Evaluating potential effects of solar power facilities on wildlife from an animal behavior perspective

Solar power is a renewable energy source with great potential to help meet increasing global energy demands and reduce our reliance on fossil fuels. However, research is scarce on how solar facilities affect wildlife. With input from professionals in ecology, conservation, and energy, we conducted a research‐prioritization process and identified key questions needed to better understand impacts of solar facilities on wildlife. We focused on animal behavior, which can be used to identify population responses before mortality or other fitness consequences are documented. Behavioral studies can also offer approaches to understand the mechanisms leading to negative interactions (e.g., collision, singeing, avoidance) and provide insight into mitigating effects. Here, we review how behavioral responses to solar facilities, including perception, movement, habitat use, and interspecific interactions are priority research areas. Addressing these themes will lead to a more comprehensive understanding of the effects of solar power on wildlife and guide future mitigation.

including perception, movement, habitat use, and interspecific interactions are priority research areas. Addressing these themes will lead to a more comprehensive understanding of the effects of solar power on wildlife and guide future mitigation.

K E Y W O R D S
animal behavior, concentrating solar power (CSP), conservation, conservation behavior, photovoltaic (PV) cells, research prioritization process, solar power, utility-scale solar energy (USSE)

| INTRODUCTION
As the global human population continues to grow, energy demand increases (IEA, 2019;Pazheri, Othman, & Malik, 2014). Although fossil fuels still dominate energy production, renewable energy sources are a rapidly expanding sector of the global energy market (Islam, Huda, Abdullah, & Saidur, 2018;USEIA, 2019). Renewable resources can help combat climate change, and with falling production costs, serve as an economical alternative to fossil fuels (IRENA, 2019). Most U.S. states now have Renewable Portfolio Standards and other policies that further incentivize production of renewable energy (NCCETC, 2020;NREL, 2019).
The number and size of utility-scale (e.g., >20 MW) solar energy facilities (hereafter solar facilities) have dramatically increased during the past 20 years (Figure 1; Hernandez et al., 2014); for example, the average utilityscale photovoltaic (PV) system installation size increased over 80% from 2010 to 2019 in the United States (NREL, 2020). Solar energy technologies typically fall into two main categories: (a) PV cells that convert sunlight into electrical current (Figures 1a and 2) concentrating solar power (CSP) which uses mirrors to focus sunlight to heat fluids that power steam turbines or generators (Figure 1b,c).
Our current understanding of the impacts of solar facilities on wildlife is limited, despite the pace and scale of its development. Environmental effects, such as soil erosion, changes in water use, and increases in local temperature, are well documented (Barron-Gafford et al., 2016;Hernandez et al., 2014;Moore-O'Leary et al., 2017). A few studies suggest that solar facilities could affect wildlife through exclusionary fencing, habitat destruction or alteration, and direct mortality (Table 1; Northrup & Wittemyer, 2013;Walston, Rollins, LaGory, Smith, & Meyers, 2016), but their relative scarcity highlights the need for additional research (see also Agha, Lovich, Ennen, & Todd, 2020). In particular, studies of wildlife behavioral response to solar facilities have been called for, including by working groups focused on bird interactions with solar facilities (ASCWG, 2020; ASWG, 2020); but such studies are largely still lacking from the literature (Lovich & Ennen, 2011;Northrup & Wittemyer, 2013).
Behavioral responses are often the most visible signs of detrimental effects, as behavioral shifts are usually an animal's first response to environmental change (Dimitri & Longland, 2018;Northrup & Wittemyer, 2013). Although direct mortality is the most obvious sign of negative impacts, large energy facilities may also impact individual fitness, as measured by survival and reproduction (hereafter "fitness"), resulting in population-level impacts that are harder to quantify without long-term demographic studies or using behavioral observations. For example, individuals could decrease mating behavior in response to increased disturbance (Holloran, Kaiser, & Hubert, 2010), stress levels (Lovich & Ennen, 2011), and pollution (Peterson et al., 2017). In addition, behavioral studies can offer approaches to understand the mechanisms leading to negative effects and to provide mitigative strategies. Animal behavior has been successfully utilized by wildlife and natural resource managers to mitigate problems and improve management strategies (Berger-Tal et al., 2011;Dimitri & Longland, 2018). For example, animal behavior has been used to understand and develop approaches to mitigate avian collisions at airports (Blackwell & Fernández-Juricic, 2013). It is imperative for the solar industry to incorporate behavioral research now, in a relatively early stage of the solar boom, to ensure solar power is sustainable for local wildlife populations and to avoid similar developmental and legal pitfalls that plagued the wind industry in its early boom (Brown & Escobar, 2007).
Using a multiphase research-prioritization process (see Supporting Information 1 for detailed methods) we implemented an online survey to ask professionals in the fields of ecology, conservation and energy to identify key behavioral research questions related to potential wildlife conservation issues at solar facilities (see Supporting Information 2 for full survey). We reduced and prioritized these questions at a 2019 workshop held by the Animal Behavior Society Conservation Committee (Supporting Information 1), and summarize here the emerging themes that resulted from this process ( Table 2).

