Abundance and Population Structure of Podocnemis erythrocephala (Testudines, Podocnemididae) in the Unini River, Amazonas
Abstract
We studied the population structure, sex ratio, and abundance of a locally endangered species of freshwater turtle, the red-headed river turtle (Podocnemis erythrocephala), in the Unini River in Negro River Basin in Brazil with capture–recapture of marked individuals in 1 yr, including both dry and rainy seasons. We used trammel nets of 3 different inner mesh sizes in the lakes. The nets were revisited at 3-hr intervals from 0600 to 1800 hrs each day. All turtles were released after being marked, measured, and sexed. During the study, 352 individuals were marked including 162 males, 150 females, and 40 immatures. The majority of the turtles (79%) were captured in the dry season when the water level of the river was declining. Mean ± standard deviation straight-line carapace length was 210 ± 14 mm (range 163–262 mm) for males and 251 ± 18 mm (221–303) for females. Most turtles captured were in the intermediate size classes: 200–210 mm for males and 230–270 mm for females. The sex ratio of adults in this population was 1.05 males per female, not significantly different from 1∶1. Only 12 turtles were recaptured, each once: 5 males, 5 females, and 2 immatures. The population showed recruitment and the adults are in equilibrium; however, few turtles were captured in the smallest or largest size classes.
The red-headed river turtle (Podocnemis erythrocephala; Fig. 1) is the smallest member of its genus, reaching 244 mm in males and 322 in females for straight-line carapace length (Rueda-Almonacid et al. 2007; Vogt 2008; Bernhard 2010; Bernhard and Vogt 2012). The Rio Negro Basin contains the largest concentration of locality records for the species. The distribution of this species spans the blackwater tributaries in the Amazonian regions of Colombia, Venezuela, and Brazil in the Nhamundá, Trombetas, Branco, Amazonas, and Tapajós rivers (Hoogmoed and Ávila-Pires 1990; Vogt et al. 1991; Rueda-Almonacid et al. 2007). The species prefer lakes, small streams, and tributaries and main channels of major rivers (Rueda-Almonacid et al. 2007).
Citation: Chelonian Conservation and Biology 13, 1; 10.2744/CCB-1039.1
The unsustainable consumption of turtles and their eggs by the local human population has resulted in the decline of populations of this species in many areas in the Rio Negro Basin (Vogt 2001; Rueda-Almonacid et al. 2007). This factor contributed, in part, to the decision to include this species in the IUCN Red List (VU A1bd; The World Conservation Union 2012). This turtle is also listed in Appendix II of CITES (Convention on International Trade in Endangered Species of Wild Fauna and Flora 2008).
Because the life history of turtles is characterized by a long reproductive period, if loss of adults exceeds recruitment for a long period of time, the population will decline (Moll and Legler 1971; Klemens 1989; Moll and Moll 2004). Baseline information describing natural population sizes and structure for P. erythrocephala is urgently needed in order to monitor the impacts of human use and habitat modifications (Brito et al. 2009). However, detailed data on the population dynamics of turtles, which have delayed sexual maturation and life cycles longer than the people who study them, are not easy to obtain (Gibbons 1990). However, short-term studies, which present snapshots of the population structure at a specific period in time, can be valuable to help to evaluate the conservation status of a species at a specific locality.
The sex ratio of some species of Podocnemis is susceptible to alterations caused by the disproportionate collection for food of nesting females (Fachín-Terán et al. 2003). Evaluations of abundance can have implications for conservation of viable populations in the future as well as for the human river communities that utilize and depend on this resource for food (Conway-Goméz 2007). If the collection of turtles is realized in a sustainable manner (without the collection of reproductive females or immatures), the use of a portion of the eggs or males should not cause severe impacts on the population (Conway-Goméz 2004). Direct actions are necessary to combat the uncontrolled consumption of these turtles so as to reduce population declines in the family Podocnemididae in the Rio Negro Basin.
Our objective was to determine the distribution of P. erythrocephala in the Unini River, documenting the abundance and population structure of the subpopulations and the possible use of these data to evaluate the use of these turtles by the river communities in this area. Our specific objective was to undertake a preliminary evaluation of population size and structure in order to determine how the use of this species by the local community affects its population structure in a Sustainable Use Reserve.
