Ecology and Status of the Magdalena River Turtle, Podocnemis lewyana, a Colombian Endemic
ABSTRACT
We studied distribution, habitat selection, and reproductive ecology of the Magdalena river turtle, Podocnemis lewyana, from January to June 2003 and from January to March 2005 in the mid and low Sinú River, Córdoba, Colombia. The southern limit of its distribution occurs in the town of Carrizola, and the northern limit is near La Doctrina. We found more turtles of bigger size classes in the mid basin than in the low basin. Ninety-two individuals were observed basking in the river. Larger turtles were mostly observed in deep pools (> 3 m) with loamy or sandy bottoms, with easy access to basking sites, such as riverbanks, fallen trees, or beaches, and in areas with all types of shoreline vegetation but especially with riparian forest. Smaller turtles were found in colder, shallower, and more turbid waters than were adults, mainly in areas without riparian forest or in pastures with few trees, usually hidden in pastures flooded by the river. Adults ventured farther from shore than did juveniles. Along the study area, 5 nesting beaches were located: 3 in the mid basin and 2 in the low basin. Main nesting sites were sand and gravel beaches, but riverbanks were also used. Nesting occurred in small groups or individually in 2 reproductive seasons: one from December to March and the second from June to August. The main threats identified for P. lewyana in the Sinú River were floods caused by water release by Urrá Dam, human consumption and exploitation, nest damage caused by beach mining, and riparian forest destruction for pasture establishment.
There is worldwide concern for the future of river turtles because this group has proportionately more species listed in the most threatened categories than any other nonmarine turtle group (Moll and Moll 2000). In Colombia, all 6 species in the genus Podocnemis are regionally threatened: Podocnemis expansa and Podocnemis unifilis are critically endangered; Podocnemis lewyana is endangered; Podocnemis erythrocephala is vulnerable; Podocnemis vogli is near threatened; and Podocnemis sextuberculata, although with data deficiency, is suspected to be threatened (Castaño-Mora 2002). In general, the decline of river turtle populations in Colombia is related to human activities in a direct way by overexploitation, indirectly by habitat alteration, or both (Castaño-Mora 2002).
Podocnemis lewyana, locally known as “river turtle” or just “turtle”, is subjected to strong human pressure. Because it is the largest continental turtle of northern Colombia, it is an important source of protein for the inhabitants of the Sinú, San Jorge, and Magdalena rivers (Castaño 1986). Besides being hunted for its meat, it also is pursued for its eggs; hatchlings also are collected to be sold illegally as pets (Castaño 1986). This turtle's distribution in poor rural areas makes it a good source of income and food for local fishermen. In addition, throughout the region, eating turtles during the Easter holiday is an obligation rather than a tradition (Castaño-Mora and Medem 2002).
Forty years ago, in the Sinú basin, P. lewyana was abundant, but overexploitation of adults and eggs contributed to its decline in several regions (Dahl and Medem 1964). Castaño (1986) reported that fisher communities in the area were unanimous in declaring that its population decreased year after year. Today, this reduction is extreme and evident. In this region, P. lewyana is not only affected by direct exploitation but by the possible effects of river damming such as flooding of nesting beaches, fragmentation of populations, obstruction of migrations, and changes in habitat quality (Moll and Moll 2000). These factors place this species at a higher level of threat. Furthermore, because P. lewyana is endemic to the drainages of the Magdalena and Sinú rivers in Colombia (Ernst and Barbour 1989), research and conservation is the responsibility of Colombia. However, no measures have been taken for its protection because of a lack of data on its natural history. With the exception of a study carried out in the Cocorná River in the Magdalena basin (Hurtado 1973) and another study conducted in captivity at the Estación Roberto Franco in Villavicencio, Meta, Colombia (Castaño 1986), there are no other studies about its natural history.
In this paper, we report on some basic aspects of the ecology of P. lewyana such as distribution, habitat selection, and reproductive ecology. We also identify the main threats in the region, which are indispensable to produce a management and conservation plan for this species. Our study provides baseline data for future work to reduce threats. Because so few data are available on this species, we also report anecdotal observations on P. lewyana by local people of the area to provide a baseline for other studies.
METHODS
Study Area
The study was conducted from January to June 2003 and from January to March 2005 in the Sinú River, Department of Córdoba, northern Colombia, on a 266-km river section, enclosed by the Urrá Dam (at the beginning of the mid basin) and the river's mouth in the Caribbean Sea (Fig. 1). It was not possible to study the high basin because it is a restricted area.
