Fidelity to nesting sites and orientation of Trachemys dorbigni (Duméril & Bibron, 1835) (Testudines: Emydidae) female in southern Brazil

Abstract

Orientation and fidelity to nesting areas were analyzed in a population of Orbigny's slider turtle, Trachemys dorbigni (Duméril & Bibron, 1835), in southern Brazil. Monitoring was carried out over a 5-year period using the mark–recapture technique for females moving overland during the nesting season. The positions of 90 females were analyzed. We also tested the capacity to return to the nesting site after the female was translocated to another point. The overall mean distance between two consecutive captures for females that had not been translocated to the release point was 545.0 m, whereas for females that were translocated the mean distance was 523.8 m. Translocated females showed a high capacity to return to the nesting site, but 2 years were necessary.

Keywords:

• fidelity
• movements
• nesting site
• orientation
• Emydidae

Introduction

Habitats selected by individual organisms can affect population and community dynamics (Blakesley et al. 2010; Hodson et al. 2010). Factors that determine the selection of a nesting site in freshwater turtles vary widely according to the species. Investigations of these factors have demonstrated the influence of soil temperature and humidity, vegetation cover, slope, and substrate granulometry, in addition to countless others (Ross and Anderson 1990; Janzen 1994; Bodie et al. 1996; Wood and Bjorndal 2000; Fagundes and Bager 2007). Turtles' nest-site characteristics have a large impact on reproductive success because offspring phenotypes produced under varying nest sites have differential fitness (Valenzuela 2001) and may determine several population parameters. The most studied are the influence of the sex ratio on species in which the sex depends on the incubation temperature (Schwarzkopf and Brooks 1987; Janzen 1993; Roosenburg 1996), variation in the rate of predation on nests (Temple 1987; Buhlmann and Coffman 2001), change in hatching success rates (Fowler 1979; Horrocks and Scott 1991; Janzen 1994), and variation in body size (Janzen 1993).

Considering that turtles are able to identify better nesting sites, the capability of returning to the same area in the future can result in a large energy saving for the female, and a significant increase in the survival potential of the hatchlings. Marine species of chelonians show strong fidelity to nesting sites, even though the turtles may travel for long distances to reach them (Lohmann et al. 1999; Richardson et al. 1999), but data about the locations of freshwater turtle nests in successive years are rare, and show higher variability than marine turtles.

The Brazilian slider turtle (Trachemys dorbigni Duméril & Bibron, 1835) is the southernmost occurring species of the genus. It inhabits marshes, ponds, and slowly flowing rivers of southern Brazil, Uruguay, and Argentina. Information on most biological and ecological aspects of T. dorbigni in the field is lacking (Bager et al. 2007), mainly in nature. Despite T. dorbigni being abundant in some regions, it is listed as near-threatened in Brazil, and several direct and indirect humans impacts affect its populations in Brazil, and probably in other parts of its distribution area. These impacts include occupation of nesting sites for agriculture (mainly rice; Bager and Rosado 2010), road mortality during the breeding season (Bager 2003), and trade of hatchlings for pets (A. Bager, unpublished data). The majority of the nests of T. dorbigni are farthest from the shore, extending up to 250 m from the water (Bager and Rosado 2010), in an aggregated pattern (Fagundes et al. 2010). Nests are identified in both sand and clay soils, and females were found digging on roadsides in areas exposed to the direct sun and with little or no vegetation (Bager et al. 2007).

There is no information about the fidelity to the nesting area or orientation capacity of Orbigny's slider turtle, T. dorbigni. In this study, we used female location data recorded during 5 years to determine whether these species exhibit fidelity to the nesting area, analyzing information about translocate and non-translocate females. We tested the hypotheses that (1) females of T. dorbigni have fidelity to nesting area during the same and in subsequent breeding seasons and (2) females translocated to areas outside their nesting site can return to the same site.

Material and methods

Study area

The study was conducted in an area of approximately 180 ha, located in the limit zone between the Taim Ecological Station, a federal protected area, and extensive plantations of irrigated rice, in the state of Rio Grande do Sul, Brazil (UTM 22H 346185 mE 6365666 mN; Figure 1). Intensive mechanized rice farming dominates the study area, and it is devoid of perennial vegetation except for two contiguous stands of eucalypts (Eucalyptus sp.) totaling 1.5 ha. We monitored turtles along a 5000 m stretch of the western shore of Mangueira Lagoon. This lagoon has 100 km of extension and maximum width of 11.5 km. The climate is humid subtropical (Koeppen 1948). Rainfall regularly occurs throughout the year (annual mean precipitation = 1252 mm) and temperatures range between − 3°C and +18°C for most of the year, rising above 22°C in the warmest months (Moreno 1961).

Figure 1 Location of the Taim Ecological Station in the extreme southern part of the state of Rio Grande do Sul.

Data collection

We captured adult female turtles over five consecutive years (1997 through 2001), from September to December, during the nesting season (Bager et al. 2007; Fagundes et al. 2010). Sampling efforts varied among years (1997: 43 days; 1998: 50 days; 1999: 115 days; 2000: 37 days; 2001: 65 days), when one or two persons sampled for 10 h/day.

