New data on abundance of lemon shark Negaprion brevirostris (Poey, 1868) at Lama Bay, Rocas Atoll, Brazil

Abstract The abundance and movement patterns of lemon sharks ( Negaprion brevirostris) at Lama Bay, Biological Reserve Rocas Atoll, were reassessed by visual census. We considered tides and daylight periods to plan our observations during two expeditions in 2015. Using daily visual counts, the mean abundance of individuals was 29 in austral summer (maximum 35) and 31 in winter (maximum 41). The results indicated that the local lemon shark population might have recovered after 18 years of a substantial drop in mean abundance. In addition, the movement pattern of the species corroborated previous studies related to their fidelity to the birthplace. These results justify the need of continuous monitoring of lemon sharks over the course of time at Rocas Atoll using non-lethal and non-invasive techniques.


INTRODUCTION
Anthropogenic actions such as overfishing, by-catch mortality, pollution, habitat destruction, and climate change are threats to marine ecosystems (Dulvy et al., 2008(Dulvy et al., , 2021;;Simpfendorfer et al., 2011).As a direct consequence, fish populations from deep sea, coral reefs, and coastal regions are declining (Jones et al., 2004;Baker et al., 2009, Dulvy et al., 2021).However, it is unknown whether decreases occur in isolated populations or if they might cause widespread extinction across all populations of a species (Dulvy et al., 2014).
As an important component of the ecosystem, sharks influence the structure and the functionality of marine communities (Camhi et al., 2008).Major populations of oceanic predators, such as large sharks, are impacted by anthropogenic threats due to their low fecundity, low growth rates, late sexual maturation, and reproductive aggregations (Baum et al., 2003;Hammerschlag and Sulikowski, 2011;Worm et al., 2013;Barreto et al., 2017).Furthermore, there are limitations and challenges in data availability to investigate population decays, extinction threats, and potential consequences for marine ecosystems (Worm et al., 2013;Ward-Paige and Worm, 2017).Thus, the investigation of threats and the development of tools for shark population research are essential (Dulvy et al., 2017).
The lemon shark, Negaprion brevirostris (Poey, 1868), inhabits tropical waters on continental shelves and oceanic islands in Western Atlantic, Eastern Atlantic, and Eastern Pacific Oceans (Compagno, 1984;Ebert et al., 2013).Juvenile individuals have philopatric behavior to their birthplace and they commonly use it as nursery and growth areas (Morrissey and Gruber, 1993;Freitas et al., 2006;Garla et al., 2009).Even though adult individuals have a pelagic phase, which hinders studies (White et al., 2014), some adult female lemon sharks also have natal philopatry, which means they return to give birth at the same location where they were born (Feldheim et al., 2014;Brooks et al., 2016;Smukall et al., 2019).Therefore, the conservation of this important species is possible with the protection of its key habitats (Tavares et al., 2016).
According to the IUCN (2021), N. brevirostris is classified as a vulnerable species worldwide.In Brazil, however, this species is classified as endangered (Ministério do Meio Ambiente, 2022).The country ranks as the first importer and 11 th producer of shark meat in the world (Barreto et al., 2017), assuming an important role in the global shark trade (Rangel et al., 2021), which may justify the endangered classification.
Lemon sharks are found in the Marine Protected Area (MPA) of Rocas Atoll (Oliveira et al., 2011), the unique atoll in South Atlantic and a Ramsar site since 11 th December of 2015 (http://www.ramsar.org).This MPA is a no-take marine reserve established on July 5 th of 1979.Despite being an area of integral conservation of biodiversity, scientific and educational activities are allowed under permission of the ICMBio (Instituto Chico Mendes de Biodiversidade) (Granville et al., 2012).
Marine protected areas are important for helping shark species worldwide (Knip et al., 2012;Goetze and Fullwood, 2013;Speed et al., 2016).Freitas et al. (2009) reported a clear decline in lemon shark abundance at Rocas Atoll, but the knowledge about the current status of this species is limited.In this context, this study aims to estimate lemon shark abundance, particularly at the nursery/growth area of Lama Bay, in different seasons and tides using a non-invasive method.This approach provides new data on abundance of newborn and juvenile lemon shark local populations and a comparison with previous data.Finally, seasonal differences in movement patterns of this species were identified at Lama Bay.

