Risk Classification of European Lampreys and Fish Species

Home , Asp (fish), Cactus roach, Carpathian brook lamprey, Common barbel, Luciobarbus bocagei, Monkey goby

Ref. Ares(2018)5582208 - 31/10/2018

Fishfriendly Innovative Technologies for Hydropower

Funded by the Horizon 2020 Framework Programme of the European Union

D1.2 Risk classification of European lampreys and fish species

Project Acronym FIThydro Project ID 727830 Work package 1 Deliverable Coordinator Christian Wolter Author(s) Christian Wolter (IGB), Ruben van Treeck (IGB), Johannes Radinger (IGB), Nicole Smialek (TUM), Joachim Pander (TUM), Melanie Müller (TUM), Jürgen Geist (TUM) Deliverable Lead beneficiary TUM, FVB.IGB Dissemination Level Public Delivery Date 31 October 2018 Actual Delivery Date 31 October 2018

Acknowledgement This project has received funding from the European Union’s Horizon 2020 research and inno- vation program under grant agreement No 727830.

Executive Summary

Aim

After having analysed the species-specific sensitivity against mortality, i.e. the species’ intrinsic resilience, this study aimed to characterise the potential impacts of hydropower on fish. The main objective was to analyse the principal operation related impacts of hydropower and identify species most at risk. The results provide the operation related component or baseline for developing a fish population hazard index for the European fish fauna.

Methods

Three principal operation related impacts of hydropower have been identified: i) migration barrier, ii) mortality, and iii) habitat loss due to impoundments. Migration barriers notably affect diadromous fish, which are obligatory migrants, moving between freshwater and marine environments. Mortality has been quantified and classified primarily during turbine passage. Other types of hydropower induced mortality, e.g. at trash racks, could not be quantified. Habitat loss due to impoundments has been attributed to hydromorphological processes and related habitat characteristics. Correspondingly, lithophilic, i.e. gravel spawning species were identified experiencing the highest impacts of habitat loss due to impounding rivers.

Species were classified at very high risk from hydropower operations, if at least three of the four following conditions were fulfilled: i) belonging to the high or highest sensitivity class, ii) having high or highest mortality risk during turbine passage, iii) being diadromous, and iv) being lithophilic.

Species were classified as high risk in hydropower environments, if two of the conditions mentioned above were fulfilled, and they were classified as lower risk, if only one or none of the conditions were fulfilled.

Results

The sensitivity matrix contains 148 native European fish and lamprey species occurring in Eu- ropean waters. Of these, 18 species were classified as having the highest sensitivity and 29 with high sensitivity. Twenty two of the classified species are diadromous. The habitat degradation and loss in the impoundment particularly affect lithophilic fish, which represent a total of 76 species.

The mortality was empirically derived from turbine passage studies. Data were gathered for 42 species in total, of which 20 had a sufficient sample size. The data could be used to derive a model to assess the length dependent potential mortality risk for data deficient species. However, here only the empirical data for the 36 species occurring in more than one study with more than one specimen provided in Table 4 were used for identifying species’ mortality risk.

Combining the intrinsic sensitivity of species and the operation related impacts of hydropower, as mentioned above, 21 and 25 species face a very high and high risk, respectively. Among the species at very high risk, there are several candidate species to be considered for environmental impact assessment, as well as mitigation for various river types, e.g., the diadromous salmonids or the large bodied cyprinids, like barbel, nase, asp, and ide, but also the smaller bodied dace.

Specific spatial and habitat requirements have been compiled for a range of riverine species, although data availability was very heterogeneous and for most species very limited and data quality highly incoherent among different studies.

727830 FIThydro - Deliverable 1.2 - Page 2 of 39

Table of Contents Executive Summary ...... 2 List of Figures ...... 3 List of Tables ...... 4 1. Background ...... 5 2. Methods ...... 6 2.1 Sensitivity matrix ...... 6 2.2 Fish mortality at hydropower plants ...... 6 2.3 Habitat alteration in impoundments ...... 8 3. Results ...... 9 3.1 Sensitivity matrix ...... 9 3.2 Fish mortality at turbines ...... 13 3.3 Habitat requirements ...... 22 3.4 Synthesis ...... 33 4. Conclusions ...... 36 5. References ...... 37

List of Figures Figure 1: Observed species-specific turbine mortality rates at hydropower plants...... 14 Figure 2: Robust empirical species-specific turbine mortality rates at hydropower plants ..... 15 Figure 3: Observed species-specific turbine mortality rates at hydropower plants ...... 16 Figure 4: Observed range of substrate particle sizes in spawning grounds and juvenile habitats ...... 25 Figure 5: Observed depths of eggs in the interstitial space...... 26 Figure 6: Observed incubation times in degree days for egg development of gravel-spawning fish species...... 27 Figure 7: Observed range of preferred surface flow velocities at spawning grounds and juvenile habitats...... 28 Figure 8: Observed range of preferred water depths at spawning grounds and juvenile habitats ...... 29 Figure 9 Reported ranges of individual spawning sites...... 30 Figure 10: Reported ranges of spawner densities at gravel spawning grounds...... 31 Figure 11: Reported juvenile densities in nursing grounds...... 33

727830 FIThydro - Deliverable 1.2 - Page 3 of 39

List of Tables Table 1: Sensitivity classification of 148 native European fish and lamprey species...... 10 Table 2: Calculated length of fish facing a mortality risk during turbine passage...... 17 Table 3: Species-specific potential mortality risk at hydropower plants during turbine passage...... 18 Table 4: Species intrinsic sensitivity versus empirical species-specific mortality at hydropower turbine passage...... 21 Table 5: Overview of reproduction traits and abiotic environmental conditions in spawning and nursing habitats...... 22 Table 6: Metadata of available information on environmental conditions in spawning and nursing habitats ...... 23 Table 7: Overview of reported traits for Salmo trutta...... 32 Table 8: Summary of native European lamprey and fish species of high and very high risk during hydropower operation...... 35

727830 FIThydro - Deliverable 1.2 - Page 4 of 39

1. Background Assessing the impact of hydropower plants on different fish species and communities is a key component of supporting decisions during the commission and operation of hydropower facilities. Therefore, work package 1 “Fish population development in hydropower effected environments” aims to: i) compile a metadata overview of existing data and review the literature on factors determining the response and resilience of fish to river fragmentation and hydropower, ii) perform meta-analyses on intrinsic biological responses of fish to hydropower development, and iii) identify species most at risk to iv) develop a European Fish Population Hazard Index as a decision and management tool for hydropower planning and environmental impact assessment. These different tasks have been conceptualised as three distinct components: the biological resilience of species, the operation-related impacts and the site- or group-specific impacts of hydropower.

Biological resilience is the ability of a species to recover from natural mortality. Resilient species have evolved traits that promote recovery, e.g. high fecundity and short generation intervals. Therefore, with increasing resilience species should be able to better recover from human- induced mortality. Following this logic, highly resilient species are considered less sensitive against human disturbance. In contrast, species with fewer needs to evolve buffer mechanisms against high adult mortality are considered highly sensitive against human disturbance. A sensitivity classification comprising 192 fish and lamprey species occurring in European waters has been developed in the first year of FIThydro and provided in deliverable D1.1 “Metadata overview on fish response to disturbance” (van Treeck et al. 2017).

The second step is now analysing the operation-related impacts of hydropower, one of the known human uses and alterations of aquatic systems. Operation-related impacts comprise those principal impacts which generally occur, but which might differ in amplitude and signifi- cance depending on site-specific modifications. Here we focus in particular on the production of electricity from hydropower, which has three main operation-related impacts: 1) Migration barrier, 2) Mortality, primarily at the energy converter, 3) Habitat alteration due to impoundments

The migration barrier will not be further considered here, because the migration type of a species has been considered already in the sensitivity classification. However, a barrier some- where in the river system is inherent to each hydropower plant, because a certain head difference between the headwater and the energy converter used for power production is essential to each hydropower plant. The species most affected by barriers are diadromous lampreys and fishes, which are obligated to migrate between freshwaters and the sea to fulfil their life cycles, followed by potamodromous species with obligatory migrations within river systems for spawning, then large-body facultative migrants with larger home ranges and finally small-body facultative migrants with small home ranges. With increasing migration or movement needs of a species, the probability of impacts by migration barriers increases as well as the probability of encountering any kind of risk within the system. Therefore, this life history trait has been considered relevant to resilience in the sensitivity classification.

The analyses presented here address the two remaining operation-related impacts, mortality and habitat alteration in the impoundment. Mortality occurs primarily at the hydropower plant, especially during passage through the energy converter, most often a turbine, but also at trash

727830 FIThydro - Deliverable 1.2 - Page 5 of 39 racks, in various bypasses and due to migration delays or increased predation. This study focused on the compilation and analyses of empirical mortality data at hydropower facilities, which allowed the assessment of species-specific mortality rates.

Impoundments typically increase the water depth and reduce flow velocity, stream power and turbulence. As a result, sediment transport is reduced and settlement of fines is increased. A conceptual overview on impoundment effects on hydromorphological processes and variables is provided by Garcia de Jalón et al. (2013). The risk classification for European lampreys and fishes primarily focuses on identifying species that are impacted by typical impoundment effects, such as deposition of fines, loss of coarse substrates and flow velocity patterns.

