The Effect of Light Pollution on the Sea Finding Behavior of Green Turtle Hatchlings on Lanyu Island, Taiwan

© 2023 Academia Sinica, Taiwan Open Access The Effect of Light Pollution on the Sea Finding Behavior of Green Turtle Hatchlings on Lanyu Island, Taiwan Chih-Hao Yen1, Yin-Ting Chan1, Yao-Chi Peng2, Kuo-Hui Chang2, and I-Jiunn Cheng1,* 1Institute of Marine Biology, National Taiwan Ocean University, Keelung, Taiwan. *Correspondence: E-mail: [email protected] (Cheng) E-mail: [email protected] (Yen); [email protected] (Chan) 2Institute of Marine Biology, Ieotek Corporation, Neihu 114, Taiwan. E-mail: [email protected] (Peng); [email protected] (Chang) Received 29 November 2022 / Accepted 24 July 2023 / Published 22 September 2023 Communicated by Benny K.K. Chan Nocturnal animals use their vision and acute hearing to adapt to the nighttime environment. Light pollution has become a serious problem for nocturnal animals in coastal areas, especially nesting sea turtles and sea turtle hatchlings. Hatchlings use visual clues to find the ocean. However, when the artificial light is stronger than the natural light, hatchlings become either misoriented, disoriented or both. Due to rapid tourism development on Lanyu Island, new sources of light pollution, especially streetlights, pose a serious threat to sea turtle hatchlings. In this study, we used a portable lamp constructed by Liteon Inc. on a circular area of a turtle nesting beach to see how artificial light sources could affect green turtle hatchlings’ sea finding behavior. In the experiments, we tested hatchling behavior under different lamp settings (strong or weak light intensity; white or yellow light; lamp shield presence or absence) and moon visibilities (moonlit or moonless). The hatchlings’ crawling tracks and locations at the end of the trials were recorded. Results showed that the light intensity had no effect on hatchling sea finding behavior. White light had a stronger impact on hatchling sea finding behavior than yellow light. When the lamp shield was installed on moonlit nights, more hatchings were able to find the sea under both white and yellow lights. Thus, it is recommended that light shields be installed on the streetlights of Lanyu Island in order to protect the sea turtle hatchlings effectively. Key words: Lanyu Island, Green sea turtle hatchling, Artificial lights, Lamp shield, White and yellow lights Citation: Yen CH, Chan YT, Peng YC, Chang KH, Cheng IJ. 2023. The effect of light pollution on the sea finding behavior of green turtle hatchlings on Lanyu Island, Taiwan. Zool Stud 62:47. doi:10.6620/ZS.2023.62-47. BACKGROUND Natural light comes from the sun, moon and stars and is important to many animals because it is a source of energy and information. It can regulate animals’ physiological features, such as biological clocks, and is important to the survival of many organisms (Kramer and Birney 2001; Verutes et al. 2014). Nocturnal animals use their vision and acute hearing to adapt to the nighttime environment and to survive and forage in order to maintain their physiologies, adjust their metabolisms, growth and other relevant behaviors (Silva et al. 2017). The population of humans living on the coast is increasing at a higher rate than anywhere else (Nicholls 1995). This results in coastal areas becoming one of the most vulneable areas to anthropogenic stresses (Bird et al. 2004). Sea turtles are one of the most affected species of wildlife in coastal areas (Salmon et al. 2000; Salmon 2003). Many nesting beaches are close to residental or developed areas, such as malls, cities, factories and power plants. With the increasing population, development and light emission, light pollution becomes a serious problem. Light pollution occurs when artificial lights impact the physiology (e.g., foraging, reproduction, metabolism, orientation, migration) Zoological Studies 62:47 (2023) doi:10.6620/ZS.2023.62-47 1 © 2023 Academia Sinica, Taiwan and behaviors of animals. It increases threats to their survival and decreases their ability to adapt (Perry and Fisher 2006). Skyglow from nearby cities can also affect sea turtles negatively (Garrett et al. 2019). Coastal artificial lights can deter female turtles from nesting, decrease nesting success, concentrate nests, and increase egg predation (Pilcher