Citation: Chelonian Conservation and Biology 7, 1; 10.2744/CCB-0643.1
The Sinú River basin's climatic regime is unimodal, with a dry season from December to March and a rainy season from April to November (when more than 80% of the precipitation falls). The mean annual precipitation is 1560 mm, and the mean annual temperature is 27.2°C (Romero et al. 1996). The mid basin has an elevation of 50 to 100 m, and the low basin has less than a 50-m altitude.
The upper part of the mid basin has the strongest currents in the river because of its slope; the water is more transparent than downstream, the river bed is mainly composed of rocks and pebbles, and there are no sand beaches. The trees Guazuma ulmifolia, Cecropia peltata, Tectona grandis, and Gynerium sagittatum prevail in the riparian forest. There are few human communities along the river margins; therefore, human activity is lower in the upper mid basin than it is downstream, despite extensive plantations of banana (Musa paradisiaca) and papaya (Carica papaya). In the middle part of the mid basin, the slope decreases, which results in a slower current; water turbidity is higher; the river bed is mainly sandy and loamy; and sand beaches begin to appear. The riparian forest is fragmented by banana and papaya plantations and pastures for cattle ranching. Prevailing tree species are Samanea saman and Inga affinis. The lower part of the mid basin has many human settlements and more human activity. Most of the riparian forest was destroyed for the establishment of pastures. What forest remains is divided into small patches of I. affinis, S. saman, and Sapium glandulosum.
In the low basin, the water is deep and turbid. Although the current is slow, considerable agitation is observed because of the vicinity of the delta. The river bed is also sandy and loamy, and sand beaches can still be observed. This area has more human activity because main towns are located there. The river margins are mostly used for pastures and for rice (Oryza sativa) plantations. The riparian forest is greatly reduced, but some isolated patches with I. affinis and mango (Mangifera indica) still remain.
The Sinú River has a hydroelectric dam located in the municipality of Tierralta in the area known as “Angostura de Urrá”, which gives the dam its name (Fig. 1). Since the Urrá Dam began operation in 1999, the river's natural flow has been regulated. Before construction of the dam, the river's flow was relatively low during the dry period from December to April and high during the rainy season. This flow regime was variable, depending on annual rainfall (INDERENA 1968). Floods could follow prolonged heavy rains in the high basin, but damaging floods seemed not to occur more often than once in a decade (Parsons 1952). Thus, beaches began to emerge at the beginning of the dry period or according to the normal fluctuations of the river. Currently, beach emergence is not determined by the river's seasonal changes but by the flows released by the dam. During the dry season, flows can be equal to those of the rainy period, flooding all beaches. The duration of the floods are very variable and depend on the amount of hydroelectric energy needed.
Capturing, Marking, and Measuring
We captured Individuals of P. lewyana at 3 different sites (Fig. 1): in the mid basin, near the mouth of Caño Betancí (lat 08°11′25″N, long 076°04′04″W), an affluent of the river, and near the town of Santa Elena (lat 08°30′02″N, long 75°58′29″W), and in the low basin, near the town of Cotocá Arriba (lat 09°07′23″N, 075°50′25″W). At the first 2 sites, turtles were captured by local people with casting nets and by hand by using green plantain (Musa paradisiaca) as bait. At the third site, we used double-funnel aquatic traps (Feurer 1980) with green plantain as bait. The traps were not used at the first 2 sites because we did not want to provide turtle hunters new ideas to capture turtles.
All trapped turtles were measured and individually marked by filing notches in the marginal scutes with a system that resembled that described by Ferner (1979). We then released all captured turtles at their site of capture. In addition to turtles captured during the study, those hunted by fishermen for their consumption or found in captivity also were measured. We took the following dimensions to the nearest 1 mm (Medem 1976): straight-line carapace length (SCL), curved carapace length, straight-line carapace width, plastron length, plastron width, and anal suture length. Turtles were sexed based on morphologic characteristics (Hurtado 1973).