METHODS
Rocas Atoll (3˚52'S; 33˚48'W) (Figure 1) is located at 269.5 km east from Natal city, Northeast coast of Brazil (Granville et al., 2012), in the South Atlantic Ocean.It is located in the mid-Atlantic ridge and rises 4,000 m from the ocean floor (Kikuchi and Leão, 1997).This atoll presents two sand islands, Farol Island (FI) and Cemitério Island (CI), which are influenced by marine and terrestrial biota, waves, currents, and winds (Soares et al., 2011).Farol Island, where Lama Bay lies, spans approximately 700 m in length and 300 m in width (Pereira et al., 2010).This bay is an elongated tidal creek, exhibits low hydrodynamics, and is influenced by the tidal regime, which makes it covered by water only during high tide (Pereira et al., 2010).
Two field trips to Rocas Atoll were conducted in 2015, 23 days in austral summer (January/ February) and 22 days in austral winter (June/ July).Data were obtained using visual census from a natural elevated point in Farol Island, between Lama Bay and Barretinha (Figure 1).Lama bay was divided in three zones: entrance (E), middle (M), and inner bay (I).Each zone measured 30×30 m and prominent landmarks were used for orientation.
The tidal table from Fernando de Noronha Santo Antonio Bay (PE) (DHN, 2015) was used to plan observations.Thus, the time of data collection occurred at the third hour before and at the third hour after high tide.Considering the two types of tidal currents (flood and ebb) and the three zones (E, M, and I), six visual counts were conducted per day.During some days it was not possible to perform all six observations due to tidal time and variation in sunlight period.
The parameters recorded by visual census were abundance, by individual counting, and life cycle stage, by visual evaluation of individual total length (TL).To avoid double counting, the abundance was determined as the maximum number of individuals in each zone at the same time.TL was considered as a measure from the tip of the snout to the end of the upper lobe of caudal fin.TL measurements obtained by visual census are not usually considered accurate but were suitable for this study.According to their length, lemon sharks were divided in four life cycle stages: newborn (less than 65 cm), juvenile (between 65 and 150 cm), sub-adults (between 150 and 220 cm), and adults (more than 220 cm) (Agra, 2009).Statistical analyses were performed with Past version 4.03 (Hammer et al., 2001).Chi-square test (χ 2 ) was used to verify the difference between the number of observations between flood and ebb tidal currents in the same day, assuming 1:1.The mean number of individuals obtained by each season was tested by analysis of variance using a non-parametric test (Kruskal-Wallis, considering statistically significant differences for p < 0.05).
Shark abundance (total, newborn, and juveniles) was compared by season (austral summer or austral winter), tidal current (flood or ebb), and tide (spring or neap).

RESULTS
A total of 2,306 sightings were recorded in 2015, with 1,168 during austral summer and 1,138 during austral winter.Tidal current Corrêa et al. and abundance (± standard deviation) of lemon sharks per day rejected the hypothesis of different number of sightings according to tide since only two days exhibited significant differences between the number of individuals sighted at flood and ebb currents (d.f.= 1, χ2 =3.841, p=0.95).
During the austral summer trip, a total of 40 visual censuses were conducted, with 21 at flood current and 19 at ebb current, amounting to 240 visual counts.Mean abundance of lemon sharks was 29.20 (± 5.04) per day, ranging from a minimum of 12 to a maximum of 35 specimens, and variance of 25.39.On the other hand, 37 visual censuses were conducted during the austral winter trip, 19 at flood current and 18 at ebb current, amounting to 222 visual counts.Mean abundance of lemon sharks was 30.76 (±6.04) per day ranging from a minimum of 16 to a maximum of 41, and variance of 36.52 (Figure 2).Kruskal-Wallis test was used to compare the mean between two seasons, showing that there were no significant differences between the mean values of total shark abundance between the two field trips (p=0.1564).Nevertheless, there were significant differences in abundance of newborns (p=1.11x 10 -15 ) and juveniles (p=3.12 x 10 -14 ) between the seasons.
The analysis of variance regarding shark abundance considered the sum of data from two field trips.There were significant differences in the mean total abundance of individuals between flood and ebb currents (p=0.002716).However, there were no significant differences in mean abundance of newborns (p=0.5733) and juveniles (p=0.236) between flood and ebb currents.Additionally, there were no significant differences in abundance of newborns (p=0.6937) and juveniles (p=0.05372), and no significant differences in total abundance (p=0.2141) between spring and neap tides.
Lemon shark abundance in different tides showed the highest average in neap tide in both austral summer and winter (Table 1).Maximum number of individuals per day was in neap tide in both seasons.There were significant differences among data from observations in spring and neap tide in the first field trip (p=0.01172),but not in the second field trip (p=0.3686).Mean lemon shark abundance was higher at flood of spring and neap tides compared to ebb current in both field trips.At flood current, low standard deviation values showed that the number of individuals sighted was numerically close to average.However, the standard deviation values of ebb current were higher, which means that the number of sharks sighted fluctuated throughout the day (Table 1).
The behavior of lemon sharks was similar concerning both tide and season.Newborn and juvenile lemon sharks enter and leave the bay following flood and ebb currents, respectively.It was observed that some sharks leave the bay before all water has flowed out taking three different ways: to Laguna (Southeast), to Barretinha (Southwest of Lama Bay), or around Cemitério Island (South).However, some sharks stayed in the mouth of Lama Bay, often in neap, when tide amplitude was lower.
Concerning life cycle stages, lemon shark sightings during the two field trips were quite different: newborns and juveniles were observed in the first field trip (Figure 3), whereas only juveniles were sighted in the second field trip (Figure 4), some reaching 1 m TL.Moreover, two sub-adults (180 cm) were observed out of the bay on different days during summer expedition, but they were not considered for overall analysis.