The findings of both impact analyses were merged with the sensitivity classification of species (van Treeck et al. 2017) to obtain a matrix of species’ mortality at hydropower plants and species’ sensitivity against habitat alterations in impoundments. The resulting matrix serves as a risk classification tool for European lampreys and fishes in relation to hydropower. The final classification of European fish species according to their susceptibility towards hydropower provides a set of highly sensitive candidate species from different biogeographic regions of Europe considered for environmental assessments and further analyses of impacts at the level of populations or communities, together predicting group-specific threats and effects of hydropower operation on fish.

In particular for species most at risk but generally also for all other lamprey and fish species, additional data on spawning and juvenile habitat characteristics and carrying capacities were compiled to provide qualitative and quantitative advice for habitat improvements as potential compensation measure for fish losses at hydropower plants.

2. Methods 2.1 Sensitivity matrix The sensitivity classification has been described in detail in deliverable D1.1 (van Treeck et al. 2017). It is based on a compilation and analysis of fish life history traits providing resilience against natural mortality and supporting recovery after a disturbance.

The trait data compilation has been continued and improved. General relationships have been derived from the data or provided by Fishbase (Froese & Pauly 2018) to calculate certain traits based on the maximum total length reported for a species, respectively in particular on the derived asymptotic length of the von Bertalanffy Growth Equation (von Bertalanffy, 1938). This enabled us to fill certain gaps in the trait database, especially for data-poor species and to derive a generally more robust classification. For a subset of species with numerous observation data the calculated data had been verified. There were no significant differences between classification results of single species based on calculated data only and of empirically observed data only.

2.2 Fish mortality at hydropower plants The literature (grey and peer-reviewed) has been extensively reviewed and analysed for studies and reports of fish mortality at hydropower plants using the search engines Web of

727830 FIThydro - Deliverable 1.2 - Page 6 of 39

Knowledge and Google Scholar with fish and hydropower (also in German) as keywords. In addition, reports were directly requested based on an extensive bibliography of mortality studies compiled by the German “Forum Fischschutz”. Studies have been selected if the methods reported were suitable to assign fish mortality or injury during the passage of the power plant, e.g. by stow net catches in the tail water of the turbines (e.g. Holzner 2000, Schmalz 2010, 2016, Schneider et al. 2012, Ebel 2013, Mueller et al. 2017). Injections of experimental fish behind the trash rack have been also considered, except if the fish were directly injected with a pipe into the turbine. The latter experiments were excluded because the moment when a fish enters the turbine is critical for injury and mortality and this important moment is linked to the pipe injection, resulting in biased mortality rates. Telemetric studies were also not considered, because they do not allow final conclusions about death or injury of a specimen (Havn et al. 2017).

Fish mortality at hydropower plants occurs mostly during turbine passage (e.g. Hadderingh & Bakker 1998, Holzner 2000, Ebel 2013, Mueller et al. 2017), but also at the trash racks (Wagner et al. 2013a, b, c), within the impoundments (Thorstad et al. 2017, Havn et al. 2018), in bypasses (Havn et al. 2018) and due to increased predation downstream the powerhouse (Havn et al. 2018). One problem is that there are several kinds of mortality that particularly affect diadromous fish, which have to reach the sea to complete their life cycle, but cannot be quantified for potamodromous or facultative migrating species, e.g. extra mortality in the impoundment, which can reach 7-17% (Havn et al. 2018). Similarly, dead fish at trash racks cannot be related to the number of approaching fish to obtain mortality rates (Wagner et al. 2013a, b, c).

Therefore, this study focused on mortality at the energy converter, where the reported numbers of injured or dead fish in relation to all passed fish allowed the assessment of species-specific mortality rates. Our data survey aimed to obtain as many investigations as possible to collate sufficient sample sizes and information on a maximum number of species. The studies usually reported observed injuries at a scale of three (Wagner et al. 2013a, b, c) or five (Holzner 2000, Schneider et al. 2012) severity classes from no visible injuries to lethally injured, as e.g. de- capitation. Some studies also reported delayed mortalities occurring in control holding tanks up to 48-96 h after turbine passage (e.g., Edler et al. 2011). Further, in mortality studies also handling related mortality and injury were detected (Pander et al. 2018).

Here all fish with injuries classified substantial or higher, dead fish or reported delayed mortality were summed up to the total number of fish killed during turbine passage and related to all passing fish to calculate mortality rates. These calculations were performed for each species aiming to obtain at least five studies per species with a minimum of ten specimens each.

Basic data comprised the total number of specimens caught, the number of specimens in the different injury or mortality classes and any information provided on the hydropower plant and especially on the energy converter. All data were recorded in a database as species by the different investigation data sets allowing for species specific analyses including the calculation of median values and percentiles from the distribution of reported mortality rates to be viewed. Mortality rates were classified according to the median value as very high, if median mortality was ≥8%, high ≥4%-<8%, moderate ≥2%-<4%, low ≥1%-<2% and very low <1%. To account for the variability of mortality rates between studies, they were additionally classified according to the 75% percentile of all observations. The corresponding mortality classes were very high,

727830 FIThydro - Deliverable 1.2 - Page 7 of 39 if 75% percentile mortality was ≥16%, high ≥8%-<16%, moderate ≥4%-<8%, low ≥2%-<4% and very low <2%.

2.3 Habitat alteration in impoundments Impoundments significantly impact hydromorphological processes, such as flow velocity, stream power, sediment transport and sorting (Garcia de Jalón et al. 2013) causing significant habitat loss for typical riverine species (e.g., Mueller et al. 2011, Pander & Geist 2018). The main hydromorphological features and structures are primarily determined by the natural flow regime of the river and the nature of the sediments available for erosion, transport and deposition. The interaction between flowing water and the size and quantity of available sediment leads to diverse substrate patterns emerging from flow-induced sorting, which are type-specific for river systems and thus, indicative of hydromorphological integrity. Impoundments impact this hydromorphological integrity, which among other changes, results in increased sedimentation of fine materials and correspondingly the loss of coarser substrates.

While a certain stream power is needed to sustain coarse, well oxygenated, permeable gravel beds, fish species unconditionally depending on such substrates for spawning principally provide sensitive, diagnostic indicators for hydromorphological integrity. Gravel spawning is com- monly considered as fish’s adaptation to faster flowing environmental conditions by protecting eggs and hatchlings from washing away. Lithophilic fish bury their eggs in or lay them on coarse gravel and larvae are benthic, living in interstitial spaces (Balon 1975, 1981). Eggs and larvae of lithophilic fish develop in the gravel layer (Pander et al. 2009, Duerregger et al. 2018). There- fore, species of this ecological guild essentially depend on interstitial flow, permeability and oxygen supply in the gravel space for their successful reproduction and population development (Sternecker et al. 2013). Accordingly, lithophilic species have been identified as most sensitive to impacts on sediment structure and hydromorphological alterations and thus, most affected by habitat loss from impounding rivers (Mueller et al. 2011). In contrast, plant spawners and eurytopic fish usually inhabiting low energy rivers might even benefit from impoundments.

Independent of the spawning substrate, nearly all fish species essentially depend on the availability of shallow, low flowing shoreline refuges for feeding and shelter as nurseries for successful recruitment. Freshwater fish larvae hatch at a total length of 2.7-9.5 mm and swim free at 6-15 mm. In this stage, their critical swimming speed is about 0.05-0.13 m/s (Wolter & Arling- haus 2003). Fish larvae emerge from the spawning substrate to the water column when the yolk sac absorption is almost complete (Bardonnet 2001). During emergence, the larvae be- come exposed to the flow and because of their low swimming abilities early larvae easily be- come entrained and transported by flow, leading to downstream displacement. A heterogeneous flow structure, especially the presence of low-transit zones and backwaters, controls this downstream displacement of fish and determines the availability of shelter and nursing habitats (Sukhodolov et al. 2009, Pander et al. 2017). Therefore, a complex mosaic of flow-protected habitats, gravel bars, large wood deposits, diverse sediment structures, and scour pools, is pivotal for maintaining diverse, self-recruiting, native fish assemblages in rivers (Jungwirth et al. 2000, Bardonnet 2001, Schiemer et al. 2003).

This study considers lithophilic, i.e. gravel spawning lampreys and fish species as the species most impacted by impoundments and thus, at the highest risk for this type of impairment from

727830 FIThydro - Deliverable 1.2 - Page 8 of 39 hydropower plants. Therefore the reproductive traits of European lampreys and fishes were compiled to align them with the sensitivity and mortality classification to the risk assessment.

In addition, an extensive literature search was performed to compile data on habitat quality of preferred spawning and nursery habitats, as well as on observed densities of eggs and larvae within these habitats. The main objective of this compilation is to provide guidance for mitigation measures in terms of habitat improvements and habitat provision to compensate for habitat losses in the impoundments as well as for adult fish mortality at the turbines. This compilation includes in particular critical abiotic reproduction parameters like substrate particle size, water depths and flow velocity in spawning and nursing habitats, but also considers species-specific spatial requirements of spawners, eggs and juveniles. The metrics used observed densities of eggs, juveniles and spawners in different habitats as well as count data like individual spawning site dimensions and egg numbers. However, the species-specific data reported were so highly variable and heterogeneous that consequently median values were used to calculate space- density-conversions.