Distribution and Habitat Use
In 2003, we traveled the study area in a boat a total of 4 times; each journey lasted 5 days (6 –10 February, 27–31 March, 1–5 May, and 16 –20 May). By using binoculars, we searched for turtles on riverbanks and basking sites. At each sighting point, the location was recorded with a global positioning system unit to ascertain species' distribution limits. We estimated turtle SCL to the nearest 1 cm by using as a reference point the basking sites where they were sited because the minimum distance to which the boat could approach the turtles without causing them to dive was 5 m. Each individual sited by boat was placed into one of the following size categories: <10 cm (SCL1), 11–20 cm (SCL2), 21–30 cm (SCL3), 31–40 cm (SCL4), > 40 cm (SCL5). At sighting points, we took the water temperature (20 cm below the surface), water turbidity, and depth, and estimated to the nearest 1 m the turtle's distance from the nearest shore. Shoreline vegetation type was grouped into 1 of 3 categories: 1) no riparian forest (including plantations, pastures, and/or areas with low vegetation), 2) pastures with few trees, or 3) riparian forest. Basking sites were grouped into 1 of 3 categories: 1) fallen trees, 2) riverbanks, or 3) beaches. Inclination and height of the basking site were measured when possible.
Reproductive Ecology
Throughout the study area, we obtained anecdotal information on nesting sites and nesting ecology by interviewing people in communities along the river. We used this information to find potential nesting sites. These sites were visited in the mornings to look for turtle tracks that could potentially lead to nests. We excavated each nest site and measured the depth to the uppermost egg. We then extracted all eggs and measured the nest chamber's maximum depth, clutch size, egg dimensions (greatest and smallest axis length), and mass. Egg volume (V) was calculated with the equation: V = xy2/6, in which x is the length of the greatest axis and y is the length of the smallest axis (Vanzolini 1977).
Status and Main Threats to Population
To evaluate the effect of irregular changes of the river's flow on turtle reproduction, we recorded days when beaches were flooded during the turtle's reproductive seasons of 2003 and 2005 and compared them with daily flows released by the Urrá Dam (data provided by the project Urrá S.A.). Our objective was to estimate the minimum flow required to overflow all beaches and to observe the frequency and duration of floods in other years. Also, by interviewing riverside communities and from direct observation, we obtained information on the status and main threats of P. lewyana in the Sinú River.
Analyses
We compared measurements of males and females by using a Student t-test. A Kruskal-Wallis test was used to compare turtle size distributions between the mid and low basin and between the 3 categories of shoreline vegetation type. Water temperature, turbidity, and depth, and the turtle's distance from shore were analyzed with the Spearman's correlation test. Statistical analyses were performed by using SAS V.8 (SAS Institute 1990), and significance was accepted at the 0.05 level.
RESULTS
During our study, we measured 59 females and 19 males (47 turtles were captured in the river, and 31 were obtained from markets and hunters). Females had longer and wider carapaces and plastrons and longer anal sutures than did males (Table 1). Maximum SCL recorded was 50 cm for females and 36 cm for males. Both of these measurements were the maximum sizes reported for this species.
Distribution
In the study area, the southern limit of the turtle's distribution was the town of Carrizola (mid basin) and the northern limit was the town of La Doctrina (low basin; Fig. 1). In Córdoba, the turtle is not only found in the main river but also in secondary tributaries such as Caño Betancí and Caño de Aguas Prietas y Quebrada Flores and in marshes, e.g., Ciénaga Grande del Bajo Sinú (Fig. 1). By using our 5 size-class categories, we found significant differences in turtle size among locations: more individuals of the larger size classes were found in the mid basin than in the low basin (H = 9.25, p = 0.0024) (Fig. 2).
Citation: Chelonian Conservation and Biology 7, 1; 10.2744/CCB-0643.1
Habitat Use
We estimated sizes for 92 individuals in the Sinú River (4 SCL1, 20 SCL2, 34 SCL3, 26 SCL, and 8 SCL5). Larger turtles were mostly observed in deep pools (>3 m), with loamy or sandy bottoms with easy access to basking sites such as riverbanks, fallen trees, or beaches, in areas with all types of shoreline vegetation, but especially with riparian forest (H = 12.69, p < 0.01) (Fig. 3). Smaller turtles were mostly found in areas without riparian forest or in pastures with few trees (Fig. 3), usually hiding in the pastures flooded by the river. Smaller turtles were also found in colder, shallower, and more turbid waters than were larger turtles (Table 2). Larger turtles ventured farther from shore than did smaller ones (Table 2). Fallen trees appear to be important for this species because they are used as subaquatic refuges and basking sites. P. lewyana usually basks for relatively long periods, individually or in big groups, and form arrays where the forelimbs of an individual lean on the posterior part of the carapace of another. Of 94 individuals that we observed basking, 81 were found on fallen trees (86.2%), 9 on riverbanks (9.6%), and 4 on beaches (4.2%). The maximum height above the water to which a turtle was observed basking was 48 cm, and the maximum perch inclination was 75°, which suggests that the species has a good ability to climb.
