Journal list menu
The influence of artificial light on stream and riparian ecosystems: questions, challenges, and perspectives
Corresponding Editor: D. P. C. Peters.
Abstract
Artificial light at night is gaining attention for its potential to alter ecosystems. Although terrestrial ecologists have observed that artificial light at night may disrupt migrations, feeding, and other important ecological functions, we know comparatively little about the role artificial light might play in disrupting freshwater and riparian ecosystems. We identify and discuss four future research domains that artificial light may influence in freshwater and associated terrestrial ecosystems, with an emphasis on running waters: (1) dispersal, (2) population genetics and evolution, (3) ecosystem functioning, and (4) potential interactions with other stressors. We suggest that future experimental and modeling studies should focus on the effects of different spectral emissions by different light sources on freshwater organisms, the spatial and temporal scale over which artificial light acts, and the magnitude of change in light at night across the landscape relative to the distribution of running and standing waters. Improved knowledge about the effects of artificial light on freshwater ecosystems will inform policy decisions about changes to artificial light spectral emissions and distributions.
Introduction
Human activities influence and have modified the majority of the Earth's ecosystems (Vitousek et al. 1997). Freshwater ecosystems are especially affected, both because they accumulate and integrate the effects of activities within their catchments, and because they have always been preferred sites for human activities (Ricciardi and Rasmussen 1998, Dudgeon et al. 2006, Balian et al. 2008).
The effects of chemical pollution (Likens et al. 1996), alteration to natural flows (Poff et al. 1997) and nutrient cycles (Turner and Rabalais 1991), invasive species (Ricciardi and Rasmussen 1998), increasing urbanization (Morely and Karr 2004), and loss of riparian margins (Sweeney et al. 2004) on freshwater ecosystems have influenced policy decisions for the past 40–50 years (e.g., the USA Environmental Protection Agency's Clean Water Act of 1972). In contrast, the influence of artificial lighting as a human-induced impact affecting freshwater systems has only been recognized in the past 10 years or so (Moore et al. 2000, Longcore and Rich 2004, Moore et al. 2006, Nightingale et al. 2006), and there are still many gaps in empirical knowledge. This is despite the fact that the use of artificial lighting is now widespread and has increased over the past century (Holden 1992). While Cinzano et al. (2001) reported that approximately 67% of Americans and 20% of people world-wide now live in locations where Milky Way is no longer visible due to interference from artificial light sources, the wider effects of artificial light on other organisms and on ecosystems are poorly quantified. While many studies have focused on the control of natural light on biorhythms (Bishop 1969, Grau et al. 1981), few have looked at the potential of artificial light as a disrupter of these rhythms (Moore et al. 2000). This is surprising as approximately 30% of vertebrates and 60% of invertebrates are nocturnal (Hölker et al. 2010a) and could, therefore, be highly influenced by the presence of artificial light.
Longcore and Rich (2004) and Navara and Nelson (2007) presented broad reviews of artificial light and summarized a range of evidence, yet over two thirds of their examples relate to terrestrial organisms. Both Moore et al. (2006) and Nightingale et al. (2006) identified some effects of artificial light on lakes and fish, but in general, freshwater ecosystems are poorly represented in the current literature. An initial search of Web of Science (13 October 2011) of peer-reviewed literature using various terms relating to human alterations and ecosystems revealed a noticeable lack of research on artificial light and freshwater systems, especially when compared to other common pressures to which these systems are subjected (Table 1). This is despite freshwaters having high biodiversity and being disproportionally affected by species loss. Globally, freshwaters are inhabited by more than 125,000 known species, and even though freshwaters cover only about 0.8% of the Earth's surface, they are home to about 9.5% of all animal species, and one-third of all vertebrates (Balian et al. 2008). Nevertheless, there have been some seminal contributions to our understanding, for example in the use of artificial lights to: (1) increase fish growth rates in hatcheries (Boeuf and Le Bail 1999), (2) understand how it influences zooplankton movements (Moore et al. 2000), and (3) guide fish around dangerous in-stream structures (Johnson et al. 2005).
Here, we attempt to redress the balance in available literature to date by focusing on freshwaters, and in particular streams with their associated riparian margins, defined as areas that are “transitional semiterrestrial areas regularly influenced by fresh water, usually extending from the edges of water bodies to the edges of upland communities” (Naiman et al. 2005:2). We give special attention to adult aquatic insects, as they represent a key in the exchange of nutrients between stream and riparian systems (Richardson et al. 2010).