| WILDLIFE PERCEPTION OF SOLAR FACILITIES
Solar facilities have the potential to deter, attract, or be imperceptible to individuals, all of which can lead to negative consequences for a variety of species (Kagan et al., 2014;Smith & Dwyer, 2016). Avoidance of solar facilities may lead to use of lower quality habitat or population fragmentation (Hernandez et al., 2014;Saunders, Hobbs, & Margules, 1991) and species attracted to solar facilities might be victims of ecological traps (Robertson & Hutto, 2006). When species attracted to facilities experience low survival or reproduction onsite, regional population dynamics could follow a source-sink pattern, affecting populations beyond site boundaries (Delibes, Gaona, & Ferreras, 2001). Alternatively, solar facilities may attract and provide high quality habitat for non-native or urban adapted species (Hufbauer et al., 2011;Tuomainen & Candolin, 2011). High population density of a few species could have cascading effects, potentially reducing food web integrity (Jessop, Smissen, Scheelings, & Dempster, 2012) or altering species' interactions (see below). Species unable to detect or avoid structures (e.g., power lines, glass windows) are at risk of collision and direct mortality (Bevanger, 1994).
At the core of the problem, we do not fully understand the mechanisms involved in wildlife perception of solar facilities or all the factors that influence avoidance or attraction (but see work by Horváth et al. (2010) and others on aquatic insect attraction to polarized light and solar panels). Individuals deterred by noise pollution might avoid facilities during construction and operation (Halfwerk & Slabbekoorn, 2015) and could also be affected by road noise from traffic associated with them. Individuals might be attracted to these sites because of microclimatic conditions, cover, water availability (e.g., evaporative cooling ponds; Walston et al., 2016), enhanced prey density, lighting, confusion of visual cues, or other potential factors (Dominoni et al., 2020). We also need to know if there is variation in perception and response to solar facilities within and between species and at different temporal scales, both seasonal and daily.
We can identify key behavioral responses by studying how species perceive solar facility structures (Kagan et al., 2014) relative to surrounding landscape elements. Ultimately, this process can allow for manipulation of F I G U R E 1 (a) An example of photovoltaic (PV) solar panels at topaz solar (550 MW; 4,700 acres). Photo by Pacific Southwest Region from Sacramento, U.S.-Solar Panels at topaz solar 1, Public Domain, https://commons.wikimedia.org/w/index.php? curid=36895794. Inset: aerial photo by Earth Observatory image by Jesse Allen, using EO-1 ALI data provided courtesy of the NASA EO-1 team. Public Domain, https://commons.wikimedia.org/w/ index.php?curid=38864327. (b) An example of a concentrating solar power (CSP) tower at Ivanpah Solar Electric Generating System (377 MW; 3,500 acres). Photo by Craig Dietrich-Flickr: Ivanpah Solar Power Facility, CC BY 2.0, https://commons. wikimedia.org/w/index.php?curid=28676343. Inset: aerial photo by Jllm06-Own work, CC BY-SA 4.0, https://commons.wikimedia. org/w/index.php?curid=42975801. (c) An example of a CSP parabolic trough at Solar Energy Generating Systems (SEGS; 354 MW; 1,600 acres). Photo by USA.Gov-BLM-Public domain stimuli and associated behavior to reduce mortality (sensu Blackwell et al., 2009 and citations therein). Birds, for example, can experience risk of mortality due to collision (i.e., direct contact with the solar facility), solarflux (i.e., birds are either burned or singed by exposure to the solar facility; Figure 2a), or become stranded F I G U R E 2 (a) Concentrating solar power (CSP) facilities can cause direct mortality to aerial species that fly into solar flare, such as this yellow-rumped warbler burned mid-air at Ivanpah (photograph by U.S. Fish and Wildlife Service, 2013, public domain). (b) CSP or PV facilities can create a "lake effect" (photograph by Kerry Holcomb, used with permission, Ivanpah Solar Electric Generating System, CA); water birds that mistakenly land on the hard surfaces can die on impact, become injured, or are unable to take off from terrestrial surfaces and ultimately die of exposure T A B L E 1 Examples of direct injury and mortality effects, as well as secondary mortality effects, on wildlife species that use the airspace and land covers at solar energy facilities. Noted effects are based on a select number of government and peer-reviewed literature sources, but not a complete survey or synthesis of the current literature (i.e., water birds that cannot take off due to lack of water; ANL & NREL, 2015). It is therefore important to understand how birds and other wildlife perceive solar facilities and why they are attracted, deterred, or fail to detect them. In addition to individual responses to cues generated by solar facilities, vulnerability will vary according to species' ecology and behavior. We discuss below how animal movement, breeding, foraging behavior, and interspecific interactions may influence population level responses to solar facilities.

