*1.1. Freshwater Biodiversity and Dams*

Although freshwater ecosystems cover less than 1% of the earth's surface, freshwater habitats are home to more than 10% of all known animals and about one-third of all known vertebrate species [1]. These ecosystems are also the most threatened: They are strongly affected by habitat modification, fragmentation and destruction, invasive species, overfishing, pollution, disease, climate change, etc. Freshwater ecosystem health is defined by its water quality and quantity [2], connectivity to other parts of the system and landscape, habitat condition [3], and diversity and abundance of plant and animal species [4]. Infrastructure development—especially dams—has caused a dramatic decline in the number of connected, free-flowing rivers [5]: Currently, there are more than 50,000 large dams

worldwide [6]. Some of these dams are used for hydropower, which is the largest contributor to global renewable electricity generation, supplying 16.4% of the world's electricity from all sources [7]. During the process of hydropower project development, insufficient attention is paid to impacts on the environment (such as the disruption of the natural flow regime, fragmentation of the single river ecosystem, suppression of migration paths and the changing habitats of species, greenhouse gas emissions from reservoirs, changes in sediment flow and channel processes, changes in the microclimate, transformation of biological and chemical properties of the water body) and measures to minimize these [8].

#### *1.2. Assessment Hierarchy and Sequencing for Hydropower*

Hydropower projects should seek to minimize their impact on natural ecosystems and ecosystem services while optimizing the project's energy generation potential. Adherence to the mitigation hierarchy from the earliest planning phase and throughout the project life cycle can achieve more sustainable hydropower generation [9].

The evolving context of available energy alternatives is implicitly present in hydropower discourse. From the beginning of the 21st century, wide recognition of the urgent need to reduce greenhouse gas (GHG) emissions was the reason for proponents of large hydro and nuclear power to promote these technologies as the most promising climate-friendly options [10] that provided unique benefits outweighing negative impacts on ecosystems and local communities [11]. With the emergence of massive wind and solar production having the potential to outcompete more expensive technologies, the same groups now describe hydropower as an important enabler and stabilizer for the mass deployment of intermittent renewable technologies [12]. Other industry-led entities, such as the Global Energy Interconnection and Development Organization, dare to propose world-wide renewable energy systems, where a unified high-voltage grid supplies load centers from remote solar, wind, and hydropower "energy bases" [13]. Therefore, an analysis of alternatives at the energy system level is no longer a formality, but a necessary first step when considering impact assessments in the energy sector.

The strategic environmental assessment (SEA) is accepted worldwide as the instrument to facilitate a more comprehensive analysis of energy development, either standing alone or in combination with integrated planning approaches, such as integrated water resources management (IWRM) or integrated river basin planning [14]. According to the Netherlands Commission for Environmental Assessment (NCEA), about 30 SEAs have been executed supporting decision-making on hydropower development, with most of those completed in Asian countries [15].

In theory, three distinct phases can be identified in the assessment and decision-making that leads to large dam development:


In practice, only phase three is subject to an environmental assessment with formalized public participation, and hydropower project proponents will usually perform the previous phases without public consultations.

The objective is to compare environmental risks that have been mitigated at different stages of hydropower development and analyze the effectiveness of conservation measures. The hypothesis is that initial avoidance and mitigation of hydropower impacts undertaken at early planning stages are more effective in reducing damage to ecosystems in a given basin than measures undertaken during and after a dam is constructed.

The effectiveness of application of specific assessment methodologies can be judged only in the institutional and social context of its use. Governance systems as such may require formalized assessment before decisions to invest in hydropower project are made. According to the Netherlands Commission for Environmental Assessment (NCEA), drivers (or root causes) for negative effects of dam development decisions are mostly related to deficiencies in the wider governance context such as neglect of strategic and system-based studies, favoring large dams above potentially more sustainable options without justification, and national public governance system incapable of correcting this situation due to late or insufficient involvement in the development process [16].