Our goal is to illustrate how artificial light influences species interactions and processes in stream and riparian ecosystems, and to stimulate research in an area that we consider of major importance for their future conservation and management. Ecologists have only recently started to acknowledge the alteration of the nightscape as a major concern in conservation policy and freshwaters are no exception (Rich and Longcore 2006, Hölker et al. 2010a).
Research Domains
We begin by presenting four major research domains relating to the ways artificial light can act on stream and riparian ecosystems, through altering: dispersal, population genetics and evolution, ecosystem functioning, and interactions with other common stressors; and then outline a range of key research questions which need addressing.
Dispersal
There is evidence that artificial lights located near streams change the behavior of adult aquatic insects as they disperse through the terrestrial environment. Eisenbeis (2006) proposes three different ways for artificial lights to trap flying insects (Fig. 1). The first is through fixation or captivity effects (Fig. 1A). Here insects located near lights fly directly to them and are killed immediately, or they circle close to the light and are unable to leave eventually dying from exhaustion, predation, or heat. The lights may also induce settling behavior that incapacitates the insects, rendering them easy targets for predators. The second mechanism is the crash barrier effect (Fig. 1B), where insect dispersal and migration are impeded by running into a “barrier” of lights, such as a row of street lights. The final mechanism is termed the vacuum cleaner effect, whereby insects from a large area are attracted to a nearby light source. However, these are only hypotheses and carefully designed experiments are needed to determine how much of an effect these mechanisms actually play in disrupting aquatic insect dispersal.
Eisenbeis (2006:281–304) proposes three different ways for artificial lights to trap flying insects. Two are shown here: fixation (dashed line), captivity (dotted line), and directly lethal (solid line) effects (A) and the crash barrier effect (B). Fixated insects do not suffer mortality directly from artificial light, but are stunned by it and are easy targets for predators or fail to engage in basic behaviors, such as reproduction. Captive insects fly to the light and circle around it endlessly until they die of exhaustion or are consumed by predators. The crash barrier effect is the result of a row of lights (like those lining a street) preventing the dispersal of insects through their attractive properties.
Studies comparing different trapping techniques provide evidence for the vacuum cleaner effect. These studies illustrate that light traps differentially capture certain insects (e.g., Trichoptera) more readily than other kinds of traps (e.g., Collier and Smith 1998). While the height of streetlights is designed to maximize safety for car drivers, lights that are used along walking and bike paths, as well as those used for decorative purposes could be adjusted to attract fewer insects, if we can predict which heights have the highest concentrations of insects based on landscape features and insect species. Svensson (1974) found that light traps at 11 and 50 m height captured fewer trichopterans than those at 1 m due to the propensity of several species to fly low to the ground, suggesting that higher lights might trap fewer insects than lower lights, but this has yet to be generalized across taxa and habitats.
Overall, research on insect dispersal, especially adult aquatic insect dispersal, is extremely limited. Part of this lack is that it is very difficult to rigorously study insect dispersal. Stable isotope and elemental markers are potentially valuable tools, as is the increasing use of genetic analysis (Smock 2007); however, capturing insects in substantial numbers generally requires the use of light or pheromone traps. While these methods are adequate for most studies of aquatic insect dispersal, light traps cannot be used in studies of artificial light as they obviously create a confounding factor. In addition, while population genetic analyses hold promise for longer-term studies and determining if populations adapt to artificial light, they are not really useful as a tool for short-term dispersal studies.
It is also unclear how much the dispersal of adult aquatic insects matters for the population dynamics of these organisms. Masters et al. (2007) found that the recovery of the benthos from acidification is not limited by adult dispersal. Furthermore, Bunn and Hughes (1997) calculated that it is likely that populations of Tasiagma spp. (Trichoptera) in a reach are maintained through the reproduction of only 3–12 females per generation. It is not obvious how much of an effect land use changes have on adult aquatic insect dispersal (Petersen et al. 2004); however, studying the effects of artificial light on insect dispersal will likely further this field.
Clearly, we need to come up with new and innovative ways to study aquatic insect dispersal. One possibility is to use Malaise traps to capture individuals marked with fluorescent dyes or stable isotope tracers (Macneale et al. 2005). Conducting more basic studies of aquatic insect dispersal will help those studying the effects of artificial light to develop hypotheses (e.g., the effect of light height, light distances from streams) more effectively.
