Moonlight diminishes seabird attraction to artificial light

Rescue programs aiming to mitigate light‐induced mortality of seabird fledglings have reported that fewer birds are grounded (and rescued) during full moon nights. Two non‐mutually exclusive hypotheses have been proposed as explanations: (1) reduction of strandings because birds are less attracted to and disorientated by light pollution during full moon nights; and (2) reduction of fledging activity, that is, chicks avoid departing the colony during nights with increased moonlight. We argue that evidence from rescue programs and other studies supports the first but not the second hypothesis. The evidence supports the conclusion that a higher proportion of fledglings make it safely to sea during full moon nights than during moonless nights. Thus, there is a decrease in the severity of light pollution on seabirds around the full moon.


| INTRODUCTION
Alteration of natural, night time light levels produced by the introduction of anthropogenic light, that is, so-called light pollution, have detrimental effects on terrestrial and marine ecosystems (Gaston & de Miguel, 2022;Marangoni et al., 2022).Light pollution can lead to mass mortality events among insects (Owens et al., 2020), sea turtles (Salmon, 2003), and seabirds (Rodríguez et al., 2017).Seabirds are one of the most endangered avian groups and light pollution can cause high mortality, especially among fledglings upon leaving their nest cavities during their nocturnal, inaugural flights to the sea (Dias et al., 2019;Rodríguez et al., 2019).During these nocturnal flights, fledglings are attracted, disoriented, and forced to land by artificial lights, crashing into human-made structures (e.g., buildings, wires, or antennas; Rodríguez et al., 2022).Light-induced grounding can be fatal because of collisions, but even if uninjured, grounded birds may be unable to take off again and are vulnerable to predation, vehicle collisions, starvation, dehydration, and poaching (Rodríguez et al., 2017).Departing at night, true as well among sea turtles, is a suggested strategy, along with cavity-nesting, to reduce predation (Gaston, 2004;Warham, 1990).At least 73 seabird species have been documented in light pollutioninduced mortality events, mostly belonging to cavity-nesting Procellariiformes species (Gilmour et al., 2023).Approximately 60% of these 73 species have exhibited population declines and 28 are globally threatened according to the International Union for Conservation of Nature (IUCN) criteria (Gilmour et al., 2023).
For many years, our best knowledge of the phenomenon of attraction and disorientation of seabirds caused by artificial lights (fallout) results from information compiled by rescue programs (Rodríguez et al., 2017).One initial observation was that fewer birds were found stranded during full moon nights (Imber, 1975;Telfer et al., 1987).Since then two non-mutually exclusive hypotheses have been proposed by several authors as explanations: (1) reduction of strandings because birds are less attracted and disorientated by light pollution during full moon nights; and (2) reduction of fledging activity, that is, chicks avoid departing the colony during nights with increased moonlight (a possible anti-predator behavior) (Chevillon et al., 2022;Deppe et al., 2017;Le Corre et al., 2002;Reed et al., 1985;Rodríguez & Rodríguez, 2009;Syposz et al., 2018;Telfer et al., 1987).In a recent review paper, Brown et al. (2023) identified the need for additional research "to confirm if fewer strandings observed around the full moon are due to reduced fledging activity at the colony (as opposed to the hypothesis that bright moonlight diminishes the attractiveness of point sources of artificial light)."Here, we argue that evidence supports the first hypothesis, that is, fledglings are less attracted and disorientated by light pollution during full moon nights, with evidence lacking to support the reduction of fledging activity hypothesis.