The main objective of this compilation is providing guidance for mitigation measures in terms of habitat improvement and habitat provision to compensate for habitat losses in the impoundments as well as for adult fish mortality at the turbines.

3. Results 3.1 Sensitivity matrix Sensitivity scores have been assigned to a total of 165 fish and lamprey species occurring in European waters. Seventeen species did not belong to the native European fish fauna and were excluded from further risk analyses. The two ecotypes of trout Salmo trutta, the anadromous and the resident form were treated as the same species here.

Of all analysed taxa, 18 were classified of “highest sensitivity” (class 5), 29 of “high sensitivity” (class 4), 33 of “average sensitivity” (class 3), 46 of “moderate sensitivity” (class 2), and 22 of “low sensitivity” (class 1), see Table 1.

727830 FIThydro - Deliverable 1.2 - Page 9 of 39

Table 1: Sensitivity classification of 148 native European fish and lamprey species.

Sensitivity class Species – Common name Sensitivity class Species – Common name Acipenser gueldenstaedtii – Russian sturgeon Squalius pyrenaicus – Cacho

Acipenser naccarii – Adriatic sturgeon Squalius tenellus – Livno masnica Acipenser nudiventris – Ship sturgeon Squalius zrmanjae – Zrmanja chub

Average Average

Acipenser oxyrinchus – Baltic sturgeon Telestes souffia – Riffle dace

–

sensitivity

Acipenser ruthenus – Sterlet 3 Vimba vimba – Vimba

Acipenser stellatus – Stellate sturgeon Zingel streber – Streber Acipenser sturio – Atlantic sturgeon Achondrostoma arcasii – Bermejuela Alosa fallax – Twaite shad Achondrostoma occidentale – Western ruivaco Anguilla anguilla – European eel Achondrostoma oligolepis – Ruivaco Barbus barbus – Barbel Achondrostoma salmantinum – Sarda Chondrostoma nasus – Nase Alburnus alburnus – Bleak

Highest sensitivity Highest

- Coregonus maraena – Maraena whitefish Alosa immaculata – Pontic shad

5 Hucho hucho – Danube salmon Alosa tanaica – Azov shad Huso huso – Beluga Anaecypris hispanica – Jarabugo, Saramugo Leuciscus idus – Ide Babka gymnotrachelus – Racer goby Salmo labrax – Black Sea trout Barbatula barbatula – Stone loach Salmo salar – Atlantic salmon Carassius carassius – Crucian carp

Moderate sensitivity Moderate

Salmo trutta (anadromous) – Sea trout – Carassius gibelio – Prussian carp

Abramis brama – Common bream 2 Chondrostoma knerii – Dalmatian nase Alburnus chalcoides – Caspian shemaya Cobitis calderoni – Lamprehuela

Alburnus mento – Seelaube Cobitis elongata – Balkan spined loach

tivity Alosa alosa – Allis shad Cobitis elongatoides – Danube spined loach

High sensi- High

– Ballerus ballerus – Blue bream Cobitis narentana – Neretva spined loach

4 Ballerus sapa – Zobel Cobitis paludica – Colmillea

727830 FIThydro - Deliverable 1.2 - Page 10 of 39

Sensitivity class Species – Common name Sensitivity class Species – Common name

Barbus plebejus – Padanian barbel Cobitis taenia – Spined loach Blicca bjoerkna – Silver bream Cottus poecilopus – Siberian sculpin Chelon ramada – Thinlip mullet Gasterosteus aculeatus – Three-spined stickleback Cyprinus carpio – Common carp Gobio obtusirostris – Danube gudgeon Esox lucius – Pike Gymnocephalus baloni – Danube ruffe Lampetra fluviatilis – River lamprey Iberocypris alburnoides – Calandino, Bordalo Leuciscus aspius – Asp Lampetra planeri – Brook lamprey Lota lota – Burbot Lethenteron zanandreai – Po brook lamprey Luciobarbus bocagei – Iberian barbel Neogobius melanostomus – Round goby

Luciobarbus sclateri – Andalusian barbel Padogobius bonelli – Padanian goby

Mugil cephalus – Flathead mullet Petroleuciscus borysthenicus – Bobyretz chub Pelecus cultratus – Razorfish Ponticola kessleri – Bighead goby Petromyzon marinus – Sea lamprey Ponticola syrman – Syrman goby Platichthys flesus – Flounder Pungitius pungitius – Ten-spined stickleback Rutilus frisii – Vyrezub Rhodeus amarus – Bitterling Rutilus heckelii – Taran Romanogobio belingi – River gudgeon

High sensitivity High

– Rutilus meidingeri – Pearlfish Romanogobio uranoscopus – Stone gudgeon

Moderate sensitivity Moderate

4 Salmo obtusirostris – Soft-muzzled trout – Romanogobio vladykovi – Danube whitefin gudgeon

2 Salmo trutta (resident) – Brown trout Sabanejewia baltica – Golden loach Sander lucioperca – Pikeperch Sander volgensis – Volga pikeperch Thymallus thymallus – Grayling Scardinius plotizza – Neretvan rudd Tinca tinca – Tench Squalius carolitertii – Bordallo Zingel zingel – Zingel Squalius illyricus – Illyrian chub

727830 FIThydro - Deliverable 1.2 - Page 11 of 39

Sensitivity class Species – Common name Sensitivity class Species – Common name Barbus meridionalis – Mediterranean barbel Squalius janae – Istrian chub

Barbus tauricus – Crimea barbel Squalius microlepis – Imotski masnica Chelon aurata – Golden mullet Squalius svallize – Neretva chub

Moderate Moderate

Cottus gobio – Sculpin – Syngnathus abaster – Shore pipefish

sensitivity Gobio gobio – Gudgeon 2 Umbra krameri – Mudminnow Gymnocephalus cernua – Ruffe Alburnoides bipunctatus – Spirlin Gymnocephalus schraetser – Yellow pope Atherina boyeri –Sand smelt Leuciscus leuciscus – Dace Clupeonella cultriventris – Black Sea sprat Luciobarbus microcephalus – Small-head barbel Eudontomyzon danfordi – Carpathian brook lamprey

Mesogobius batrachocephalus – Knout goby Eudontomyzon mariae – Ukrainian brook lamprey Misgurnus fossilis – Weatherfish Eudontomyzon vladykovi – Danube brook lamprey Osmerus eperlanus –Smelt Gobio lozanoi – Iberian gudgeon Perca fluviatilis –Perch Iberochondrostoma almacai – Mira pardelha

Phoxinellus alepidotus – Naked minnow Iberochondrostoma lemmingii – Pardilla Phoxinellus dalmaticus – Cikola minnow Iberochondrostoma lusitanicum – Portuguese pardelha Phoxinus phoxinus –Minnow Knipowitschia longecaudata – Longtail dwarf goby

Average sensitivity

– Pseudochondrostoma duriense – Douro nase Leucaspius delineatus – Sunbleak

Pseudochondrostoma polylepis – Tagus nase Low sensitivity Neogobius fluviatilis – Monkey goby

Pseudochondrostoma willkommii – Guadiana nase – Pomatoschistus microps – Common goby

1 Rutilus pigus – Pigo Proterorhinus semilunaris – Tubenose goby Rutilus rutilus – Roach Romanichthys valsanicola – Asprete Rutilus virgo – Cactus roach Romanogobio albipinnatus – Whitefin gudgeon Sabanejewia romanica – Romanian golden loach Romanogobio benacensis – Italian gudgeon Salmo marmoratus – Marble trout Romanogobio kessleri – Sand gudgeon Scardinius erythrophthalmus – Rudd Sabanejewia balcanica – Balcan golden loach Silurus glanis – Wels Squalius aradensis – Arade chub Squalius cephalus – Chub Squalius torgalensis – Torgal chub

727830 FIThydro - Deliverable 1.2 - Page 12 of 39

3.2 Fish mortality at turbines The search for mortality data yielded in total 37 studies covering 101 different sites with a total of about 124,500 caught lampreys and fish from 42 species. From these studies, a total of 643 species-specific datasets (species * study) could be derived. The species that most frequently occurred were eel Anguilla anguilla (N= 73 datasets), trout (N= 58) and roach Rutilus rutilus (N= 36), while for most European species no empirical mortality data at hydropower plants were found.

Most of the studies used stow nets (N= 601 data sets) and investigated the natural downstream migration of fishes (N= 516 data sets). In total 127 datasets were obtained by fish injections in front of but not directly into the turbine.

According to the type of energy converter, the database includes 194 species-specific datasets for Kaplan turbines, 187 for Francis turbines, 119 for waterwheels, 107 for Archimedes screws, and 36 for other turbine types (Ossberger, Pelton). The observed mortalities ranged from 0- 23% for waterwheels, 0.02-23% for Archimedes screws, 0.9-47% for Kaplan turbines, and 1.1- 50% for Francis turbines. The complete database including the references is available at IGB and becomes publicly available at the project website after publication.