METHODS
Study Area
The study was conducted in the Unini River about 200 km northwest of Manaus. The river is a blackwater tributary of the right margin of the Rio Negro, in the Municipality of Barcelos in Amazonas State (Borges et al. 2004; Fundação Vitória Amazônica [FVA] et al. 2005). This river is the boundary between two federal conservation units: the left bank is the southern boundary of Rio Unini Extractive Reserve (RESEX) and the right bank is the northern boundary of Jaú National Park (PNJ; Fig. 2). The RESEX was created in 2006 and has an area of 865,210 ha. In the RESEX, the inhabitants have permission to use the natural resources of the forest and rivers for commercial and sustenance purposes, supposedly in a sustainable manner (FVA et al. 2005). The people who live in Reserves of Sustainable Development (RSD) such as Amana and Mamiraua are allowed to consume turtles, but they are not allowed to sell them commercially. Within biological reserves, turtles are protected by law, and people living within these reserves are not allowed to eat turtles. Outside of reserves throughout the Brazilian Amazon, it is illegal to capture, kill, sell, buy, or eat freshwater turtles or their eggs, except those purchased from legal commercial farms or in case of starvation when no other food is available. Generally, this law is only enforced in city markets, in public riverboats and airline transportation, and in urban restaurants, not in rural areas where turtles are taken for home consumption.
Citation: Chelonian Conservation and Biology 13, 1; 10.2744/CCB-1039.1
We sampled the river near 3 communities in RESEX (Fig. 2): Tapiira (106 km upstream from the mouth of the Unini River), the fifth community upstream of the mouth, with 30 families; Manapana (156 km upstream), the community farthest upstream in the Reserve, with 10 families; and Vila Nunes in the Amana Reserve (267 km upstream from the mouth of the Unini River), with 8 families. We chose these study sites for the following reasons: 1) A pilot study conducted in September 2009 documented that inhabitants of these communities used turtles for commerce and food, 2) we had cooperation from the communities to conduct the study, and 3) this section of the river was well defined and the sample areas were more than 50 km apart.
We conducted the study in the areas used by the communities: lakes, channels, and oxbow lakes. The inhabitants indicated these areas were important in that they had the greatest concentration of nesting turtles. We sampled turtles in 7 expeditions between October 2009 and September 2010 (Fig. 3). These sampling periods covered both wet and dry seasons (Fig. 3), with 7 collecting trips total.
Citation: Chelonian Conservation and Biology 13, 1; 10.2744/CCB-1039.1
We used trammel nets to capture the turtles (Vogt 1980). The nets used were 100 m long with 2-m deep, 30-cm-mesh walling on each side of a finer diameter 11- or 18-cm mesh that was 3 m deep. The differential sizes of the inner mesh allowed the capture of different size classes of turtles.
We set the trammel nets in areas with little or no current at 0600 hrs and checked them every 3 hrs until 1800 hrs. The number of nets per locality varied from 1 to 4, depending on the size of the area being sampled. We sampled each area for a mean of 4 d (3–5 d) per sampling period with a mean total of 28 d (21–35 d) per site. We calculated an abundance estimate based on catch per unit effort (CPUE), defined as the number of turtles caught per number of net hours used in the sample (CPUE = Nturtles/Nnets/12 hrs/d). River level data were taken from the National Water Agency database (Agência Nacional das Águas [ANA] 2009).
We verified the sex of the turtles captured using external secondary sexual characteristics; males having red cephalic spots and longer tails than females. We measured straight-line maximum carapace length (CL) and width (CW) using a digital caliper (± 1 mm). We also used a tape measure to record the minimum plastron length (MPL) and plastron width (MPW). We weighed the turtles using Pesola® spring scales (1000 ± 1 g and 10 ± 0.1 kg). After all the measurements were taken, we marked each turtle individually with a combination of rectangular cuts in center of the marginal scutes, using a hacksaw following Cagle (1939), and then released the turtles at the site of capture. Rectangular cuts are preferred in that they are less likely to be confused with piranha bites, as would the “v” cuts made with a triangular file.
We classified turtles as adults if males were > 161 mm CL and females > 221 mm (Bernhard and Vogt 2012). Gravid females were also identified by the presence of oviducal eggs, as noted by inguinal palpation. Mean CL is reported ± 1 SD. All turtles that were not mature adults or hatchlings were called immatures; the names juvenile or subadult are not used because there is no identifiable, concrete biological property that can separate more than 1 age group of turtles in the range from hatchlings to adults, except perhaps yearlings.
We characterized the population structure by the distribution of size classes of CL in intervals of 1 cm. We defined sex ratio as the proportion of males in relationship to the total capture. We used the chi-squared (χ2) test to compare the sex ratios in the 3 different sampling areas and among the seasons. We used Yates' Correction of Continuity in this analysis (Zar 1999).
The distribution of size classes was compared among the sampling periods in different seasons and among localities. We used an analysis of variance (ANOVA) to compare carapace length among seasons and localities, with pairwise multiple comparison procedures (Holm-Sidak method) with an overall significance level p = 0.05.
