Global Proliferation of Small Hydropower Plants

REVIEWS REVIEWS REVIEWS Global proliferation of small hydropower 91 plants – science and policy

Thiago BA Couto* and Julian D Olden

Large-scale electricity policies that embrace renewable resources have led to continued investments in hydropower. Despite evolving viewpoints regarding the sustainability of large hydropower installations, there has been a major increase in support for the widespread development of small hydropower plants (SHPs). A global synthesis reveals that 82,891 SHPs are operating or are under construction (11 SHPs for every one large hydropower plant) and that this number is estimated to triple if all potential generation capacity were to be developed. Fueled by considerable political and economic incentives in recent decades, the growth of SHPs has greatly outpaced available ecological science. We provide evidence for not only the lack of scientifically informed oversight of SHP development but also the limitations of the capacity-based regulations currently in use. The potential indiscriminate expansion of SHPs under the pretense of promoting sustainable energy is concerning, and we identify several important steps to help ensure new scientific advances, effective management, and policy reform in the future.

Front Ecol Environ 2018; 16(2): 91–100, doi: 10.1002/fee.1746

ne of the greatest challenges of this century is to expected to be built and to commence operation in the Osupport growing human societies with electricity coming decades (Zarfl et al. 2015). These activities are most policies that embrace environmentally sustainable prevalent in developing countries where governments are resources. Hydropower, the power derived from the prioritizing new large dams as the centerpiece of energy energy of flowing water, is the world’s primary source of plans (Winemiller et al. 2016). Unfortunately, these dams renewable energy, contributing almost three-quarters of have substantial socioecological costs (Rosenberg et al. the global renewable supply and nearly one-fifth of all 1995). Indeed, public support for new large dams has electricity production (REN21 2015). Global invest- diminished as the high socioeconomic costs, emissions of ments in hydropower peaked in the second half of the greenhouse gases, and negative consequences for valued 20th century, partly in response to a growing desire to ecosystem services, including water quantity and quality, diversify away from thermoelectric facilities and avoid biodiversity, and fisheries, are increasingly realized (Ansar greater dependency on fossil fuels (Oud 2002). et al. 2014; Kahn et al. 2014; Deemer et al. 2016). In response to the mounting demand for renewable Amid evolving conversations regarding large hydro- energy, thousands of new hydropower installations are power sustainability, a marked surge of support for small hydropower plants (SHPs) has emerged (Panel 1). The term SHP broadly refers to facilities that produce less In a nutshell: electricity and operate in smaller rivers as compared to • We estimate that 82,891 small hydropower plants (SHPs) large hydropower plants (LHPs). However, because SHPs are operating or are under construction in 150 countries have a diversity of operation modes (eg diversion and and that another 181,976 new plants may be installed if all potential capacity were developed non-diversion with or without storage), flow control • Country-level definitions of SHPs vary immensely with structures, and environmental impacts (Panel 1), this respect to maximum generation capacity and include limited definition is ambiguous. Political and economic incen- consideration of potential descriptors of ecological impacts tives for sustainable energy development initially fueled (eg dam height, reservoir area, and operating procedures) the growth of SHPs, particularly in Europe and China • Environmental policy and the existing body of scientific knowledge are insufficient to inform SHP development, (Paish 2002; Tang et al. 2012), but now countries across leading to a chronic underappreciation of the potential the world consider SHPs to be a critical component of socioecological consequences of their proliferation future energy strategies. For instance, more than half of • We outline the need for stricter and better-informed reg- US states have renewable portfolio standards that ulations, integrated watershed management that considers disallow electricity from LHPs, yet they embrace power SHPs in aggregate, and information systems to support strategic planning of new construction into the future produced from SHPs (Kao et al. 2014), due in part to a perception that “smaller” equates to lower socioecological impact (Gleick 1992). As this viewpoint continues to gain traction in public and political arenas, the lack of School of Aquatic and Fishery Sciences, University of Washington, scientifically informed regulations (Gleick 1992; Erlewein Seattle, WA *([email protected]) 2013) and the potential indiscriminate expansion of