Citation: Chelonian Conservation and Biology 7, 1; 10.2744/CCB-0643.1
Reproductive Ecology
We found 5 nesting beaches along the study area: near the mouth of Caño Betancí (lat 08°31′55″N, long 075°58′40″W) on 23 January 2003, upstream from the mouth of Caño Betancí (lat 08°30′33″N, long 075°57′38″W) on 29 January 2003, near the town of Cotocá Arriba (lat 9°07′45″N, long 075°50′07″W) on 15 July 2003, near the town of Caño Viejo (lat 09°14′37″N, long 075°52′21″W) on 30 January 2005, and in the municipality of Valencia (lat 08°04′13″N, long 076°00′49″W) on February 2005 (Fig. 1). Unfortunately, all nests but two were extracted by fishermen before we were able to obtain clutch data. One nest had 17 eggs, and the eggs' dimensions were the following: length range 44.1–49.6 mm, mean 47.8 mm; width range 33.1– 45.3 mm, mean 35.0 mm; weight range 35– 40 g, mean 30.96 g; and volume range 27.0–51.8 cm3, mean 30.9 cm3. The second nest had 30 eggs, and the eggs' dimensions were the following: length range 41.2– 44.9 mm, mean 43.4 mm; width range 33.5–35.6 mm, mean 34.4 mm; weight range 53– 40 g, mean 30.96 g; and volume range 24.9–29.6 cm3, mean 26.9 cm3. Both nest chambers were spherical, with a maximum diameter of 15 cm and small in proportion to clutch size; eggs were tightly packed, which caused them to indent at the point of contact with adjacent eggs. Depths to the uppermost egg were 15 cm and 13 cm, and maximum depths were 30 cm and 33 cm, respectively. Besides observing nests on beaches, we also found nests in the upper parts of the riverbanks and in the lower parts of the riverbank's slope. According to local people, the former nest locations are more frequent than the latter and less frequent than those on beaches.
Status and Main Threats to Population
We found that, when Urrá Dam releases a water flow of 400 m3/s, all shoreline beaches in the mid basin and all but high (> 3 m) beaches of the low basin are submerged. The intensity and duration of each flood is variable and depends exclusively on the necessities of the dam: more than 5 floods could happen in the same reproductive season, with intervals of less than 15 days and with durations of more than 3 days (Fig. 4). Egg mortality caused by the Urrá Dam may be the main threat to this species in the Sinú River; however, more data are needed to determine the condition of eggs after a flood. In addition, human predation on P. lewyana and their eggs in the Sinú River area is strong. This species is mainly hunted for its meat, given that, for its great size, more meat can be obtained from it than from the other turtles consumed in the region, such as Trachemys callirostris, Kinosternon scorpioides, Kinosternon leucostomum, and Chelonoidis carbonaria. Both females and males are consumed; although, the former is preferred for its bigger size and the eggs they may have inside. Turtles are mostly hunted during the dry season because in the region eating turtles during Easter Week is a deeply rooted tradition. Turtles are stored in deposits where they remain until Easter's eve, either for sale or for consumption. The hunting season coincides with the turtle's reproductive season, which facilitates trapping of nesting females and egg gathering. Capturing turtles in the strong current is very difficult, and only specialist fishermen can perform this task; therefore, trapping nesting females on beaches has become more common. Egg gathering is also very common in the region because the eggs are considered delicacies. Nests that are laid on beaches hardly escape human predation, which aggravates population declines. Trade of hatchlings as pets was not observed.
Citation: Chelonian Conservation and Biology 7, 1; 10.2744/CCB-0643.1
Another threat for this species in the region is sand and gravel mining from the river and the beaches. Because all nesting beaches have not been identified and mining is a frequent activity along the river, it is probable that nest damage is significant. Many miners have found nests during beach excavation; however, this has not deterred them from continuing to extract sand. Also, motor craft are used to transport sand along the river as much in the day as in the night; traffic, noise, and waves probably disturb turtle nesting.