We formulated the capture index using the database of the sampling periods, which reflects the relative abundance of the species in the area sampled. The index of recapture was calculated for each sex and for immature turtles, dividing the number of captures by the number of turtles recaptured and dividing by 100.
RESULTS
We captured 352 P. erythrocephala: 162 males, 150 females, and 40 immature (37 females, 3 unknown sex). All of the males captured were adults according to the maturity size designated as 161 mm CL (Bernhard and Vogt 2012). Fourteen gravid females were captured during the nesting season (August–September 2010), with the smallest being 221.8 mm in CL. Adults (n = 312) represented 88.6% of the turtles sampled.
We recaptured 5 males, 5 females, and 2 immatures. Because the number of recaptures was < 10%, it was not possible to estimate population size, growth rate, survivorship, or rate of migration (O'Brien et al. 2005). Males showed a normal distribution with a modal interval of 200–210 mm. Females had a bimodal distribution within the interval 230–260 (Fig. 4). Mean carapace length was 0.2 ± 14.02 mm (163–262) for males and 251.4 ± 17.75 mm (221–303) for females (Table 1). The immatures were 202.3 ± 28.21 mm (113–219). In Manapana the mean CL was 211 ± 15.1 mm (163–262) for males and 246 ± 14.7 mm (221–283) for females. At Vila Nunes, we captured fewer turtles than at the other localities (23.8% of the total from the 3 localities; Fig. 5).
Citation: Chelonian Conservation and Biology 13, 1; 10.2744/CCB-1039.1
Citation: Chelonian Conservation and Biology 13, 1; 10.2744/CCB-1039.1
The mean values of male carapace length did not differ among localities (ANOVA, males: n = 162, df = 2, F = 1.084, p = 0.341). In contrast, carapace length of females was significantly different among localities, with Manapana females being smaller than those of Tapiira and Vila Nunes (ANOVA, females: n = 150, df = 2, F = 4.78, p = 0.010).
Sex Ratio and Percent Immature Turtles
The general sex ratio of the turtles captured in the Unini River was not significantly different from 1∶1 (1.05♂∶1.00♀). The sex ratios ranged from 0.77–1.31♂∶1.00♀ among the localities sampled. However, in Manapana and Vila Nunes we found a significant skew in the sex ratios (1.20♂∶1.00♀ and 1.31♂∶1.00♀, respectively).
We captured 40 immature turtles in the Unini River, of which 3 (8%) did not have secondary sexual characteristics and the remainder were females. The capture index for immatures varied from 3.03% to 14.81% among the localities sampled. The highest incidence of immatures recorded was in the first low-water period (22.6% of all captures).
Seasonal Variation in Captures
We caught more turtles during the parts of the year when water levels were declining (“vazante”; Fig. 6), when nesting by P. erythrocephala takes place: October 2009, August–September 2010, November–December 2009, and January 2010. When the water levels were rising (“enchente”, April) we caught more females (n = 32) than males (n = 14). During vazante in August and September 2010, even though the river only lowered 2092 m, the number of turtles captured was much higher than in other sampling periods, 90 males and 71 females. There was no significant variation in CL of males, females, or immatures collected during different hydrologic regimes. (ANOVA; males: n = 162, df = 3, F = 0.69, p = 0.56; females: n = 150, df = 3, F = 0.80, p = 0.83; immatures: n = 40, df = 3, F = 0.18, p = 0.90). Greater numbers of large turtles were collected during the vazante: 31 females with a CL above 250 mm and 22 males with a CL above 230 mm (n = 22) (Fig. 7).
Citation: Chelonian Conservation and Biology 13, 1; 10.2744/CCB-1039.1
Citation: Chelonian Conservation and Biology 13, 1; 10.2744/CCB-1039.1
Capture Index
We used 4752 hrs of total capture effort in the Unini River, 865 hrs at each site, except that at Manapana it was not possible to sample during enchente. Mean CPUE for the Unini River was 0.07 ± 0.08 turtles/net/hr (0–0.32). At Tapiira, mean CPUE was 0.07 ± 0.06 turtles/net/hr (0–0.20); at Manapana, the mean was 0.12 ± 0.12 turtles/net/hr (0–0.32); and at Vila Nunes, the mean was 0.05 ± 0.05 turtles/net/hr (0–0.18; Fig. 7).
DISCUSSION
The number of P. erythrocephala captured at Manapana was greater than that at the other 2 sites studied (41% more turtles were captured). This could be attributed to the fact that this community is relatively small compared with the others, with only 10 families, and they might cause less negative pressure on the turtle population in their area. Conway-Goméz (2004) found statistically significant differences showing that the abundance of turtles near communities was lower compared with river sections not inhabited by people.