Aquatic insects are not the only stream organisms that may have their dispersal interrupted by the addition of artificial lights. It is well established that the migration of Pacific salmon species (Oncorhynchus spp.) can be slowed or stopped by the presence of artificial lights (Nightingale et al. 2006). Furthermore, exposure to constant light can decrease smoltification and increase the deterioration in body condition associated with smoltification in chinook salmon (O. tshawytscha) (Hoffnagle and Fivizzani 1998). This might be due to the synchronization of downstream migration with the new moon; however, it is possible that the lunar timing of downstream migration is stock-dependent (Hoffnagle and Fivizzani 1998). It is likely that any species that uses lunar cycles to cue migration or dispersal will be disrupted by the addition of artificial lights (see Key Research Questions; Fig. 2).
Sky glow eliminates monthly variation in light levels. The solid line shows the natural light provided by a full moon in a temperate region. The dashed line is the light level measured in the center of Berlin on a clear night and is roughly equivalent to a full moon, while the dotted line is the light level in the center of Berlin on a cloudy night and is roughly four times greater (Kyba et al. 2011). The y-axis is logarithmic.
Population genetics and evolution
To our knowledge, no one has yet experimentally investigated the possibility that artificial light can act as an evolutionary force in freshwater or riparian species. However, its potential to influence evolution has received attention from Moore et al. (2006) and Nightingale et al. (2006).
Artificial light at night could reduce effective population sizes through the direct loss of individuals, reproductive failure, or changes to sex ratios. The direct mortality of individuals is probably most likely in the case of aquatic insects; either through the attraction of the adults to lights (Scheibe 2003, Eisenbeis 2006), or increased predation through improved predator vision. However, mid-trophic fish species could also suffer higher rates of predation under artificial light (see: Ecosystem functioning: Food webs). Reproductive failure could be due to the inability to locate suitable mates, as in the case of several amphibian species (Longcore and Rich 2004). Aquatic insects are again likely to suffer from changes to sex ratios, as there are often biases in light trap catches, depending on the species (Waringer 1989).
Reduction in effective population sizes will lead to less genetic diversity and possibly genetic drift; leaving a population with insufficient variation to adapt to future stressors, and therefore is a major concern for species conservation (Lande and Barrowclough 1987). If some populations are eliminated, it could result in reduced gene flow across the range of some species, with the potential to lead to the diversification of populations and potentially even speciation.
There is already some evidence that other environmental stressors alter genotype frequencies in a population. Populations of a common aquatic insect (Chironomus riparius) that were exposed to a chemical stressor (tributyltin) in a laboratory study had increased rates of larval mortality and reduced genetic variation (Nowak et al. 2009). This result was especially significant because the changes were seen in neutral markers, not in response genes, and therefore represent a true reduction in effective population size. Conversely, mosquitoes living on an arid slope showed increasing diversity (due to higher rates of recombination and mutation) as a result of exposure to greater environmental stress, such as increased temperatures and solar radiation, than those living on a humid slope of the same valley (Nevo 2001). Furthermore, females from the arid slope showed an increased tendency to mate with males that were also from the arid slope, potentially leading to sympatric speciation between the two groups (Nevo 2001). While it might be difficult to forecast which species will have increased or decreased genetic diversity, artificial light could also change the frequency of heritable behaviors that could influence the evolution of organisms.
Mating and reproductive behaviors in freshwater species are likely to be influenced by artificial light (Moore et al. 2006, Nightingale et al. 2006). Sexual selection for traits that are visually stimulating could increase or decrease with exposure to artificial light, depending on the spectral qualities of the light and species' visual sensitivities. For instance, cichlid fishes undergo strong sexual selection that favors brightly colored individuals and has driven speciation events in populations in clear water that allows plenty of light (Seehausen et al. 1997). The effects of artificial light on sexual selection could be especially interesting and unpredictable, given the common use of high pressure sodium lamps, which have a very limited emission spectra and could prevent females from recognizing male color patterns (Fig. 3). This has taken place in Lake Victoria, where turbidity from eutrophication reduces the spectral range of light entering the water to wavelengths that are similar to the emission spectra of high pressure sodium lamps, and reduces female selectivity based on color (Seehausen et al. 1997). Similarly, guppy habitat specialization has been driven by a combination of diverse ambient light conditions, predation, and sexual selection (Endler 1992). The introduction of artificial light to these streams could lead to the visual homogenization of these environments, which could lead to reduced speciation as well as increasing susceptibility to predation.
