Distribution and hotspots of the feeding areas of jaguars on sea turtles at a national park in Costa Rica

A poorly described aspect of the trophic relation between sea turtles and jaguars is the distribution and hotspots of the feeding areas of jaguars on the nesting beaches. It is very important to identify the areas where sea turtles are predated because we could concentrate conservation and management efforts in these areas. Therefore, the aim of this work is to describe the spatial distribution and hotspots of the feeding areas of jaguars at Nancite beach, Santa Rosa National Park, Costa Rica. We recorded a total of 76 predated carapaces of sea turtles, of these, 54 (71%) were of Lepidochelys olivacea and 22 (29%) of Chelonia mydas. Two major feeding hotspots areas were identified within the Nancite beach. Both hotspots are located at the extremes of the beach, one is at the southern edge and the other is at the northern extreme. Human activity and the distribution of nesting turtles influence synergistically to determine the sites where the sea turtles are predated at Nancite beach. Based on the information of predation hotspots, the environmental authorities should regulate the monitoring activities within those areas to avoid interfering with the trophic relation between sea turtles and jaguars.


Introduction
The effect of predators over adult marine turtles has been overlooked because it is difficult to observe and quantify (Heithaus et al., 2008).On the sea, the most common predators of adult sea turtles are sharks and killer whales Orcinus orca (LINNAEUS 1758) (Heithaus et al., 2008).When female adult sea turtles are nesting on the beach, only four different predators have been recorded, namely jaguars Panthera onca (LINNAEUS 1758), American crocodiles Crocodylus acutus (CUVIER 1807), coyotes Canis latrans (SAY 1823), andhumans (Ortiz et al., 1997;Drake et al., 2001;Aguirre et al., 2006;Heithaus et al., 2008;Alfaro et al., 2016).Recently, the first record of consumption of sea turtle by a cougar Puma concolor (LINNAEUS 1771) has been reported but it is not clear whether the cougar was acting as a predator or as a scavenger (Escobar-Lasso et al., 2016a).
It is important to highlight that the jaguar and the American crocodile (Crocodylus acutus) are the only predators recorded in Nancite beach capable to kill sea turtles (Cornelius, 1986;Carrillo et al., 1994;Ortiz et al., 1997;Figure 1).However, predation events by American crocodiles on sea turtles are extremely rare in this beach (Ortiz et al., 1997).Sea turtles are important food sources for jaguars because they are easy prey and they represent a big amount of biomass (Veríssimo et al., 2012;Arroyo-Arce and Salom-Pérez, 2015;Guilder et al., 2015).Additionally, they can be key resources when other prey availability is low (Veríssimo et al., 2012).Recently, it has been discovered that many species of vertebrate scavengers could be indirectly benefited by the predator-prey relationship between jaguars and sea turtles (Escobar-Lasso et al., 2016b).Therefore, the sea turtles could also be key resources for scavengers in periods when the availability of other types of carcasses is low (Escobar-Lasso et al., 2016b).
The trophic relation between sea turtles and jaguars is one of the most interesting and controversial because it involves six species threatened with extinction (Veríssimo et al., 2012;Arroyo-Arce et al., 2017).On the other hand, sea turtles and jaguars are widely recognized as conservation flagship species (Caro et al., 2004;Eckert and Hemphill, 2005) and large-scale projects have been developed to conserve them (Bjorndal et al., 1999;Ceballos et al., 2002).Historically, four species have been recorded nesting on Nancite beach, namely Lepidochelys olivacea, Chelonia mydas, Eretmochelys imbricata and Dermochelys coriacea (Cornelius, 1986;Fonseca et al., 2009).However, D. coriacea and E. imbricata are extremely rare and their occurrence in Nancite is very low (Cornelius, 1986).It must be highlighted that L. olivacea has been catalogued as vulnerable and C. mydas as endangered by the IUCN, also, these species are decreasing worldwide, according to their population trend (Seminoff, 2004;Abreu-Grobois and Plotkin, 2008).In regard to the jaguar, this species is considered to be near threatened by the IUCN, and its populations are also decreasing (Caso et al., 2008).For their critical state, knowledge about these species and its predatory interactions is imperative.The knowledge of the ecological aspects of jaguar-sea turtle relationships may contribute to their conservation and to take management decisions (Alfaro et al., 2016).The temporal and spatial trends of jaguar predation on turtles have been recently studied (Veríssimo et al., 2012), as well as carcass utilization rates (Tröeng, 2000;Guilder et al., 2015;Alfaro et al., 2016), the intraspecific interactions between jaguars (Escobar-Lasso et al., 2016c), the influence of the scavengers on the feeding behavior of the jaguar on sea turtles (Escobar-Lasso et al., 2016b) and the impact on sea turtle populations (Arroyo-Arce and Salom-Pérez, 2015).However, a poorly described aspect is the distribution of the feeding areas of the jaguar on sea turtles on its nesting beaches (Veríssimo et al., 2012;Arroyo-Arce and Salom-Pérez, 2015;Alfaro et al., 2016).
There have been reported predations of sea turtles by jaguars in eight beaches throughout America (Fretey, 1977;Autar, 1994;Carrillo et al., 1994;Keeran, 2013;Cuevas et al., 2014;Guilder et al., 2015;Alfaro et al., 2016;Arroyo-Arce et al., 2016), but in only two beaches, both in Costa Rica, the distribution and hotspots of the feeding areas of jaguars on sea turtles were reported: (a) Tortuguero National Park (Veríssimo et al., 2012;Arroyo-Arce and Salom-Pérez, 2015) and (b) Naranjo beach, Santa Rosa National Park (Alfaro et al., 2016).These works have found that the distribution of the feeding areas varies spatially and there are distinct hotspots across the beaches.Identify the distribution and hotspots of the feeding areas of the jaguars on sea turtles is very important because all conservation and management efforts should be concentrated in these areas to maintain the trophic relationship between jaguars and sea turtles.Therefore, the aims of this work are: (a) to document the number of predation events of jaguars on sea turtles at Nancite beach and (b) to describe the spatial distribution and hotspots of the feeding areas of jaguars on sea turtles at Nancite beach, Santa Rosa National Park, Costa Rica.

