The giant Amazon river turtle (Podocnemis expansa) and yellow-headed river turtle (P. unifilis) are two common freshwater turtles of the Araguaia River Basin in Brazil. The turtles nest on vast sand bars that emerge close to the margins of the rivers during the dry season. Although they nest on the same beach, overlapping of actual nesting areas for the two species is rare. In the Javaés River, P. expansa chooses the most elevated points of the beach, with as much as 330 cm height in relation to the river level on the nesting date; P. unifilis prefers the lowest parts of the beach, where heights are less than 150 cm above river level (Ferreira Júnior 2003).

Podocnemis expansa and P. unifilis are species whose sex determination depends on environmental conditions during the incubation (Alho et al. 1985; Souza and Vogt 1994; Valenzuela et al. 1997). In general, high incubation temperatures induce the production of females, and low temperatures the production of males (Bull 1980; Ewert and Nelson 1991; Janzen and Paukstis 1991).

Our work investigates the thermal characteristics of P. expansa and P. unifilis nesting areas in the Javaés River, along the east border of Bananal Island. The beaches chosen for nesting here are wide point bars formed by sandy sediments. These point bars lack vegetation and receive direct insolation during most of the day. Daily variations in subsurface heat transmission were analyzed on one beach. The influence of the beach height and the water table on the subsurface temperature were also analyzed.

METHODS

The sampling points were chosen based on P. expansa and P. unifilis nest distributions and geologic characteristics of the beach, such as grain size distribution and beach morphology. The sampling points were located in the central portion of Canguçu beach, along the right margin of the Javaés River (50°05'11”W, 9°59'12”S) (Fig. 1). Canguçu beach has irregular morphology; in a distance of less than 15 m, beach height varied more than 200 cm in relation to the Javaés River level. The grain size distribution of the beach was analyzed using standard grain-size charts, with the size analyses of the samples collected based on sieving according to the Folk-Wentworth classification (Folk 1974).

View Figure

• Download as Powerpoint
• Download Figure

To minimize humidity-influenced variation, the sample points on the beach were selected using two criteria: first, they had to have similar grain size distributions; second, they had to be close to each other. Thus, variation in heat transmission profiles may be attributed to humidity variation caused by differences in the proximity of the water table, regardless of the weather. Because the distances among the three sample points were less than 15 m, insolation, relative air humidity, rainfall, wind speed, and air temperature at these points were assumed to be similar.

The height of the beach at Point 1 (P₁, Fig. 1) was 85 cm above the Javaés River level on 2 November 2001, when the experiments were initiated. P₁ was positioned at the base of an accretionary dune front that defines two topographical and morphological domains. At the base of this accretionary dune front, 42 P. unifilis nests were laid in 2001. Point 2 (P₂) was located at the top of the accretionary dune front, and was 245 cm high at the beginning of the experiment. Point 3 (P₃) was situated 280 cm above the Javaés River level, at the top of the accretionary dune front. The platform where P₂ and P₃ were situated had 25 P. expansa nests in 2000. P₂ was near the steep slope of the accretionary dune front, while P₃ was 10 m away from this slope.

The air and ground surface temperatures were collected at the Canguçu Meteorological Station (50°02'11”W, 9°58'45”S), which is located about 10 km south of Canguçu beach. Data were collected in 60-min intervals. The ground temperature sensor was placed in a position that avoided direct insolation most of the time.

Three Novus Logbox dataloggers were used to quantify subsurface temperatures on the beach. They were programmed to register the temperature every 30 min. The temperature registered was the average of three readings.

Placement of temperature probes considered biological criteria. Ongoing studies at Bananal Island indicate that the average depth of P. unifilis nests is about 15 cm, whereas P. expansa nests have floors situated at a mean depth of 50 cm, and the top of the eggs chamber is located at a mean depth of 35 cm. Depths of 15, 35, and 50 cm correspond to mean points of P. unifilis and P. expansa egg chambers.

Three experiments were performed to analyze the influence of beach height relative to the water table on heat transmission with increasing depth (Table 1). The depth of the water table throughout the beach followed the variations of the level of the river, and the humidity of the beach is directly influenced by the height of the beach in relation to the level of the river (Ferreira Júnior 2003).

Table 1. Time period in each plot, location and depth of temperature probes, and beach height.^a

View Fullscreen

RESULTS

The grain size distribution analysis showed that grain size differences of the sediments from P₁ and P₃ were minimal (Table 2). The P₂ sediments were significantly coarser. In the three points, the sediments were sandy, poorly selected, and dominantly medium to coarse sized.

Table 2. Grain size fractions of the sediment samples, according to the Folk-Wentworth classification (Folk 1974). (Dimensions are given in mm.)