et al. 2000; Salmon et al. 2000; Salmon 2006; Wyneken et al. 2000). They can also disturb the sea finding behavior of hatchlings (Berry et al. 2013; Hamann et al. 2007; Salmon 2006). Sea turtle hatchlings use their vision to search for the brighter horizon when crawling towards the ocean (Limpus and Kamrowski 2013; Limpus 1971). However, nearby artificial lights disturb their ability to search for the ocean and decrease their chances of survival (Tuxbury and Salmon 2005; Witherington and Martin 2000). Artificial lights can create two kinds of disturbances: misorientation, where hatchlings move in the a circular motion or remain motionless without a clear direction; and disorientation, where hatchlings crawl in the opposite direction of the ocean and move towards the artifical lights (Salmon and Witherington 1995; Verheijen 1985). The degree to which hatchlings are disturbed by artificial light depends on its intensity and wavelength (Witherington 1991). Green turtle hatchlings show positive phototaxis towards short-wave blue light with a weaker response to long-wave red light within the 350 to 540 nm wavelength range (Levenson et al. 2004; Mrosovsky 1972). An ERG (electroretinogram) study showed that long-wave red light has to be approximately 600 times stronger than the blue light before hatchling green turtles will have the same response behavior (Mrosovsky 1972). Both adult and hatchling green turtles can be influenced by light wavelengths between 440 to 700 nm, but hatchlings are particularly sensitive to the 350 to 450 nm wavelength range (Horch et al. 2008; Witherington and Bjorndal 1991; Witherington and Martin 2000). There are two main nesting sites in Taiwan: Wanan Island of Penghu Archipelago and Lanyu Island of Taitung County. Long-term studies have shown that the nesting population on Lanyu Island has increased since 1997 and become the major nesting island in Taiwan (Cheng et al. 2008 2018). However, the rapid development of tourism and the accompanying facilities, such as bars, restaurants and streetlights by the beach, create a major threat to nesting female and hatchling sea turtles. Lanyu Island is located in the Pacific Ocean, approximately 145 km southeast of Taiwan. The size of the island is 45.7 square kilometers (Fig. 1). The number of tourists visiting the island per year in 2016 was 110,000, which then increased to approximately 140,000 in 2019 and reached 160,000 in 2020 (Taitung County Report 2020). Due to COVID-19, most tourists visited domestic islands instead of traveling abroad. This resulted in the rapid increase of tourists to this island. In order to create safer roads and attract tourists, many of the bars and restaurants were built near the nesting beaches. Most streetlights around the island that face the beaches use bright LED lighting. This resulted in more artificial light spilling onto the nesting beaches. Ko (2020) found that nearly 90% of the green turtle hatchlings on this island failed to conduct sea finding behavior, even under the full moon. The streetlights thus pose serious threat to the hatchlings on this island. In order to understand the impact of streetlights on the sea finding behavior of green turtle hatchlings, as well as possible solutions, we cooperated with the Liteon Inc. to construct a portable lamp to conduct experiments on the nesting beach. There are two purposes of this study: first, to determine how the sea finding behavior of the hatchlings is affected under the lamp with strong and weak light intensities, with and without the lamp shield installed and on moonlit or moonless nights, and second, to determine the sea finding behavior of hatchlings under long-wave yellow light (440 to 700 nm) and short-wave white light (350 to 540 nm). Fig. 1. Map of Lanyu Island with marked nesting beaches, 1: Big Bai-Dai Beach, 2: Small Bai-Dai Beach, 3: Donchin Beach. “A” and the red circle on the inset map on the lower left indicate Lanyu Island. N page 2 of 13Zoological Studies 62:47 (2023) © 2023 Academia Sinica, Taiwan MATERIALS AND METHODS Study site and period There are three nesting beaches on Lanyu Island: Big Ba-Dai, Ba-Dai and Donchin (Fig. 1). Ba-Dai hosts the most turtle nests, Big Ba-Dai receives the most artificial light, and Donchin is a relatively large and undisturbed beach. We moved the nests from