Major Ecological Threats for Young Fish in Regulated Rivers: Modified Migrations Cause Reduced Fitness and Increased Mortality

1. Introduction

Humans have extensively altered river systems through impoundments and diversions to meet their water, energy, and transportation needs. Dam construction is one of the most fundamental impacts humans can have on natural river systems [1,2,3,4]. Along with the positive effects, dams and other hydrotechnical constructions obstruct the dispersal and migration of organisms, and these and other consequences have been directly linked to impaired fitness, biological production, loss of populations and entire species of freshwater fish [3,5,6,7,8,9,10].

Both natural and anthropogenically transformed riverine habitats are heterogeneous [11,12,13,14]. Fish populations often occupy or travel through a diversity of habitats [15,16] wherein selection pressures are shaped by local environments. Periods of migration are physically challenging, often exposing individuals to elevated levels of predation, starvation and other risks [17,18]. Young fish migrating from the breeding to nursery habitats are especially vulnerable due to their poorly developed locomotion, orientation, foraging and defense abilities [19,20]. Flow regulation in rivers reduces the abundance of larval fish in nursery habitat, alter taxonomic composition, and disrupt microhabitat relations [21].

Contrary to the recent opinion that downstream migration of the young fish is a predominantly passive process [22], more evidence shows the key role of fish behavior in controlling their migration at the very early stages of the life cycle [23,24].

To date, there has been too much emphasis on upstream migration which is largely the domain of adult fish, and too little on downstream migration, which involves predominantly early stages [13,25]. Patterns and mechanisms governing the downstream migration (DSM) of young fish in natural and regulated rivers need special studies [but see [24,26,27]].

Here we review and analyze our published and unpublished results together with the literature on the patterns and processes associated with the downstream migration of young fish in natural and anthropogenically transformed rivers. Our primary goal is to determine ecologically significant newly emerged structures and habitats which influence migration of the young fish - modify migration patterns, cause enhanced mortality, and transform foraging conditions in the whole reservoir and within its sections from upper to lower reaches. More specifically, we consider impacts of altered topography, hydraulic structures, and biotic and abiotic conditions on 1) spatial and temporal patterns of downstream migration at different scales, 2) emigration rate of young fish out of the manmade reservoir, and 3) response of fish of different ecological groups to the variable conditions of water withdrawal from reservoirs.

The review is divided into four sections. In the first, we consider riverine fish migrations in natural rivers focusing mainly on the patterns and governing mechanisms of the downstream migration of young fish. In the second section, we highlight processes and patterns associated with modified habitat heterogeneity in regulated rivers, which control downstream migrations and their ecological consequences. In the third section, we consider and exemplify impacts of habitat structure and migratory behavior on trophic relations and fitness of young fish in natural and altered rivers. A brief overview of the main concepts and approaches to conservation and restoration of fish from different ecological groups in anthropogenically transformed rivers is provided in the fourth section.

2. Habitat Structure and Migrations of Young Fish in Natural (Undisturbed) Rivers

Natural riverine ecosystems are characterized by a high level of heterogeneity across a range of spatiotemporal scales [28,29]. The River Continuum Concept was proposed which provides a framework for integrating predictable and observable biological features of lotic systems [30]. The physical structure coupled with the hydrologic cycle form a templet [31] for biological responses, and result in consistent patterns of community structure, function, transport etc., along the length of a river. Fish make movements to gain access to patchily distributed resources, critical habitats (e.g., spawning and nursery grounds), and capitalize on favorable environmental conditions [32]. These movements influence demographic processes such as immigration and emigration, and functionally connect habitats [33,34].

Most of riverine fish migrate upstream towards the spawning grounds providing their descendants with opportunity to move (drift) downstream saving energy expenditures for swimming in viscous medium [26,35,36,37]. However, such a drift – downstream migration (DSM) - is not a safe and easy travel, but rather a serious behavioral and ecological challenge associated with various risks of passage through unknown habitats [24,27,32].

How young fish, especially at the early stages of their ontogeny, when they usually undertake risky DSM, respond to the conditions and physical structures of natural (undisturbed) rivers? The unidirectional drift along the main stream is the most evident and clearly pronounced part of DSM. According to the simplified view of the DSM of young larvae as a passive process [22,38], the migration is due to a lack of orientation in the dark and washing out from the inshore zone to the main channel of the river. If their transition to the flow is completely passive, the relative abundance of larvae of different species has to be basically similar in both habitats of residence and migratory habitats.