DISCUSSION
Non-invasive methods of observation from land points and underwater images are important tools for conservation of biodiversity.These methods are being adopted to investigate the ecology and behavior of elasmobranchs as a cheap and simple technique (Castro and Rosa, 2005;Agra, 2009;Rada, 2010).Thus, this study's proposed technique, based on visual census from emerged fixed points, reinforces that non-invasive methods can be effective alternatives and can be applied for continuous monitoring with a group of trained people.
From captures with gillnets and longlines to tag individuals with rototags (a plastic mark for tag-recapture studies) and transmitters, several studies have reported declines on lemon shark abundance at Rocas Atoll (Freitas et al., 2006(Freitas et al., , 2009;;Wetherbee et al., 2007).Additionally, Freitas et al. (2009)  Similarly, a fluctuation in numbers of juvenile lemon shark populations through years was reported in Bimini nursery (White et al., 2014).
Our observations resulted in a mean number of 29 individuals per day (maximum 35) in January/ February and 30 (maximum 41) in June/July, approximately twice the number of sharks observed since 2003.Considering the information added by this study and the previous data, we could verify a great variation in the number of lemon sharks, with a higher decline from 1998 to 2003, and an increase reported in 2015 from 14 to 29-30 individuals.Although the methods adopted in both studies were not the same and despite a temporal lag of registered data, there was a consistent increase in the average number of individuals observed in this period.Freitas et al. (2009) attributed different causes that may explain the decrease in the number of juvenile lemon sharks sighted at Lama Bay: high mortality due to previous tag research; natural fluctuation in number of reproductive females and, consequently, a low birth rate; and, finally, even though Rocas Atoll is a Biological Reserve, illegal fishery activities around the area, could influence the resident fauna.Corrêa et al.Since this site is a full protection marine area since 1979, fishing practices have been prohibited for a long time.In addition, shark finning has been prohibited in Brazil since November 2012 by "Instrução Normativa Interministerial MPA/MMA number 14."However, as Brown and Gruber (1988) and Feldheim et al. (2001) pointed out, adult lemon sharks exhibit a nomadic movement pattern, doing incursions in deep water and, sometimes, migrations.This is the reason why they could be captured off the MPA, reducing females and births due to direct adult catches.
Our results showed an increasing abundance trend that may provide insights concerning the recovery potential of lemon shark populations.Data from a non-exhaustive literature-based study present cases of elasmobranch population increase (Ward-Paige et al., 2012).According to that study, the main causes listed for population recovery were the reduction of fishing mortality by traditional fishery management and incidental by-catch, along with shark finning prohibition.In addition, ecosystem-based strategies, such as the creation of shark sanctuaries and restoration of their natural habitat, nature-based education, and species-specific policies of conservation may have contributed to the positive results (Ward-Paige et al., 2012;Ward-Paige and Worm, 2017).
According to Costa et al. (2016), the tides play an important role in driving the circulation at Rocas Atoll and the waves drive morphological changes on the reef islands.Additionally, the extreme tidal flux results in an exposed shallow area during low tide (Gherardi and Bosence, 1999;Silva et al., 2002).As a consequence, the tidal regime and local depth influence the movement patterns and the behavior of juvenile lemon sharks at Lama Bay (Wetherbee et al., 2007).This area is described as an unusual refuge area for growth and feeding for this species due to absence of mangroves or seagrass habitats (Freitas et al., 2006).According to Agra (2009), there was segregation by size in the lemon shark population at Rocas Atoll: juveniles were more frequent in closed shallow waters far from the ocean connection.
In this study, we corroborate that the behavior of juvenile lemon sharks follows the strong tidal currents.We observed that some sharks left the bay before all water flowed out, taking different ways: to Laguna, Barretinha, and around Cemitério Island.Some sharks stayed in the mouth of Lama Bay, often in neap tide, when the tide amplitude was lower.These movements could explain the differences in the number of lemon shark sightings at flood and ebb currents, the highest average in neap tide comparable with spring, and the maximum number of sharks per day in neap tide in both seasons.During high tide, this current flows in ocean-atoll direction and during low tide it flows in opposite direction, allowing sharks to move in and out of Lama Bay, which is considered a primary and secondary nursery area (Oliveira et al., 2011).
Previous studies of lemon shark at Rocas Atoll reported that their copula and birth probably occurred at the beginning of the year, leading to a gestation period ranging from 10 to 12 months.This finding is based on the co-occurrence of newborn, juvenile, and adult females with bite marks in December, February, and March (Oliveira et al., 2011).The results of our study corroborate the previous results since newborn and juveniles were sighted in the first field trip (January/February), whereas only juveniles were seen in the second trip (June/July), four months later.Parturition of lemon sharks was observed from November to April at Fernando de Noronha Archipelago (Garla et al., 2009), with mating behavior during the austral summer (Clapis- Garla et al., 2022).In Rocas Atoll, Freitas et al. (2006) observed lemon sharks' births from February to May.
Based on data collected in this study associated with the previous information in the area, it is possible to indicate that the birth of the individuals was between December 2014 and February 2015.Furthermore, by visual census, all the newborns from the first field trip (January-February) already had attained their juvenile period in the second field trip (June-July) since the estimated growth rate of this species is 24.7 cm year -1 (±3.4) in total length (Freitas et al., 2006).
Our results showed no significant difference in abundance of lemon sharks between summer and winter and it was not observed a pattern associated with spring and neap tides.Therefore,