#### *1.3. Importance of Stakeholder Engagement*

Stakeholder analysis and proper engagement are crucial for environmental impact assessments (EIAs) because outcomes depend on the timely participation of stakeholders in the process, which can increase the willingness of the project developers to implement recommendations resulting from assessments. The social and cultural dimensions of dam development are widely documented and recognized [17] but are still rarely considered in assessments, especially at early stages of project identification [18]. The most neglected dimensions are an analysis of stakeholders involved in and affected by hydropower construction, and the development of tools to promote early stakeholder involvement into decision-making. In emerging economies, strategic planning is often expected to be based on objective science [19], rather than on the demands of various stakeholders [20].

While negative impacts on local communities partly resulting from incomplete assessments are well described in the literature [21], analyses of response strategies available to various stakeholder groups excluded from the assessment process are limited. One research work compared the more robust indigenous participatory mechanisms in Canada to those in Russia, where the development of the large Evenkiiskaya dam was halted partly due to active resistance from the indigenous community that was denied access to the official decision-making process [22]. Similar research on the Mekong [23] and rivers of Myanmar [24] has shown that the various facets in civil society that are denied access to decision-making may engage in developing their own disruptive strategies, including alternative assessment frameworks, and this may lead to drastic adjustments to project development processes driven by project proponents. Even companies strictly following national guidelines (and even best international practices) for public consultations during an EIA process still often experience fierce opposition from local communities, mainly when communities believe that the consultation process is being used as a tool to force hydropower projects in areas where communities do not agree in principle with dam building, or they see dams as a tool used by outside forces (e.g., central government) to impose control over territory/resources that local communities want to manage themselves [25]. Therefore, caution regarding public participation among project developers is quite understandable and underlies the strategy to make investment plans public only at the last stages of decision-making. However, a study that modeled such a process in Nepal still concluded that conflicts with local populations (inevitably embedded in dam projects) should be revealed and addressed at early stages of development and that by doing so, developers will save money and decrease risks [26]. The authors emphasized the lower implementation costs and higher chance for success when plans are changed at an early stage in the project cycle.

Many local stakeholders potentially affected by hydropower development are also unprepared to devote their resources to assess strategic planning documents. The majority of strategic planning documents, such as basin management plans, are not viewed by the public in developing/emerging economies as the real foundation for future decision-making, which is subject to the discretion of various government officials. Local actors most readily address problems that have already materialized and threatened their livelihoods, such as proposals to build a specific dam.

#### *1.4. Amur River Basin: Biodiversity and Hydropower*

An evaluation of the impacts of hydropower dams in the transboundary Amur basin is used to support the hypothesis that initial avoidance and mitigation of hydropower impacts can lead to better conservation outcomes. Amur is the largest transboundary river system in Northeast Asia, flowing through Mongolia, China, and Russia and forming a natural border between China and Russia (see Figure 1). The river basin is famous for rare waterfowl, big cats, and endemic fish. The basin is still an arena for massive fish migration along the main stem and tributaries, with salmonids, sturgeon, and lamprey being three important examples [27]. The floodplains of the Amur and its tributaries create a belt of wetlands with high biodiversity value. The basin area is included in the list of Global 200 Ecoregions of the World, which are a priority for conservation efforts [28].

**Figure 1.** The Amur River basin and its protected areas. © WWF-Russia.

In 2019, there were approximately 100 dams in the basin, including 19 large dams, of which 3 are located in the Russian part on the Zeya and Bureya rivers. The total generation is almost 22 GWh per year, with the three large Russian dams producing 13.7 GWh per year.

In China, all the most promising sites suitable for large hydropower had already been used by the beginning of the 21st century. Recently built medium-sized dams (e.g., Hadashan on the Second Songhua River) have pursued multiple purposes, with hydropower being subordinate to irrigation, water supply, navigation, etc. Since 2015, Chinese authorities have imposed restrictions on new small hydropower construction and have started the assessment, reconstruction, and removal of smaller dams in many river basins, including the headwaters of the Second Songhua River in the Amur basin [29].