| MOVEMENT AND HABITAT USE IN AND AROUND SOLAR FACILITIES
Many animals, particularly those living in arid environments where solar facilities are more common, are living at their physiological limits; any added movement may thus be costly (Vale & Brito, 2015). Whether and how movements are influenced by a solar facility will be determined by: (a) the trade-off of associated benefits and T A B L E 2 Key themes in animal behavior research that could improve our understanding of impacts of solar facilities on wildlife and potential solutions. These themes emerged from a multiphase research prioritization process (see Supporting Information 1) and the final list of priority research questions ( costs, (b) whether species are attracted or deterred by solar facilities, (c) whether a species is residential or migratory, and (d) the fitness impact of the responses.

| Resident species
Solar facility construction and operation directly and indirectly alter habitat use via functional habitat fragmentation, dispersal limitations, population isolation, and altered habitat quality (as previously reviewed in Lovich and Ennen (2011) 2019) or prey species that experience reduced predation , in most cases, modifications are likely to have negative impacts.

| Migratory species
Migratory animals are under escalating threat due to growth in human activity (Hardesty- Moore et al., 2018;Wilcove & Wikelski, 2008). Compared to other groups of species, migratory birds appear to suffer disproportionately higher mortality from solar facilities, particularly those located on migration routes and/or near breeding and wintering grounds (Walston et al., 2016). The greater abundance of insect prey attracted by the high structures and light (Diehl et al., 2016) likely attracts aerial insectivores, resulting in a higher risk to burning via solar flux from concentrated solar power (Figure 2a; McCrary et al., 1986;Kagan et al., 2014). Migratory water bird species are also susceptible because solar facilities may be perceived as waterbodies (a hypothesized "lake effect"), attracting them to land and injuring, killing, or stranding them in the process (Figure 2b; Kagan et al., 2014).

| Facility siting
The effects of solar facilities on wildlife may be exacerbated or mitigated through decisions about where to build them. Models have been developed at regional scales to identify areas that have both high potential for solar energy development and suitability for species of special concern (Phillips & Cypher, 2019), or high species richness (Thomas et al., 2018), representing potential conflict areas that should be avoided. These and other studies also identify priority areas for facility siting that minimizes the loss of high quality habitat (DRECP, 2020;Stoms, Dashiell, & Davis, 2013). While these models provide greatest benefit to resident species, research on migratory routes for aerial and terrestrial wildlife is critical to improve siting recommendations (e.g., Ruegg et al., 2014). The infrastructure necessary to operate solar facilities often extends far into the habitat, and effects of these structures on migratory wildlife have been documented in other energy sectors. For instance, mule deer (Odocoileus hemionus) abandoned former migration corridors as a result of oil and gas exploration and moved into suboptimal habitat, resulting in migration bottlenecks with no observed acclimation over several years (Sawyer et al., 2009). Reindeer (Rangifer tarandus) actively avoid power lines (Reimers et al., 2007;Vistnes et al., 2004), a behavioral response that could similarly alter migration routes for other ungulates. Gene flow in populations of desert bighorn sheep (Ovis canadensis nelsoni) is impeded by the presence of barriers, including roadways and large mining operations, resulting in rapid declines in genetic diversity (Epps et al., 2005). Minimizing these off-site impacts by siting facilities closer to existing infrastructure is important for mitigating effects on wildlife (Stoms et al., 2013).