Study site
Santa Rosa National Park is located in the Guanacaste Province, Northwestern Pacific coast of Costa Rica.Santa Rosa comprehends one of the best-preserved dry forests of Central America.This National Park has an extension of 38,628 ha and it is within a block of 163,000 ha of protected land within the Guanacaste Conservation Area.In this park there are several important sea turtles nesting beaches (e.g.Naranjo, Colorada, Nancite, Isla San Jose, Potrero Grande, among others).One of the most important nesting beaches for sea turtles is Nancite (Cornelius, 1986), located in the Southwestern part of Santa Rosa National Park (10º48´N and 85º39´W; Figure 2); it has a length of approximately 1050 m and it preserves mainly coastalmarine ecosystems (besides mangroves, lagoons and dry deciduous and semi-deciduous forests).
Nancite beach has been fully protected from intrusive human activities, including tourism and it supports a well preserved wild ecosystem (Figure 3).This beach is important because is one of the few places in the world that presents the olive ridley arribada phenomenon (Cornelius, 1986;Fonseca et al., 2009).This phenomenon consists of the massive synchronous nesting of hundreds or thousands of olive ridley over a few consecutive nights (Cornelius, 1986;Bernardo and Plotkin, 2007;Fonseca et al., 2009).This phenomenon also occurs in other beach in Costa Rica (Ostional beach) and some beaches in Mexico, India, Nicaragua and Panama (Fonseca et al., 2009).