In Russia, the Amur basin was the first region where intensive hydropower construction was resumed after the economic crisis of the 1990s: The Bureya hydropower dam was put into operation in 2003 to complement the Zeya dam built in 1975, and the Lower Bureya dam started generation in 2017. The flow regime of the Zeya and Bureya have changed significantly, which has resulted in the alteration of the natural floodplain ecosystems on both rivers. This has caused a decline in typical floodplain communities [30], habitats of cranes and storks [31], and refuges for fish species [32]. Dams have become a barrier to migration, and chum salmon (keta), lamprey, and whitefish have disappeared above the dams [33]. The continuing degradation processes of the floodplain system of the Amur under the cumulative influence of the Zeya and Bureya hydropower dams has been further exacerbated below the mouth of the Songhua River, where the flow regime has undergone additional anthropogenic changes due to the construction of hydro-engineering structures in China [34].

#### *1.5. Initial Dialogue between CSOs and Energy Industry*

The hypothesis is that the basin-wide assessment of hydropower impacts is an effective tool in early planning when analyzing environmental costs of hydropower development and comparing possible development scenarios.

In Russia, the official environmental impact assessment methods used for hydropower projects do not allow for the analysis of complex impacts from dam construction on the ecosystem of the basin as a whole and do not compare different scenarios to optimize development [35]. For comparison, in 2011, Chinese regulations were issued for basin-wide hydropower schemes that require a basin-wide strategic assessment to precede individual dam EIAs [36]. However, the application of such assessments with meaningful public participation is yet to be fully implemented.

The lack of legal requirements and established practices for comprehensive basin-wide assessments of hydropower impacts in early 21st century Russia has become a serious obstacle to productive dialogue between the newly created state-owned Hydro-OGK hydropower company, in charge of 70% of hydropower dams (later renamed RusHydro Co.), and a wide coalition of non-governmental organizations (International Socio-Ecological Union, Greenpeace, WWF-Russia, Russian Bird Conservation Union, and other NGOs.). In the aftermath of the World Commission on Dams Report [17], the two sides sought to agree on safeguards to prevent negative impacts from new hydropower projects under specific conditions for Russia. NGO and industry experts compiled a "comprehensive list of impacts from hydropower dams and reservoirs on the environment", which identified 9 groups of problems and more than 100 specific potential impacts. An analysis of causal relationships between a multitude of effects and issues demonstrated that the majority of these were determined or strongly correlated with three fundamental environmental changes brought about by large reservoir development. These changes included (1) freshwater ecosystem fragmentation/blocking of the natural movement of biological organisms, sediments, nutrients, etc.; (2) the augmentation of natural flow dynamics of matter and energy in the river network; and (3) the creation of vast lake-like habitats serving various human activities that replaced natural river valleys. Discussions between NGOs and hydropower corporations highlighted mutual interest in identifying and ranking hydropower development options in each large river basin according to the potential severity of local and basin-wide impacts on ecosystems. This could assist both the industry and NGO community in defining long-term priorities for action and predicting and preventing possible conflicts and negative consequences. Therefore, it was suggested that the optimization of hydropower development in a large river basin first requires an assessment of impacts resulting from these three fundamental changes under different hydropower development scenarios. Such an assessment scheme was later designed by NGO experts and was initially welcomed by RusHydro management, although it was not explicitly used by the company [37].

#### **2. Materials and Methods**

The dialogue between RusHydro and leading civil society organizations (CSOs) held in 2007 helped to identify and introduce into the national policy debate key tools for preventing hydropower's negative impacts on nature and people who became the foundation for CSO activities in this field for the following decade. In this paper, we explore how some of those tools were applied to the Amur basin.

The tools to prevent adverse dam impacts employed in the Amur basin by civil society included the following:


The methodology underlying each of the tools designed for trial use in the Amur River basin is described in this section.

#### *2.1. Strategic Basin-Wide Rapid Assessment of Hydropower Impacts*

During preliminary phases of hydropower development planning, it is important to determine the scale of impacts from each potential dam and rank proposed projects and multidam development scenarios according to the degree of their environmental impacts. A basin-wide assessment of hydropower options allows for a comparison of dam sites in terms of potential effects on connectivity and downstream flow regimes. Somewhat similar basin-wide optimization approaches proposed by The Nature Conservancy (TNC) [38] and other research papers on the global footprint of dams [39] inspired the formulation of specific minimum requirements for the rapid assessment of basin-wide environmental impacts of existing and planned hydropower dams. Such an approach could yield the most significant gains if used to guide development in large basins not yet significantly altered by hydropower, and therefore the transboundary Amur basin was considered the priority among basins of Russia.