Cattle ranching is another potential threat that affects the P. lewyana population. Ranchers are rapidly destroying riparian vegetation for the establishment of pastures, which alters turtle habitat and probably diminishes food resources. Cattle, in addition, destroy nests by trampling as they traverse beaches to drink water from the river. Because cattle ranching is the most widespread activity in the region, this has become an important threat for the species.
DISCUSSION
No turtles were found upstream of the town of Carrizola (southern limit of P. lewyana distribution). We speculate that the river's physical conditions are not appropriate for turtle movements: the water is shallow, currents are strong, and the river bed is stony. Dahl and Medem (1964) suggest that P. lewyana may be able to handle strong currents in deep waters but not in shallow waters because more obstacles (e.g., boulders) are present in shallow waters than in deep waters. Stony river beds also may cause damage to the turtle's plastron because we observed that individuals maintained in captivity in stone or concrete ponds, with no current, had scratches and lesions on their plastrons. Also, there are no sand beaches in this area, which prevented turtle nesting. Downstream of the town of La Doctrina (northern limit of P. lewyana distribution), no turtles were observed, probably because of the brackish waters and the influence of the sea because this is where the river delta begins. More work is needed to ascertain the physical limitations of this species in this area.
We found that turtles in the mid basin were, on average, larger than those found in the low basin. Similar findings were reported for this species in the Cocorná River, where more juveniles were found near the mouth of the river, and more adults were in the upper part (Hurtado 1973). Hurtado (1973) suggests that the presence of nesting beaches upstream may explain the differences in the distributions of size classes. In the Sinú River, in contrast, beaches are not restricted to the upper part of the river but rather are distributed along the mid and low basins. It is possible that the size difference we observed is caused by other factors, e.g., a disparity in the number of nesting females hunted in both basins rather than in the beach location. We recommend studies on the species' movements, as well as a quantification of the predation pressure on the reproductive population to explain this pattern.
We observed that the habitat used by juveniles differs from that used by adults. This same difference has also been observed in other aquatic turtles, such as Graptemys geographica (Pluto and Bellis 1986), Emydoidea blandingii (McMaster and Herman 2000), Pseudemys scripta (Hart 1983), and Malaclemys terrapin (Roosenburg et al. 1999). In these species, as in our study, a positive correlation between turtle size and water depth, turbidity, and near-shore distance was found. Podocnemis lewyana juveniles tend to be near the shores, possibly because these areas provide better conditions and protection than the center of the channel. Near the shores, the currents are slower, which allows juveniles easier maneuvering in the water to get food or to avoid predators. Adults of G. geographica have a greater swimming ability than juveniles in terms of maintaining locomotion and orientation in fast currents, which allows them to occupy faster and/or deeper waters than those waters in the center of the channel (Pluto and Bellis 1986). In addition to the slow current, water near the shore is more turbid because of the erosive effect of the river on banks. Turbid waters may reduce predator visibility and/or increase the turtle's camouflage and decrease predation (Reese 1996).
Most of the observed individuals in our study were basking on branches, and only a few were observed basking on other substrates. The perching substrate is important for aquatic turtles because the energy absorbed by an insolated perch depends on the color and reflectivity of its surface. Also, the amount of heat transmitted by the perch to the turtle depends on the conductivity of the material and the proximity or contact surface between the individual and the substrate; because stone has higher conductance than wood and transmits heat faster, turtles spend less time basking on beaches than on wood perches (Boyer 1965). Beaches may, in some cases, be too warm to serve as long-term basking sites, and given that basking serves functions other than heating (e.g., drying), wood perches might be preferred in some circumstances (Reese 1996). Probably for this reason so few turtles were observed basking on beaches. However, we observed that turtles generally use riverbanks as basking sites when there are no other perches available, especially during high-water periods when fallen trees and beaches are submerged. Aquatic turtles prefer basking sites away from shore to avoid terrestrial predators and to have a better panoramic view (Boyer 1965). They may also prefer overwater basking sites for fast escape by dropping quickly into the water.