92 Panel 1. What is small hydropower? There is great diversity in the type and size of small hydropower plants (SHPs). The general principle of hydropower generation is the conversion of water pressure into mechanical shaft power by hydroturbines (Paish 2002). Despite the diversity of operation modes, hydropower can be classified according to the level of flow control (ie proportion of water stored) and the presence of diversion structures (McManamay et al. 2016). The degree of flow control ranges from installations that store water in reservoirs to schemes that do not retain much water (ie run-of-river) and are subjected to natural fluctuations of river discharge (Figure 1). Some facilities are intermediate to these classes, in the sense that they control the flow only during specific periods of the year or times of day (McManamay et al. 2016). Reservoirs are created to minimize variability in water supply, storing water Figure 1. Schematic depicting the primary classification of operation modes of for periods of low flow or high electric- hydropower according to the presence of storage and diversion structures. Any of these ity demand (Egré and Milewski 2002). In operation modes can be found in installations classified as SHPs. Non-diversion with multi-purpose hydropower dams, res- storage (a) and without storage (b). Diversion with storage (c) and without storage (d). ervoirs are also used to supply other human needs such as irrigation, flood control, and urban consumption (Lehner et al. 2011). Whereas (Wu et al. 2010). Dewatered sections range from very short operations involving water storage cause major alteration to distances to several kilometers of reduced flow, depending on natural flows of rivers, run-of- river schemes use weirs or small the SHP project (Anderson et al. 2006). dams to block the river channel, in theory maintaining more SHPs can be operated as complexes of multiple installations natural flow conditions (Csiki and Rhoads 2010). both in close proximity and distributed broadly within a water- Non-diversion schemes use water from an impoundment shed. For example, some run-of-river SHPs are located down- structure that is adjacent to the powerhouse, whereas diver- stream of large reservoirs that are used to store water and sion schemes transport water via canals or pipes to a distant control flow (McManamay et al. 2016). Others can use water powerhouse (Figure 1). Thus, mainstream sections in diversion transported from remote basins via interbasin transfer, causing schemes are dewatered until the water is returned after pass- major alterations to the ecosystem (Cada and Zadroga 1982; ing through the turbines. The amount of water abstracted and Anderson et al. 2006). These examples emphasize the com- transferred to the powerhouse dictates the magnitude of river plexity of the environmental repercussions from the imple- dewatering, which can be very high during drought periods mentation of multiple SHP facilities.

SHPs under the pretext of sustainable energy promotion policies and the scientific knowledge that currently shape is concerning (Premalatha et al. 2014). the management and regulatory practices of SHPs. We provide the first global overview of the expansion Finally, we provide a synopsis of the scholarly literature of SHPs, summarizing status and trends in an interna- examining the known or potential ecological impacts of tional context, and encouraging the bridging of critical SHPs, and identify the main challenges to ensure that science and policy gaps. First, we explore the rapid environmental policies inform the worldwide expansion growth of the sector and quantify the current and future of SHPs. hotspots of SHP development based on a dataset compiled from the most up-to-date international energy policy JJ Global proliferation of SHPs reports. The dataset includes national definitions of SHPs, current numbers of SHPs and LHPs, and estimates Our synthesis of the energy policy reports reveals that of future SHP numbers based on countries’ untapped at least 82,891 SHPs are operating or are under con- hydropower potential (methods described in WebPanel struction in 150 countries (Figure 2a; WebPanel 1). 1). Second, we expose the global inconsistencies in how We estimate a tripling of this number to include an SHPs are defined, and argue for the removal of the additional 181,976 plants that could be installed if all “small” modifier. Third, we discuss the environmental potential capacity were to be developed – of these,

www.frontiersinecology.org © The Ecological Society of America TBA Couto and JD Olden Small hydropower plants

10,569 new projects already appear in 93 national plans (Figure 2b; WebPanel 1). Distinct geographic patterns in SHP numbers are apparent, reflecting differences in socioeconomic conditions, varying regulations and incentives, and contrasting hydrologic potential. China is the global front- runner, with 47,073 SHPs currently operating; the surge was propelled by private investments, technology lead- ership, and rural electrification programs (Tang et al. 2012; WSHPDR 2016). An additional 26,877 SHPs are operating in Europe, where SHP development not only has a long history but has also received recent attention associated with the need to satisfy international agreements promoting clean energy (Paish 2002; Hermoso 2017). Future plans for SHPs are concentrated in Asia, the Americas, Southern and Eastern Europe, and East Africa. Considering that national dam inventories generally underestimate the number of smaller facilities, we expect that the numbers of SHPs reported here are conservative. Even based on these values, the current Figure 2. Map depicting (a) the current number of SHPs (82,891) operating or under number represents more than 91% of construction across 150 countries and (b) the estimated number of potential SHPs the total hydropower installations, (181,976), of which at least 10,569 are included in national plans to be implemented dwarfing the number of LHPs n( = in the coming decades. See WebPanel 1 for description of methods. Country-level 7721; Figure 3a). On average, there statistics are available at: https://figshare.com/s/e118ecae964140f16f80. are close to 11 SHPs for every one LHP in the world (WebPanel 1). Although representing (Kao et al. 2014), and that new hydropower would come the vast majority of hydropower plants, SHPs contribute predominantly in the form of hundreds of new SHPs. just 11% of the global electricity generation capacity based on hydropower (WebPanel 1). SHPs are proliferating around the world (Figure 3b; JJ Just how “small” are SHPs? WebPanel 1; IRENA 2017) as a response to governmental incentives in the form of competitive tariffs, as well as Our assessment revealed notable disparities in how coun- investments by the private sector (ie trading energy with tries classify hydropower plants as “small” (WebPanel 1). the grid or providing lower-cost energy for local indus- The size of a given hydropower facility is commonly tries), programs promoting rural electrification in remote defined according to its generation capacity in watts regions, simplified licensing processes, and quick con- (hereafter termed “capacity”), which is the maximum struction time (Egré and Milewski 2002; Tang et al. 2012; capacity of hydropower production assuming optimal hy- WSHPDR 2016). For example, aggressive policies in sev- drologic conditions and turbine efficiency. Capacity-based eral European countries have encouraged rapid sector definitions of SHPs vary substantially, ranging from up development since the 1980s (Paish 2002; Jesus et al. to 1 megawatt (MW) for facilities in Germany and Burundi 2004). Numerous other countries are now duplicating to up to 50 MW for facilities in Canada, China, and such policy initiatives. In Brazil, a series of incentives and Pakistan (Figure 5). About 70% of countries with formal new regulations contributed to a fivefold increase in the definitions classify SHPs as installations with less than number of SHPs over the past two decades (Panel 2). 10 MW, which is increasingly recognized as the inter- Even in the US, where current conversations typically national standard (WSHPDR 2016). Of those countries focus on the removal of aging dams (Bellmore et al. 2016), that currently operate SHPs, 15% do not have a formal a recent US Department of Energy report highlights that (or widely accessible) national designation (Figure 5). In 65 gigawatts of potential hydropower remain untapped addition to considerable variability in SHP definitions