Rapid assessment methodology is focused on simple modeling of potential impacts from dams on river ecosystems and is intended to guide developers on how to choose options with the least negative environmental effects. When assessing the cumulative impacts of several dams on the ecological condition of the basin, first and foremost, we considered the following broad impact factors, which have been jointly identified as the most important by NGOs and industry experts:


As demonstrated by previous assessments of river basins around the world, these three factors are associated with the majority of observed and predicted consequences of dam building for aquatic ecosystems, e.g., mapping the world's free-flowing rivers [5], global threats to human water security and river biodiversity [40], and restoring environmental flows by modifying dam operations [41].

To assess and compare the impacts of multiple scenarios on the river basin, three pairs of proxy indicators for impact were designed, so that each of the main factors could be expressed both in absolute and relative values.

## 2.1.1. Flow Regime Alteration and Floodplain Transformation Downstream from Dams

The most important objective should be to protect the Amur River and its floodplains, which contain the main biological resources and ecological services as well as the natural support base for the local communities of the Southern Far East. The methodology for evaluating the socioecological impacts and criteria for future dam construction should be developed based on the main concern [42]. The degree of floodplain ecosystem transformation downstream from dams reflects the consequences of altering the natural hydrological regime of a river. The index is based on the relationship between the storage volume (live volume) of a reservoir(s) and the total annual river flow volume at the dam's location, which is a consensual index used in many studies. However, for the Amur River, it is expressed not just as the percentage of the flow volume that can be withheld in the dam's reservoirs upstream of a given river's reach (index commonly known as the degree of regulation (DOR), see Reference [5]), but is further multiplied by the area of the floodplain ecosystem at a given reach. The resulting index, *Imp\_fl* (km2), is proportional to both the area of vulnerable habitats and the degree of impact in a given river reach (stretch). To characterize the basin-wide situation, it is summed across all river reaches of the basin/sub-basin (*ALT\_fl* (km2)). When related to the original unaltered floodplain area of a given river reach, it approximates the share of affected floodplains as a percent: *Imp\_fl* (%). For each scenario, it is summed across all river reaches of the basin/sub-basin. To characterize the basin-wide alteration of floodplains, the ratio between *ALT\_fl* (km2) and total area of natural floodplains (*Sfl*) in the given basin/sub-basin is used and is expressed as a percent: 100 × *Imp\_fl*/( *Sfl*) = *ALT\_mean* (%).

The degree of floodplain ecosystem transformation provides an index to measure how strongly a dam or set of dams can affect the natural floodplains and their ecosystem services.

Floodplains are singled out as the most ecologically important habitat affected by flow regulation, which is valid for the Amur basin, but applying the assessment methodology in a different river ecosystem may require focusing on different key habitat types. The biota and ecosystems of the rivers in the Amur catchment are strongly dependent on the floods that are regulated by dams.

#### 2.1.2. Transformation of Riverine Habitats by Reservoirs

Any reservoir is an anthropogenic feature replacing the essential socioecological landscapes: river valleys. It is assumed that the larger the surface of the water reservoirs and the greater their share in all water surfaces of the river system, then the more they transform aquatic and terrestrial ecosystems. The surface of the existing river system and the modeled surface of planned reservoirs as well as the surface of all freshwater ecosystems upstream from the dam preceding reservoir formation (during the low-flow period) is calculated. Reservoir surface areas are expressed as *Imp\_res* (km2). Transformation by the reservoir is calculated as the percent ratio between reservoir surface area and surface area of all aquatic ecosystems of a given sub-basin before reservoir formation *IMP\_reservoir* (%).

The value of this measure, when applied to the Bureya and Zeya dams, can be seen on the impacts of fish species and stocks. The Bureya dam and Zeya dam reservoirs together occupy 3160 km2, which is roughly equal to 45% of the total water surface in the Middle Amur Freshwater Ecoregion in Russia. In China, all 12 hydropower reservoirs of the basin occupy only half of that area. The Zeya and Bureya reservoirs have low-quality water, in part due to inundation by massive volumes of vegetation, soil, and peat. Before the Zeya dam construction, the composition of the fish fauna of the Upper Zeya basin in 1970 included 38 species, but by 2007 the fish fauna of the Zeya reservoir was reduced to 26 species [43]. Fish stocks of the Zeya reservoir have been seriously depressed for many years.