The light sensitivities of various animals are displayed against a background of wavelengths that humans perceive as visible light (A). The black ovals represent the peak sensitivities for each organism; note that some organisms have sensitivities in the UV range. The dashed vertical lines designate the limits of the UV (10–400 nm) and human-visible light (390–750 nm) (Menzel and Blakers 1976, Smith and Macagno 1980, Lythgoe 1984, Loew and Wahl 1991, Fratzer et al. 1994, Hawryshyn and Hárosi 1994, Vorobyev and Osorio 1998, Gal et al. 1999, Wiltschko and Wiltschko 1999, Sillmann and Dahlin 2004). The wavelengths of light emitted from various artificial light sources (B) are highly variable, with some emitting light over a broad spectrum and others having only a few narrow peaks.
Other behaviors that could be influenced by artificial light that are potentially important to evolution are feeding behaviors. Some spiders are more likely to build their webs in close proximity to artificial light to take advantage of the increased densities of insects found at lights (Heiling 1999). If there is a genetic basis for this behavior, then the presence of artificial light could very well contribute to the evolution of this species. Ultimately, any behavior that could be altered by artificial light and is under genetic control could allow artificial light to change the evolution of a species exhibiting such a behavior.
It is also important to consider the effect of artificial light in combination with species interactions in driving rapid evolutionary change, which could lead to altered ecological dynamics, e.g., different guppy phenotypes result in altered ecosystem structure and function (Schoener 2011). To test if artificial light causes rapid evolution of exposed organisms, researchers could hatch diapausing copepod eggs that were laid before artificial light became widespread. The feeding and diel vertical migration (DVM) behavior of pre-lighting and modern copepods could then be compared (Hairston et al. 1995). To determine what percentage of the behavioral change is really due to evolution, and not some other ecological factor, genetic techniques should be used to identify genes that are likely responsible for the observed behavior and then tested to ensure that they are responsive to altered light conditions and change organism behavior (Hairston et al. 2005, Fussmann et al. 2007). Furthermore, it will be beneficial to establish if any of these rapid evolutionary changes results in genetic isolation, and eventually, speciation (Hendry et al. 2007).
Ecosystem functioning
As previously addressed by Moore et al. (2000), Longcore and Rich (2004), and Moore et al. (2006), we expect that modified lighting regimes will lead to a range of whole freshwater ecosystem changes and also influence the linkages between freshwater and riparian ecosystems. Of particular interest is how artificial light could alter the exchange of organic matter between stream and riparian systems. Artificial light could influence ecosystems in ways that might be unexpected from single species studies, e.g., by changing species interactions, especially predator-prey interactions, and therefore have important conservation implications (Wooten et al. 1996).
Primary production
Primary production is a key ecosystem process controlled by light. To our knowledge, only one study has found evidence that riparian vegetation could be influenced by the presence of artificial light at night (Cathey and Campbell 1975). Their work illustrated that trees and shrubs exposed to streetlamps, particularly incandescent or high pressure sodium luminaires, may have longer growing periods, earlier leaf-out and later leaf fall times than those in darker environments (Cathey and Campbell 1975). This may have a range of bottom-up effects. For example, earlier leaf-out could cause earlier inputs of terrestrial insects (that use riparian vegetation as habitat) to freshwater systems, but only if terrestrial insects are able to use this new habitat resource. Later leaf fall could result in a mismatch of resources and consumers, as detritivorous aquatic invertebrate taxa might have evolved to match the timing of the allochthonous inputs of leaves with critical life stages (Hershey and Lamberti 1998:169–199). However, substantial changes in leaf-out/fall and growth are unlikely unless artificial lights are present with warmer temperatures that allow for a longer growing season (Cathey and Campbell 1975). While this situation is currently unlikely in temperate climates, global temperatures are projected to increase by 0.6–6.4°C in the next 90 years, with greater warming in northern temperate regions (IPCC 2007), which would increase the chances that artificial light might influence riparian vegetation. The effects of increased temperatures and light could be studied in urban areas that not only have increased levels of artificial light, but also artificially high temperatures due to the heat island effect (Oke 1973).