94 and Hart 2002; Mbaka and Mwaniki 2015). For instance, two SHPs in Brazil (São Sebastião, Braço Norte II) have the same capacity (10 MW), but differ thirtyfold with respect to their reservoir areas (0.2 and 6.0 km2, respectively). In addition to the limited and inter- nationally inconsistent definitions of SHPs, there are considerable discrep- ancies in hydropower classification within countries; this means that some nations may support policies that are based on different capacity thresholds. In Russia, contrasting regulations classify SHPs as facilities generating less than either 25 or 30 MW (WSHPDR 2016). SHP definitions remain highly dynamic else- where, including in the US and Brazil, where recent amendments increased the upper limits from 5 to 10 MW Figure 3. (a) Comparison of the number of SHPs and LHPs per country. Continents (Hydropower Regulatory Efficiency are represented by different colors and symbols. (b) Trends in installed capacity of SHPs Act of 2013) and from 30 to 50 MW (<10 MW) by continent from 2000 to 2016. See WebPanel 1 for description of methods. (Law 13.360/2016), respectively. Some countries further delineate across countries, facilities designated as SHPs may have SHPs into small, mini, micro, and pico hydro classifications substantially different dam sizes, reservoir areas, storage (WebPanel 1). For the purposes of this review, we consider capabilities, outlet structures, and operating procedures SHPs as all those with maximum capacity defined accord- (Panel 1; Figure 6), all of which are strongly correlated ing to the country-specific definition or 10 MW for coun- with environmental impacts of dams (Gleick 1992; Poff tries without national definitions.

Panel 2. Brazil as a model of the rapid expansion of SHPs In Brazil, 1007 SHPs are currently operating, an additional 35 are under construction, and 156 are approved and awaiting final licensing (ANEEL 2016). On average, 33 SHPs have been constructed per year from 2001 to 2016, a growth rate 14 times as fast as that witnessed in the 1990s (Figure 4). By contrast, increases in LHPs have remained constant in recent decades. Massive investments by the private sector after 1995 were stimulated by new regulations within the energy market and by economic incentives, resulting in hundreds of new SHPs (Ferreira et al. 2016). Simultaneously, there was a refor- mulation of the environmental licensing process (CONAMA 237/1997) that later included the introduction of simplified licensing procedures for projects with perceived “little potential of causing environmental disturbance”, which were defined as SHPs with less than 10 MW (CONAMA 279/2001). This licensing reform was followed by a re-definition in 2003 of what constitutes an SHP, which quadrupled the maximum reservoir Figure 4. Number of Brazilian hydropower facilities in area from 3 to 13 km2 while maintaining the maximum limit operation since 1970. Although LHPs have demonstrated a of 30 MW (Ferreira et al. 2016). In November 2016, a new bill constant construction rate for decades, a rapid proliferation in was approved, increasing the upper limit of SHPs from 30 to the construction of SHPs commenced in the early 2000s 50 MW (Law Nº 13.360/2016). following new regulations and incentives (WebPanel 2). This plot includes 937 hydropower facilities (out of 1227 total) reporting the first date of operation (ANEEL 2016).

www.frontiersinecology.org © The Ecological Society of America TBA Couto and JD Olden Small hydropower plants