#### 2.1.3. Blocked Sub-Basins: A Measure of Basin Fragmentation

River fragmentation indices typically measure the degree to which river networks are fragmented longitudinally by infrastructure, such as hydropower dams. Fragmentation prevents effective ecological processes that depend on longitudinal river connectivity, including the transport of organic and inorganic matter and upstream and downstream movements of aquatic and riparian species [5,39].

A simple measure of the fragmentation of the river basin is a percentage of the basin area that is cut off from the sea by dams. This study takes as a proxy measure the area (*Imp\_bl* (km2)) or share (*IMP\_block* (%)) of the basin disconnected by dams from the sea. In other more sophisticated versions of this methodology (not presented in this paper), additional fragmentation indices have also been tested that measure the degree of partition in many disconnected sub-basins [44]. For purposes of this rapid assessment, those additional indices do not reveal new trends but reinforce the emphasis on fragmentation. In a more nuanced study with a detailed comparison of multidam scenarios, the use of such indices would be more justified.

A blocked basin index was justified as a proxy index in our analysis because the global value of the Amur Freshwater Ecoregion was primarily attributed to the abundance of diadromous fish species that suffer from disruption of migratory routes to the sea [27]. For example, upstream from the Zeya and Bureya dams, the Amur sturgeon, kaluga, keta salmon, Japanese lamprey, and other migratory species have already disappeared. Taken together, by 2011, the Zeya and Bureya dams blocked 9% of the Amur catchment area, while all of the existing dams in China blocked an additional 22–23%. This means that nearly one-third of the Amur River system has already been isolated from the sea and can no longer sustain key migratory species, e.g., diadromous fish.

For calculations, the analyzed river system was divided into reaches (stretches) limited by existing and potential dams and confluences with tributaries. Since the analysis was applied only to large dams/reservoirs, the inquiry was limited to streams having a catchment area of more than 10,000 km2 (with 3–4 exceptions where large dams were planned in catchments of a smaller size). Each stretch with a complex of natural characteristics was defined as a key basic unit in the multifactor analysis. The whole Amur basin was delineated into 214 units/stretches (see Figure S1), to which we attributed characteristics of the respective elementary sub-basins. To define stream direction, borders of watersheds, the size of planned reservoirs, hydrographic models, and data from the Shuttle Radar Topographic Mission (SRTM) [45] were used.

Satellite imagery was processed using Erdas Image software, while expert interpretation of water bodies and floodplains was performed using ArcGIS 10.5. The sources of cartographic information were Landsat Aster and Sentinel-2 satellite imagery, SRTM water body data (SWBD) [46], and vector topographic maps (1:500,000–1:100,000).

Characteristics of hydropower engineering projects have been derived from planning documents, river basin development schemes, technical documentation for dam design, etc. Annual discharge at each reach was derived from official Soviet hydrological bulletins and literature on hydropower projects and plans, e.g., hydropower master plans [47], natural resource allocation studies [48], bilateral water management schemes [49,50], and five-year plans [51]. The remaining data gaps were filled by modeling/approximation.

From all available sources, a list of 84 possible large dam locations in the Amur River basin was compiled, from which 45 projects had data sufficient to perform a basin-wide impact analysis (see Figure 2).

The characteristics were attributed to each river reach (stretch) and related sub-basin as appropriate. The three proxy indicators of impact were calculated at each river stretch and/or each sub-basin for each assessed basin-wide hydropower development scenario (see Table S1).

The values of the three indicators at the river mouth (or at the terminal point of any sub-basin/catchment) were taken as a proxy for a basin-wide score for a given scenario in this catchment. To characterize the degree of alteration from one value (following the advice of Dr. Alexander Martynov), an integrated index was also developed:

$$INT3 = \sqrt[3]{(ALT\\_mean) \times (IMP\\_reservoir) \times (IMP\\_block)} . $$

For comparisons between basin-wide scenarios, the three proxy indicators were used as well as the integrated index calculated at the Amur River estuary.