Females were captured in the field by hand during their reproductive activity (digging, nesting, or covering nest) because we needed to exactly identify the nest position. All females were palpated to confirm eggs' presence. We measured carapace length on all turtles upon initial capture and permanently marked them by notching on the marginal scutes (Bager et al. 2007, 2010). We recorded the Global Positioning System (GPS) location (10 m precision) for each individual and then released them at the site of capture within 20 min. All the turtles captured in 1999 were translocated and released at the midpoint of the monitored area (released point), on the shore of Mangueira Lagoon.

Data analysis

Nest-site fidelity is the tendency for an individual to return to the same geographic location or microhabitat type to nest (Valenzuela and Janzen 2001). We defined “fidelity” as the capacity of returning consecutive times to the same nest site in lower distances than 20% of total potential area (5000 m). In this study, this capacity was evaluated considering the distance variation of capture and recapture of females. Non-translocated females are animals that were captured and released in the same point, whereas translocated females are the ones that were captured during reproductive activity, transported to the “release point” and released there. We compared the distance between the positions of capture and recapture in five different time periods (Years 0–4) for the non-translocated females and three different time periods (Years 0–2) for translocated females. Year 0 corresponds to females captured and recaptured in the same nesting season, Year 1 corresponds to females captured in one nesting season and recaptured in the following one, and so on for the remaining years.

Orientation was tested taking into account non-translocated females as the control group, whereas translocated females were the experimental group. The position of the recaptured females was compared to the one of the first capture even when the female had been captured for more than one time. When recapture distance was shorter than the transference distance, there would be a non-randomized movement of active searching through the original nesting area. Complementary orientation analysis used two compass bearings, one for capture and other for recapture. The compass bearing for capture was calculated between the straight line between the geographic north and the straight line between the capture point and the release point. The compass bearing for recapture originated at the release point, and was compared with the angle formed between geographic north and the recapture position. Data were analyzed with circular analysis (Rayleigh test) using software Bioestat (Ayres et al. 2007). The Rayleigh test verified whether the angular distribution was uniform around the circumference. Circular analysis only considered females captured from the south to the release point (n = 35) because north had few specimens (n = 8).

Normal data were compared through ANOVA, whereas non-normal or heteroscedatic distribution data were compared by the Kruskal–Wallis median test. We eliminated outliers of distances to fidelity analyses using statistics, and repeated them until the upper and lower limits established by the software included all of our data. We considered a 5% confidence level as significant in all analyses.

Results

We analyzed capture and recapture data for 90 females, totaling 226 contacts. The mean number of captures per individual was 2.5 (minimum = 2; maximum = 6) and carapace length was from 203 to 278 mm ( = 234.4 mm; 1 SD = 11.8; n = 89).

Non-translocated female nesting site fidelity

The time interval between two consecutive captures of non-translocated females was from one day to 4 years (1518 days). Mean for the distance between the first and second captures was 545 m (minimum: 8.1 m; maximum: 2160.9 m). Twenty-six percent of these females returned from distances less than 200 m from the first capture point, and 80% were recaptured at distances less than 600 m from that point. Turtles recaptured in the same year were those which showed the least mean distance between two consecutive captures (451 m; Table 1). However, there was no significant difference in the distances calculated for turtles recaptured at different year intervals (Kruskal–Wallis χ² = 2.54; df = 4; P = 0.636).

Table 1 Analysis of the distances between two consecutive captures of females of T. dorbigni which were translocated and non-translocated to the release point, in relation to time interval (outliers excluded).

	Non-translocated					Translocated
	Mean	Minimum	Maximum	SD	n	Mean	Minimum	Maximum	SD	n
Year 0	451.0	19.2	1595.2	376.63	34	657.9	38.1	2794.1	680.10	23
Year 1	731.2	8.1	1149.0	404.18	7	949.4	484.0	1524.2	528.66	3
Year 2	552.5	211.5	1139.0	339.37	7	233.1	16.0	414.0	107.74	15
Year 3	456.6	83.7	1195.0	508.38	4	–	–	–	–	–
Year 4	701.6	44.8	2160.9	655.30	14	–	–	–	–	–
Mean	545.0	8.1	2160.9	457.89	66	523.8	16.0	2794.1	572.75	41
Notes: Year 0 incorporates the data for capture and recapture of females in the same breeding season; Year 1 includes the data for capture and recapture of females in two consecutive breeding seasons; and so on. Values are in meters.

