evident that the prototype hydraulic barrier had a selective effect at 130 L/s water discharge because no round goby passed the barrier, while gudgeon and bullhead passed. Because the hydraulic forces differed between species and corresponded to the live fish swimming behaviour, and the live fish passage behaviour differed between water discharges, it is highly probable that the selective effect of the barrier was mediated by the created flow field. As this evidence merged from a laboratory experiment, long term field studies are necessary to assess the performance of the barrier in the field and its impact on the ecosystem.

From a basic research perspective, this thesis developed an approach to compare the hydraulic burden comparable benthic fish encounter when swimming in flowing water. In addition, this thesis provides the first empiric assessment of the hydraulic forces that benthic fish experience in flow fields of such ecological importance. An important insight was that the fish indeed experience hydraulic burden differing between species and that these hydraulic burdens correspond to the individual swimming styles of the corresponding species. The fish respond to these hydraulic burden individually, depending on their own species biology. Nevertheless, there was a general adaptation of the passage behaviour to increased water discharge: All species swam faster with more speed variation and with straighter paths upstream the barrier.

This thesis is a further step towards species selective fragmentation of rivers for ecosystem conservation purposes. These findings open an avenue to fish pass design adapted to the fish community of specific ecosystems and will inform fish pass engineers, decision makers and researchers who work about the behavioural response of fish to flowing water.

Outlook and perspectives

Assessment of the barrier for more species

The barrier performed promisingly in the experiments and the need to perform field studies to describe the long term effect of the barrier and its ecosystem effects was already mentioned. Indeed, the effect of the barrier on further species is still unknown. Therefore, special attention should be spent on the effect of the barrier on the passage performance of all further species that were not included in the present study to ensure their passage will not be negatively affected by the barrier.

Assessment of fish passes for their suitability of target species

The force measurement enabled a close description of the hydraulic burden the fish have to overcome when swimming upstream a vertical slot fish pass. Thereby, the force measurements provided a more direct flow field assessment for the requirements of the specific species than common flow measurements, such as acoustic Doppler-velocimeter applications. Measuring hydraulic forces provides a valuable technique enabling the assessment of fish pass designs for the suitability for species of relevance for the corresponding ecosystem. This method can be applied in the laboratory at prototype fish pass models or in the field at constructions that exist already. Nevertheless, because it is still not possible to measure the forces fish experience while swimming, future research should focus on the refinement of this measurement technique.

Development of the force measuring technique

There is potential to further develop and improve the force measurement technique. In the reported approach, the preserved fish were connected to the multi-axis force-torque sensor via a fixation stick. The stick was five times smaller compared to flow channel study (Paper I), which increased the precision of the measurement, but the stick was still exposed to the flow and thereby affected the forces measured. In addition, because the stick represented a lever, it was not possible to account for all measurement channels provided by the sensor because it was not possible to account for the torque experienced by the fish. If integrating the sensor in the fish, these problems would have been avoided and all sensor channels could have been used including force in X-, Y-, Z-direction and torque in X-, Y-, and Z-direction. Indeed, the size of both, fish and sensor impeded an integration of the sensor. With continued development of such sensors to smaller size or testing larger fish, these problems could be solved.

Automated fish pass monitoring

With tracking the live fish and predicting the species based on swimming trajectory features that trained a random forest machine learning approach, this thesis provides the basics for automated fish identification based on swimming patterns. Similar approaches based on automated fish contour recognition (Mandal et al., 2018; Shafait et al., 2016) are already available, but these approaches require high resolution images of the fish contours. When identifying the fish via individual swimming patterns, also poor image quality without sharp contours could be sufficient for species identification because only tracking the trajectories of the fish is required. To further develop this approach, it is promising to replace the supervised random forest machine learning algorithm by an unsupervised approach, such as Convolutional Neural Networks (Wang & Gupta, 2015), that account for patterns in the swimming paths that were not accounted by the trajectory features included in the present study. This model could be implemented in a tracking program that automatically records fish trajectories (e. g. Rodríguez et al., 2015) and feeds the trajectories to the fish identification model. All technical tools for such an approach are available. The largest challenge might be to obtain sufficient training data of fish migrating upstream with known identity.