the other beaches to Donchin, since it offered the most space and fewest disturbances. The experiments lasted from midJune to mid-September of 2019 and 2020. Nest and hatchling collection Beaches were patrolled every two hours from 7 pm to 3 am each night. In order to prevent nests from overcrowding the main nesting beach and protect them from light pollution, nests were relocated from Big Ba-Dai and Ba-Dai beaches to Donchin Beach and reburied at the undisturbed site within three hours of depositing the eggs. According to a previous study, the nest incubation periods on the island ranges from 50 to 55 days (Cheng et al. 2009). A wired cage was installed from 7 pm to 7 am daily starting at day 45 of incubation to collect the emerged hatchlings. Caged nests were visited according to the beach patrol schedule, every two hours from 7 pm to 3 am. Hatchlings were collected if they had naturally emerged from the nests and avoided if they were stunted, weak or yet to be hatched (Pendoley and Kamrowski 2016). The collected hatchlings were stored in a ventilated light-tight insulation box until the experiment began. Design of the portable lamp Based on the new design of a turtle-friendly light proposed by Robertson et al. (2016), a portable lamp was constructed by Lite-On Inc. The lamp was compliant with Taiwan Road Lighting Specifications. Experiments were designed to mimic the light pollution of streetlights and its effect on hatchling crawling behavior by exposing the hatchlings to different light sources, intensities, installed light shield or uninstalled light shield, and different levels of moonlight. Due to the fact that hatchlings have a strong phototaxis to short-wave light, the second set of experiments were designed to determine hatchling crawling behavior under yellow and white light. The lamp was 2 meters tall and composed of 40 LED light bulbs which could be replaced depending on the color of light needed for the experiment. In order to prevent any vibration created by the generator, the lamp used lifespan (Fig. 2). Lamp shield and light intensity determination The lamp shields were designed to shield light from the front and sides of the lamp. The light reached up to 12 meters from the source without a shield, while it only reached up to 6 m with the shield installed. Light intensity was determined using a portable spectroscopic spectrometer (Chunyua Scientific Technology, model MK350N Premium) 9 meters from the light source. Light intensity reached 37.45 lx with the strong white light without the shield and decreased to 4.153 lx with the shield. Light intensity reached 14.53 lx with the strong yellow light without the shield and decreased to 2.187 lx with the shield. Liteon Inc. used a DC24V, 70W, 2.5A, 1-10V eight-segment current selector to determine the light intensity. Light intensity increased from segment 1 to 8. In this study, we used segment 1 for the weak light and segment 8 for the strong light intensity. Moon phase determination No moon and new moon were defined as “moonless”. The other moon phases were defined as “moonlit night” (Witherington and Martin 2000). Fig. 2. The portable lamp installed on the Donchin nesting beach of Lanyu Island. page 3 of 13Zoological Studies 62:47 (2023) © 2023 Academia Sinica, Taiwan Design of circular arena The circular arena was designed based on Bertolotti and Salmon (2005). A circle with a diameter of 3 meters was drawn on the beach and divided into 12 sections. Among these sections, degrees 345 to 15 faced the ocean, and degrees 165 to 195 faced the lamp. The lamp was placed 9 meters from the center of the arena (Fig. 3). Conducting the experiment During the experiment, hatchlings were placed in a shallow depression about 10 cm deep in the center of the arena and covered with a light-proof box to protect them from the effects of the light. A 5 cm depression was made around the edge of the arena to indicate the end of the experiment. At least 10 hatchlings were used in each experiment. They were allowed to crawl for 4 minutes. For hatchlings that reached the edge of the arena within 4 minutes, their crawl tracks and final locations in degrees were recorded. Hatchlings that failed to reach the edge within 4 minutes were considered “incomplete” and removed from the test. All the hatchlings were released back to the