Food webs
Light is an important cue for both predator avoidance and feeding in freshwater systems. Aquatic invertebrates in lotic systems drift at light levels below 10−3 lux (at 400–535 nm) to avoid predation by fish (Bishop 1969). However, Atlantic salmon have been shown to change foraging strategies below light levels of 10−1 lux, moving to areas of slow-moving water that, while not as rich in prey, allow more time for identification of prey items and night-time foraging (Metcalfe et al. 1997). Light adaptations are also evident in lentic environments, where zooplankton engage in DVM in the water column to feed on phytoplankton during the night when they are less visible to predators (Young and Watt 1996). Moore et al. (2000) were able to detect a decrease in the amplitude of DVM in Daphnia retrocurva as a result of artificial light from a nearby city, by monitoring the vertical migration inside darkened versus clear enclosures. Light intensity also had a significant influence on the ability of vendace (Coregonus albula) to feed on Daphnia magna, with declining efficiency down to a threshold of 0.05 lux (Ohlberger et al. 2008). On the other hand, a decrease in feeding movements to avoid artificial light has been observed in vendace (Schmidt et al. 2009). These studies suggest that artificial light can result in altered food webs in lentic systems, leading to increased algal biomass as zooplankton spend less time in the upper euphotic water column feeding on algae (Moore et al. 2000, Moore et al. 2006). Lotic systems could see higher relative abundances of armored grazers, such as glossosomatid caddisflies or snails, as invertebrates with less physical protection, such as mayflies, are eliminated through heavy predation (McNeely et al. 2007). In this case, there would eventually be a reduced number of invertebrates available to fish predators, but if there are adequate numbers of protected invertebrate grazers, they would likely control lotic algal standing biomass.
We expect that artificial light at night not only influences freshwater food webs (Fig. 4), but also the exchange of materials between stream and riparian environments (Richardson et al. 2010; Fig. 5), which can be mediated by predators (Baxter et al. 2004). Accordingly, one key question here is how artificial light changes predator-prey relationships. Some species might be able to exploit artificial light to extend foraging opportunities, at least in the short-term (Moore et al. 2006, Nightingale et al. 2006). One example of this is the spiders who build their webs near light sources (Heiling 1999). However, foraging benefits, if they exist, may be short-lived due to resulting reductions in prey populations (Beier 2006). This will probably depend on the trophic structure of specific food webs, as apex predators will benefit more than mid-trophic species that have to avoid predation themselves.
A schematic of a stream ecosystem under natural nighttime conditions (A), and the same system under the influence of artificial light (B). Note the predicted shifts as light is introduced from A to B in the positions of small fishes seeking cover, suppression of streaminvertebrate drift, adult insects attracted to lights, and bats shifting their foraging efforts near the lights. Refer to Ecosystem functioning: Food webs in the text for details.
Theoretical changes in carbon flux from freshwater to terrestrial systems, in the form of insects. Aquatic insects will be attracted to lights very near aquatic bodies in high densities, while lights that are farther away will increase the inland dispersal of aquatic insects, up to some threshold (A). Similarly, there will be higher inputs of terrestrial insects to the aquatic system when lights are close to the aquatic environment but these inputs will decrease quickly as the light location moves inland (B).
Patterns of invertebrate drift and fish feeding are both likely to change under the influence of artificial light (Moore et al. 2006, Nightingale et al. 2006). If fish are able to feed much more efficiently on drifting insects, it could result in a decrease of emerging aquatic insects. However, light is known to depress drift rates (Bishop 1969); if fish are more active under artificial lights but prey is less available, fish could suffer from increased energetic demands. Conversely, the number of terrestrial invertebrates entering the stream and available for fish to prey on could also change. Under natural conditions, terrestrial insects are an important allochthonous resource for fish (Fig. 4A). Kawaguchi and Nakano (2001) found that terrestrial insects contribute about 50% of the total annual prey consumption of salmonids in some Japanese streams, while about 84% of the consumption in a cyprinid (Alburnus alburnus) in a German lake comes from terrestrial sources (Mehner et al. 2005). In the presence of artificial light near a waterbody, terrestrial insects could become an even more important food source for fish. On the other hand, juvenile and other vulnerable fish might retreat to overhangs and reduce foraging efforts in order to avoid predation (Nightingale et al. 2006; Fig. 4B).