J Limited environmental regulations for SHPs 95

The rapid expansion of SHPs is due in part to weak regulatory oversight that has encouraged investments in the private and public sectors. According to a global policy compilation, at least 44 out of 160 countries require a formal environmental licensing process to construct and operate SHPs (WSHPDR 2016), leaving potentially more than two-thirds of the countries without recognized environmental requirements. Certain countries have regulations requiring mitigation actions – such as fishway structures that attempt to facilitate fish movement and standards of minimum flow releases aimed at sustaining human and environmental water needs – when SHPs are constructed. In France, all hydropower installations, regardless of size, must have facilities that promote passage of migratory fish species (Larinier 2008). By contrast, countries such as Costa Rica lack formal rules of minimal flow releases or else adopt arbitrary standards based on Figure 5. Frequency of upper limits of generation capacity used the policies of other countries (Anderson et al. 2006). to define SHPs in 150 countries and territories (22 countries lack When present, environmental licensing processes asso- formal definitions). See WebPanel 1 for description of methods. ciated with SHPs vary substantially among countries, but Data available at: https://figshare.com/s/e118ecae964140f16f80. are often less rigorous than licensing of LHPs. Installation of LHPs frequently requires Environmental Impact Assessments (EIA), public consultations, management Turkey include reservoir area in their SHP definitions (as plans, and monitoring programs (Erlewein 2013), whereas large as 13 km2 and 15 km2, respectively) but these are the installation of SHPs may involve only a simplified exceptions (Dursun and Gokcol 2011; Ferreira et al. 2016). EIA if anything at all. In India, SHP construction (<25 MW) does not necessitate an EIA; instead, only a JJ Science lags behind the rapid rise of SHPs Detailed Project Report must be approved, and even these evaluations are commonly inadequate (Erlewein 2013). The widespread increase in SHPs continues to greatly Similarly, in Brazil, only a Simplified Environmental outpace advances in scientific knowledge in the environ- Report is required for constructing SHPs with a capacity mental field. Of the publications that we reviewed, fewer of less than 10 MW, but larger SHP projects (10 to 30 than 5% explicitly investigated SHPs, even though SHPs MW) are subjected to a licensing process similar to that of represent more than 91% of existing hydropower instal- LHPs (Panel 2). In the US, SHPs of less than 10 MW are lations (WebPanel 2). Research on SHPs grew in the exempt from some licensing requirements; similar exemp- 2000s when the sector flourished in many countries, a tions are also granted to hydropower projects up to 40 trend that started more than three decades earlier for MW placed along canals associated with agricultural, LHPs. Research on LHPs has been fundamental to un- municipal, or industrial water consumption (Hydropower derstanding, regulating, and attempting to mitigate the Regulatory Efficiency Act of 2013). Prior to 2008, EIAs ecological impacts of hydropower (Kahn et al. 2014; Olden were not mandatory in Turkey for installations up to 50 2016); however, we question whether such knowledge is MW, but now projects between 0.5 and 20 MW require transferable to inform policy and management decisions full EIAs (Dursun and Gokcol 2011). These examples for SHPs. Furthermore, although generation capacity is reinforce the notion that the “small” in SHP policies con- the benchmark used to regulate and manage hydropower, tinues to be equated with negligible environmental only 15% of the papers included in our synthesis reported impacts, and a streamlined approval process with inade- the capacities of the studied installations (WebPanel 2), quate regulatory oversight is applied in many instances. making generalizations between SHPs and LHPs even Perhaps most concerning is that environmental regulations more difficult. In short, the rise in the number of scholarly for SHPs are mandated on a single and arbitrary capacity publications has not been commensurate with the rapid threshold that does not necessarily equate with the magni- increase in SHPs observed globally. tude of environmental impacts (Gleick 1992). By pooling diverse hydropower projects, these regulatory frameworks JJ Emerging evidence for the ecological overlook considerable variability in modes of operation, impacts of SHPs degree of flow modification, dam size, and inundated area, which ultimately define the ecological impacts of dams The diversity of sizes and operation modes of SHPs (Egré and Milewski 2002; Poff and Hart 2002). Brazil and (Panel 1) produce a myriad of ecological consequences

96 (a) (b) N Wu

Figure 6. Examples of SHPs in (a) Brazil (CGH Abrasa, 1.5 MW), (b) China (Cangpinghe, 9.1 MW), (c) Brazil (PCH Ludesa, 30 MW), and (d) Canada (Rutherford, 49 MW).