| EVIDENCE AND SUPPORT ON HYPOTHESES
Evidence from rescue programs supports the hypothesis that increased moonlight reduces the attraction and disorientation of seabirds, consequently reducing fallout magnitude.In species with synchronized breeding cycles, most fledglings leave nests within a period of time that is shorter than one lunar cycle, for example, Cory's shearwater Calonectris borealis (Figure S1; Rodríguez & Rodríguez, 2009), Barau's petrels Pterodroma baraui and wedge-tailed shearwaters Ardenna pacifica (Chevillon et al., 2022), short-tailed shearwater Ardenna tenuirostris (Rodríguez et al., 2014), or Newell's shearwater Puffinus newelli (Ainley et al., 2001;Raine et al., 2023).When plotting the number of stranded fledglings per night against the rescue date, we observe clear differences in the rescue patterns and the magnitude of the fallout: full moon is linked to shifts in the peak fallout date by decreasing the number of fledglings grounded during full moon nights (Figure 1).If fledglings avoided departure during full moon nights, we would expect a similar number of rescued fledglings across years irrespective of the timing of the moon cycle.Birds would fledge earlier or later than full moon nights, with the annual totals being the same (other than as affected by breeding success or population declines).However, data obtained for each of these well-studied species reveals that the annual number of stranded fledglings is less in years when the full moon coincides with the mean peak fledging date.Therefore, the premise of the fledging reduction hypothesis is not met (Figure 1c; we consider mean peak rescue date as a proxy of the mean peak fledging date; Ainley et al., 2001;Chevillon et al., 2022;Rodríguez & Rodríguez, 2009).
Longer wing length and lower body mass found in Barau's petrel fledglings rescued after a full moon was considered evidence of delayed burrow departure (Le Corre et al., 2002).However, this analysis was based on a single year.To pursue this idea further, we used data from 8954 Cory's shearwater fledglings grounded and rescued by the SOS Cagarro campaigns conducted in the Azores during 6 years (2016-2021).In addition, we re-analyzed data from 267 Cory's shearwater fledglings grounded on Tenerife, Canary Islands, in 2009 and 2010 (Rodríguez et al., 2012).We plotted wing length and body mass on the rescue date and ran local polynomial regressions using the method loess in the function geom_smooth of the package ggplot2 in R (R Foundation for Statistical Computing, Vienna, Austria, v. 4.2.2).We found that potential delayed fledglings did not have longer wings or lighter body masses after full moon nights and trends are similar across the years regardless of the moon cycle (Figures 2 and S2).Fledglings are still growing their outermost primary feathers during the first days of the fledging season, but growth stops around mean peak fledging date.Body mass decreases throughout the fledging season, a well-known trend in fallout rescue birds (Cuesta-García et al., 2022 and references therein), and no interaction with the moon cycle is apparent (Figures 2 and S2).So, biometric data do not support the hypothesis of reduced fledging activity around full moon.
To our best knowledge, fledging of cavity-nesting seabirds is not correlated to the moon cycle.Since the first studies, several authors stated that Newell's shearwaters from cross-fostered and natural colonies fledged irrespective of the moon phase (Byrd et al., 1984;Telfer et al., 1987), which has been recently corroborated by the use of camera traps (Raine et al., 2023).Preliminary data suggest that Leach's storm-petrels Hydrobates leucorhous also fledge across the moon cycle (Miles et al., 2010) and a radar survey conducted near a Barau's petrel colony at Réunion Island very convincingly demonstrated that the number of fledglings detected was not influenced by moonlight (Gineste, 2016).Thus, hundreds of birds fledged during full moon but they were not grounded by urban lights (Gineste, 2016).Additionally, Global Positioning System (GPS) tracking of fledglings of Cory's shearwater upon nest departure supports the hypothesis that birds are less susceptible to light attraction during bright moonlight conditions.Experimentally released birds flew straighter over light-polluted areas during full moon and were able to reach the ocean successfully (Rodríguez et al., 2022).Yet, during nights with low moonlight, birds were more attracted to lighted areas and were grounded in higher numbers (see the flights of birds NX07393 and NX07402 in animations of Rodríguez et al., 2022).Consequently, there is no support to the hypothesis that inclinations of birds to fledge are reduced during the full moon.