Figure 1 provides an overview on all empirically observed species-specific mortalities at hydropower turbines. A more detailed view of studies with sufficient sample sizes (at least five investigations with a minimum of ten species each) is illustrated in Fig. 2.

727830 FIThydro - Deliverable 1.2 - Page 13 of 39

Figure 1: Observed species-specific turbine mortality rates at hydropower plants. Grey areas = frequency of mortality rates, thin lines = 10-90% percentiles, thick bars = 25-75% percentiles, dots = median, rhomb = 75% percentile, colour codes according to the mortality class from very high in red to very low in green. Species with insufficient empirical data for mortality classification are shown in light grey.

727830 FIThydro - Deliverable 1.2 - Page 14 of 39

Figure 2: Robust empirical species-specific turbine mortality rates at hydropower plants based on a minimum of five studies with ten specimens each. Grey areas = frequency of mortality rates, thin lines = 10-90% percentiles, thick bars = 25-75% percentiles, dots = median, rhomb = 75% percentile, colour codes according to the mortality class from very high in red to very low in green.

727830 FIThydro - Deliverable 1.2 - Page 15 of 39

There was surprisingly little difference in mortality between the two main turbine types, while only waterwheels caused significantly lower mortality rates (Fig. 3).

Figure 3: Observed species-specific turbine mortality rates at hydropower plants depending on the turbine type. Grey areas = frequency of mortality rates, thin lines = 10-90% percentiles, thick bars = 25-75% percentiles.

Because the available data allowed an assessment of robust mortality rates for only 20 species, these datasets were used to derive more general mortality functions. Several previous studies provided significant evidence that fish mortality at hydropower turbines is primarily determined by fish length (reviewed e.g. by Monten & Hill 1985, Ebel 2013). Larger fish generally face a higher mortality risk. Correspondingly, all empirical and physical models of turbine mortality listed by Ebel (2013) contain fish length as a main predictor. The significant positive correlation between total length of a fish and probability of death during turbine passage was also evident in the database compiled here. Therefore, a regression model was set up to analyse the relation between mortality rate and fish length using the robust data set of species with sufficient sample size only. In particular, a generalized linear mixed model (GLMM) was used with

727830 FIThydro - Deliverable 1.2 Page 16 of 39 the primary predictor fish length as a fixed effect, and species, investigation, turbine type, author, study design, and sampling method as random effects: logit(M) = a + bL * Total length + {random effects}, with a= intercept, bL= length dependent model parameter, and total length = average species-specific length in the respective investigation.

The results show a correlation between fish length and turbine mortality of bL= 0.0478 independent of random effects, i.e. independent of the above-mentioned factors including turbine type. The probability of turbine-induced mortality is logarithmically correlated with fish length. During turbine passage the ratio of becoming lethally injured to the possibility to survive increases by the factor 0.0478 for each cm gain in total length.

The final statistical model logit(M) = -4.65+ 0.0478 * Total length of the correlation between fish length and mortality risk was used to calculate, which body length of fish is associated with a mortality rate corresponding to the five mortality classes from very high to very low (Table 2).

The body length (LM) associated with a mortality rate of 1%, 2%, 4% and 8% was calculated according to: LM= (logit(M) + 4.65)/0.0478

Table 2: Calculated length of fish facing a mortality risk during turbine passage according to the classes from very high to very low (in parentheses corresponding median mortality rates) based on empirical observations following LM= (logit(M)+4.65)/0.0478.

Very high (≥8%) Total fish length (TL) > 46.5 cm High (≥4%-<8%) 31.4 cm < TL < 46.5 cm Moderate (≥2%-<4%) 16.7 cm < TL < 31.4 cm Low (≥1%-<2%) 2.2 cm < TL < 16.7 cm Very low (<1%) TL < 2.2 cm

The length-mortality risk ratio displayed here allowed for the calculation of potential mortality risk of all species when passing a hydropower turbine based on length data. For most of the species classified for their sensitivity against mortality, no empirical observations were found or exist (compare Figs 1 & 2). Several reasons may explain this, from simply non-reporting, to the lack of investigations of turbine mortality in some regions, to the serious decline or even extirpation of species, which lowers their contemporary encounter probability with hydropower plants.

To account for potential future risks (e.g. after recolonization or rehabilitation), for all species a potential mortality risk during hydropower turbine passage was calculated based on their reported maximum length (Table 3).

727830 FIThydro - Deliverable 1.2 Page 17 of 39

Table 3: Species-specific potential mortality risk at hydropower plants during turbine passage based on reported total length and empirically determined length ranges for mortality classes according to Table 2. Risk classes based on median mortality rate thresholds; for common names refer to Table 1.

Mortality risk Species Mortality risk Species Abramis brama Alburnoides bipunctatus

Acipenser gueldenstaedtii Alburnus alburnus Acipenser naccarii Alosa tanaica Acipenser nudiventris Atherina boyeri Acipenser oxyrinchus Barbatula barbatula Acipenser ruthenus Barbus carpathicus Acipenser stellatus Barbus meridionalis Acipenser sturio Barbus petenyi Alosa alosa Chondrostoma knerii Alosa fallax Cottus gobio Anguilla anguilla Cottus poecilopus Ballerus sapa Eudontomyzon danfordi Barbus barbus Eudontomyzon mariae Barbus plebejus Eudontomyzon vladykovi Barbus tauricus Gobio gobio

Carassius carassius Gymnocephalus baloni Chelon aurata Gymnocephalus cernua

Chelon ramada Gymnocephalus schraetser Chondrostoma nasus Iberochondrostoma lemmingii

Moderate

Coregonus maraena – Iberocypris alburnoides

3 Cyprinus carpio Lampetra planeri

Very high

– Esox lucius Lethenteron zanandreai

5 Hucho hucho Misgurnus fossilis Huso huso Neogobius fluviatilis Lampetra fluviatilis Neogobius melanostomus Leuciscus aspius Parachondrostoma toxostoma Leuciscus idus Ponticola kessleri Lota lota Ponticola syrman Luciobarbus bocagei Rutilus aula Luciobarbus comizo Rutilus basak Mugil cephalus Squalius carolitertii Pelecus cultratus Squalius pyrenaicus Perca fluviatilis Syngnathus abaster Petromyzon marinus Telestes souffia Platichthys flesus Umbra krameri Rutilus frisii Zingel streber

Rutilus heckelii Achondrostoma arcasii Rutilus meidingeri Achondrostoma occidentale

Low

Rutilus pigus – Achondrostoma oligolepis Rutilus rutilus 2 Achondrostoma salmantinum

727830 FIThydro - Deliverable 1.2 Page 18 of 39

Mortality risk Species Mortality risk Species Salmo labrax Anaecypris hispanica Salmo marmoratus Babka gymnotrachelus Salmo obtusirostris Benthophiloides brauneri Salmo salar Benthophilus stellatus

Salmo trutta (anadromous) Clupeonella cultriventris Salmo trutta (resident) Cobitis calderoni Sander lucioperca Cobitis elongata Scardinius erythrophthalmus Cobitis elongatoides

Very high

– Silurus glanis Cobitis narentana

5 Squalius cephalus Cobitis paludica Thymallus thymallus Cobitis taenia Tinca tinca Gasterosteus aculeatus Vimba vimba Gobio lozanoi Zingel zingel Gobio obtusirostris Alburnus chalcoides Iberochondrostoma almacai Alburnus mento Iberochondrostoma lusitanicum Alosa immaculata Knipowitschia caucasica Ballerus ballerus Knipowitschia croatica Blicca bjoerkna Knipowitschia longecaudata Leuciscus leuciscus Knipowitschia panizzae

Luciobarbus microcephalus Knipowitschia radovici Luciobarbus sclateri Knipowitschia mrakovcici

Low

Mesogobius batrachocephalus – Leucaspius delineatus Osmerus eperlanus 2 Padogobius bonelli

Petroleuciscus borysthenicus Phoxinellus alepidotus

High

Pseudochondrostoma duriense Phoxinellus dalmaticus

–

4 Pseudochondrostoma polylepis Phoxinus phoxinus Pseudochondrostoma willkommii Pomatoschistus microps Rutilus virgo Proterorhinus semilunaris Sander volgensis Pungitius platygaster Scardinius plotizza Pungitius pungitius Squalius illyricus Rhodeus amarus Squalius janae Achondrostoma arcasii Squalius microlepis Romanogobio albipinnatus Squalius svallize Romanogobio belingi Squalius tenellus Romanogobio benacensis Squalius zrmanjae Romanogobio kessleri Romanogobio uranoscopus Romanogobio vladykovi Sabanejewia balcanica Sabanejewia baltica Sabanejewia romanica Squalius aradensis Squalius torgalensis

727830 FIThydro - Deliverable 1.2 Page 19 of 39

In contrast to the empirical observations, where several species passed turbines with a very low median mortality rate (<1%, dark green dots in Fig. 2), this mortality class is empty in the calculated probabilities (Table 3). Accordingly, the calculated potential mortalities during turbine passage based on maximum reported length might tend to overestimate the species-specific mortality, despite some significant exceptions for small-bodied species. For example, both bleak Alburnus alburnus and bitterling Rhodeus amarus were empirically found at very high mortality risk (Fig. 2), while their calculated potential risk was moderate and low, respectively (Table 3). Nonetheless, the calculated potential risks seem very helpful in environmental risk assessments to make predictions for species with insufficient data or in view on developmental goals, like stock enhancement and recolonization of species.