CONCLUSION
The estimate of lemon shark abundance at Lama Bay, Rocas Atoll, in 2015 during austral summer and winter suggests a possible population recovery based on the previous reported data.The consistency of observations between two periods in this study reinforces and qualifies the non-invasive method as an alternative tool to understand population dynamics that is independent of fisheries data.Our results reported newborns and juveniles in January/February, but only juveniles in June/July, suggesting birth of lemon sharks during the summer and their growth afterwards, which is compatible with estimated growth rates.No tidal pattern and abundance relationship was identified.As a sharp decline in the abundance of this species was detected in previous publications, new studies are needed to understand the causes of its numerical fluctuation.

Figure 2 .
Figure 2. Median of the number of N. brevirostris across tides, the interquartile range, range and maximum and minimum values, from both field trip, at Lama Bay, Rocas Atoll.Austral summer field trip (January/February 2015): flood and ebb currents at spring tide; and flood and ebb currents at neap tide.Austral winter field trip (June/July 2015): flood and ebb currents at spring tide and flood and ebb currents at neap tide.Mean number of individuals presented in red.

Figure 3 .
Figure 3. Number of sightings of lemon sharks (N.brevirostris) per life cycle stage (newborn and juveniles) sighted on the austral summer (January/February 2015) at Lama Bay, Rocas Atoll (F = flood; E = ebb).

Figure 4 .
Figure 4. Number of sightings of lemon sharks (N.brevirostris) per life cycle stage (newborn and juveniles) sighted on the austral winter (June/July 2015) at Lama Bay, Rocas Atoll (F = flood; E = ebb).
observed juvenile lemon sharks in Lama Bay for six years (from 1998 to 2003) in seven expeditions of 20 days each.Using visual census, they reported a decrease in mean numbers of individuals from 64 in March 1998 to 58 in March 1999; from 41 in March 2000 to 11 in August 2002; and stable values of 15, 15, and 14 individuals in March, May, and October 2003, respectively.

Table 1 .
Mean lemon shark abundance (± standard deviation; maximum), in flood and ebb currents of spring and neap tides in both austral summer (January/February 2015) and winter (June/July 2015) field trip at Rocas Atoll.Bold numbers showed the highest average of each field trip.
Corrêa et al. studies on other parameters are necessary to clarify the influence of hydrography in the demography of N. brevirostris.The proxies based on records of population abundance from research surveys must use standardized observations to obtain consistent long-time results.Although movement patterns were directly influenced by tide and depth, other variables may influence biotic interactions.Continuous monitoring studies are fundamental to verify how different parameters could influence movement patterns.Currently, management plans for lemon sharks are inexistent in Brazil.In this context, this study and future research should provide knowledge to develop conservation plans for lemon shark preservation.