Special thanks is for Patricia Holm for supervision of this thesis. I want to thank for your trust in me and my work during the project. You taught me a lot and helped to improve my work by providing numerous challenging but very helpful comments for my manuscripts. Thank you!

I also want to thank Peter Huggenberger for being my co-supervisor and his advice during our meetings.

Furthermore, I thank Jost Borcherding for being the external expert of my PhD. Our discussion during the Goby-meeting in Starnberg was very helpful for the further development of the force measurements approach.

I have special thanks for Philipp Hirsch, who functioned as a mentor during the project and provided fundamental support, also during the hard times. Without you, Philipp, I would not have been able to complete my thesis. In light of our recent discussion about the future, please stay who you are, whatever the future holds!

Daniel Oesterwind, my master-thesis supervisor, provided important advice from an outer perspective during my PhD. It was such a great experience to discuss my project with you and I am pleased that we stayed in contact all the years. Thank you, Oesi!

I acknowledge the contributions from Bernd Egger, who performed the live fish experiments in the flow channel study and helped fundamentally with the live fish experiments in Karlsruhe.

Furthermore, I want to thank Frank Seidel for providing the test rig at the KIT, the acoustic Dopplervelocimeter, and important support with analysing and interpreting the flow data.

There was a great cooperation with Georg Rauter, Department of Biomedical Engineering, Basel, who provided the force measurement device with the ATI-sensor. Thank you for your trust in me, your effort when providing the measurement setup and your advice for the force data analysis, as well as your advice for the random forest model!

I also acknowledge the contribution from Peter Reimann, Laurent Marot and Marco Martina, department of Physics, University of Basel, for their support with the force measurements and providing the measurement setup of the flow channel study.

Gabriel Erni Cassola contributed by proof-reading the manuscripts. Thank you!

Our master-students Nandhakumar Govindasamy, Manuela Flattich and David Scherrer spent numerous hours watching the videos and recording the fish behaviour events, as well as tracking the trajectories manually. Thank you for your support and conscientious work!

I want to thank Christine Gogel for administrative support, as well as Nicole Seiler and Heidi Schiffer for their help in the laboratory.

I want to thank Karen Bussmann for her important advice during our meetings. Thank you for your support!

I want to thank the whole team of the MGU for their support, fruitful meetings, and such a great time!

Johannes Hilpert performed the flow measurements in the flow channel in Basel. Thank you!

I thank Hans-Peter Jermann, Cantonal Fisheries Officer Basel Stadt, Frank Hartmann, Stephan Hüsgen, Anouk N’Guyen, Urs Kaiser for help with catching the fish and Matthias Thimm for help during maintenance of the fish in Basel.

Technical support was provided by Daniel Lüscher, University of Basel, and Michael Ritzmann, Christopher Ulrich and Adeline Pöschl from the technical staff at the KIT for kind assistance during the experiments. Thank you!

Anonymous reviewers gave critical review and helpful advice. You helped to increase the quality of the manuscripts significantly. Thank you!

I experienced support from my parents, Andrea Preuss and Matthias Wiegleb, and my brother, Timo Wiegleb, as well as my family. Thank you!

Important support was provided by Viviana Wiegleb and Niclas Wiegleb. Our calls, also during the hard times, helped me continue my PhD. Thank you!

I thank Karoline Wiegleb and Markus Wiegleb for their support. Thank you!

Jonathan Pfeuffer-Rooschütz and me connects a strong friendship since our early years in live and I am pleased that our paths crossed again in Basel during our PhDs. Thank you for your support!

Mike Becker and Ralf Müller, I am so pleased about our strong friendship since we sat in classrooms together. Please stay who you are, whatever the future holds! Thank you for your support!

There was always import support from Dirk Schultes and Denisa Britze. Also without you, I would not have been able to complete this thesis. Thank you! Ihr habt uns immer unterstützt und dafür danke ich Euch! Liebe Grüsse vom “Fisch-Doktor”!

My fantastic kids, Johannes, Cornelius, and Julius, and my wife, Franziska Wiegleb, gave me the power to create this thesis. I would not have been able to complete it without you. Thank you for being at my side! You are wonderful!