According to local people, the results of this investigation, and other investigators, P. lewyana nests in a great variety of substrates, such as fine and coarse grain sand, clay, and gravel (Hurtado 1973; Medem 1965). In our study, the main nesting sites were beaches along the river and its tributaries. Local people stated that nesting on riverbanks seems to be frequent, a behavior previously observed by Dahl and Medem (1964) in Caño Viloria, an affluent of the San Jorge River. It is possible that, now that the dam floods the beaches during the reproductive season, this behavior has become more common; however, this needs to be studied. The diversity in nesting sites is common in turtles of this genus. Podocnemis unifilis, for example, nest on sand and clay beaches and on the banks (Fachín 1992). Podocnemis expansa typically nests in sand beaches; although, it also has been seen nesting between the vegetation (Soini et al. 1997), the same as P. sextuberculata (Ernst and Barbour 1989). Podocnemis erythrocephala uses seasonally flooded savannas, clearings in the flooded forest and beaches (Vogt 2001; Castaño-Mora et al. 2003). Nesting on riverbanks also could be because of disturbances caused by mining activities. Soini and Soini (1995) pointed out that walking through nesting beaches or canoeing near them, from sunset to nightfall, generally impeded nesting events by P. unifilis. When P. expansa does not find an undisturbed place to nest, it is forced to lay its eggs in low banks or inside the riverside vegetation (Soini et al. 1997); whereas, P. unifilis waits until the beaches are free of disturbance (Soini 1994).
In the Sinú River, P. lewyana has 2 nesting seasons (Dahl and Medem 1964): the main one coincides with the dry season, December to March, and the second one coincides with a short dry period during the rainy season, known as “veranillo” or little summer, which is around July or August. The discovery of nests in January and February confirms the existence of the first reproductive season, and the discovery of a nest in July confirms the existence of the second season. These same reproductive periods also have been observed for this species in the Cocorná River (Hurtado 1973). Nesting sites found in this study were distributed along the mid and low basin. Hurtado (1973) found that, for the Cocorná River, P. lewyana migrate to the upper part of the river during the months of September to November for the first laying season and again during the months of April to June for the second laying season, given that nesting beaches are located at that site. This suggests that, in the Sinú River, turtles do not have a specific nesting area (e.g., mid or low basin) to which all migrate to nest, but rather that they lay near feeding places or within their home ranges. We found, at most, 2 nests on a single beach, which agreed with the findings of local people. It appeared that P. lewyana nesting is solitary or in small groups. These observations are similar to those of the nesting behavior of P. unifilis, which nests in groups of 2 to 6 or sometimes individually (Fachín 1992), in contrast to the aggregated nesting behavior of P. expansa (Soini et al. 1997).
Incubation period has not been assessed in situ; however, local people reported an incubation period of 48 days, and Gallego-García (2004) reported an incubation time of 66 to 70 days for eggs in a controlled environment at 32.4°C. Because intervals between beach floods are far shorter than the incubation period, eggs laid on these beaches probably do not survive. If beach flooding has been occurring since the Urrá Dam began operating in 1999, then hatchling recruitment to the population may be minimal. However, floods may be forcing females to nest in sites other than beaches. During a flood in February 2005, 4 females arrived at the backyard of a riverside house looking for a place to lay their eggs; these were captured by the inhabitants while they were digging their nests and were then killed. It is possible that the only eggs that survive the floods are those laid on the riversides; however, this is counteracted by the risk that females undergo when they move too far away from the river to nest and by the risk that hatchlings undergo when they return to the water after hatching. Because the species' reproductive ecology is not well known, as well as the effect that the floods have on the eggs, it has not been possible to estimate the impact that the dam has caused on the population. Egg mortality because of floods added to adult extraction for consumption, which suggests that this species is much more threatened than was previously believed. We hope that the information obtained in this study will help to guide management decisions for its conservation. More data are needed on this species' life history to ensure its continued existence.
Study sites (filled circles) and nesting beaches (numbers 1–5) for Podocnemis lewyana. Distribution limits are represented by squares.
Sight frequency of Podocnemis lewyana in the Sinu River, Colombia according to size class and location. Dotted bars (SCL1), stripped bars (SCL2), gray bars (SCL3), black bars (SCL4), and white bars (SCL5).
Sight frequency of Podocnemis lewyana in the Sinú River, Colombia according to size class and shoreline vegetation type. Dotted bars (SCL1), stripped bars (SCL2), gray bars (SCL3), black bars (SCL4), and white bars (SCL5).
Daily flows released by Urrá Dam during 3 reproductive seasons. Horizontal line shows minimal flow required to overflow all beaches. Data provided by the Project Urrá S.A.