However, the following matrix used to identify species at risk was based on empirical observations using only the median of all observations for risk classification (Table 4).

727830 FIThydro - Deliverable 1.2 Page 20 of 39

Table 4: Species intrinsic sensitivity versus empirical species-specific mortality at hydropower turbine passage. Risk classification is based on median mortality rates; grey indicates species with insufficient number of observations (N< 5 studies with n< 10 specimens). Common names of species are shown in Table 3.

Sensitivity class Mortality risk during turbine passage Very high High Moderate Low Very low Highest Anguilla anguilla, Barbus barbus, Chondrostoma nasus, Leuciscus idus, Salmo salar, Salmo trutta (anadromous) High Abramis brama, Blicca Sander lucioperca Thymallus thymallus Cyprinus carpio, Esox lu- bjoerkna, Leuciscus as- cius, Lampetra fluviatilis, pius, Salmo trutta (resi- Tinca tinca dent) Average Perca fluviatilis, Rutilus Rutilus rutilus, Silurus Phoxinus phoxinus Cottus gobio, Gobio go- virgo glanis, bio, Gymnocephalus cernua, Leuciscus leuciscus, Scardinius erythrophthalmus, Squalius cephalus Moderate Alburnus alburnus, Barbatula barbatula, Carassius gibelio, Carassius carassius, Gas- Rhodeus amarus terosteus aculeatus, Lam- petra planeri, Pungitius pungitius Low Leucaspius delineatus, Romanogobio albipinnatus

727830 FIThydro - Deliverable 1.2 Page 21 of 39

The overall species at risk classification revealed that several species of high and highest sensitivity against mortality face a very high mortality risk when passing turbines. Among them are the anadromous salmonids, but also typical riverine fish species like asp, barbel, ide, and nase. These species should be primarily targeted by mitigation measures.

3.3 Habitat requirements Thanks to the sentinel works by Balon (1975, 1981) and previous European projects like EFI+ (Developing a European Fish Index, at http://efi-plus.boku.ac.at/), nearly all European lampreys and fish species have been classified for their spawning guilds already. However, it must be noted that the classification as gravel spawner merges the various species-specific gravel requirements. The commonality between all lithophilic fish is that they lay their eggs on or in gravel and that their larvae are benthic, in the interstitial spaces after hatching. In total 76 of the native European species classified here for intrinsic sensitivity against mortality belong to the guild of lithophilic spawners and are thus, potentially affected by impoundments. The distribution of lithophilic species between the sensitivity classes is very similar, with 16, 16, 20, 12, and 8 species in the class of highest, high, average, moderate, and low sensitivity, respectively. Eighteen species-specific reproduction traits, spawning and nursing habitat parameters and carrying capacities for eggs juveniles and spawners were surveyed (Table 5). Only for 38 species, we obtained data for five or more of the 18 parameters. Information for all other species was too scarce and too scattered to be analysed in a meaningful way. The latter species are not further discussed here. An overview of available data provides Table 1Table 6. Table 5: Overview of reproduction traits and abiotic environmental conditions in spawning and nursing habitats that were used to explore species-specific compensation measures.

Trait label Trait description A Relative fecundity B Fecundity at maturity C Spawning events per female D Spawning site dimensions E Spawner densities F Spawning site densities G Substrate size spawning habitat H Current velocity spawning habitats I Water depth spawning habitats J Water temperature spawning K Egg densities L Depth of eggs in interstitial space M Day-degrees until hatch N Hatching rates O Substrate size juvenile habitats P Current velocity juvenile habitats Q Water depth juvenile habitats R Juvenile densities

727830 FIThydro - Deliverable 1.2 Page 22 of 39

Table 6: Metadata of available information on environmental conditions in spawning and nursing habitats as well as densities of eggs, juvenile, and spawners. Data availability refers to the number of independent observations for each trait (subscript value). Trait letters according to Table 5; in red = no data. The data quality indicates the average number of independent observations across all traits.