that may not necessarily differ from what is expected direct impacts of deforestation and land preparation, as from LHPs (Cada and Zadroga 1982). Despite relatively well as indirect impacts associated with transportation, slow progress in scientific research, evidence suggests construction materials, and energy requirements (Cada that the environmental impacts of SHPs include those and Zadroga 1982; Pang et al. 2015). When SHPs impound associated with the dam’s construction and land water, the newly created reservoirs produce greenhouse inundation, as well as post-construction alteration to gases through mechanisms such as decomposition (Deemer flow regimes, the loss of habitat connectivity, and the et al. 2016) and inundate terrestrial habitats, potentially cumulative effects of multiple installations. However, the threatening biodiversity and compromising ecosystem magnitude of such impacts may depend more on the functions (Benchimol and Peres 2015). attributes of individual projects and on the landscape Hydropower dams and diversions not only modify context in which they are located, rather than merely downstream flow regimes, channel morphology, and water being related to an arbitrary size classification (Panel 1; temperature but also affect sediment transport and deposi- Gleick 1992). Perhaps the greatest difference between tion (Baker et al. 2011; Olden 2016). These modifications SHPs and LHPs is their position in the watershed; SHPs tend to be more evident in installations with greater stor- generally occur in smaller rivers as compared to LHPs age capacities (Mbaka and Mwaniki 2015), but they are (Kibler and Tullos 2013). This is problematic given the also present for run-of-river SHP schemes (Stanley et al. importance of headwater streams in maintaining hydro- 2002; Meier et al. 2003; Anderson et al. 2015). Among logic connectivity, harboring biodiversity, and supporting the various ecological effects (Table 1), artificially warm ecosystem integrity at regional scales (Meyer et al. 2007). or cold water is released downstream from smaller dams, The construction of the main (eg dams, reservoirs, diver- depending on the mode of operation and thermal profile sion channels, penstock, powerhouses) and accompanying of the reservoir, leading to impacts on biological commu- (eg roads, transmission lines, substations) structures are the nities (Hayes et al. 2006). The transformation of flowing initial sources of impact. During construction, there are rivers to shallow, standing water can have numerous con-

Table 1. Examples of ecological effects of SHPs on freshwater biodiversity, organized according to flow control 97 (storage versus non-storage) and presence of a diverting structure (diversion versus non-diversion). Additional examples are included in WebTable 1.

Flow Structure Capacity Ecological response Country Reference control (MW) Storage Diversion 18 Fish species composition modified in response to fragmentation Costa Anderson et al. by damming and dewatering. Opportunistic life-history strategists Rica (2006) dominate over species exhibiting more complex reproductive requirements < 10 Downstream decrease in macroinvertebrate density and richness, Portugal Jesus et al. with the replacement of high-flow- adapted species by low-flow, (2004) generalist species < 1 Periphytic chlorophyll and biomass showed little difference between Spain Izagirre et al. diversion canals and mainstream, suggesting minimum effect on (2013) nutrient uptake efficiency despite alterations to hydrology Non-diversion < 1 Fish condition in reservoirs decreases with compromised integrity Brazil Ferreira et al. of nearshore vegetation (2015) Not Diversity of macroinvertebrate communities reduced by altered Portugal Cortes et al. reported downstream flow caused by hydropeaking dam operations (1998) Non- Diversion < 10 Fish communities dominated by small-bodied species in dewatered Czech Kubecka et al. storage sections of rivers. Average individual weight and total biomass of Republic (1997) fish decreased four times in the reduced discharge sections, where biomass losses were proportional to the degree of abstraction < 1 Benthic macroinvertebrate communities showed modest down- UK Copeman et al. stream responses, where high-flow-adapted mayfly and stonefly (1997) species exhibited reduced abundance in dewatered stream sections Not Benthic algae communities modified after SHP construction. China Wu et al. reported Modification in algae communities and chlorophyll concentrations (2009a), tend to be more pronounced in the dry season in response to a (2009b), (2010) higher proportion of water abstraction sequences that include effects on primary production and cles also hinders the dispersal abilities of freshwater organ- the alteration of algae, invertebrate, and fish communities isms. For instance, after facing a series of obstacles along (Kubecka et al. 1997; Jesus et al. 2004; Wu et al. 2010). their migration route, only 18% of tagged Atlantic salmon Damming rivers can create major conservation problems (Salmo salar) reached spawning grounds above a hydro- by fragmenting critical habitats (Gido et al. 2016). Dams power complex in Scotland (Gowans et al. 2003). Given and weirs associated with SHPs represent physical barriers the huge number of current and planned SHPs, determin- for migratory species that rely on connected rivers to move ing the ecological effects of multiple rather than individ- upstream to spawn, to access floodplains, and for down- ual installations deserves urgent attention. stream migrations (DuBois and Gloss 1993; Benstead et al. 1999; Ovidio and Philippart 2002). Recent advances in fishway construction seek to mitigate some of these impacts; JJ Management and policy however, scholarly literature sources suggest that fishways often fail to support fish migrations (Olden 2016). Several Environmental policies must support robust and science- issues related specifically to fishways in SHPs remain to be based licensing regulations for SHPs. Differentiating resolved (Larinier 2008). In Portugal, for example, over half “small” from “large” hydropower is a primary decision of the SHP fishways are unsuitable for fish migration due to node in most licensing frameworks when determining poor performance, insufficient structural integrity, and lack whether a full EIA (versus a simplified one, or none of maintenance (Santos et al. 2012). at all) is required for a new installation. However, a Perhaps the most underappreciated issue is that the eco- capacity-based criterion alone is insufficient to define logical impacts of SHPs may be detectable at the level of expectations regarding the potential environmental im- multiple installations, so SHPs should be considered col- pacts of hydropower, especially considering that the lectively and not in isolation. In China, the cumulative thresholds are often arbitrarily defined. Moreover, by effects of hydrologic modifications and reduced connec- focusing exclusively on a single attribute of an instal- tivity in rivers associated with 31 SHPs exceed those lation (namely its generation capacity, the desired product caused by four LHPs on a per unit of energy comparison of hydropower), current environmental regulations could (Kibler and Tullos 2013). The presence of multiple obsta- run the risk of favoring less productive, ecologically