Baki, A. B. M., Zhu, D. Z., Harwood, A., Lewis, A., & Healey, K. (2017). Rock-weir fishway I: flow regimes and hydraulic characteristics. Journal of Ecohydraulics, 2(2), 122–141. https://doi.org/10.1080/24705357.2017.1369182

Barrett, D. S., Triantafyllou, M. S., Yue, D. K. P., Grosenbaugh, M. A., & Wolfgang, M. J. (1999). Drag reduction in fish-like locomotion. Journal of Fluid Mechanics, 392, 183–212. https://doi.org/10.1017/S0022112099005455

Beal, D. N., Hover, F. S., Triantafyllou, M. S., Liao, J. C., & Lauder, G. V. (2006). Passive propulsion in vortex wakes. Journal of Fluid Mechanics, 549, 385–402. https://doi.org/10.1017/S0022112005007925

Belletti, B., Garcia de Leaniz, C., Jones, J., Bizzi, S., Börger, L., Segura, G., … Zalewski, M. (2020). More than one million barriers fragment Europe’s rivers. Nature, 588(7838), 436–441. https://doi.org/10.1038/s41586-020-3005-2

Blake, R. W. (2004). Fish functional design and swimming performance. Journal of Fish Biology, 65(5), 1193–1222. https://doi.org/10.1111/j.0022-1112.2004.00568.x

Bombač, M., Četina, M., & Novak, G. (2017). Study on flow characteristics in vertical slot fishways regarding slot layout optimization. Ecological Engineering, 107, 126–136. https://doi.org/10.1016/j.ecoleng.2017.07.008

Carlson, R. L., & Lauder, G. V. (2011). Escaping the flow: Boundary layer use by the darter Etheostoma tetrazonum (Percidae) during benthic station holding. Journal of Experimental Biology, 214(7), 1181–1193. https://doi.org/10.1242/jeb.051938

Castro-Santos, T., Cotel, A., & Webb, P. (2009). Fishway Evaluations for Better Bioengineering : An Integrative Approach A Framework for Fishway. American Fisheries Society Symposium, 69, 557–575. Researchgate PDF

Chambers, L. D., Akanyeti, O., Venturelli, R., Jezǒv, J., Brown, J., Kruusmaa, M., … Megill, W. M. (2014). A fish perspective: Detecting flow features while moving using an artificial lateral line in steady and unsteady flow. Journal of the Royal Society Interface, 11(99). https://doi.org/10.1098/rsif.2014.0467

Coombs, S., Anderson, E., Braun, C. B., & Grosenbaugh, M. (2007). The hydrodynamic footprint of a benthic, sedentary fish in unidirectional flow. The Journal of the Acoustical Society of America, 122(2), 1227–1237. https://doi.org/10.1121/1.2749455

Dashinov, D., Czerniejewski, P., Balshine, S., Synyshyn, C., Tasheva-Terzieva, E., Stefanov, T., … Uzunova, E. (2020). Variation in external morphology between the native and invasive populations of the round goby, Neogobius melanostomus (Actinopterygii: Gobiidae). Zoomorphology, 139(3), 361–371. https://doi.org/10.1007/s00435-020-00480-7

Davey, A. J. H., Hawkins, S. J., Turner, G. F., & Doncaster, C. P. (2005). Size-dependent microhabitat use and intraspecific competition in Cottus gobio. Journal of Fish Biology, 67(2), 428–443. https://doi.org/10.1111/j.0022-1112.2005.00736.x

Drucker, E. G., & Lauder, G. V. (1999). Locomotor forces on a swimming fish: Three-dimensional vortex wake dynamics quantified using digital particle image velocimetry. Journal of Experimental Biology, 202(18), 2393–2412. https://doi.org/10.1242/jeb.202.18.2393

Egger, B., Wiegleb, J., Seidel, F., Burkhardt‐Holm, P., & Hirsch, P. E. (2020). Comparative swimming performance and behaviour of three benthic fish species: The invasive round goby (Neogobius melanostomus), the native bullhead (Cottus gobio), and the native gudgeon (Gobio gobio). Ecology of Freshwater Fish, 30(3), 391–405. https://doi.org/10.1111/eff.12592