To relate electricity production to the potential impact, all proxy indices were divided by the annual electricity production expected in a given scenario. The impact per unit of energy generation is an important measure of efficiency in terms of environmental economics [52].

All existing and possible dams and hydropower development scenarios (combinations of dams described in planning documents) in the Amur River basin were ranked according to their potential environmental impacts and efficiency relative to energy production (see Table S2).

For the sustainable development of hydropower in a given river basin, the aim is to minimize both potential impacts and their relative measure per unit of energy production. This rapid assessment methodology enables a quick initial ranking of multiple hydropower development scenarios in a large river basin according to their key potential negative impacts. Therefore, it provides developers with an opportunity to concentrate further efforts on less risky development options, while conservationists can focus on the prevention of dam construction at locations where it may lead to the greatest negative impacts for natural freshwater ecosystems.

This is the first basin-wide assessment of hydropower impacts designed for the Amur basin and the first ever suggested in Russia. The main innovation of our methodology compared to earlier assessment methods designed for other basins has been to reduce the assessment to three key variables, which, once agreed upon, do not require subjective expert inputs in further scenario assessment. This enables a credible comparison of impacts from any number of hydropower development scenarios at a basin-wide scale, including existing and proposed dams. Despite simplification, the methodology still takes into account specific features of the basin in question (e.g., using the degree of floodplain

ecosystem transformation index in place of the conventional degree of regulation index (DOR)). Our model has also been shown to have certain theoretical value by successfully relating various degrees of hydropower potential utilization to measurements of potential impacts associated with such development scenarios.

**Figure 2.** Locations of the existing and potential dams in the Amur River basin. © WWF-Russia.

#### *2.2. Creation of Protected Areas at River Stretches Targeted for Hydropower Development*

In the case of the Amur, the river stretches where hydropower construction is predicted to cause the most damage to the ecosystem of a given river basin are pre-emptively protected in perpetuity to constrain the future design of hydropower at such sites by legal regulations. Since sites known to be suitable for large dams are scarce and often located in river gorges with rich biodiversity, it is often possible to justify protected area development based on local natural values even without reference to its wider safeguarding function for a basin-wide river ecosystem. Since Russian legislation does not prescribe the design of specific biodiversity conservation measures as part of basin-wide water resources management planning, after the identification of critical sites, the rest of the work follows a standard methodology for the creation of protected areas.

In Russia, legally protected natural areas of different types can be established at a provincial and national level, while in China, four levels of government have such authority. In both countries, there are long-term plans for gazetting conservation areas, in which prospective protected areas should be included in order to be considered [53]. Field research by local scientists to collect data to justify the protection of a given site is usually required. In both countries, formal reports covering the scientific, economic, and management justification for a given protected area should be presented to a special commission formed by a relevant government agency. In addition, until 2019, Russia required the socioeconomic justification for protected areas (PA) establishment to be subject to an EIA procedure, which included public consultations with potentially affected stakeholders. The overall process of PA establishment could take anywhere from 2 to 10 years. However, once preservation on a natural hydrological regime and/or prohibition for massive infrastructure development is written into management regulations of a new protected area, it becomes a legal requirement. Removing or adjusting such prohibitions requires a new and lengthy procedure involving a new EIA and public

consultations. Adjustments to a protection regime happen more often in China than in Russia, but still present serious legal and reputational challenges for proponents of dam building. In China, at least one precedent of the protected area being reconfigured to give way to dam building is known. However, this has not happened in Russia.

Protected areas have been created several times since the 1990s in the Amur basin to protect the most vulnerable sites. The Norsky Strict Nature Reserve in Amurskaya Province of Russia was the first such PA (successfully created in 1998) to safeguard the confluence of the Nora and Selemdzha rivers, where development of the Dagmarskaya dam was proposed in the early 1990s. The method was pioneered by a team of local conservation biologists led by Dr. Yury Darman [54].

This team of authors (to the best of our knowledge) is the first to apply this approach to protected areas for freshwater ecosystem conservation at the basin scale, namely to establish conservation areas at potential damming sites in a systematic manner.