While adult aquatic insect flight in a dark riparian forest might normally be restricted to areas immediately adjacent to streams (Petersen et al. 1999), insects may cluster around artificial lights located in floodplains (Figs. 1, 5). Many aquatic insects emerge at night (Tobias 1967, Jackson 1988, Pinder et al. 1993), and are therefore vulnerable to attraction to artificial lighting while in their adult phase. We hypothesize that as the distance of an artificial light source from a water body increases, the proportion of freshwater carbon transferred to the terrestrial ecosystem increases relative to a riparian system that does not have lights, as aquatic insects are attracted further into the terrestrial system (Fig. 5A). Preliminary support for this hypothesis comes from Kovats et al. (1996) who found adult caddisflies 5 km inland when using light traps. Conversely, we predict the amount of terrestrial carbon contributed to a freshwater system through terrestrial invertebrates will decrease as the distance of an artificial light source to a water body increases. For instance, a light situated on a dock will draw terrestrial insects to the water body, while lights from a road running parallel and several hundred meters away from a water body will draw terrestrial insects away from the water (Fig. 5B). This will create areas that are highly dense in resources for insectivorous organisms, while creating other areas that are depauperate. Outcomes of this process may be an increase in competitive interactions between insectivores (Rydell 2006) and also an increased transfer of freshwater resources to terrestrial consumers.
Interaction with other stressors
There is a growing concern about how environmental stressors might interact with each other, and in fact, an entire issue of the journal Freshwater Biology (see Ormerod et al. 2010) was dedicated to this topic. However, the specific ways that artificial light might interact with other common urban stressors have not yet been described in the peer-reviewed literature. As artificial light most frequently occurs in urbanized areas, its effects may be confounded with other urban stressors, making it impossible to determine how much a role artificial light has played in declines in biodiversity and ecosystem functioning. Artificial light may already play a major role in changing organism behavior and ecosystem functioning. However, to fully understand its importance, we must elucidate how it interacts with other stressors in freshwater and riparian ecosystems. Does light pollution act synergistically with other stressors to increase the stress experienced by organisms, or does it potentially lessen the effect of some stressors? How artificial light interacts with other stressors will help prioritize what areas are most important to protect. Dudgeon et al. (2006) enumerated the five major threat categories to freshwater ecosystems as overexploitation, water pollution, habitat degradation, species invasion, and flow modification. Of course, another major threat to freshwater ecosystems is climate change. Artificial light has the potential to interact with all of these threats. By conducting carefully designed studies to understand the interaction between artificial lighting and the threats mentioned in Dudgeon et al. (2006), we will be able to develop a model for when artificial light is likely to do the most harm and be carefully controlled, or conversely, when it could be used as a mitigating factor for some other stressor.
In this section, we explain the ways artificial light could combine with changes to temperature regimes, increased chemical pollution and urban development, altered flow regimes, and increased nutrient concentrations. We also describe how the effects of artificial light might be masked by the presence of other stressors and may not become apparent until the other stressors are removed.
One potential concerns is for light to interact with other common urban stressors, such as temperature and pollution, to interfere with migration and dispersal. For example, some fish have been shown to become disoriented when swimming near lights (Tabor et al. 2004, Nightingale et al. 2006), which they are more likely to encounter when traversing urban areas that also contain other stressors. In the absence of light, migratory fish, such as salmonids, travel quickly through large rivers (Økland et al. 2001) that are more likely to have sub-optimal temperatures or increased pollutants, but the disorientation caused by urban lights could increase the time these fish spend in polluted environments and, as a result, increase their risk of mortality (McCormick et al. 1998).
The interaction of artificial light and other urban stressors could also alter patterns of the dispersal of riparian obligates, such as adult aquatic insects. For instance, the presence of culverts has been shown to reduce the upstream flight of adult caddisflies (Blakely et al. 2006). These culverts are usually installed to allow roads to pass over small streams, leading to a high probability of street lighting being associated with culverts. The street lighting would most likely run perpendicular to the stream (Fig. 6), leading the insects farther away from the stream. We hypothesize that this will lead to decreased dispersal and gene flow, and potentially the elimination of up-stream populations; however, it is possible that these lights could draw the insects over to a neighboring small watershed and, as a result, enhance genetic exchange. Similarly, Málnás et al. (2011) found that a bridge reduced the upstream flight of the mayfly Palingenia longicauda on a river in Hungary. At least part of the disruption was caused by polarized light reflecting off the surface of the bridge, which enticed gravid females to oviposit there (Horváth et al. 2009, Málnás et al. 2011).