Facey, D. E., & Grossman, G. D. (1992). The relationship between velocity energetic costs and microhabitat use in four North American stream fishes. Hydrobiologia, 239, 1–6. https://doi.org/10.1007/BF00027524

Fish, F. E., & Lauder, G. V. (2006). Passive and active flow control by swimming fishes and mammals. Annual Review of Fluid Mechanics, 38, 193–224.https://doi.org/10.1146/annurev.fluid.38.050304.092201

Franssen, N. R., Stewart, L. K., & Schaefer, J. F. (2013). Morphological divergence and flow-induced phenotypic plasticity in a native fish from anthropogenically altered stream habitats. Ecology and Evolution, 3(14), 4648–4657. https://doi.org/10.1002/ece3.842

Fuentes-Pérez, J. F., Tuhtan, J. A., Carbonell-Baeza, R., Musall, M., Toming, G., Muhammad, N., & Kruusmaa, M. (2015). Current velocity estimation using a lateral line probe. Ecological Engineering, 85, 296–300. https://doi.org/10.1016/j.ecoleng.2015.10.008

Gilbert, M. J. H., Barbarich, J. M., Casselman, M., Kasurak, A. V., Higgs, D. M., & Tierney, K. B. (2016). The role of substrate holding in achieving critical swimming speeds: a case study using the invasive round goby (Neogobius melanostomus). Environmental Biology of Fishes, 99(10), 793–799. https://doi.org/10.1007/s10641-016-0514-9

Gosselin, M. P., Petts, G. E., & Maddock, I. P. (2010). Mesohabitat use by bullhead (Cottus gobio). Hydrobiologia, 652(1), 299–310. https://doi.org/10.1007/s10750-010-0363-z

Habersack, H., & Piégay, H. (2007). 27 River restoration in the Alps and their surroundings: past experience and future challenges. Developments in Earth Surface Processes, 11(07), 703–735. https://doi.org/10.1016/S0928-2025(07)11161-5

Havel, J. E., Kovalenko, K. E., Thomaz, S. M., Amalfitano, S., & Kats, L. B. (2015). Aquatic invasive species: challenges for the future. Hydrobiologia, 750(1), 147–170. https://doi.org/10.1007/s10750-014-2166-0

Hirsch, P. E., Thorlacius, M., Brodin, T., & Burkhardt-Holm, P. (2017). An approach to incorporate individual personality in modeling fish dispersal across in-stream barriers. Ecology and Evolution, 7(2), 720–732. https://doi.org/10.1002/ece3.2629

Hoover, J. J., Adams, S. R., & Killgore, K. J. (2003). Can Hydraulic Barriers Stop the Spread of the Round Goby? Environmental Laboratory (U.S). https://hdl.handle.net/11681/5086

Imre, I. (2002). Phenotypic plasticity in brook charr: changes in caudal fin induced by water flow. Journal of Fish Biology, 61(5), 1171–1181. https://doi.org/10.1006/jfbi.2002.2131

Kalchhauser, I., Mutzner, P., Hirsch, P. E., & Burkhardt-Holm, P. (2013). Arrival of round goby neogobius melanostomus (Pallas, 1814) and bighead goby ponticola kessleri (Günther, 1861) in the high rhine (Switzerland). BioInvasions Records, 2(1), 79–83.https://doi.org/10.3391/bir.2013.2.1.14

Katopodis, C., & Williams, J. G. (2012). The development of fish passage research in a historical context. Ecological Engineering, 48, 8–18. https://doi.org/10.1016/j.ecoleng.2011.07.004

Kieffer, J. D., & Colgan, P. W. (1992). The role of learning in fish behaviour. Reviews in Fish Biology and Fisheries, 2(2), 125–143. https://doi.org/10.1007/BF00042881

Lauder, G. V., & Madden, P. G. A. (2007). Fish locomotion: Kinematics and hydrodynamics of flexible foil-like fins. Experiments in Fluids, 43(5), 641–653. https://doi.org/10.1007/s00348-007-0357-4