No. Scientific name Common name Data availability Data quality

1 Acipenser baeri Siberian sturgeon A8, B1, C1, D, E1, F, G2, H2, I4, J3, K2, L, M2, N3, O, P, Q, R 2.54

2 Acipenser gueldenstaedtii Russian sturgeon A1, B1, C1, D, E1, F, G2, H2, I3, J3, K2, L, M1, N2, O, P, Q, R 2.73

3 Acipenser nacari Adriatic sturgeon A1, B1, C1, D, E1, F, G2, H3, I3, J3, K2, L, M2, N2, O, P, Q, R 1.91

4 Acipenser nudiventris Ship sturgeon A1, B1, C1, D, E1, F, G2, H2, I3, J3, K2, L, M1, N2, O, P, Q, R 2.00

5 Acipenser oxyrinchus Baltic sturgeon A1, B1, C1, D, E1, F, G3, H3, I4, J3, K2, L, M1, N2, O, P, Q, R 1.82

6 Acipenser ruthenus Sterlet A1, B1, C1, D, E1, F, G3, H3, I4, J4, K2, L, M1, N2, O, P, Q, R 2.09

7 Acipenser stellatus Stellate sturgeon A2, B1, C1, D, E1, F, G3, H3, I4, J2, K2, L, M1, N2, O, P, Q, R 2.00

8 Acipenser sturio Atlantic sturgeon A2, B1, C1, D, E1, F, G3, H3, I3, J3, K2, L, M1, N2, O, P, Q, R 2.00

9 Alburnoides bipunctatus Spirlin A1, B1, C1, D, E, F, G4, H4, I, J4, K, L, M2, N, O, P, Q, R 2.43

10 Barbus barbus Common barbel A2, B1, C3, D4, E2, F, G14, H16, I15, J13, K1, L, M8, N2, O1, P13, Q15, R6 7.25

11 Barbus bocagei Iberian barbel A1, B1, C1, D, E, F, G, H, I1, J1, K, L, M, N, O, P, Q, R 1.00

12 Chalcalburnus chalcoides Caspian shemaya A1, B1, C, D, E, F, G, H2, I2, J4, K, L, M1, N1, O, P, Q, R 1.71

13 Chondrostoma nasus Nase A1, B1, C2, D5, E10, F, G30, H19, I26, J18, K2, L3, M10, N3, O2, P16, Q11, R2 9.53

14 Cobitis taenia Spined loach A4, B1, C1, D, E, F, G, H, I, J3, K, L, M1, N1, O, P, Q, R 1.83

15 Coregonus maraena Maraena whitefish A3, B1, C, D, E, F, G, H, I, J2, K, L, M1, N, O, P, Q, R 1.75

16 Cottus gobio Sculpin A3, B1, C4, D, E1, F, G7, H11, I4, J3, K, L, M7, N, O9, P8, Q4, R 5.17

17 Eudontomyzon vladykovi Danube brook lamprey A1, B1, C, D, E, F, G1, H1, I1, J1, K, L, M, N, O, P, Q, R 1.00

18 Gobio gobio Gudgeon A1, B1, C, D, E, F, G1, H2, I1, J1, K1, L, M, N, O, P3, Q3, R6 1.14

19 Hucho hucho Danube salmon A4, B1, C, D10, E, F5, G3, H7, I7, J10, K, L4, M14, N4, O, P, Q, R 6.27

20 Huso huso Beluga A1, B1, C1, D, E1, F, G2, H3, I4, J3, K2, L, M1, N2, O, P, Q, R 1.91

21 Lampetra fluviatilis River lamprey A2, B1, C, D3, E, F, G, H1, I2, J1, K1, L, M, N2, O6, P3, Q, R1 2.09

22 Lampetra planeri Brook lamprey A1, B1, C1, D4, E, F, G3, H1, I2, J2, K, L, M1, N, O1, P3, Q, R 1.82

23 Leuciscus aspius Asp A1, B1, C, D, E, F, G1, H1, I, J15, K1, L, M4, N, O, P2, Q2, R10 3.80

727830 FIThydro - Deliverable 1.2 Page 23 of 39

No. Scientific name Common name Data availability Data quality

24 Leuciscus borysthenicus Bobyretz chub A1, B1, C1, D, E, F, G, H, I1, J1, K, L, M, N, O, P, Q, R 1.00

25 Leuciscus idus Ide A1, B1, C, D, E, F, G, H, I, J5, K, L, M, N2, O, P1, Q1, R4 2.14

26 Leuciscus Leuciscus Dace A3, B1, C, D, E, F, G2, H5, I5, J11, K1, L, M14, N7, O, P11, Q10, R8 6.50

27 Lota lota Burbot A1, B1, C, D, E, F, G1, H, I1, J3, K, L, M1, N1, O, P, Q1, R 1.25

28 Osmerus eperlanus Smelt A1, B1, C, D, E, F, G2, H2, I3, J4, K, L2, M2, N, O, P, Q, R 2.12

29 Petromyzon marinus Sea lamprey A3, B1, C, D3, E, F1, G3, H1, I3, J2, K, L, M, N, O1, P, Q, R1 1.90

30 Phoxinus phoxinus Minnow A2, B1, C4, D, E1, F, G14, H7, I3, J12, K, L3, M4, N1, O1, P3, Q7, R 4.93

31 Rutilus frisii Vyrezub A1, B1, C, D, E, F, G, H, I, J1, K, L, M2, N2, O, P, Q, R 1.40

32 Salmo salar Atlantic salmon A9, B1, C5, D13, E1, F, G25, H18, I19, J4, K3, L9, M5, N5, O6, P16, Q16, R5 9.41

33 Salmo trutta Brown trout, sea trout A2, B1, C1, D18, E, F5, G21, H13, I12, J4, K4, L14, M2, N16, O, P11, Q8, R4 8.50

34 Squalius cephalus Chub A2, B1, C1, D, E, F, G3, H8, I8, J11, K, L, M3, N1, O1, P12, Q11, R10 5.54

35 Telestes souffia Riffle dace A1, B1, C1, D, E, F, G3, H4, I1, J3, K, L1, M3, N, O, P, Q, R 2.00

36 Thymallus thymallus Grayling A2, B1, C, D, E2, F, G15, H19, I11, J10, K, L5, M12, N3, O2, P9, Q17, R1 7.79

37 Vimba vimba Vimba A3, B1, C, D, E, F, G, H1, I14, J6, K, L, M2, N2, O1, P, Q1, R 3.44

38 Zingel streber Streber A2, B1, C, D, E, F, G, H, I, J2, K, L, M2, N, O, P1, Q1, R 1.50

727830 FIThydro - Deliverable 1.2 Page 24 of 39

Data availability and quality varied strongly between species and traits: Common species, ubiq- uitous species or species with high economic value are comparably well studied in contrast to small-bodied or endemic species with constrained geographic dispersal ranges. Further, for the well-studied species, there are usually several independent studies and data sources available, while information on rare species is often based on a single source.

A general lack of data was determined for the dimensions of individual spawning sites (trait label D) and densities of spawning sites in a spawning area (F), the densities of eggs (K), the depth of eggs in the interstitial space (L) and characteristics of juvenile habitats like substrate particle size (O), current velocity (P), water depth (Q), and average densities of juveniles in their nursing grounds (R) (compare Table 6). Per trait between 1 and 10 independent observations were found. It must be noted that minimum and maximum values reported by a study were counted as one observation, whereas two values from independent sampling periods or sites were treated as two observations.

It has been already described above, why especially gravel spawning species experience the most significant habitat loss due to impoundments (e.g., Mueller et al. 2011, Pander & Geist 2018). Suitable spawning gravel beds have grain sizes between 4 mm and 69 mm (Kondolf & Wolman 1993, Mann 1996, Kondolf 2000, Figure 4) and very low (<2%) contents of fine materials (<1 mm grain size) (Jensen et al. 2009).

Figure 4: Observed range of substrate particle sizes in spawning grounds (light blue) and juvenile habitats (orange).

727830 FIThydro - Deliverable 1.2 Page 25 of 39

Accordingly, their maintenance requires a certain stream power that flushes finer materials away. The burial depth of eggs in the gravel ranges between 3 cm and 30 cm, sometimes up to more than 50 cm (e.g., DeVries 1997, Duerregger et al. 2018, Figure 5). Thus, it is not surprising that the interstitial water quality plays a key role for successful egg development in lithophilic (gravel spawning) species (e.g., Sternecker et al. 2013).

Figure 5: Observed depths of eggs in the interstitial space. Species whose eggs only stick to stones on the surface of the stream bed weren't included.

The availability of oxygen in the interstitial spaces depends on the exchange of channel water with the gravel riverbed. The hydro-morphologic processes driving this exchange include bed permeability and surface roughness effects, while the flux of oxygenated water through riverbed gravels is controlled by gravel permeability, coupling of surface-subsurface flow and oxygen demands imposed by materials infiltrating riverbed gravels (Greig et al. 2007). The pro- portion of fine sediments is inversely correlated to permeability and therefore, detrimental to the survival of eggs and embryos in the interstitial spaces.

However, for a successful reproduction, it is also critical that the required environmental conditions in spawning grounds in terms of oxygen supply and flow velocities are maintained throughout the whole egg incubation period. The latter is species-specific and temperature dependent and can last for more than 400 degree days until the larvae hatch (Figure 6). Short-

727830 FIThydro - Deliverable 1.2 Page 26 of 39 term water level alterations and fluctuations caused by hydropower operation regimes like hy- dropeaking or water abstraction in residual water stretches may further threaten the reproductive success of these species. This is also relevant for the spatial arrangement and placing of habitat rehabilitation measures.

Figure 6: Observed incubation times in degree days (days * water temperature) for egg development of gravel-spawning fish species.

Preferred flow velocities at the spawning sites of lithophilic fish range between 0 m/s and 2 m/s, but are usually well below 0.8 m/s. Preferred flow velocities in the juvenile habitats are even substantially lower (Figure 7). Given these preferred flow velocities at spawning and nursing sites, the reduced flow velocities due to impoundments do not directly impact on fish recruitment. However, as mentioned above spawning and incubation of fish is adjusted to the natural flow regime, in particular to the periods of more stable flows. Nonetheless, the maintenance of the spawning gravel quality requires also the flow peaks of the natural flow regime for flushing fines and organic deposits. Impoundments permanently reduce the flow peaks and stream power of the river, which supports the sedimentation of fine materials and thus the permanent deterioration of the quality of spawning gravel substrates and juvenile habitats. Accordingly, species which entirely depend on high-quality gravel substrates for successful recruitment, suffer most from the impact of impoundments on river hydromorphology. Therefore, the species

727830 FIThydro - Deliverable 1.2 Page 27 of 39 at risk classification presented here focused on lithophilic fish and renders this spawning guild as an appropriate indicator of impacts from impoundment effects.

Besides these impacts on hydromorphological processes, impoundments usually have steep bank slopes and increased water depth. Accordingly, they are very limited in shallow (less than one meter deep), slow-flowing bank habitats, which are essential for larvae and juvenile recruitment (Pander et al. 2017, Figure 8).

Figure 7: Observed range of preferred surface flow velocities at spawning grounds (light blue) and juvenile habitats (orange).

727830 FIThydro - Deliverable 1.2 Page 28 of 39

Figure 8: Observed range of preferred water depths at spawning grounds (light blue) and juvenile habitats (orange).

Besides the displayed suite of abiotic conditions required for successful reproduction, the spatial dimension and arrangement of suitable habitats as well as the effective population of spawners determine recruitment and have to be considered as boundary conditions for mitigation by habitat improvement or rehabilitation.

The size of an individual spawning site has been reported for very few species only (Figure 9) and is best documented for Atlantic salmon and trout, for which the size of an individual red can be rather easily measured. For those gravel spawners who do not dig redds estimates of the abundance of spawners and the total dimension of the spawning area might be used to determine individual spatial requirements for spawning. Reported median sizes of individual spawning spots went up to 7 m² per female but ranged mostly below one meter (Figure 9). Information on the individual size of a spawning spot serves habitat rehabilitation in two ways: based on the average fecundity of a species, quantitative benchmarks for desired egg production and larvae recruitment can be derived for mitigation measures providing spawning gravel. At the same time, such information serves deriving minimum sizes of gravel provision measures to avoid superimposition of redds due to space limitations resulting in increased egg and larvae mortality.

727830 FIThydro - Deliverable 1.2 Page 29 of 39

Figure 9 Reported ranges of individual spawning sites.

The search for information on effective spawner populations at spawning sites to derive spawner densities yielded even less results than the individual spatial requirements for spawning (Figure 10). Despite some correspondence between the reported spawning spot size of a female (Figure 9) and the spawner density (Figure 10) especially for barbel and nase, other data lack plausibility. For example, the reported spawner density for the minnow is surprisingly low (Figure 10).

727830 FIThydro - Deliverable 1.2 Page 30 of 39

Figure 10: Reported ranges of spawner densities at gravel spawning grounds.

In combination with species-specific fecundity parameters and hatching rates, these data can provide an estimate of the number of spawners and the size of spawning and nursing grounds needed to achieve rehabilitation targets for selected riverine fish species or to compensate for fish mortality at hydropower plants. The latter has been addressed here to derive some rough guidance for mitigation measures in form of additional habitat provision and rehabilitation.

To illustrate this compensation idea, a hypothetical scenario is calculated for resident Salmo trutta (brown trout) based on the reported data given in Table 7. In its simplest form, the number of larvae produced by a square meter of spawning gravel is estimated from the reported median egg density of 4.355 eggs per square meter of spawning gravel. At a typical hatching rate of 88% one might expect 3.83 (4.355 * 0.88) larvae produced on a square meter of spawning gravel. This number dramatically drops at emergence from the substrate (50 % mortality) and becomes further reduced to much less than one larvae per m² spawning gravel at the end of the year.