Translocated female nesting site fidelity

The distance over which the females were relocated from the capture location to the release point varied from 71.3 to 2550 m ( = 1005.7 m; 1 SD = 542.6; n = 43; Table 1). The distance between the points of the first and second captures varied from 16 to 2794 m ( = 523.8 m; 1 SD = 572.7; n = 41), and was significantly less than the transfer distance (ANOVA F _1.84 = 16.670; P < 0.01). Approximately 86% of the translocated females were recaptured at distances shorter than the respective transfer distances, indicating a tendency to return toward the capture point. A total of 35% of the individuals were recaptured at distances less than 200 m from the location of the first capture, and 60% were recaptured less than 400 m away. There was no difference between the means of the distances between the first capture and the recapture, and between the translocated (523.8 m) and non-translocated (545 m) individuals (ANOVA F _1.107 = 0.06; P = 0.804). There was no significant relationship between the transfer distance and the distance between the first and second capture (r ² = 0.03; P = 0.283).

Female orientation

Angular mean from capture point to release point was 9.85 (CI 95%: 0.049–19.658) and 187.47 (CI 95%: 172.97–201.97) from release point to recapture point. Both azimuth of transference and azimuth of capture were different from uniform circular distribution (P < 0.01). Analysis of two related samples, comparing azimuth from the same female, also results in a significant difference (F _1.37 = 70.226; P < 0.001).

Discussion

Reproduction involves a high utilization of energy, and this energy is finite (Rasmussen and Litzgus 2010). Considering the resources spent in producing the eggs, the search for favorable locations for nesting shows a positive trade-off. Thus, it rests on the female to maximize the probability of a successful incubation by seeking areas where environmental impacts (e.g. inundation), predation, and human activity are low, and eclosion and size of offspring are high. Studies on freshwater turtles have shown that females oviposit non-randomly, suggesting the existence of preferences on their part (Christens and Bider 1987; Tucker 2001; Restrepo et al. 2006). Our results demonstrated that T. dorbigni has nest-site fidelity in both translocated and non-translocated females, returning to shorter distance than 600 m in subsequent ovipositions even after several years.

Fidelity to the nesting area in non-translocated females

Greater nesting movements than those mentioned for other authors were found, but it was concluded that they support nest-site fidelity. This statement is based on the fact that Mangueira Lagoon is more than 800 km² and that suitable nesting sites occur along their entire length, then our observation that most females return to a location less than 600 m from the first capture is a strong indication of nest-site fidelity.

The return of female marine turtles in different years to the same beaches is considered as fidelity to the same nesting site, even when two succeeding nests are situated kilometers apart (Talbert et al. 1980). This is because of the difference in scale between the migratory travels and the distance between nests. The mean distance measured between two consecutive nesting, incorporating information from the four nesting seasons, was 545 m. Only four of them moved more than 1500 m, and of these, two moved more than 2000 m. Considering that these distances do not correspond to the normal behavioral pattern of the species, and are isolated cases or possibly sampling errors, we excluded these values from the analysis and obtained a mean distance of 463 m. Congdon et al. (1983) considered that females of Emydoidea blandingi (Holbrook, 1838), which lay their eggs at distances more than 258 m among subsequent nests, do not show fidelity to the nesting site. Previous reports of nest-site fidelity include Kinosternon baurii (Garman, 1891) (Wilson et al. 1999) and Chrysemys picta marginata (Schneider, 1783) (Christens and Bider 1987). Kinosternon baurii move 65.7 m from first capture, and C. p. marginata return to points between 10 cm and 10 m from the position of the previous nest. Christens and Bider (1987) reported that one of the main nesting areas was < 1 ha, which certainly influenced the proximity of consecutive nests. Females may select areas favorable for oviposition without deliberately returning to the exact location of the previous year's nest. Rowe et al. (2005) reported that C. p. marginata moved 125.2 m between consecutive nests. These investigators worked in much smaller areas than ours (∼1500 m), which were composed of matrix of marshes and nesting habitat. This may favor short movements due to available suitable nesting areas. Because many reports do not mention the size of the nesting area and its influence on the inter-nest distance, it is difficult to make comparisons. Thus, distances between consecutive nests must be influenced by the size of favorable nesting areas, environmental conditions (i.e. temperature, humidity, and rainfall) at time of nestings, and by the capability of females to return to the previously selected site.

Nest-site fidelity and orientation in translocated females

The females were translocated over a mean distance of 1006 m from their capture point; nevertheless, they were recaptured at approximately 520.5 m from this point. The transference and recapture distances were significantly different, indicating that the females' movements are not random and they are capable of discerning their direction of origin. This was reinforced by comparing the transference angles and the recapture angles: 72.1% of the females captured south of the release point were recaptured in their region of origin. This result may indicate that there was a non-random effort by the females of T. dorbigni to return to the initial nest site, demonstrating a high capacity for orientation. The means of orientation in this species is unknown. Although the translocated females were easily capable of moving in the reverse direction to their transference, they only began to show a high rate of fidelity to their nesting sites 2 years following the transfer. The lack of a significant difference between the distances of consecutive captures of non-translocated and translocated females is another factor that reinforces the results for the orientation and fidelity to the nesting sites.

Acknowledgements

The first author is grateful to the Brazilian agency CNPq for a PhD scholarship which allowed this study to be carried out. The authors are also grateful to many students who helped either by gathering data in the field or by processing the data in the Laboratory of Management and Environmental Conservation (LAMCA).