However, the results of field and experimental studies demonstrate that this is far from the case and suggest that behavioral mechanisms are at play [24,27,39,40]. We observed that despite a lower concentration in the habitat of residence, roach Rutilus rutilus larvae were more common migrants than another cyprinid species Phoxinus phoxinus, which were most abundant in the habitat of residence [41]. An increase in the number of roach larvae in the migratory habitat compared with the number in the inshore zone can be explained only by behavior. More examples illustrating the role of behavior in the control of the young fish DSM: 1) DSM takes place during the polar day, when illumination allows orientation for 24 h [24]; 2) the behavioral responses of migrants are more pronounced near interfaces, where small-scale topography and hydraulic gradients serve as cues for orientation [42,43]; 3) using an experimental model of the hydraulic interface between habitats of residence and migratory habitats, we discovered that fish larvae with the appropriate internal state (migrants) choose a suitable velocity depending on illumination. In the darkness, migrants choose zones with higher velocity, thus actively entering the drift (DSM), while at dawn fish were leaving the migratory habitat for entering the low velocity habitat of residence [44]; 4) we suggested that one of the important aspects of DSM, synchronization of migratory behavior, may result from the similarity of individual responses to changes in illumination and/ or could be related to shoaling behavior, which is typical of late-stage larvae and fry [45,46,47]. More synchronized responses of older larvae to diel changes in illumination support the concept of interactive behavioral responses [24]. Thus, we believe that the major temporal and spatial parameters of DSM are influenced by fish behavior (locomotion, orientation, decision-making), 3D flow structure, and diel cycles of illumination. Although we emphasize the importance of behavior in the control of DSM, we do not ignore passive drift. The whole process of DSM is composed of a sequence of active and passive components.

Short periods of decision-making when migrants enter the DSM at dusk and return to the inshore retention zone at dawn are of particular importance in controlling of DSM. We suggest that such key events are associated with high gradient/heterogeneous zones (interfaces between habitats of residence and migration habitats) [24,48]. Throughout DSM, migrating fish move across the transition zone many times when they enter the drift (migratory habitats) and return to the shore (habitat of residence).

The overall success of DSM depends not only on the migratory behavior of fish in the main channel of the river (migratory habitat) but also on other vital activities, such as foraging, defense, and grouping, associated with both habitats of residence and migratory habitats [42,43,49]. Migrating fish both drifting in the main channel, and moving from one inshore site to another constantly find themselves in unfamiliar habitats. The micro-scale heterogeneity and local aggregation of fish in these habitats within a natural river seem to have a significant influence on the synchronization of activity and DSM of the migrating fish.

The risk of dilution (drop of local concentration) is great for migrating populations of young fish, which have high feeding demands, vulnerability to predators, and a tendency to group [24,49,50]. Without shoaling, which is impossible at a low local population density [51], young fish would not be able to efficiently avoid predators, search for food, and navigate [46,47,52]. We hypothesized that DSM is adaptive if dispersal and widening of nursery grounds work together with mechanisms for maintenance of local aggregations to ensure efficient intraspecific interactions (“dispersal with low dilution” hypothesis) [24].

One of the most pronounced impacts of the river regulation on young fish is creation of enlarged water bodies, reservoirs, with reduced flow velocity, modified habitat heterogeneity, and high risk of fish mortality in some localities of reservoirs. In the following section, we consider most important physical traits and patterns of habitat heterogeneity in water reservoirs, as well as their effects for young fish migrations.

3. Multiscale Influence of Modified Habitat Heterogeneity on Downstream Migration of Young Fish in Regulated Rivers

Humans all over the planet are assisting with the development of new ecosystems. Such ecosystems are not emerging de novo. Instead, they are emerging from ‘within’ pre-existing systems that are naturally dynamic, both over long and short time-scales [53]. Habitat heterogeneity influence fish migratory behavior at different scales, from the local patterns of landmarks and gradients to the configuration of the whole river system [10,54]. Natural rivers with gradually changing characteristics ensuring connectivity provide conditions for the DSM along the length of a whole river [24,26,27].