However, this rough estimate does not consider the spawning requirements of an individual female. The absolute fecundity of an average brown trout is roughly 1300 eggs, which are released in on average 5.7 spawning events with roughly 228 eggs each. The reported median size of a brown trout redd is 0.75 m² (Table 7). Assuming one spawning event per redd, a single female spawns in 5-6 redds (accounting for 304 eggs per m² redd area). Median reported densities of brown trout redds are 0.05 per m², which means 20 m² of suitable gravel bars contain a single redd (which would result in a substantially higher average number of 15 eggs/m² gravel compared to 3.83 eggs/m² used above). If one female releases there eggs into 5-6 redds at that spatial arrangement, the spawning gravel requirements of a single female would amount to 114 m² (20 * 5.7). While the final egg densities roughly correspond to the basic estimate, the

727830 FIThydro - Deliverable 1.2 Page 31 of 39 total habitat requirements are much higher when using the spawner approach for modelling. The amount of habitat provision has to be multiplied by the desired spawner population to avoid additional mortality due to redd superimposition, a common phenomenon in habitats in which the size of available spawning area is insufficient for the number of spawners.

Table 7: Overview of reported traits for Salmo trutta. Data points shown here are median values.

Trait Median value Relative fecundity (eggs/g body weight) 2 Fecundity at maturity 1299 Spawning events per female 5.7 Spawning site dimensions (m²) 0.75 Spawner densities - Spawning site densities 0.05 Substrate size spawning habitat (cm) 2.4 Flow velocity spawning habitats (m/s) 0.4 Water depth spawning habitats (m) 0.5 Water temperature spawning 8.75 Egg densities (eggs/m² spawning gravel) 4.35 Depth of eggs in interstitial space (cm) 16.3 Day-degrees until hatch 269.5 Hatching rates (%) 88 Substrate size juvenile habitats (cm) - Current velocity juvenile habitats (m/s) 0.1 Water depth juvenile habitats (m) 0.3 Juvenile densities 2

However, such estimates can just serve as a rough guidance, because they have significant uncertainties. Here we must first mention the wide lack of basic autecological data for most fish species and significant inconsistencies between available studies and datasets. Most of the variation results from regional differences between river systems in terms of food supply, habitat complexity, temperature, climate, hydrologic regime, regional species pools and biotic in- teractions (e.g. predation), which affect the study results. Further, studies from sites with different levels of human alteration will lead to underestimates of the natural reproduction potential. Last not least, individual fecundity is well known to be highly determined by maternal effects with larger and older females having more eggs of better quality. All these potential sources of uncertainty could not be accounted for here, due to the general lack of studies.

Furthermore, even at the same site spawning gravel differs in quality depending on the position at the gravel bar, to the main flow, and to the bank, which all influences the water flow through the interstitial space and the accumulation of fines. Therefore, densities of eggs and larvae as well as emergence from the substrate and mortality are not evenly distributed all over the spawning gravel. Usually, deeper and edge areas of gravel bars have a much lower offspring productivity.

727830 FIThydro - Deliverable 1.2 Page 32 of 39

Next to spawning gravel, shallow, slow-flowing nursing areas are required for the survival and growth of the juveniles in close proximity. There is no habitat choice of fish larvae emerging from the substrate due to their low swimming abilities (Wolter & Sukhodolov 2008). When the larvae swim up to the water surface to fill their swim bladder, the typical flow will displace them and it remains a function of habitat complexity and the availability of slow flowing littoral areas, how quick they will end up in suitable nursing habitats and start feeding (Sukhodolov et al. 2009). The time being displaced and prevented from feeding is directly related to larvae mortality.

Reported densities of juveniles in nursing areas are illustrated in Figure 11. According the data reported the juvenile densities correspond well to the expected larvae densities after initial mortality during emergence. That means that available nursing areas have to be of similar size to suitable spawning areas. For mitigation measures it therefore has to be considered that accompanying with gravel provision also similar amounts of nursing habitats for juveniles have to be provided. Here it is essential to realise the functional connectivity between spawning and nursing areas; otherwise habitat improvements addressing fish recruitment will fail.

Figure 11: Reported juvenile densities in nursing grounds. Not shown are reported juvenile densities of asp (Leuciscus aspius; min=0.002, median=10, max=122) and chub (Squalius cephalus, min=0.75, median=57.2, max=586).

However, it must be noted that several studies of juvenile densities come from large rivers, which suffer from significant habitat deficits especially in the littoral zone. Therefore, the lower estimates might reflect heavily degraded conditions rather than usual densities of fish larvae.

3.4 Synthesis Beside the natural sensitivity of species against mortality, there are three typical impacts of hydropower operation in general: 1) migration barrier, 2) mortality, and 3) habitat degradation.

727830 FIThydro - Deliverable 1.2 Page 33 of 39

Migration barriers affect primarily diadromous fish, which migrate between freshwaters and the sea and thus must pass a barrier to complete their life cycle. These species are considered most sensitive against the barrier effect. The data compilation for the sensitivity analysis com- prises 22 diadromous lampreys and fish species.

The mortality was empirically derived from turbine passage studies. Data were gathered for 42 species in total, 20 of them with sufficient sample size (N< 5 studies with ≥10 specimens). The data could be used to derive a model to assess the length dependent potential mortality risk for data deficient species. However, here only the empirical data for the 36 species occurring in more than one study with more than one specimen were used for identifying species at risk (Table 4).

The habitat degradation and habitat loss in the impoundment particularly affects lithophilic, which represent a total of 76 species. For 38 of the species more than average data were collated on habitat requirements and carrying capacity for eggs, juveniles and spawners to guide mitigation measures addressing habitat provision and rehabilitation to enhance recruitment.

In detail, there are four factors determining the risk of a fish species from hydropower operation: i) high sensitivity to natural mortality, ii) being diadromous, iii) lithophilic, and iv) experiencing a high empirical mortality during turbine passage. Accordingly, species were classified as very high risk from hydropower operation, if at least three of the four following conditions were fulfilled: i) belonging to the high or highest sensitivity class (Table 1), ii) having high or highest mortality risk during turbine passage (Table 3), iii) being diadromous, and iv) being lithophilic.

Species were classified as high risk in hydropower environments, if two of the conditions mentioned above were fulfilled, and they were classified as lower risk if only one or none of the four conditions were fulfilled.

A comprehensive list of native European lamprey and fish species of high and very high risk from hydropower operation is presented in Table 8.

However, it has to be mentioned that this risk classification is a first attempt, which needs further evidencing and practical testing. It is a rather conservative approach, well aware of the tremendous variability and high uncertainties in the underlying data. There is a surprising lack of basic autecological knowledge for most of the European lampreys and fish species leading to a compilation of all available data over wide regions, climatic zones, river types and also study objectives, which resulted in substantial heterogeneity of the database. We accounted for this variability by using median and percentile values only for further estimates.

The same holds true for the empirical mortality rates of fishes at hydropower turbines obtained. Sufficient data were gathered for 20 species subjected to various random effects due to turbine type, study design, species considered, handling of fish and several more. We accounted for variability by using very conservative estimates based on empirical data only and using median and 75% percentile of the mortality rates for classification. Further data and future research will improve the classification.

However, so far the risk classification reflects the state of knowledge as not better and more comprehensive system is available.

727830 FIThydro - Deliverable 1.2 Page 34 of 39

Table 8: Summary of native European lamprey and fish species of high and very high risk during hydropower operation.

Very high risk High risk Acipenser gueldenstaedtii – Russian sturgeon Abramis brama – Common bream Acipenser naccarii – Adriatic sturgeon Acipenser ruthenus – Sterlet Acipenser nudiventris – Ship sturgeon Alburnus mento – Seelaube Acipenser oxyrinchus – Atlantic sturgeon Alosa immaculata – Pontic shad Acipenser stellatus – Stellate sturgeon Alosa tanaica – Azov shad Acipenser sturio – Common sturgeon Ballerus ballerus – Blue bream Alburnus chalcoides – Caspian shemaya Ballerus sapa – Zobel Alosa alosa – Allis shad Barbus plebejus – Padanian barbel Alosa fallax – Twaite shad Blicca bjoerkna – Silver bream Anguilla anguilla – European eel Chelon ramada – Thinlip mullet Barbus barbus – Barbel Hucho hucho – Danube salmon Chondrostoma nasus – Nase Leuciscus idus – Ide Coregonus maraena – Maraena whitefish Lota lota – Burbot Huso huso – Beluga Luciobarbus bocagei – Iberian barbel Lampetra fluviatilis – River lamprey Luciobarbus sclateri – Andalusian barbel Leuciscus aspius – Asp Mugil cephalus – Flathead mullet Petromyzon marinus – Sea lamprey Osmerus eperlanus – Smelt Salmo labrax – Black Sea trout Platichthys flesus – Flounder Salmo salar – Atlantic salmon Rutilus frisii – Vyrezub Salmo trutta (anadromous) – Sea trout Rutilus heckelii – Taran Salmo trutta (resident) – Brown trout Rutilus meideringi – Pearlfish Salmo obtusirostris – Soft-muzzled trout Sander lucioperca – Pikeperch Thymallus thymallus – Grayling Zingel zingel – Zingel

727830 FIThydro - Deliverable 1.2 Page 35 of 39

4. Conclusions While most of the sturgeon species may potentially be selected for recovery or renewal in the future and are envisaged in river rehabilitation and species conservation programs, several species at very high or high risk from hydropower operation would greatly benefit from environmental assessments and mitigation measures. In particular, the large bodied rheophilic cyprinids, such as barbel, nase, asp or ide, are also targeted by the Water Framework Directive (WFD), which aims to achieve the good ecological status of rivers. Accounting for their risk during hydropower operation by mitigation measures at the same time contributes to achieving the environmental targets of the WFD and thus, creates win-win situations.