| HOW MIGHT MOONLIGHT AFFECT FALLOUT?
The answer to this question is unsolved, but the greater ambient light from the moon may reduce the relative glare of artificial lights diminishing seabird attraction (Reed et al., 1985) or permit fledglings to see obstructions: natural, such as cliffs, valleys, or trees, and  .The full moon date indicates the difference in days with the mean peak rescue date, which we consider a proxy of the mean peak fledging date.Thus, negative values indicate an earlier full moon than November 4. Positive values indicate a later full moon.Zero indicates that the full moon coincided with the mean peak rescue date (November 4).Dots with light tones, that is, those representing years with full moon night around the mean peak rescue date, reach relatively low numbers of rescued birds in comparison to subsequent years.Source: Data come from the La Tahonilla Wildlife Rehabilitation Centre (Cabildo Insular de Tenerife), Rodríguez andRodríguez (2009), andRodríguez et al. (2023).
anthropic infrastructure, such as wires, antennas, or buildings (Telfer et al., 1987).Another non-mutually exclusive explanation is that fledglings orient toward the moon and hence do not come down to artificial lights (Reed et al., 1985), in a similar way to that proposed for insects (Fabian et al., 2023).
Moonlight might also help the development of the visual system of fledglings.Being typically cavity-nesting species, the limited light coming into nests prevents a growing eye to achieve emmetropia, that is, chicks fledge with an underdeveloped and untrained visual system (Atchoi et al., 2020;Mitkus et al., 2018).However, nearto-fledging procellariform chicks emerge from burrows at night to stretch and exercise their wings in the weeks before their departure (Raine et al., 2023;Yoda et al., 2017).So, moonlight during these nights might be enough to provide the needed visual stimuli to develop their visual systems.In this sense, a higher exposure to artificial light during ontogeny has led to a higher response to light stimuli in experimental shearwater chicks (Atchoi et al., 2023).Clearly, further studies are needed to explain how moonlight affects the light attraction of seabirds.

| CONCLUSIONS
Considering the arguments above, the reduction of strandings during full moon nights appears to be due to birds being less attracted and disorientated by light sources, that is, moonlight counters the effects of artificial light, and not because fledging is avoided when there is increased moonlight.This has conservation and management implications regarding the effects of light pollution on seabirds.During full moon nights, a higher proportion of fledglings safely fly to sea than during moonless nights, which is translated into a reduction of the mortality induced by artificial night lighting.We hope this contribution clarifies the current evidence, and derived hypothesis, regarding the observed low fallout numbers during full moons, while discouraging the unsupported use of reduced fledging as an explanation, unless new evidence is presented.

AUTHOR CONTRIBUTIONS
Airam Rodríguez conceived the idea.Airam Rodríguez, Elizabeth Atchoi, Beneharo Rodríguez, Tania Pipa, Matthieu Le Corre, and David G. Ainley contributed to the development of the argumentation and writing.

F
I G U R E 1 (a) Mean number (± standard deviation [SD]) of Cory's shearwater Calonectris borealis fledglings admitted to the rescue program in Tenerife, Canary Islands, during 31 years (1990-2021).Dots and error bars indicate the mean and SD of the number of rescued birds.The vertical red line represents the mean peak rescue date (November 4).(b) Cory's shearwater fledglings rescued in Tenerife, Canary Islands, during recent fledging seasons (2015-2021).The black line and the shaded gray area on each graph represent the nightly numbers of birds rescued.Vertical blue lines represent the full moon.Dots and error bars in gray indicate the mean and SD of the number of rescued birds during 31 years.Note the different Y-axis scales for 2015-2017 in relation to 2018-2021.(c) Annual numbers of Cory's shearwater fledglings rescued in Tenerife, Canary Islands, during 31 years

F
I G U R E 2 Wing length and body mass variations across the years (2016-2021) as a function of the date Cory's shearwaters Calonectris borealis are rescued by the SOS Cagarro program in the Azores.Vertical blue lines represent the full moon.Vertical red lines represent the mean peak rescue date, which we consider a proxy of the mean peak fledging date (October 30; Figure S1).Gray areas indicate the 95% confidence intervals of the predicted effects according to local polynomial regression (loess smoothing in the function geom_smooth of the package ggplot2).Each gray dot represents a fledgling.The darker the dots the higher overlap of dots.