Methods
The distribution and hotspots of the feeding sites of jaguars on sea turtles was assessed through diurnal surveys.These surveys were made by Escobar-Lasso be- Distribution and hotspots of the feeding areas of jaguars on sea turtles at a national park in Costa Rica tween October 1 st 2015 and February 29 th 2016 at Nancite beach.It is noteworthy that during the performance of this investigation the area was presenting the climate conditions called "El niño".During these surveys we look for carapaces of sea turtles (Figure 4) inside the beach vegetation and nearby forests (Figure 3).The following variables were recorded: (i) the species of sea turtle, (ii) the geographic location (coordinates), and (iii) the distance in meters that the jaguar dragged the sea turtle carcasses from the beach to the forest (see Appendix 1).In order to avoid duplication of carcasses, we mark every carapace with red painting when its location was recorded.Therefore, when a carapace was found it was examined to identify jaguar predation signals (e.g.bite marks on the anterior part of carapace; Figures 3 and 4).We assume that every carapace found in the beach vegetation and nearby forest corresponds to a predation event by the jaguar.
The identification of the species of sea turtle was based on its carapace (Cornelius, 1986).The olive ridley sea turtle has ovate shaped carapace, 5 to 9 lateral scales and an average length of 65 cm (Figure 4a).The green sea turtle has a drop shaped carapace, 4 lateral scales and an average length of 80 cm (Figure 4b).
The distribution and hotspots of the feeding sites of jaguars on sea turtles were computed and mapped using the plugin Heatspots of Qgis version 2.14.The hotspots size was calculated using the plugin Measuring areas of Qgis version 2.14.We used t-Student test to evaluate the differences in distance of drag of the carcasses of green versus olive ridley sea turtles.The statistical analysis was performed using R language with Rcmdr interface (Fox, 2005) and the graphics were made using the ggplot2 package (Wickham, 2009).

Results
We recorded a total of 76 carapaces of sea turtles predated by jaguars, 54 (71 %) of which were on olive ridley sea turtles and 22 (29%) on green sea turtles.The carapace of the olive ridley sea turtles presented a length of 67.1 ± 2.94 cm (61-73.5 m) and a width of 71.4 ± 3.41 cm (64-82.2m) (Figure 5).The green turtle carapaces had a length of 83.5 ± 4.92 cm (65-89 m) and a width of 78.6 ± 4.23 cm (68-86 m) (Figure 5).The jaguars dragged the sea turtles' carcasses from the beach to the forest 78.39 m in average (± 77.46 m, range = 3-336 m).For the green sea turtles the average distance of drag was 60.72 ± 56.69 m (3-220 m) and for olive  ridley sea turtles it was of 85.59 ± 83.88 m (8-336 m).However, there were no statistical differences in the distance of drag between both species (t = 1.49, p = 0.140).
By analyzing the distribution and hotspots of the feeding areas of jaguars on olive ridley and green sea turtles, we found spatial variation and separated feeding hotspots areas across the Nancite beach (Figure 6).Furthermore, two major hotspots of feeding were identified within the study area.The hotspots are located on the southern and northern extremes of the beach (Figure 6).The southern hotspot is slightly greater (19 events; 837 m 2 ) compared with the northern hotspot (11 events; 570 m 2 ).
Although the numbers of predation events were different for green and olive ridley sea turtles, the distribution and hotspots of the feeding areas of jaguars were similar; therefore, two major feeding hotspots areas were identified regardless of the species (Figure 7).In the species-specific distribution, for green sea turtles the southern hotspot is slightly greater (6 events; 600m 2 ) and denser compared with the northern hotspot (two events; 50m 2 ) (see Figure 7a).On the contrary in olive ridley sea turtles the northern hotspot is slightly greater (17 events) but highly dispersed   Neotropical Biology and Conservation Distribution and hotspots of the feeding areas of jaguars on sea turtles at a national park in Costa Rica (480 m 2 ) compared with the southern hotspot (13 events) that is less dispersed (410 m 2 ) (Figure 7b).