Liu, M., Rajaratnam, N., & Zhu, D. Z. (2006). Mean flow and turbulence structure in vertical slot fishways. Journal of Hydraulic Engineering, 132(8), 765–777. https://doi.org/10.1061/(ASCE)0733-9429(2006)132:8(765)

Lothian, A. J., & Lucas, M. C. (2021). The role of individual behavioral traits on fishway passage attempt behavior. Ecology and Evolution, 11(17), 11974-11990. https://doi.org/10.1002/ece3.7964

Lupandin, A. I. (2005). Effect of flow turbulence on swimming speed of fish. Biology Bulletin, 32(5), 461–466. https://doi.org/10.1007/s10525-005-0125-z

Lutz, E., Hirsch, P. E., Bussmann, K., Wiegleb, J., Jermann, H.-P., Muller, R., … Adrian-Kalchhauser, I. (2020). Predation on native fish eggs by invasive round goby revealed by species-specific gut content DNA analyses. Aquatic Conservation: Marine and Freshwater Ecosystems. https://doi.org/10.1002/aqc.3409

Mallen-Cooper, M., & Brand, D. A. (2007). Non-salmonids in a salmonid fishway: What do 50 years of data tell us about past and future fish passage? Fisheries Management and Ecology, 14(5), 319–332. https://doi.org/10.1111/j.1365-2400.2007.00557.x

Mandal, R., Connolly, R. M., Schlacher, T. A., & Stantic, B. (2018). Assessing fish abundance from underwater video using deep neural networks. Proceedings of the International Joint Conference on Neural Networks, 2018-July, 1–6. https://doi.org/10.1109/IJCNN.2018.8489482

Meyers, P. J., & Belk, M. C. (2014). Shape variation in a benthic stream fish across flow regimes. Hydrobiologia, 738(1), 147–154. https://doi.org/10.1007/s10750-014-1926-1

Ohlberger, J., Staaks, G., & Hölker, F. (2006). Swimming efficiency and the influence of morphology on swimming costs in fishes. Journal of Comparative Physiology B: Biochemical, Systemic, and Environmental Physiology, 176(1), 17–25. https://doi.org/10.1007/s00360-005-0024-0

Porreca, A. P., Hintz, W. D., & Garvey, J. E. (2017). Do Alluvial Sand Dunes Create Energetic Refugia for Benthic Fishes? An Experimental Test with the Endangered Pallid Sturgeon. River Research and Applications, 33(5), 690–696. https://doi.org/10.1002/rra.3132

Quicazan-Rubio, E. M., Van Leeuwen, J. L., Van Manen, K., Fleuren, M., Pollux, B. J. A., & Stamhuis, E. J. (2019). Coasting in live-bearing fish: The drag penalty of being pregnant. Journal of the Royal Society Interface, 16(151), 20180714. https://doi.org/10.1098/rsif.2018.0714

Rahel, F. J., & McLaughlin, R. L. (2018). Selective fragmentation and the management of fish movement across anthropogenic barriers. Ecological Applications, 28(8), 2066–2081. https://doi.org/10.1002/eap.1795

Rodríguez, Á., Bermúdez, M., Rabuñal, J. R., & Puertas, J. (2015). Fish tracking in vertical slot fishways using computer vision techniques. Journal of Hydroinformatics, 17(2), 275–292. https://doi.org/10.2166/hydro.2014.034

Roje, S., Drozd, B., Richter, L., Kubec, J., Polívka, Z., Worischka, S., & Buřič, M. (2021). Comparison of behavior and space use of the european bullhead cottus gobio and the round goby neogobius melanostomus in a simulated natural habitat. Biology, 10(9). https://doi.org/10.3390/biology10090821

Sagnes, P., Champagne, J. Y., & Morel, R. (2000). Shifts in drag and swimming potential during grayling ontogenesis: Relations with habitat use. Journal of Fish Biology, 57(1), 52–68. https://doi.org/10.1006/jfbi.2000.1288

Sagnes, Pierre, & Statzner, B. (2009). Hydrodynamic abilities of riverine fish: a functional link between morphology and velocity use. Aquatic Living Resources, 22(1), 79–91. https://doi.org/10.1051/alr/2009008