98 harmful projects over ones that are more productive lights the need to more rigorously quantify their cumula- and ecologically benign. SHP regulations must look be- tive impacts (Kibler and Tullos 2013). By addressing these yond capacity and incorporate flow alteration, inundation knowledge gaps, the role of science in informing energy area, impacts on habitat connectivity, and the cumulative policy and watershed management will be strengthened. effects of multiple installations. SHPs and LHPs are not competing components of the We recognize the many challenges of conducting hydropower sector. SHPs are broadly located in different detailed EIAs for every potential hydropower project, espe- parts of a basin (smaller rivers), financed by different cially in countries with limited financial and human entities (private investment), regulated by different resources. However, we advocate that EIAs should be com- policies, and respond to different energy demands (sup- pulsory during the licensing process for all SHP projects plying out-of- grid consumers) as compared to LHPs. Con given the diversity of facility sizes, operation modes, and sequently, in environmental policy making, there are geographic locations. Adjusting site-specific characteristics political and social motivations to retain the SHP versus to maximize electricity generation requires a wide range of LHP classification. For instance, capacity classifications technical solutions, which may create a myriad of environ- play a role in regulating the energy market and prices of mental changes (Egré and Milewski 2002). By considering electricity. By contrast, we argue for the removal of the SHPs as a homogeneous group in licensing frameworks, “small” and “large” modifiers describing hydropower policy makers are likely to underestimate the facilities’ plants when conducting environmental assessments and potential environmental impacts. EIAs can help to iden- mandating potential mitigation. A hydropower project tify specific impacts on the affected ecosystem, quantify artificially labeled as “small” should not automatically threats to biodiversity, address disruptions to coupled lead to a faster and less comprehensive licensing process. human–natural systems (eg subsistence fishing, access to agricultural lands, social and cultural values), and propose JJ Conclusions mitigation actions. Environmental assessments must additionally go beyond individual SHP projects and consider Electricity is a basic human need in modern societies, the broader watershed context before new construction is and SHPs have a key role to play in supporting electricity approved. The continued proliferation of SHPs necessi- policies that mandate the use of renewable resources. We tates systematic evaluations that take into account the have described the widespread proliferation of SHPs across cumulative effects of current and planned installations, a the world, and highlighted the potential for continued process that is overlooked by most present-day environ- growth of this sector into the future. The process by mental regulations. By incorporating elements of inte- which hydropower plants are classified as “small” is based grated watershed management (Wang et al. 2016) into exclusively on the measure of generation capacity, which hydropower strategic planning, managers can ensure adap- is arbitrarily chosen and inconsistently applied across tive, multifaceted, and watershed-scale evaluations based countries. Additionally, scientific evidence indicates that on the location and attributes of proposed projects. the environmental impacts of SHPs are substantial, but Ecological data are critically needed to support manage- existing policies and regulations appear to underestimate ment strategies and policy reforms that attempt to mini- these impacts. We argue for stricter and more informed mize the impacts of SHPs. The vast and decentralized regulations, integrated watershed management plans that nature of SHPs imposes major challenges to the acquisi- consider SHPs in aggregate, and readily available infor- tion and compilation of fundamental information, espe- mation systems to support strategic and impartial planning cially when compared to LHPs (Lehner et al. 2011; Zarfl for new construction. Until such regulatory mechanisms et al. 2015). We argue that the collation of basic attributes are in place, it would be wise to reconsider the current of present and future SHPs is fundamental to guide any pace of SHP expansion globally. meaningful strategic planning. Information on geographic location, dam height, reservoir area and volume, capacity, JJ Acknowledgements and operation mode could foster a baseline inventory without requiring substantial investment. This informa- We thank E Whattam for assisting in the literature tion can be used to prioritize projects that meet electricity search and R Lee for illustrations. JDO was supported demands while minimizing harm to the environment. by H Mason Keeler Endowed Professorship from the Several research questions need to be addressed in order School of Aquatic and Fishery Sciences at the University to understand the full impacts of SHPs. First, scientists of Washington (UW). TBAC was supported by UW must determine which factors, in addition to capacity, are and by a grant from CNPq (Science without Borders able to reliably delineate the environmental impacts of 203991/2014-1). hydropower. Second, given the recognized importance of headwater streams for watershed integrity (Meyer et al. JJ References 2007), strategic plans must identify the species and ecolog- Anderson D, Moggridge H, Warren P, and Shucksmith J. 2015. ical processes likely to be at risk from SHP development. The impacts of “run-of-river” hydropower on the physical and Third, the extreme density of SHPs in some areas high- ecological condition of rivers. Water Environ J 29: 268–76.