| Foraging
Foraging involves a complex suite of behaviors, including detection of food sources, perceiving temporal and spatial cues about food availability, and food searching, choice, retrieval, and processing. Solar facilities might alter cues and predation risk assessment or disrupt normal search patterns via habitat change or construction of novel obstacles. Therefore, we must understand a species' trophic level (Fauvelle, Diepstraten, & Jessen, 2017;Moore-O'Leary et al., 2017) and the mechanisms underpinning its foraging decisions (e.g., olfactory cues; Schmitt, Shuttleworth, Ward, & Shrader, 2018) to estimate the impact of landscape alteration caused by solar facilities. Spatial knowledge, which is critical in foraging behavior, increases individual fitness (Spencer, 2012), and changes in spatial distribution of resources may impact species depending on their capacity to update such information. Assessments on the plasticity of cognitive mapping and role of memory in animal foraging decisions would contribute to our understanding about the impact of solar facilities. For example, bison (Bison bison) remembered and used information about location and quality of meadows to make movement decisions, building individual cognitive maps of their environment (Merkle, Fortin, & Morales, 2014). Studies of species affected by solar facilities measuring the effect of changes in the distribution and availability of resources on animal behavior can help predict impacts of development at a population level.

| Predation, antipredator behavior, and competition
Habitat modification can affect predator-prey dynamics (Dorresteijn et al., 2015;Hawlena, Saltz, Abramsky, & Bouskila, 2010) and competitive interactions between species (Berger-Tal & Saltz, 2019). At solar facilities, reflective surfaces of buildings and PV panels create polarized light pollution that attracts polarotactic organisms, including many insects (Horváth, Kriska, Malik, & Robertson, 2009). Insectivorous species might benefit from the increased availability of prey but trade off potential danger from collisions with reflective surfaces and increased competition for food. In the Mojave Desert, the population of urban-associated common ravens (Corvus corax) has increased with development, and they exert high predation pressure on threatened desert tortoise (Kristan & Boarman, 2003), which also face other impacts due to solar development (Lovich & Ennen, 2011).
Alternatively, PV panels or mirrors could serve as shelter for some animals against predators, especially aerial ones, and solar facility buildings and fences can also provide shelter and escape routes for smaller prey by excluding larger terrestrial predators . Increased vegetation near structures due to runoff (BLM & DOE, 2012) may be perceived as protective cover from predators (Jacob, 2008), but the vegetation may also make it more difficult to detect predators. Peripheral visibility has been shown to be valued by both mammals (Bednekoff & Blumstein, 2009) and birds (Bednekoff & Lima, 1998); in areas with reduced peripheral visibility, animals perceive a greater risk of predation and may modify their behavior in potentially maladaptive ways, such as increasing time allocated to vigilance over foraging.