ocean immediately after the experiment. Four conditions were defined in the experiment: (1) hatchlings that crawled in a directionless, motionless or circular pattern and had end points located in the artificial light sections were defined as “disoriented”; (2) hatchlings that crawled towards artificial light or away from the ocean and had end points located in the artificial light sections were defined as “misoriented”; (3) hatchlings that crawled towards the ocean and had end points located in the ocean sections were defined as “sea finding”; (4) hatchlings that crawled towards neither light nor ocean sections and had end points located in sections other than the artificial light or ocean sections were defined as “other directions”. Data analyses Oriana 4 software was used to conduct a Rayleigh z test (Zar 1999) to determine whether the end points of the hatchlings in each experiment were distributed in a certain direction or randomly. In the analysis, r denotes the crawling direction of the hatchling (with high r values indicating that the hatchlings crawled in a similar dirction), and µ denotes the mean direction of the end point of crawling. A Watson-Williams F-test (Dimitriadis et al. 2018) was used to determine the differences between light intensities, use of lamp shield or not, moon visibilities, and lamp colors. RESULTS A total of 21 nests and 534 hatchlings were collected for the experiments. Among them, 22 hatchlings failed to crawl and were subsequently removed from the test. Thus, 512 hatchlings were used. In order to determine the effect of background light, crawl experiments were conducted with the lamp turned off during both moonlit and moonless nights prior to the experiments that used the lamp turned on. Light turned off A Rayleigh test indicated that none of the hatchlings in the sample size used for these trials showed disoriented crawling, and that all behaved either “sea finding” or “other direction” crawling (Table 1, Fig. 4). The Watson-Willions F-test showed no difference between the moonlit and moonless nights (p > 0.05). Most hatchlings crawled towards the ocean. Strong and weak white light (350 to 540 nm) 261 hatchlings were used for this part of the experiment. Among them, 11 did not complete the experiment, 13 exhibited sea finding crawling (5%), 173 exhibited misoriented crawling (69%), 56 crawled in other directions (22%) and 8 exhibited disoriented Fig. 3. Design of the experimental arena. page 4 of 13Zoological Studies 62:47 (2023) © 2023 Academia Sinica, Taiwan crawling (3%). The crawling behavior of hatchlings under the strong and weak white lights, with the lamp shield installed or uninstalled and on moonlit and moonless nights, along with statistical analyses, are shown in table 1. Results of the distribution of end points in strong and weak light, with and without lamp shield installed, and on moonlit and moonless nights are shown in figures 5 and 6. These results showed that more than 68% of hatchlings displayed misoriented crawling under both the strong and weak white light. Only in two cases, under both bright and weak light settings, with both the lamp shield installed and on moonlit nights, did a portion of hatchlings either crawl towards the sea or in other directions (Figs. 5 and 6; Table 2). Rayleigh tests showed that, under the strong light setting, the hatchlings crawled directionally in both moonlit and moonless nights, with and without the lamp shields installed (p < 0.05 in all cases). This test also showed that, under the weak light, the hatchlings crawled directionally without the lamp shield on both moonlit and moonless nights and with the lamp shield on moonless nights (p < 0.05 in all cases). When the lamp was set to the weak light setting and the shield was installed on moonlit nights, the hatchlings’ crawling direction was random (Z = 2.943, p = 0.052). The average r values were more than 0.87 on the moonless nights with both strong and weak intensities and with the lamp shield installed or uninstalled. The r values decreased to less than 0.41 in both strong and weak light settings with the lamp shield installed on moonlit nights. The average μ values were more than 170° in all cases (Table 2). These results suggest that most hatchlings crawled towards the artificial light. However, the moon did attract