Anderson EP, Freeman MC, and Pringle CM. 2006. Ecological Hermoso V. 2017. Freshwater ecosystems could become the big- 99 consequences of hydropower development in Central America: gest losers of the Paris Agreement. Glob Chang Biol 23: 3433– impacts of small dams and water diversion on neotropical 36. stream fish assemblages. River Res Appl 22: 397–411. IRENA (International Renewable Energy Agency). 2017. ANEEL (Agência Nacional de Energia Elétrica). 2016. Banco de Renewable Energy Capacity Statistics. Abu Dhabi, United Informações de Geração. www.aneel.gov.br. Viewed 08 Aug Arab Emirates: IRENA. 2016. Izagirre O, Argerich A, Martí E, and Elosegi A. 2013. Nutrient Ansar A, Flyvbjerg B, Budzier A, and Lunn D. 2014. Should we uptake in a stream affected by hydropower plants: comparison build more large dams? The actual costs of hydropower meg- between stream channels and diversion canals. Hydrobiologia aproject development. Energy Policy 69: 43–56. 712: 105–16. Baker D, Bledsoe B, Albano C, and Poff N. 2011. Downstream Jesus T, Formigo N, Santos P, and Tavares G. 2004. Impact evalu- effects of diversion dams on sediment and hydraulic conditions ation of the Vila Viçosa small hydroelectric power plant of Rocky Mountain streams. River Res Appl 27: 388–401. (Portugal) on the water quality and on the dynamics of the Bellmore JR, Duda JJ, Craig LS, et al. 2016. Status and trends of benthic macroinvertebrate communities of the Ardena River. dam removal research in the United States. WIREs Water 2: Limnetica 23: 241–56. e1164. Kahn J, Freitas C, and Petrere M. 2014. False shades of green: Benchimol M and Peres CA. 2015. Widespread forest vertebrate the case of Brazilian Amazonian hydropower. Energies 7: extinctions induced by a mega hydroelectric dam in lowland 6063–82. Amazonia. PLoS ONE 10: 1–15. Kao SC, McManamay RA, Stewart KM, et al. 2014. New stream- Benstead JP, March JG, and Pringle CM. 1999. Effects of a low- reach development: a comprehensive assessment of hydro- head dam and water abstraction on migratory tropical stream power energy potential in the United States. Washington, DC: biota. Ecol Appl 9: 656–68. Wind and Water Power Program, US Department of Energy. Cada GF and Zadroga F. 1982. Small-scale hydroelectric power for GPO DOE/EE-1063. developing countries: methodology of site-selection based on Kibler KM and Tullos DD. 2013. Cumulative biophysical impact of environmental issues. Environ Conserv 9: 329. small and large hydropower development in Nu River, China. Copeman VA. 1997. The impact of micro-hydropower on the Water Resour Res 49: 3104–18. aquatic environment. J Chart Inst Water Environ Manag 11: Kubecka J, Matena J, and Hartvich P. 1997. Adverse ecological 431–36. effects of small hydropower. Regul Rivers Res Manag 13: 101–13. Cortes RMV, Ferreira MT, Oliveira SV, and Godinho F. 1998. Larinier M. 2008. Fish passage experience at small-scale Contrasting impact of small dams on the macroinvertebrates of hydro-electric power plants in France. Hydrobiologia 609: 97– two Iberian mountain rivers. Hydrobiologia 389: 51–61. 108. Csiki S and Rhoads BL. 2010. Hydraulic and geomorphological Lehner B, Liermann CR, Revenga C, et al. 2011. High-resolution effects of run-of-river dams. Prog Phys Geogr 34: 755–80. mapping of the world’s reservoirs and dams for sustainable Deemer BR, Harrison JA, Li S, et al. 2016. Greenhouse gas emis- river-flow management. Front Ecol Environ 9: 494–502. sions from reservoir water surfaces: a new global synthesis. Mbaka JG and Mwaniki MW. 2015. A global review of the down- BioScience 66: 949–64. stream effects of small impoundments on stream habitat conditions Dubois RB and Gloss SP. 1993. Mortality of juvenile American and macroinvertebrates. Can J Fish Aquat Sci 23: 257–62. shad and striped bass passed through Ossberger crossflow tur- McManamay RA, Oigbokie CO, Kao S-C, and Bevelhimer MS. bines at a small-scale hydroelectric site. North Am J Fish Manag 2016. Classification of US hydropower dams by their modes of 13: 178–85. operation. River Res Appl 22: 1–19. Dursun B and Gokcol C. 2011. The role of hydroelectric power and Meier W, Bonjour C, Wuest A, and Reichert P. 2003. Modeling contribution of small hydropower plants for sustainable devel- the effect of water diversion on the temperature of mountain opment in Turkey. Renew Energy 36: 1227–35. streams. J Environ Eng 129: 755–64. Egré D and Milewski JC. 2002. The diversity of hydropower pro- Meyer JL, Strayer DL, Wallace JB, et al. 2007. The contribution of jects. Energy Policy 30: 1225–30. headwater streams to biodiversity in river networks. J Am Erlewein A. 2013. Disappearing rivers – the limits of environmen- Water Resour Assoc 43: 86–103. tal assessment for hydropower in India. Environ Impact Assess Olden JD. 2016. Challenges and opportunities for fish conserva- Rev 43: 135–43. tion in dam-impacted waters. In: Closs GP, Krkosek M, and Ferreira FC, Souza UP, and Junior MP. 2015. Presence of riparian Olden JD (Eds). Conservation of freshwater fishes. Cambridge, vegetation increases biotic condition of fish assemblages in two UK: Cambridge University Press. Brazilian reservoirs. Acta Limnol Bras 27: 289–300. Oud E. 2002. The evolving context for hydropower development. Ferreira JHI, Camacho JR, Malagoli JA, and Guimarães SCJ. 2016. Energy Policy 30: 1215–23. Assessment of the potential of small hydropower development Ovidio M and Philippart JC. 2002. The impact of small physical in Brazil. Renew Sustain Energy Rev 56: 380–87. obstacles on upstream movements of six species of fish: synthe- Gido KB, Whitney JE, Perkin JS, and Turner TF. 2016. sis of a 5-year telemetry study in the River Meuse basin. Fragmentation, connectivity and fish species persistence in Hydrobiologia 483: 55–69. freshwater ecosystems. In: Closs GP, Krkosek M, and Olden JD Paish O. 2002. Small hydro power: technology and current status. (Eds). Conservation of freshwater fishes. Cambridge, UK: Renew Sustain Energy Rev 6: 537–56. Cambridge University Press. Pang M, Zhang L, Wang C, and Liu G. 2015. Environmental life Gleick P. 1992. Environmental consequences of hydroelectric cycle assessment of a small hydropower plant in China. Int J development: the role of facility size and type. Energy 17: 735– Life Cycle Assess 20: 796–806. 47. Poff NL and Hart DD. 2002. How dams vary and why it matters Gowans ARD, Armstrong JD, Priede IG, and McKelvey S. 2003. for the emerging science of dam removal. BioScience 52: 659–68. Movements of Atlantic salmon migrating upstream through a Premalatha M, Tabassum-Abbasi, Abbasi T, and Abbasi SA. 2014. fish-pass complex in Scotland. Ecol Freshw Fish 12: 177–89. A critical view on the eco-friendliness of small hydroelectric Hayes DB, Dodd H, and Lessard J. 2006. Effects of small dams on installations. Sci Total Environ 481: 638–43. coldwater stream fish communities. Am Fish Soc Symp 49: REN21 (Renewable Energy Policy Network for the 21st century). 1791–806. 2015. Renewables 2015 global status report. Paris, France: REN21.