Regulation of the rivers, mainly with dams, creates reservoirs and other structures which are considered movement barriers for riverine and anadromous species [48,54,55]. Flow regulation endangers migrating fish, especially those at the early ontogeny, predominantly by entrainment into water abstraction devices [56,57]. The vast majority of studies on downstream migration of young fish in regulated rivers is focused on passage through the turbines of power plants and other water intakes associated with dams [58,59,60]. These constructions for water discharge from reservoirs are considered the most dangerous for migrating fish in terms of high mortality and damage rate [37,61,62,63]. The role of transformed parts of the river located upstream the dam (proper reservoir) is still underestimated from the viewpoint of mortality and foraging conditions for migrating fish.

Reservoirs are usually created by damming rivers or streams and flooding the surrounding floodplain and tributaries. Many reservoirs may have properties that reflect both lakes and flowing systems. Two major types of water reservoirs are commonly distinguished, lake-like and channel-like reservoirs [48]. The main physical traits of these manmade objects, which are of primary importance from the point of influence on the DSM intensity, include a) annual water exchange rate, and b) morphological complexity of the reservoir (Table 1). The index of morphological complexity was calculated as the ratio of the total length of impounded tributaries to the length of impounded part of the main river [48]. We hypothesized that the water exchange rate is the main physical factor influencing the intensity of DSM and emigration of fish from reservoirs. Negative correlation between biomass of the pelagic (open water) fish inhabiting several reservoirs of the Volga River and water exchange rate supported this suggestion [64]. It means that withdrawal of fish is expected to be the highest from reservoirs with the largest water exchange rate. However, later we found another, probably not less important factor, morphological complexity of the reservoir, which hampers emigration of fish, predominantly 0+ juveniles, from reservoirs (Table 1). Taking into account only water exchange factor, we can expect the highest withdrawal (intensity of DSM) from the Ivan’kovskoe Reservoir. However, the highest annual withdrawal was observed from another, channel-like, reservoir of the Volga River cascade, Volgogradskoe, where morphological complexity was at minimum (Table 1). The lowest withdrawal rate was recorded for the Tsimlyanskoe Reservoir with relatively low exchange rate and rather high morphological complexity.

River regulation radically transforms environment for all types of fish migrations at different scales [5,54,65,66]. Reservoirs change stream conditions throughout impounded watersheds. Immediately upstream of the dam, the river gains lentic properties: flow is reduced, width, depth, and sedimentation increase [67].

Before considering ecological consequences of the river regulation and novel physical structures for the fish DSM, we have to provide a general scheme of the altered habitats in the reservoir emerged upstream the constructed dam. Modified parts of the river include: 1) lotic-limnetic interface (transient region in the upper part of reservoir); 2) reservoir itself with stretched gradients and changed heterogeneity; 3) water-intake zone at the dam.

In natural rivers, migrating fish are basically preadapted to habitat heterogeneity (hydraulics, topography, shelters, foraging sites) and spatiotemporal distribution of ecological risks [24,27]. Novel physical structures emerging in regulated rivers either fully block or significantly hamper DSM of young fish [48,54].

First newly emerged ecological barrier/filter, which seriously influences DSM of young fish, migrating from the upper reaches of the regulated river, is the lotic-limnetic interface in the upper part of a reservoir [70]. In this area (backwater zone), the river continuum is disrupted, which leads not only to changes in the living conditions of riverine fishes, but also dramatically changes conditions for DSM of young fish. The main features of this zone are: sharp drop in the flow velocity, intense sedimentation, modified habitat heterogeneity (flow structure, bottom topography, water transparency), high amount of juvenile piscivorous fish (Table 2).

Large water bodies of impounded parts of a river (reservoirs) are the most voluminous obstacles for migrating fish [48,54]. Their varying depth and currents, and modified transition zones between habitats of residence and migratory habitats radically change conditions of the DSM for young fish, creating ecological barriers of different permeability. To understand the relative influence of biotic and abiotic factors on the patterns and mechanisms of DSM either in undisturbed rivers, or in reservoirs, we have to consider the characteristics of not only migratory habitats (main channel of the river, or open water of the reservoir) but also habitats of residence (sites of stopovers in the inshore habitats) and the transition zone (interface) between these two types of habitats. The structure, topography, and flow parameters of the transition zone are especially important for fish undertaking DSM [22,24,69,71] (Figure 1).