View Fullscreen

There was a clear association among air temperature, surface temperature, and subsurface temperature of Canguçu beach at different depths (Fig. 2). The average temperatures in P₁, P₂, and P₃ at depths of 15, 35, and 50 cm were greater than air temperatures. The maximum and minimum peaks of air temperature and of ground temperature were out of phase. This asynchrony increased with depth. Daily subsurface temperature variations decreased with depth.

View Figure

• Download as Powerpoint
• Download Figure

Plot 1

Air temperature was lower than the subsurface temperature. The temperature at 15 cm depth was not affected by nest position (F_2,843 = 0.74; p = 0.4770), and temperature differences in P₁, P₂, and P₃ were not significant (Table 3). An exception to this occurred between 1200 and 1800 hours, on 3, 6, and 7 November, when the temperatures in P₁ exceeded by 2°C the temperatures in P₂ and P₃ (Fig. 2). On those days, solar radiation was greater than on other days. Slight rainfall occurred after noon on 2, 4, and 5 November. Otherwise, temperatures were practically constant in the three points during the experiment.

Table 3. Summary of values corresponding to the three plots.^a

View Fullscreen

Plot 2

Subsurface temperatures did not vary significantly with depth (F_2,429 = 0.49; p = 0.6164), but at greater depths, daily temperature oscillations were lower (Table 3). There was a delay in heat transmission with depth. For example, on 11 November, the highest air temperature occurred around 1500 hours. The highest temperatures at 15, 35, and 50 cm depths were reached at 1700, 1900, and around 2100 hours, respectively (Fig. 2). The difference between the daily air temperature maximum and the subsurface temperature maximum decreased with depth. But the minimum difference between daily air temperature minimum and subsurface temperature minimum increased with depth (Table 3). The daily maximum difference was calculated taking into account the daily maximum values of both the air temperature and the subsurface temperature. Conversely, the daily minimum difference was calculated considering the daily minimum values of both the air temperature and the subsurface temperature.

Plot 3

The subsurface temperature varied significantly with depth (F_2,426 = 4.32; p = 0.0139). Mean temperature and daily temperature variation decreased with depth. The heat transmission delay followed the same pattern as in plot 2. The difference between the highest air temperature and the highest subsurface temperature was smaller than in plot 2, but the difference between minimum air temperature and minimum subsurface temperature was greater than that documented in plot 2.

DISCUSSION

The results of plot 1 showed that temperature at 15 cm depth remains practically the same in all points, regardless of the depth of the water table. In this case, P. unifilis nests that are situated on average at 15 cm depth will be subject to the same thermal environment throughout the beach. The relative height of the water table does not influence the mean temperature of substrates at 15 cm. Small differences in grain size distribution of the sediments did not affect heat transmission. Podocnemis unifilis lays its eggs throughout the beach, but concentrates its nesting in areas with beach height lower than 150 cm above river level (Ferreira Júnior 2003). The topographical irregularities of the beach do not hinder females from wandering all over the beach; the concentration of nests in the lowest beach areas suggests a preference for nesting in more humid areas closer to the river channel.

Plot 2 indicated that the ground temperature does not vary with depth in the low portion of the beach, where P. unifilis nests at 15 cm. The extension of these results to the thermal environment of the nests, and the consequences on turtle embryogenesis require extreme care. Table 3 and Fig. 2 indicate that temperature variation decreased with depth. The upper beach layers were subject to temperature oscillations of more than 10°C daily. Thus, although average temperature was practically the same at the monitored depths, temperature oscillations produced a differentiated thermal environment in different portions of the nest. This is an important point to the understanding of the influence of temperature on turtle embryogenesis.

According to Georges et al. (1994), the daily average temperature of natural nests of freshwater turtles, whose sexual determination depends on environmental conditions, is not a good predictive tool of the sex ratio when nest temperatures vary. Natural Emys orbicularis nests, which were monitored by Pieau (1982) and incubated most of the time under temperatures that favor production of males, produced basically females due to the effect of the daily incubation temperature fluctuations. The average incubation temperature was inadequate to explain the sex ratio of Graptemys (Bull 1985). The sex ratio of Chrysemys picta was better explained by the number of hours above the critical temperature than by the mean incubation temperature (Schwarzkopf and Brooks 1985). The critical or pivotal temperature is defined as a temperature that produces 50% males and 50% females (Yntema and Mrosovsky 1982). The embryonic development is faster when the temperature is more elevated (Bull and Vogt 1981; Mrosovsky et al. 1984; Bull 1985). Daily temperature variations do not influence the average development rate of embryos, because it is absorbed by the total period of incubation. The daily temperature variations affect sex ratios. Based on these observations, Georges et al. (1994) concluded that the most important factor in sex determination of Caretta caretta was daily temperature variation, not the embryonic development rate, which directly influences the incubation duration. These examples and field data suggest that, despite the incubation duration, which is a direct consequence of the fact that embryonic development rate does not significantly change along the beaches, the sex ratios of P. unifilis can be affected by the depth of the nests, because the daily temperature oscillations depend on depth. For sex determination, P. unifilis nest depth is more important than the nest position in relation to the water table.