Using trammel nets of different mesh sizes, we were able to capture both adult and immature individuals but we were not able to capture the smaller size classes of immatures. However, the proportional number of these smaller size classes in natural populations has been shown to be very low (Bernhard and Vogt 2012), so it is not necessary to describe the population structure in these populations. Most of the turtles (93.5%) were captured in the low-water season, and fewer in the high-water season, probably because there were more habitats available for the turtles to disperse into the flooded forest where it was difficult to set nets. Even though we captured immature turtles of 120 mm CL, this does not indicate that there is recruitment in these localities from eggs laid by the resident females; the smaller juveniles could have migrated in from the protected areas upstream. Only by marking thousands of hatchling turtles with passive integrated transponder (PIT) tags will we be able to understand the population dynamics of these systems of sustainable-use sinks and preserve production to maintain population levels (Schneider et al. 2011). In the Unini River, the inhabitants collect nearly all of the eggs from the nests near their communities. This could help to explain the low level of recruitment (percentage of immatures in a population): 11.3% of the turtles captured in this population of P. erythrocephala were immature, whereas Bernhard and Vogt (2012) encountered immatures in 40% of captures in another tributary of the Rio Negro without extractive communities and in 78% of the turtles captured in Colombia (Castaño-Mora 1997).
We found a seasonal fluctuation in the sex ratio of the P. erythrocephala captured in this study, skewed principally during enchente when more females were captured. This could be the result of greater postnesting movements of females in the channel of the river toward the interior of the lakes.
We noted that all of the nests encountered on the beaches had been depredated for consumption by the inhabitants of the communities, lowering the reproductive success and recruitment in these areas. At least some of the hatchlings produced by a female must reach maturity to maintain stable turtle populations. Exactly how many are needed or how many would naturally reach maturity without human intervention is unknown for this species. However, Valdez Guilhon et al. (2011) found that, after 12 yrs, at least 8 of 2000 hatchling Podocnemis unifilis released with PIT tags reached maturity in the wild. In addition to human consumption, eggs, hatchlings, immatures, and adults suffer natural depredation from other animals as well, pressures which are exceeding the ability of river turtles to compensate for the loss of reproductive stock, thereby lowering their fecundity (Mittermeier et al. 2009). Here on the Rio Unini there is pressure at both ends of the life table via consumption of both the eggs and the mature females in these populations (Ferrara et al. 2011). If these practices are not modified, these populations of turtles we studied are destined for extirpation in the near future.
Only in the stretch of river above the last community (Vila Nunes) is there an area of preservation where the local inhabitants do not collect eggs or adult turtles. This area is preserved because it is within the Amana RSD, and there are no people living in this area along the headwaters of the Rio Unini. The inhabitants of the communities we studied on the Unini River think that they can maintain a scorched earth policy, eating all stages of the life cycle of P. erythrocephala in their section of the RESEX without lowering populations of turtles, because turtles are continually repopulating from upstream (Ferrara et al. 2011). To an extent, this could be the case, but studies are lacking to confirm if this is true or to what level they can consume all the adults and eggs they find and still maintain the population. A continuation of this study is necessary to determine at what level extractive-use communities can consume turtles and their eggs on a sustainable basis. We need to determine what percentage of the eggs and adults can be used, not just report what the inhabitants are doing at this time.
Our study is the first to give a quantitative approach to the state of the population structure of a species of turtle that is being utilized by the communities in a RESEX. Hopefully it will be of value to compare use in other areas and set up guidelines for establishing quotas for the use of turtles in these reserves.
Male (left), female (middle), and immature (right) Podocnemis erythrocephala. (Photos by Virgínia Campos Diniz Bernardes).
Study area, Unini River Basin, Barcelos, Amazonas, Brazil, showing the location of the 10 communities within the reserve.
Water level of the Unini River, monthly mean values (connected points), standard deviations (rectangles), and ranges (vertical bars) recorded daily from August to July 2010 at Manapana (lat 01°53′14.5″S, long 62°27′52.9″W), Barcelos, Amazonas (Agência Nacional das Águas 2010).
Frequency distribution of maximum straight-line carapace length (CL) of adult Podocnemis erythrocephala in the Unini River.
Number of Podocnemis erythrocephala captured at each locality: immature (n = 40), male (n = 162), female (n = 150).
Number of Podocnemis erythrocephala captured in the Unini River throughout the study period, demonstrating the variation in the number caught during different hydrological cycles of the river.
Frequency distribution of carapace length of Podocnemis erythrocephala during different hydrological seasons (vazante, water levels were lowering; seca, lowest water level; enchente, the water levels were rising; cheia, highest water level).
Contributor Notes