Young fish, migrating in a reservoir, face enormously extended transition zones between the migratory habitat (open part of the reservoir) and the habitat of residence (shallow-water inshore zone). Compared with natural rivers, the extended transition zone makes repeated lateral diel movements between the habitats much more difficult. A significant number of migrants retain for the whole day in the open water of the reservoir (migratory habitat) instead of returning at dawn to the inshore zone (habitat of residence) (Figure 1 and Figure 2).

In the Upper Volga reservoirs, the period of downstream migration is more prolonged than in the river. Within this period, i.e., June–September, a much higher concentration of young percids was recorded in the open water (migration habitat) than in shallow water (habitat of residence) (Figure 3 and Figure 4). The opposite pattern was observed in cyprinids—they were much more abundant in the inshore zone (zone of residence). This difference produces a prerequisite for more intensive drift of the young percids than cyprinids towards the dam [14].

In homogeneous environments of the open part of the reservoir with few landmarks and highly variable flow structure, most of the main vital functions and biotic interactions of migrating fish are hampered and difficult to be fulfilled, especially for young fish [24,42]. Maintenance of local aggregations (shoals) is important not only for fish in habitats of residence in the inshore zone but also for drifting fish, as drifting is the most dangerous part of migration. Increased density of migrant conspecifics dramatically improved the survival rate of sockeye salmon Oncorhynchus nerka smolts [72].

Besides the impacts on the diel characteristics of DSM, changed conditions of regulated rivers significantly influence seasonal patterns of DSM. Compared to the DSM of 0+ perch, Perca fluviatilis, in the natural river (upper Volga River), peak of DSM of this species in the reservoir of the same river appears later and the migration lasts much longer. Thus, not only most numerous in the midsummer early stages are involved in migration, but also grownup juveniles in autumn – winter period (Figure 3). Several factors of different scale we mentioned above, namely, morphological complexity of the reservoir, variable hydraulic structure, extended transition zone between migratory habitat and habitat of residence, homogeneity of open water area are supposedly important factors, which modify seasonal dynamics of DSM in fish juveniles.

In the sequence of modified parts of the impounded river, a 3-D hydrological structure with a downstream flow through the Hydro-Power-Plant passage in the dam is the most dangerous area, where the impact on DSM and survival of young fish may be especially harmful. The total concentration of fish in this area is higher than that at the adjacent upstream area of the reservoir, largely due to the drift and the accumulation of migrants [48]. The area associated with the water intake constitutes an eco-hydraulic template where the DSM of fish is under control of the two interacting structures: (1) transformed water flow, which is like a hydraulic 3-D funnel, and (2) a set of adjacent ecological zones (habitats). Depending on the parameters of these structures, different types of Eco-Hydraulic Patterns of Interaction (EHPI) between water flow and ecological zones were observed [4]. Larvae and juveniles of pelagic fishes were significantly more vulnerable to be entrained by the outflow, especially through the deep-water intakes, than littoral fishes.

DSM of cyprinid and percid fishes are strongly influenced by novel ecological structures associated with different parts of water reservoirs [14]. While the intensity and patterns of the DSM of cyprinids and percids are rather similar in free-flowing natural rivers, a much more intensive emigration of juvenile percids, especially zander, Sander lucioperca, was observed from reservoirs with deep-water intakes. Modified conditions of regulated rivers, especially those at the lower part of water reservoirs (water intake zone), induce a new type of impact of the downstream water flow on fish inhabiting adjacent ecological zones [4].

Removal of migrants through the water intakes is under control of hydraulic and ecological impacts at different spatiotemporal scales: the whole-lake, between-habitats and within-habitat scales [48,71.4]. The pelagic zone of many reservoirs has lower fish diversity and populations numbers. Intensive (selective) removal of fish from this area through the water intake of the limnetic–lotic interface could be one of the factors causing this effect. For example, in the reservoirs of the Volga and Kama rivers, a strong negative relationship between the fish biomass in the pelagic zone and the intensity of water exchange was observed [64].