100 Rosenberg DM, Bodaly RA, and Usher PJ. 1995. Environmental United Nations Industrial Development Organization. Vienna, and social impacts of large scale hydroelectric development: Austria/International Center on Small Hydro Power. who is listening? Glob Environ Chang 5: 127–48. Hangzhou, China: UNIDO/ICSHP. Santos JM, Silva A, Katopodis C, et al. 2012. Ecohydraulics Wu NC, Jiang WX, Fu XC, et al. 2010. Temporal impacts of a small of pool-type fishways: getting past the barriers. Ecol Eng 48: hydropower plant on benthic algal community. Fundam Appl 38–50. Limnol 177: 257–66. Stanley EH, Luebke MA, Doyle MW, and Marshall DW. 2002. Wu N, Tang T, Fu X, et al. 2009a. Impacts of cascade run-of-river Short-term changes in channel form and macroinvertebrate dams on benthic diatoms in the Xiangxi River, China. Aquat communities following low-head dam removal. J North Am Sci 72: 117–25. Benthol Soc 21: 172–87. Wu N, Tang T, Zhou S, et al. 2009b. Changes in benthic algal Tang X, Li Q, Wu M, et al. 2012. Ecological environment protec- communities following construction of a run-of- river dam. tion in Chinese rural hydropower development practices: a J North Am Benthol Soc 28: 69–79. review. Water Air Soil Pollut 223: 3033–48. Zarfl C, Lumsdon AE, Berlekamp J, et al. 2015. A global boom in Wang G, Mang S, Cai H, et al. 2016. Integrated watershed manage- hydropower dam construction. Aquat Sci 77: 161–70. ment: evolution, development and emerging trends. J Forest Res 27: 967–94. JJ Winemiller KO, McIntyre PB, Castello L, et al. 2016. Balancing Supporting Information hydropower and biodiversity in the Amazon, Congo, and Mekong. Science 351: 128–29. Additional, web-only material may be found in the WSHPDR (World Small Hydropower Development Report). online version of this article at http://onlinelibrary. 2016. World Small Hydropower Development Report 2016. wiley.com/doi/10.1002/fee.1746/suppinfo