Artificial lights located near culverts intersecting streams may strengthen their disruption of aquatic insect upstream flight (A) or mediate it (B), depending on their location. The white arrow represents the direction of streamflow, the dashed line is the flight path of the aquatic insects, the yellow circles are lights and the black line is a culvert.
The construction of dams has led to altered flow regimes, often with a dampening of pre-dam high flows. These high flows can serve as a signal to cue migration or spawning events (McCormick et al. 1998, Bunn and Arthington 2002). Normally, light is also a strong Zeitgeber for these behaviors (Grau et al. 1981, Greenstreet 1992), but where artificial lighting and flow alterations occur, there could be a complete loss of external cues for these behaviors. This could lead to asynchronous migration and spawning events, and ultimately result in lower population sizes.
While flow modifications are largely a concern of stream environments, increasing loads of nitrogen and phosphorus pollution is a common problem across all freshwater systems (Carpenter et al. 1998). Areas with increased nutrient loading that are also exposed to artificial light at night could be at an increased risk for algal blooms, largely as a result of night-time light altering the behavior of grazing macroinvertebrates (Moore et al. 2000, Moore et al. 2006). Other common pollutants in freshwater ecosystems could also interact with artificial light, most resulting in further reductions of biodiversity. However, bright artificial light could mitigate effects of pollutants that degrade under light exposure.
In restoration efforts, common urban stressors might act in concert to hide the negative effects of artificial light. For instance, water quality was the limiting factor in fish survival and reproduction in a central European river system. However, after decades of efforts to improve water quality, hydromorphological degradation then emerged as the main obstacle to further ecological improvement and freshwater diversity (Borchardt et al. 2005, European Commission 2007). Improving degraded habitats became important once pollutants and oxygen stress had been eliminated; similarly, after degraded habitats have been improved artificial lights could prevent a restoration site from achieving full functionality. This is important to consider as freshwater and riparian ecosystems that have undergone successful restoration often become attractive places for recreation (Woolsey et al. 2007). As recreational uses of these areas increase, user groups might call for the installation of artificial lights, particularly along biking and running paths in temperate zones with long periods of dark during winter months.
Key Research Questions
We have identified three main general questions facing researchers in artificial light that deserve more attention. These include understanding how different spectral qualities of various sources of artificial light, spatial and temporal scales over which artificial light acts, and the magnitude of changes in light influence organisms and ecosystems (Table 2).
Diverse organisms have sensitivities in different parts of the light spectrum, and various artificial lighting sources emit very distinctive wavelengths of light (Fig. 3). Therefore, different light sources (e.g., high pressure sodium, metal halide) with distinct color spectra are expected to elicit unique responses from different organisms (Fig. 3; Moore et al. 2006). Recently, the European Eco-Design Directive has enacted a step-by-step plan to phase out particularly energy-intensive lighting products (e.g., high-pressure mercury lamps, the European Parliament and the Council of the European Union 2009). Thus, many countries and the EU have launched a number of programs to adopt efficient lighting systems with a focus on LEDs as a promising energy-efficient lighting technique. There is some evidence that LEDs will attract fewer insects than previous bulb types (Eisenbeis and Eick 2011), but this needs to be more rigorously tested, as the light levels and luminaire construction in this study varied in addition to bulb type. Further, it is completely unknown how other freshwater organisms might respond to different wavelengths, although some fish (e.g., Acipenser baeri and Oncorhynchus mykiss) have peak sensitivities that correspond to peak emissions from LEDs (Hawryshyn and Hárosi 1994, Sillmann and Dahlin 2004; Fig. 3).
The spatial and temporal scalar influence of artificial light is also an area that requires elucidation. Scheibe (2003) showed that one street light located near a stream can attract caddisflies hatching from several hundred meters of stream, but it is unclear how applicable his results are for different habitat and ecosystem types, or what the impact of multiple light sources might be. At larger spatial scales, it is clear that the sky glow created by the cumulative lights of a large city can influence natural areas 10s and even 100s of kilometers away (Albers and Duriscoe 2001, Kyba et al. 2011). For example, Moore et al. (2000) found that artificial light from 16 km away was strong enough to alter the DVM of Daphnia. We need to know if wide-spread use of artificial lights near freshwater and riparian habitats will contribute to the decline or disappearance of sensitive species, lead to localized decreases close to bright light sources, or even be beneficial for other species. Even if sky glow does not cause extinctions, it could very likely alter food web structure either by changing predators' ability to detect prey or prey behavior (Moore et al. 2006). Another question that needs to be answered is if light-sensitive species are able to re-colonize areas when lights are removed. Mapping the occurrence of artificial light across landscapes will allow us to make better predictions about the likelihood of specific habitats being recolonized.