4. Ecological Effects of Transformed Habitats on Fish Migratory Behavior, Biotic Interactions, and Mortality in Different Parts of Reservoirs

Altered morphology, hydrology, chemical and biotic characteristics of regulated rivers dramatically change conditions of DSM and other vital activities of young fish. From the viewpoint of individual fitness and population number, mortality and foraging conditions are the major ecological factors/determinants influencing overall success of the DSM. In addition to the natural factors of mortality (predation, parasitism, starvation), newly emerged morphological and hydraulic structures create local dangerous situations in different parts of reservoirs, which directly or indirectly increase young fish mortality.

Predation by piscivorous fishes and birds is one of the major ecological threats for young fish, especially in the period of high migratory activity. Effects of predation depend not only on the numbers of interacting predators and prey, but to a large extent on the availability of prey, which is strongly related to the microscale heterogeneity of the foraging ground [49,73]. Such a heterogeneity is typical of the shallow-water inshore habitats, where residing young fish may easily find landmarks, food, and shelters. All these objects are much less available in the homogeneous open-water parts of reservoirs. Low permeability of the transient zone between inshore and open-water habitats, which prevents lateral movements of migrants, retains a significant number of young fish, especially percids [14], in the dangerous pelagic habitat for the most part of the diel cycle. Below, we consider ecological effects of staying in the open water on predator-prey interactions and feeding success of young fish. There are several sections of the regulated river (river stretches), where novel factors, besides natural factors of mortality, emerge.

Lotic-limnetic interface in the upper part of reservoir. The combined effect of abiotic and biotic factors leads to a significant inhibition of DSM of young riverine fish and an increase in their mortality in the transient zone between river and reservoir. Inhibition of DSM occurs not only due to a decrease in the average flow rate, but also due to the lateral transport of migrants from the incoming flow to the inshore shallow waters [71]. In shallow waters of the Ili River, where the flow rate is noticeably lower than in the midstream channel, and the water transparency is much higher, migrating larvae of gobies Rhinogobius brunneus, bream Abramis brama, dace Leuciscus leuciscus, and other species become easy prey for predatory fish like asp Aspius aspius, zander Sander lucioperca, and Balkhash perch Perca schrenki. Impact of predators, as well as the death of eggs and larvae in areas of intense sedimentation, determined the high total mortality of migrating riverine fish in the uppermost part of reservoir, where riverine conditions were sharply replaced by lacustrine ones [41]. This transient zone’s “filter” not only reduces the total number of migrants, but, acting selectively, reduces the abundance of certain species much more than others. Only a small number of the early stages of grass carp Ctenopharingodon idella, silver carp Hypophthalmichthys molitrix, and Aral barbel Luciobarbus brachycephalus migrating from the upstream stretch of the Ili River enter the reservoirs due to siltation of eggs and larvae, as well as predation pressure [71].

Water reservoir (impounded river stretch). The vast area of the reservoir itself is characterized by low and changeable flow velocity, and poorly structured (homogeneous) habitats [10,25,48,54,67]. These factors produce not only unfavorable conditions for fish migrations (migratory behavior) along the whole reservoir, but also influence various vital activities like foraging, defense, orientation, grouping in transformed river habitats [42,43,49].

Peak densities of migrating fish in natural rivers are usually recorded during the night. During the day, the migration habitat (main channel) is virtually free of downstream migrants (Figure 1). The diel rhythms in the behavior of migrating fish and their interactions with small scale hydrological gradients and heterogeneous topography shape the temporal patterns of DSM in young fish. Migration habitats in regulated rivers (open water of reservoirs) contain larval fish both day and night (0+ percids are often the most numerous migrants). Comparing these two contrasting water bodies, river vs reservoir, we argue that riverine migrants easily leave the migratory habitat at dawn and enter the heterogeneous inshore zone. Lateral movements allow them to cross a narrow flow gradient between the habitats, and undertake foraging under the structured, stable, and safe conditions. Such behavior seems much more difficult for larvae, which migrate in the reservoir, because extended interface between migratory and residence habitats creates a hardly permeable barrier for migrants. Thus, a large number of migrants stay in the poorly structured (homogeneous) open water during both night and day (Figure 2 and Figure 4). Under such conditions, feeding, communication, defense, and other behaviors are much less successful than in the structured (heterogeneous) inshore habitats. 0+ perch Perca fluviatilis foraging in the heterogeneous area at the macrophyte beds consumed twice more zooplankton crustaceans, than perch from the neighboring open-water area [74] (Figure 5).