Discussion
Our results show that the jaguars dragged the sea turtle carcasses from the beach to the forest a maximum distance of 336 m.Alfaro et al. (2016) found that the jaguars dragged the carcasses a maximum distance of 1025 m on Naranjo beach.It is believed that the jaguars drag the sea turtle carcasses into dense vegetation as an attempt to conceal them from scavengers (Guilder et al., 2015).The green sea turtle is heavier (70-125 kg) compared to olive ridley sea turtle (35-45 kg) (Cornelius, 1986), in any case, we did not find statistical differences in the distance of drag among green and olive ridley sea turtles.It is possible that green sea turtles are not heavy enough to hold the jaguar from harrow them and for this reason no differences were found.Apart from the scavengers, the dragging distance could be related to the vegetation cover type, where less dense cover demands a longer distance of dragging to conceal the carcass.Even when vegetation is highly homogeneous along the beach, studies addressing this variable could reveal more about dragging distance.
Our results demonstrate that Nancite beach is not used uniformly for feeding by jaguars.The distribution of the feeding areas varies spatially, and it reveals distinct hotspots across Nancite beach.Similar results also have been reported in Tortuguero National Park (Veríssimo et al., 2012;Arroyo-Arce and Salom-Pérez, 2015) and in Naranjo beach, Santa Rosa National Park (Alfaro et al., 2016).For jaguars, the three main factors that determine the distribution and hotspots of the feeding areas are: (a) distribution of the human activity along the beach, (b) distribution of the nesting female turtles along the beach, and (c) interactions and territorial behaviors among jaguars (Veríssimo et al., 2012;Arroyo-Arce and Salom-Pérez, 2015;Alfaro et al., 2016).These three factors may act independently or synergistically depending on the particular circumstances for each beach (Veríssimo et al., 2012;Arroyo-Arce and Salom-Pérez, 2015).
Several studies have recorded that the jaguar tends to avoid human-dominated areas (Cullen et al., 2013).Therefore, it is expected that the core areas of jaguar activity and its feeding hotspots areas would be located away from infrastructure and human presence (Arroyo-Arce and Salom-Pérez, 2015).For example, in Tortuguero beach the sea turtle feeding hotspots are located in the farther extreme from the Tortuguero village (Veríssimo et al., 2012;Arroyo-Arce and Salom-Pérez, 2015).Similarly, Alfaro et al. (2016) recorded that the feeding areas are located far from camping sites (which have the higher tourist concentration).At Nancite beach the tourism is strictly prohibited, however the beach is open to scientific research.
During most of the year there are groups of researchers and volunteers monitoring the nesting of sea turtles.These groups are composed in certain occasions of up to 25 people, who live on the biological station of Nancite and patrol the beach all nights, time during which the jaguars are in activity and feed on sea turtles.The human activity restricts the feeding hotspots areas of the jaguars at Tortuguero beach (Veríssimo et al., 2012;Arroyo-Arce and Salom-Pérez, 2015) and at Naranjo beach (Alfaro et al., 2016).We believe that the negative effect of human activity could be higher in Nancite considering that this beach has a length of only 1050 m, which is much smaller, compared to Tortuguero (with 29 km of length) and Naranjo (with 6 km of length) beaches.The areas without human presence are less in Nancite due to its small size.Our results suggest that human activity could restrict the sea turtle predation area in Nancite because the main path that leads to the beach and the biggest infrastructure of the station are located in the center of the beach.Therefore, the activity of researchers and volunteers is concentrated in the center of the beach, which could explain the fact that sea turtle predation hotspots are located on the southern and northern extremes, and not in the center of the beach.
In addition to human activity, the spatial distribution of the nesting sea turtles along the beach is other factor that could also be affecting the spatial arrangement of jaguar feeding areas in Nancite beach.Interestingly, at Naranjo and Tortuguero, the areas with more nesting sea turtles are also the zones with less human activity, and these concentrate the sea turtle feeding hotspots (Veríssimo et al., 2012;Arroyo-Arce and Salom-Pérez, 2015;Alfaro et al., 2016).We find something similar on Nancite beach.The areas of greatest sea turtle nesting in Nancite are in the central and northern part of the beach (Fonseca et al., 2009).Hence, in the predation hotspot located in the north of Nancite (Figure 7) the pattern recorded in Naranjo (Alfro et al., 2016) and Tortuguero (Arroyo-Arce and Salom-Pérez, 2015) is repeated.However, there is a high human activity in the center of the beach at some point of the year, and it might be possible that this prevents the jaguar from hunting.This allows the formation of a hotspot in the southern part of the beach, even when this area has less sea turtle nesting compared to the central and northern areas of the beach.This suggests that both human activity and the distribution of nesting turtles influence synergistically to determine the sea turtle hunting area and the sea turtle predation hotspots in Nancite beach.However, other factors specific to the habitat may also determine the size and distribution of the feeding sites.Habitat variables have not been evaluated in this study, nor in previous works (Verissimo et al., 2012, Arroyo-Arce and Salon-Pérez, 2015, Alfaro et al., 2016).
Traces of wild carnivores are scarce where people are allowed (Coghlan, 2008).Ideally, the research activities should not interfere in the trophic relations between jaguars and sea turtles.In this sense, to identify the sea turtle feeding hotspots is very important to guide management actions and prevent or mitigate all possible threats that may alter the trophic relationship between jaguars and sea turtles.Therefore, the environmental authorities of Nancite, based on the information of the feeding hotspots, should apply strategies to protect the trophic relations within those key zones.In this sense, research activities could be reorganized in order to be extra-cautious with these ecological interactions.A management option could be a different distribution of people in night patrols; they could wait at stable sites as far as possible from the nesting hotspots.Also, there should be less movement across the beach to avoid disturbing the jaguars.
Within Santa Rosa National Park there are several important sea turtles nesting beaches where jaguars have been recorded preying sea turtles (e.g.Naranjo, Colorada, Nancite, Potrero Grande).The sea turtle predation hotspots have been recorded only in Naranjo (Alfaro et al., 20016) and Nancite (this work).Therefore, additional research must be done at Potrero Grande and Colorada beaches to document the sea turtle predation hotspots.