While the results of Scheibe (2003) and Moore et al. (2000) suggest that artificial light can influence organisms over a relatively large spatial area, we do not know the temporal scale of this influence. Does exposure to artificial light during the night alter the behavior of organisms during the day? Perhaps some organisms have life stages that are particularly vulnerable to exposure to artificial light, but are not sensitive during the rest of their lives. These species might be able to take advantage of dark refuges for sensitive life stages and then live in artificially lit areas at other times. If populations are negatively affected by artificial light, are they able to recover quickly once artificial light is removed from their habitat? This largely depends on whether artificial light alters the genetic structure of populations. Furthermore, spatial analysis is needed to determine the overlap of artificial lights and freshwater bodies. As noted in the introduction, freshwater environments are preferred sites for human activities, which will often lead to an increase of artificial lights. We expect to find the greatest amount of lighting in already damaged urban areas, but we also need to determine if vacation homes and highways introduce a meaningful amount of light to more natural areas.
The magnitude of changes in light also needs to be better understood. While direct glare is the most conspicuous form of light pollution, sky glow is a much more wide-spread phenomenon that is likely to influence animal behavior (Longcore and Rich 2004, Moore et al. 2006, Nightingale et al. 2006). Sky glow can increase ambient light levels hundreds of kilometers away from the cities from which it emanates. This is the case in several ecologically important U.S. National Parks (Everglades, Channel Islands, and Joshua Tree), which have night skies that are substantially brighter than natural due to sky glow from nearby cities (Albers and Duriscoe 2001). One potential problem of increased light from sky glow is that it reduces or eliminates the natural monthly variation in night-time light that arises from the lunar cycle (Longcore and Rich 2004, Kyba et al. 2011; Fig. 2). If the general increase in ambient light caused by sky glow can alter behavior and harm ecosystems, then managing artificial light becomes a much more pressing conservation concern. However, it will be very difficult to study the effects of sky glow on ecosystems, as there are very few places left in North America and Europe that do not have elevated levels of sky glow to use as control sites (Cinzano et al. 2001). Furthermore, once researchers have located a promising location, how do they mimic an increase in sky glow that would normally be produced by a city of 500,000+ inhabitants that is 50 km away? While researchers may be able to introduce direct glare by introducing a few lights to an ecosystem, those interested in understanding the influence of sky glow may have to introduce artificial darkness to an already lit area, as Moore et al. (2000) did.
Conclusion
How artificial light at night might influence stream and riparian ecosystems is a relatively unexplored topic, with many possibilities for relevant research. Even though the experimental knowledge of the ecological impacts of artificial light at night is still developing, governments are creating legislation to regulate it, mostly to reduce energy costs and decrease greenhouse gas emissions (Hölker et al. 2010b). Reducing energy consumption is a desirable goal, but if it is achieved solely through changing lighting fixtures and not necessarily reducing lighting, and without knowing how different aspects of artificial light (e.g., intensity and spectral qualities) influence ecosystems, this legislation could have unintended and even negative impacts on ecosystems. We also expect that governments will not be able to regulate artificial light everywhere, but by understanding its potential consequences, we can better prepare for or mitigate them.
Carefully designed experiments are needed to determine the exact effects of artificial light on ecosystems and over what spatial and temporal scales they act. From a management perspective, it is highly important to consider and incorporate the mitigation of potential ecological impacts and losses of biodiversity and ecosystem services into new lighting concepts (Rich and Longcore 2006, Hölker et al. 2010a, b). While there are many challenges to overcome in pursuing this research, the potential for new breakthroughs in understanding ecosystems and their functioning is high and should motivate researchers to innovate new techniques.
Acknowledgments
Funding for this project was provided by the Freie Universität Berlin, the Leibniz Gemeinschaft, and the “Verlust der Nacht” project, funded by the German Federal Ministry of Education and Research (BMBF 033L038A). The quality of the manuscript was greatly improved by comments from Ann-Christin Honnen, Travis Longcore, Michael Monaghan, Steve Ormerod, Nike Sommerwerk, and four anonymous reviewers.