some hatchlings to crawl in the other directions (Table 2). The Watson-Williams F-test showed that, under the strong light, no difference was found between lamp shield installed or not, nor between moonlit and moonless nights. Also, no difference was found between moonlit and moonless nights without the lamp shield under the weak light setting (p > 0.05 in both cases; Figs. 5 and 6). Most hatchlings crawled towards the artificial light. However, the test found a significant difference between moonlit and moonless nights with the lamp shield installed under the weak light setting (p > 0.05). Some hatchlings crawled towards the ocean when the lamp shield was installed on moonlit nights (Fig. 5). Strong and weak yellow light (593 nm) 273 hatchlings were used. Among them, 11 did Table 1. Total hatchlings used and proportion of incomplete, disorieted, misoriented, correct sea finding, other directions, concentrated (r) and average degree with the lamp turned off parameter total incomplete (%) misoriented (%) disoriented (%) sea finding (%) others (%) rµ (°) moonless night 20 0 00 35 65 0.924 343.269 moonlit night 20 0 0 0 55 45 0.675 350.103° Fig. 4. The end points of hatchlings with the lamp turned off on the (a) moonless night and (b) moonlit night. The circle is divided into 12 sections of 30 degrees each. The number on the exterior of the circle represent “degree” and correspond to the section number from degree 15 (faces the ocean) towards the right direction. One can determine in which direction the hatchling crawled from the number in the square. The diameter of the circle is 3 m. Hatchlings were released from the center point of the arena (circle). (a) (b) page 5 of 13Zoological Studies 62:47 (2023) © 2023 Academia Sinica, Taiwan Table 2. Total hatchlings used and proportion of incomplete, disoriented, misoriented, correct sea finding, other directions, concentrated (r) and average degree under the strong and weak white light parameter total incomplete (%) misoriented (%) disoriented (%) sea finding (%) others (%) rµ (°) strong light with shield moonless night 35 3 74 0 0 25 0.936 185.377 moonlit night 35 0 40 6 11 43 0.407 172.857 without shield moonless night 31 10 71 10 0 10 0.911 173.595 moonlit night 50 0 92 0 0 8 0.997 181.189 weak light with shield moonless night 60 6 74 0 0 20 0.874 181.032 moonlit night 75 0 38 8 23 33 0.271 210.893 without shield moonless night 35 6 86 0 0 9 0.941 185.773 moonlit night 25 12 68 0 0 20 0.98 173.963 Fig. 5. The end points of hatchlings when exposed to strong white light with (a) no lamp shield on a moonlit night, (b) no lamp shield on a moonless night, (c) lamp shield on a moonlit night, and (d) lamp shield on a moonless night. (c) (a) (b) (d) page 6 of 13Zoological Studies 62:47 (2023) © 2023 Academia Sinica, Taiwan not complete the experiment, 81 displayed sea finding crawling (31%), 33 exhibited misoriented crawling (13%), 138 crawled in other directions (53%) and 10 demonstrated disoriented crawling (4%). The crawling behavior of hatchlings under the strong and weak yellow lights, with the lamp shield installed or not installed and on moonlit or moonless nights, along with statistical analyses, are shown in table 2. The distribution of end points under the strong and weak lights, with and without the lamp shield installed, and on moonlit and moonless nights are shown in figures 7 and 8. Results showed that, with the exception of when the lamp shield was not installed on moonless nights under both strong and weak light, more than 20% of the hatchlings crawled towards the ocean and in most cases, close to 50% or more crawled in the other directions (Table 3). Rayleigh tests showed that the hatchlings all crawled in a particular direction on both moonlit and moonless nights, with and without the lamp shield installed, under both the strong and weak light setting (p < 0.05 in all cases). Comparisons showed that, except in the strong light with no lamp shield, the average r values were higher than 0.86 in all cases on moonlit nights. The values decreased to less than 0.67 on moonless nights. The average μ values were close to 360° in all cases on moonlit nights and decreased to less than 212° on moonless nights (Table 3). These results suggest that in the strong and weak yellow light, most hatchlings crawled towards the