Sällström, E., & Ukeiley, L. (2014). Force estimation from incompressible flow field data using a momentum balance approach. Experiments in Fluids, 55(1). https://doi.org/10.1007/s00348- 013-1655-7

Sfakiotakis, M., Lane, D. M., & Davies, J. B. C. (1999). Review of fish swimming modes for aquatic locomotion. IEEE Journal of Oceanic Engineering, 24(2), 237–252. https://doi.org/10.1109/48.757275

Shafait, F., Mian, A., Shortis, M., Ghanem, B., Culverhouse, P. F., Edgington, D., … Harvey, E. S. (2016). Fish identification from videos captured in uncontrolled underwater environments. ICES Journal of Marine Science: Journal Du Conseil, 73(10), 2737–2746.https://doi.org/10.1093/icesjms/fsw106

Shao, X., Fang, Y., Jawitz, J. W., Yan, J., & Cui, B. (2019). River network connectivity and fish diversity. Science of the Total Environment, 689(19), 21–30. https://doi.org/10.1016/j.scitotenv.2019.06.340

Tierney, K. B., Kasurak, A. V., Zielinski, B. S., & Higgs, D. M. (2011). Swimming performance and invasion potential of the round goby. Environmental Biology of Fishes, 92(4), 491–502. https://doi.org/10.1007/s10641-011-9867-2

Triantafyllou, M. S., Triantafyllou, G. S., & Yue, D. K. P. (2000). Hydrodynamics of fishlike swimming. Annual Review of Fluid Mechanics, 32, 33–53. https://www.annualreviews.org/doi/full/10.1146/annurev.fluid.32.1.33

Tudorache, C., Viaene, P., Blust, R., Vereecken, H., & De Boeck, G. (2008). A comparison of swimming capacity and energy use in seven European freshwater fish species. Ecology of Freshwater Fish, 17(2), 284–291. https://doi.org/10.1111/j.1600-0633.2007.00280.x

United Nations. Convention on Biological Diversity, United Nations (1992). http://www.cbd.int/doc/legal/cbd-en.pdf

Van Liefferinge, C., Seeuws, P., Meire, P., & Verheyen, R. F. (2005). Microhabitat use and preferences of the endangered Cottus gobio in the River Voer, Belgium. Journal of Fish Biology, 67(4), 897–909. https://doi.org/10.1111/j.0022-1112.2005.00782.x

Venturelli, R., Akanyeti, O., Visentin, F., Jeov, J., Chambers, L. D., Toming, G., … Fiorini, P. (2012). Hydrodynamic pressure sensing with an artificial lateral line in steady and unsteady flows. Bioinspiration and Biomimetics, 7(3), 036004. https://doi.org/10.1088/1748-3182/7/3/036004

Wiegleb, J., Hirsch, P. E., Egger, B., Seidel, F., & Burkhardt-Holm, P. (2020). Flow field-induced drag forces and swimming behavior of three benthic fish species. Limnologica, 84, 125812. https://doi.org/10.1016/j.limno.2020.125812

Wiegleb, J., Hirsch, P. E., Seidel, F., Rauter, G., & Burkhardt-Holm, P. (2022a). Flow, force, behaviour: assessment of a prototype hydraulic barrier for invasive fish. Hydrobiologia, 849(4), 1001–1019. https://doi.org/10.1007/s10750-021-04762-z

Wiegleb, J., Hirsch, P. E., Seidel, F., Rauter, G., & Burkhardt-Holm, P. (2022b). Impact of hydraulic forces on the passage of round goby (Neogobius melanostomus), gudgeon (Gobio gobio) and bullhead (Cottus gobio) in a vertical slot fish pass. Ecology of Freshwater Fish, 00, 1–15. https://doi.org/10.1111/eff.12696

Wang, X., & Gupta, A. (2015). Unsupervised learning of visual representations using videos. Proceedings of the IEEE International Conference on Computer Vision, 2015 Inter, 2794–2802. https://doi.org/10.1109/ICCV.2015.320

Wu, S., Rajaratnam, N., & Katopodis, C. (1999). Structure of Flow in Vertical Slot Fishway. Journal of Hydraulic Engineering, 125(4), 351–360. https://doi.org/10.1061/(asce)0733- 9429(1999)125:4(351)