Figure 1 .
Figure 1.Jaguar (Panthera onca) and American crocodile (Crocodylus acutus) feeding on a sea turtle on Nancite Beach, Santa Rosa National Park, Costa Rica.Photos by Luis G. Fonseca (A) and Juan Carlos Cruz Díaz.

Figure 4 .
Figure 4. Carapaces of sea turtles predated by jaguars (Panthera onca) at Nancite beach, Santa Rosa National Park, Costa Rica.(A) Carapace of olive ridley sea turtle (Lepidochelys olivacea) and (B) carapace of green sea turtle (Chelonia mydas).Note the bite marks of the jaguar in the fore part of the sea turtles' carapaces.

Figure 6 .
Figure 6.Distribution and hotspots of the predation events by jaguars (Panthera onca) on sea turtles at Nancite beach, Santa Rosa National Park, Costa Rica.

Figure 5 .
Figure 5. Differences between the length (A) and width (B) of the carapaces of olive ridley (Lepidochelys olivacea) and green (Chelonia mydas) sea turtles, which were killed by jaguars (Phantera onca) and were found in the surrounding vegetation of Nancite beach, Santa Rosa National Park, Costa Rica.

Figure 7 .
Figure 7. Distribution and hotspots of the predation events by jaguars (Panthera onca) on green sea turtles (Chelonia mydas) (A) and olive ridley sea turtles (Lepidochelys olivacea) (B) at Nancite beach, Santa Rosa National Park, Costa Rica.
-HoyosVolume 12 number 1  january -april 2017 Records of carapaces of sea turtles found in the beach vegetation and nearby forest at Nancite beach, Santa Rosa National Park, Costa Rica.Each carapace corresponds to an independent predation event by the jaguars.Distribution and hotspots of the feeding areas of jaguars on sea turtles at a national park in Costa Rica