Most of the P. unifilis nesting in the Javaés River occurs at the base of dunes. The gradual accretion of dune sediments on these nests increases the relative depth of the nests supplying extra protection to the eggs (Ferreira Júnior et al. 2003). This improves protection of these nests from predation by vultures (Coragypus atractus and Cathartes aura) and caracarás (Polyborus plancus) that normally depredate nests intensely throughout development (Ferreira Júnior 2003). Because the temperature does not vary with the sand depth at the base of dunes in the lower part of the beach, the addition of sediments does not produce differences in the final temperature of the eggs.

Plot 3 indicated that the temperature decreases significantly with depth in those higher beach areas selected by P. expansa for nesting. In this case, mean temperature of P. expansa eggs depends on nest depth: the deeper the nest, the lower the temperature. As occurred in plot 2, temperature variation also decreased with depth.

Considering that grain size of the sediments in P₁ and P₃ were quite similar, the difference between vertical gradients of temperature of plots 2 and 3 may be explained by the height of substrate above the water table. In plot 2, the proximity of the water table at P₁ caused a higher humidity, which was responsible for lower temperature variations. According to Harrison (1985), Novak and Black (1985), and Harrison and Morrison (1993), high humidity decreases with daily heat variation. In the drier sediments of P₃ (plot 3), subsurface heat flow was inconspicuous because of the large volume of air in the spaces among the grains. The heat diffusion and the high thermal capacity of the interstitial water (humidity) eliminated the temperature change on the surface of the sediments (Harrison 1985). The lower thermal conductivity of dry sands causes less subsurface heat transfer when compared to the humid sediments. As a consequence, in dryer environments, the energy of the sun elevates the temperature of the upper layers before being transmitted, by conduction, to the lower layers of the beach (Harrison and Morrison 1993). Humidity at P₁ was higher than at P₃. This fact explains the differences of air and subsurface daily maximal temperatures, that were greater in plot 2 than in plot 3.

For P. expansa, the temperature and the humidity of the beach affect the choice of nesting sites along the Javaés River. The lowest areas of the beach are not adequate for nesting, as the rising of the level of the river often precedes emergence of the hatchlings—nests situated in these low areas would flood and the hatchlings and embryos drown. In the Javaés River, nests located in the highest points of the beach have good incubation conditions, as these areas only flood after the hatchlings emerge. The results of the reproductive seasons of 2000 and 2001, when 447 nests were monitored, showed that P. expansa preferred beaches with coarser sediments where the duration of the incubation was shorter, despite the hatching success being lower (Ferreira Júnior 2003; Ferreira Júnior and Castro 2003). This suggests a possible mechanism for control of the sex ratio by P. expansa: females choose a nest site where temperatures are more favorable to the development of females instead of beaches where the hatching success is higher.

Hatching of P. unifilis generally coincides with the beginning of the rise of the level of the Javaés River and in most seasons hatchlings emerge before the lower parts of the beach are flooded. As for P. expansa, P. unifilis females do not choose nesting sites where hatching success is higher (higher part of the beach, above the flooded area) and apparently prefer sites where the survival of females is higher (Ferreira Júnior 2003).

Subsurface temperature variation in P. expansa and P. unifilis nesting areas show the influence that the proximity of the water table exerts on heat transmission at a given depth. The results presented here, due to small sample sizes, should be taken critically, but nevertheless suggest some characteristic or trends that deserve further investigation.

RESUMO

A variação da temperatura em subsuperfície sugere que o ambiente termal das covas da tartaruga-da-amazônia (Podocnemis expansa) é influenciado pela proximidade do lençol freático e profundidade do ninho. A diferença entre os gradientes verticais de temperatura ao longo das praias fluviais indica que o local de nidificação poderá exercer influência significativa sobre a temperatura experimentada pelos ovos dos quelônios. A temperatura diminui significativamente com a profundidade nas áreas de desova de P. expansa, mas não se altera substancialmente nos pontos de nidificação de P. unifilis. As oscilações diárias da temperatura diminuem com a profundidade ao longo de toda a praia.