Moreover, considerably bigger prey prevailed in the diet of fish from the heterogeneous area. Such a successful feeding of 0+ perch was supposedly related to their confident behavior under safe conditions, where physical and “social” (shoals of conspecifics) shelters were available. We also observed similar behavior in experiments with another predator-prey interaction: the guppy Poecilia reticulata fed intensively and readily consumed large prey even under the impact of piscivorous fish kairomons, if the interaction occurred nearby the macrophyte beds (heterogeneous habitat); in the open-water (homogeneous habitat) added kairomons strongly suppressed feeding, especially on large prey, handling of which is time-costly [73]. Our field and experimental data suggest that hampered foraging and high risk of predation for young fish migrating in the open waters of reservoirs lead to their lowered fitness and increased mortality. Large size and homogeneous structure of reservoirs makes them a significant ecological factor, which can control population number and composition of fish communities.

Water-intake zone at the dam. The DSM of juvenile fish through the water intakes of different types associated with dams and other constructions is a potentially powerful factor, which influences the abundance and composition of fish populations inhabiting reservoirs. Depending on the pattern of interactions between the discharge water flow and adjacent ecological zones, the outflow through the dam selectively removes fish of different ecological groups from different habitats [4].

The deep-water flow of Hydro-Power-Plant removes much more fish from the pelagic habitats than the littoral ones. The concentration of migrants through the water intake of the HPP is much higher than that through the nearby Shipping-lock, which abstracts water mainly from the shallow-water habitats [4,14]. These effects resulted not only from the higher amount of water going through the HPP intake, but also from the physical heterogeneity of the adjacent habitats. Heterogeneous habitats with various shelters and landmarks provide higher retaining capacities for fish from flashing out from the reservoir. Young fish of ubiquitous species from the Percidae and Cyprinidae families, perch, roach, bleak, bream, emigrate in mass through the lentic-lotic interface (deep-water HPP water intake) entrained by the 3-D outflow from the pelagic habitats (Figure 6 (a), (c)); DSM of these fishes, especially cyprinids, is much less intensive, when they migrate through the shallow-water intake, which entrains water and migrants from heterogeneous (rich of shelters and landmarks) inshore habitats (Figure 6 (b), (c)).

5. Ecologically Based Approaches to Conservation and Restoration of Fish in Regulated Rivers

Management and conservation of young fish inhabiting reservoirs are predominantly focused on prevention of the entrainment of migrating fish by the water abstraction flow, which carry them into the HPP passages equipped with damaging turbines and other obstacles. Development and design of various types of physical and behavioral screens, as well as fish-friendly turbines, which prevent or mitigate the impacts of entrainment and entrapment of migrants and their mortality, are still the main approaches to protection of fish larvae and juveniles in the periods of mass migration and dispersal. Most of the efforts and developed devices are linked with the lower part of reservoir adjacent to the dam and water abstraction sites. Other parts of the reservoir, where DSM and biotic interactions directly or indirectly influence mortality, fitness and emigration of young fish received much less attention.

Ecologically based approach to efficient management of fish populations in regulated rivers is mainly related to the principle of minimizing the overlap between the downstream flow patterns in the water-intake zone near the dam and the spatiotemporal distribution of young fish during their mass migration period. Knowledge of the diel and seasonal dynamics of DSM, as well as the structure of the spatial distribution of migrants both at the whole reservoir and between-habitats levels, especially near water intakes, makes it possible to regulate flow discharge and emigration of young fish. Maximum water discharge should not be associated with periods of mass fish migration (diurnal and seasonal migration windows).

To develop specific measures to minimize mortality risks, we have to (1) determine a target species or group of fishes, taking into account their ecology and behavior, especially those related to migrations of juveniles, predator-prey interactions and responses to habitat heterogeneity; (2) assess the major periods of migratory activity of fishes on both diel and seasonal scales; (3) estimate the rate of transportation of migrants from the upper to lower part of a reservoir—this process is strongly influenced by the topographical and hydraulic complexity of the reservoir, and the rate of water exchange; (4) estimate the eco-hydraulic pattern of interaction (EHPI) between the downstream water flow and ecological zones/habitats within the water intake area; (5) assess within-habitat complexity (availability of landmarks and shelters), which may suppress the removal of fish from habitats of residence by the outflow.