| FUTURE RESEARCH AND DESIGNING SOLUTIONS
As evidenced by our research and those of others (Agha et al., 2020;Conkling, Loss, Diffendorfer, Duerr, & Katzner, 2020), more studies about the potential impacts of solar facilities on wildlife are needed to develop solutions. Documented efforts to deter wildlife from solar power facilities and other human-made structures include acoustic (Arnett et al., 2013;May, Reitan, Bevanger, Lorentsen, & Nygård, 2015;Swaddle, Moseley, Hinders, & Smith, 2016), visual (Martin, 2011;Goller, Blackwell, DeVault, Baumhardt, & Fernández-Juricic, 2018;Hausberger, Boigné, Lesimple, Belin, & Henry, 2018), and tactile deterrents (Ho, 2016;Seamans, Martin, & Belant, 2013). Evaluation of the effectiveness of such deterrents, however, is often limited or inconclusive (e.g., Dorey, Dickey, & Walker, 2019), and may not address why individuals are attracted to the facilities or collide with facility structures in the first place. A more effective approach may be to understand wildlife perception of solar facilities and minimize features that attract them (e.g., Horváth et al., 2010), or modify features so that wildlife detect them and avoid collisions, burning and singeing. For instance, we can better understand how wildlife visually or otherwise perceive solar facilities by: (a) quantifying key properties of the sensory systems of species that experience high mortality, (b) use this information to quantify the degree of conspicuousness of solar panels and other structures from the species' sensory perspective, then (c) modify the properties of the solar panels to enhance or reduce their conspicuousness, and (d) measure behavioral responses to these modifications (Blackwell & Fernández-Juricic, 2013;Fernández-Juricic, 2016). For example, Horváth et al. (2010) tested the attraction of several aquatic insect species to PV solar panels with various modified features and found that white-framed and white-gridded panels were less attractive than black panels.
Our survey identified several research priorities for designing solutions focusing on where and how solar facilities can be built to minimize influences on behavior and fitness (Table 2 and Supporting Information 1). Another overarching question identified, while not specific to behavior, was whether facility designs should be exclusionary or permeable to wildlife. Some solar facilities are currently evaluating how to co-manage wildlife and PV panels by making them more permeable (e.g., Cypher et al., 2019;Wilkening & Rautenstrauch, 2019). Nevertheless, the answer to this question is likely complex and specific to geography and species (see also Moore-O'Leary et al., 2017).
With regard to assessing and minimizing impacts of solar facilities on wildlife, our workshop identified the need for more purposeful study designs to begin addressing these priority questions (Table 2). Ideally, a before-after control-impact design is desirable; whereby, key behaviors are studied before and after the solar facility is developed, both at the facility location and at control sites (Conkling et al., 2020;Lovich & Ennen, 2011). While this rarely happens (see Agha et al., 2020), such design is the most powerful way to isolate the effects of a solar facility on behavior while controlling for other spatial and temporal variation. Experimental studies assessing impacts of different design features (such as panel height and spacing, corridor placement and size, and vegetation treatment), in addition to studying behavior at different distances from solar facilities, are also necessary to minimize detrimental effects on wildlife.

| CONCLUSIONS
Development of utility-scale solar facilities is expected to continue at a rapid pace (USEIA, 2019). There is an urgent need to address how to better locate, design, and operate solar facilities to mitigate potential negative effects on wildlife populations. We have highlighted major research themes addressing how approaches using animal behavior can be utilized to study wildlife-solar facilities interactions and how they could lead to solutions to reduce negative effects. Similar to how those in the wind energy industry have worked with animal behaviorists to reduce wildlife fatalities (e.g., Cryan et al., 2014), finding such solutions will need collaboration across industry, research, and management agencies. This can be achieved by forming working groups that can bring together entities from solar power facilities, wildlife agencies, and academia to determine shared research goals and to facilitate access to solar facilities, research permitting, and research funding opportunities (e.g., Bats and Wind Energy Cooperative, 2020).

ACKNOWLEDGMENTS
The authors thank the Animal Behavior Society for the opportunity and funding to organize a workshop (entitled "Conservation Behavior Workshop: Implications of Solar Power on Wildlife Conservation") at the Animal Behavior Society Conference 2019 (Chicago, IL), and the funding to publish this manuscript. The authors thank everyone who responded to the online survey and participated in the workshop, especially Dr Thomas Dietsch and Peter Sanzenbacher (U.S. Fish and Wildlife Service). The findings and conclusions in this article are those of the author(s) and do not necessarily represent the views of the U.S. Fish and Wildlife Service.

CONFLICT OF INTEREST
The authors declare no conflicts of interest.
AUTHOR' CONTRIBUTIONS Rachel Y. Chock, Barbara Clucas, and Elizabeth K. Peterson: Organized the workshop that resulted in this study and coordinated the manuscript. All authors participated in the workshop and extensively contributed to the writing and revision of the manuscript.

DATA AVAILABILITY STATEMENT
Survey questionnaire and results from the workshop are freely available and included as Supporting Information.

ETHICS STATEMENT
The survey was approved by the Humboldt State University Institutional Review Board (IRB# 18-161).