ocean on moonlit nights. However, in the absence of the moon, hatchlings crawled in random directions (Figs. 7 and 8). The Watson-Williams F-test showed that, under strong and weak light settings, a significant difference was found between moonlit and moonless nights with and without the lamp shield installed (p < 0.01 in both cases). Most hatchlings crawled towards the ocean on moonlit nights, while most hatchlings crawled towards the artificial light or in other directions on moonless nights (Figs. 7 and 8). This test showed that under both strong and weak light settings, hatchlings crawled towards the ocean in both moonlit and moonless nights when the lamp shield was installed (p > 0.05 in both cases). Fig. 6. The end points of hatchlings exposed to weak white light with (a) no lamp shield on a moonlit night, (b) no lamp shield on a moonless night, (c) lamp shield on a moonlit night, and (d) lamp shield on a moonless night. (c) (a) (b) (d) page 7 of 13Zoological Studies 62:47 (2023) © 2023 Academia Sinica, Taiwan Table 3. Total hatchlings used and proportion of incomplete, disoriented, misoriented, correct sea finding, other directions, concentrated (r) and average degree under the strong and weak yellow light parameter total incomplete (%) misoriented (%) disoriented (%) sea finding (%) others (%) rµ (°) strong light with shield moonless night 30 0 3 0 33 63 0.669 7.812 moonlit night 30 0 7 0 47 47 0.572 356.148 without shield moonless night 38 24 24 0 0 53 0.408 212.522 moonlit night 39 0 0 0 41 59 0.866 357.022 weak light with shield moonless night 46 0 11 17 20 52 0.457 183.329 moonlit night 33 0 0 0 76 24 0.948 357.278 without shield moonless night 30 7 33 7 0 53 0.457 183.329 moonlit night 27 0 22 0 26 52 0.948 357.278 Fig. 7. The end points of hatchlings exposed to strong yellow light with (a) no lamp shield on a moonlit night, (b) no lamp shield on a moonless night, (c) lamp shield on a moonlit night, and (d) lamp shield on a moonless night. (c) (a) (b) (d) page 8 of 13Zoological Studies 62:47 (2023) © 2023 Academia Sinica, Taiwan DISCUSSION This is the first study to use of a portable lamp that mimics real road conditions to determine the impact of streetlights on hatchling sea finding behavior. Due to the rapid development of tourism on Lanyu Island, hatchlings have been seriously impacted by artificial lights (Chang 2012; Ko 2020; Tsai 2016). Among the various sources of light pollution, streetlights are the most serious (Ko 2020). This study was designed to find a solution to decrease the disturbance of streetlights on hatchling sea finding behavior. Results found that under white light, 72% of hatchlings were misoriented and disoriented, which is similar to other studies (Lohmann et al. 2017; Salmon 2003; Tuxbury and Salmon 2005; Witherington and Martin 2000). Under yellow light, only 17% were misoriented and disoriented. Previous research supports the findings that decreasing artificial light increases the ability of hatchlings to crawl towards the sea under yellow light (Witherington and Bjorndal 1991). Before we conducted our lamp experiments, we tested hatchling sea finding behaviors with the lamp turned off. In these trials, all hatchlings crawled towards the ocean on both moonlit and moonless nights. Thus, we determined that the arena we selected was suitable for the artificial light experiments. White light We found that white light, either strong or weak in intensity and on moonlit and moonless nights, has a negative impact on hatchling sea finding behavior. In most cases, hatchlings crawled owards this artificial light. However, when the lamp shield was installed on the moonlit nights, some hatchlings crawled towards either the sea or in other directions. Pendoley and Kamrowski (2016) pointed out that the lamp shield can decrease the amount of artificial light radiating directly onto the hatchlings, and thus can decrease the chances of misoriention and disoriention. The impact of the artificial light decreased when there was a shield and there was moonlight. Fig. 8. The end points of hatchlings exposed to weak yellow light with (a) no lamp shield on a moonlit night, (b) no lamp shield on a moonless night, (c) lamp shield on a moonlit night, and (d) lamp shield on a moonless night. (c) (a) (b) (d) page 9 of 13Zoological Studies 62:47 (2023)