The species identified here most at risk from hydropower operation do not only warrant environmental impact assessment, but they also help in designing targeted successful rehabilitation and mitigation measures.

727830 FIThydro - Deliverable 1.2 Page 36 of 39

5. References Balon EK. 1975. Reproductive guilts of fishes: A proposal and definition. Journal of the Fisher- ies Research Board of Canada 32, 821-864.

Balon EK. 1981. Additions and amendments to the classification of reproductive styles in fishes. Environmental Biology of Fishes 6, 377-389.

Bardonnet A. 2001. Spawning in swift water currents: implications for eggs and larvae. Archiv für Hydrobiologie Suppl. 135, 271-291.

DeVries P. 1997. Riverine salmonid egg burial depths: review of published data and implications for scour studies. Canadian Journal of Fisheries and Aquatic Sciences 54, 1685-1698.

Duerregger A, Pander J, Palt M, Mueller M, Nagel C, Geist J. 2018. The importance of stream interstitial conditions for the early life stage development of the European nase (Chondrostoma nasus L.). Ecology of Freshwater Fish 27, 920-932.

Ebel G. 2013. Fischschutz und Fischabstieg an Wasserkraftanlagen – Handbuch Rechen- und Bypasssysteme. Büro für Gewässerökologie und Fischereibiologie 4.

Edler C, Diestelhorst O, Kock M. 2011. Untersuchungen zur Abwanderung und Schädigung von Fischen an der Wasserkraftschnecke Rhede-Krechting (Bocholter Aa, Kreis Borken) im Sommer und Herbst 2010. Landesfischereiverband Westfalen und Lippe e.V.

Froese R, Pauly D. Eds. 2018. FishBase. World Wide Web electronic publication. www.fishbase.org, version (06/2018).

Garcia de Jalón D, Alonso C, González del Tango M, Martinez V, Gurnell A, Lorenz S, Wolter C, Rinaldi M, Belletti B, Mosselman E, Hendriks D, Geerling G. 2013. Review on pressure effects on hydromorphological variables and ecologically relevant processes. Project Report REFORM D1.2 (http://www.reformrivers.eu/deliverables/d1-2).

Greig SM, Sear DA, Carling PA. 2007. A review of factors influencing the availability of dis- solved oxygen to incubating salmonid embryos. Hydrological Processes 21, 323-334.

Hadderingh RH, Bakker HD. 1998. Fish Mortality due to Passage through Hydroelectric Power Stations on the Meuse and Vecht Rivers. In: Jungwirth M, Schmutz S, Weiss S (Eds). Fish Migration and Fish Bypasses, Fishing News Books, Blackwell Science Ltd., Oxford, 315-330.

Havn TB, Økland F, Teichert MAK, Heermann L, Borcherding J, Sæther SA, Tambets M, Diserud OH, Thorstad EB. 2017. Movements of dead fish in rivers. Animal Biotelemetry 5, 7.

Havn TB, Thorstad EB, Teichert MAK, Sæther SA, Heermann L, Hedger RD, Tambets M, Diserud OH, Borcherding J, Økland F. 2018. Hydropower-related mortality and behaviour of Atlantic salmon smolts in the River Sieg, a German tributary to the Rhine. Hydrobiologia 805, 273-290.

Holzner M. 2000. Untersuchungen über die Schädigung von Fischen bei der Passage des Mainkraftwerks Dettelbach. PhD Thesis, Technical University Munich.

727830 FIThydro - Deliverable 1.2 Page 37 of 39

Jensen DW, Steel EA, Fullerton AH, Pess GR. 2009. Impact of fine sediment on egg-to-fry survival of Pacific salmon: a meta-analysis of published studies. Reviews in Fisheries Science 17, 348-359.

Jungwirth M, Muhar S, Schmutz S. 2000. Fundamentals of fish ecological integrity and their relation to the extended serial discontinuity concept. Hydrobiologia 422/423, 85-97.

Kondolf GM. 2000. Assessing salmonid spawning gravel quality. Transactions of the American Fisheries Society 129, 262-281.

Kondolf GM, Wolman MG. 1993. The sizes of salmonid spawning gravels. Water Resources Research 29, 2275-2285.

Mann RHK. 1996. Environmental requirements of European non-salmonid fish in rivers. Hy- drobiologia 323, 223-235.

Monté n E, Hill H. 1985. Fish and turbines: fish injuries during passage through power station turbines. Vattenfall, Stockholm.

Mueller M, Pander J, Geist J. 2011. The effects of weirs on structural stream habitat and biological communities. Journal of Applied Ecology 48, 1450-1461.

Mueller M, Pander J, Geist J. 2017. Evaluation of external fish injury caused by hydropower plants based on a novel field-based protocol. Fisheries Management and Ecology 24, 240-255.

Pander J, Geist J. 2010. Salmonid egg floating boxes as bioindication for riverine water quality and stocking success. Journal of Fish Biology 76, 2584-2590.

Pander J, Geist J. 2018. The contribution of different restored habitats to fish diversity and population development in a highly modified river: A case study from the River Günz. Water 10, 1202.

Pander J, Mueller M, Knott J, Egg L, Geist J. 2017. Is it worth the money? The functionality of engineered shallow stream banks as habitat for juvenile fishes in heavily modified water bodies. River Research and Applications 33, 63-72.

Pander J, Mueller M, Knott J, Geist J (2018). Catch-related fish injury and catch efficiency of stow-net-based fish recovery installations for fish-monitoring at hydropower plants. Fisheries Management and Ecology, 1–13. DOI: 10.1111/fme.12263Schiemer F, Keckeis H, Kamler E. 2003. The early life history stages of riverine fish: ecophysiological and environmental bottle- necks. Comparative Biochemistry and Physiology A 133, 439–449.

Schmalz W. 2010. Untersuchungen zum Fischabstieg und Kontrolle möglicher Fischschäden durch die Wasserkraftschnecke an der Wasserkraftanlage Walkmühle an der Werra in Meinin- gen. FLUSS - Fischökologische und Limnologische Untersuchungsstelle Südthüringen.

Schmalz W. 2016. Untersuchungen zu Fischschäden beim Fischabstieg über ein Wasserrad an der Schlossmühle in Reurieth. FLUSS - Fischökologische und Limnologische Untersu- chungsstelle Südthüringen.

727830 FIThydro - Deliverable 1.2 Page 38 of 39

Schneider J, Hübner D, Korte E. 2012. Funktionskontrolle der Fischaufstiegs-und Fischab- stiegshilfen sowie Erfassung der Mortalität bei Turbinendurchgang an der Wasserkraftanlage Kostheim am Main. Bürogemeinschaft für Fisch- & Gewässerökologische Studien.

Sternecker K, Cowley DE, Geist J. 2013. Factors influencing the success of salmonid egg development in river substratum. Ecology of Freshwater Fish 22, 322-333.

Sukhodolov A, Bertoldi W, Wolter C, Surian N, Tubino M. 2009. Implications of channel processes for juvenile fish habitats in Alpine rivers. Aquatic Sciences 71, 338-349.

Thorstad EB, Havn TB, Sæther SA, Heermann L, Teichert MAK, Diserud OH, Tambets M, Borcherding J,.Økland F. 2017. Survival and behaviour of Atlantic salmon smolts passing a run-of-river hydropower facility with a movable bulb turbine. Fisheries Management and Ecol- ogy 24, 199-207. van Treeck R, Van Wichelen J, Coeck J, Vandamme L, Wolter C. 2017. D1.1 Metadata overview on fish response to disturbance. Project Report FIThydro D1.1 (https://www.fithydro.eu/deliverables-tech/D1.1)

Wagner F, Royan M, Mirza J, Schubert A. 2013a. Überprüfung der Fischschutz- und Fischab- stiegseinrichtungen sowie der Fischschädigung an der WKA Lützschena/Weiße Elster. Institut für Gewässerökologie und Fischereibiologie Jena.

Wagner F, Royan M, Mirza J, Schubert A. 2013b. Überprüfung der Fischschutz- und Fischab- stiegseinrichtungen sowie der Fischschädigung an der WKA Wehlitz/Weiße Elster. Institut für Gewässerökologie und Fischereibiologie Jena.

Wagner F, Schmalz W, Royan M, Mirza J, Schubert A. 2013c. Überprüfung der Fischschutz- und Fischabstiegseinrichtungen sowie der Fischschädigung an der WKA Stahmeln/Weiße Els- ter. Institut für Gewässerökologie und Fischereibiologie Jena.

Wolter C, Arlinghaus R. 2003. Navigation impacts on freshwater fish assemblages: the ecological relevance of swimming performance. Reviews in Fish Biology and Fisheries 13, 63-89.

Wolter C, Sukhodolov A. 2008. Random displacement versus habitat choice of fish larvae in rivers. River Research and Applications 24, 661-672.

727830 FIThydro - Deliverable 1.2 Page 39 of 39