INTRODUCTION

The broad-snouted caiman, Caiman latirostris (Daudin, 1802), is a Neotropical crocodilian with a wide geographical distribution in South America, including Brazil, Argentina, Paraguay, Bolivia, and Uruguay (Verdade and Piña, 2006; Verdade et al., 2010). This species inhabits a variety of habitat types, such as rivers, lakes, wetlands, and floodplains (Medem, 1983; Verdade and Piña, 2006), that have historically suffered anthropogenic pressure due to urbanization and agriculture. However, its ability to colonize man-made habitats (e.g., small weirs in forest plantations) has been reported in the last decades (Scott et al., 1990; Verdade and Lavorenti, 1990; Borteiro et al., 2008; Marques et al., 2013a,b).

The broad-snouted caiman presents a considerable degree of genetic isolation between subpopulations in the São Paulo state, Brazil, even on a microgeographic scale, with a relatively low number of individuals dispersing per generation, possibly due to anthropogenic pressures (e.g., hunting; Verdade et al., 2002). The state of São Paulo is located in the central area of the geographical distribution of the broad-snouted caiman (Verdade and Piña, 2006), and the use of anthropogenic environments can hypothetically ensure gene flow between the southern and northern populations of the species (Villela et al., 2008).

Exotic timber plantations have great economic importance in Brazil, covering around 6.7 million ha (ABRAF, 2012). An increasing demand for Eucalyptus products (e.g., cellulose, paper, charcoal, and wood) has stimulated their expansion in Brazil in the last decades (SBS, 2013). The expansion of exotic timber areas provides new elements (e.g., Eucalyptus plantations) in landscape composition that directly influence the dynamics of animal populations by changing resource availability (Saunders et al., 1991; Threlfall et al., 2012).

Understanding how species use the resources available in their environments is essential for species management and conservation plans because it provides data on their spatial requirements. One of the basic life history variables that must be understood is home range. Kenward (2001: 201) defined home range as “the area repeatedly visited by an animal.” Home range and movement patterns of wildlife animals can provide relevant information about social organization, mating system, foraging strategies, territoriality, and habitat use (Powell, 2000; Kernohan et al., 2001; Jacob and Rudran, 2003).

To date, only a few studies have been conducted to determine the home range and movement patterns of Neotropical crocodilians (e.g., Magnusson and Lima, 1991; Campos et al., 2003, 2006), likely due to the difficulty of applying conventional radiotelemetry to these often cryptic animals that inhabit in remote locations and frequently avoid humans (Read et al., 2007; Franklin et al., 2009). Magnusson and Lima (1991) showed that females of Paleosuchus trigonatus (Schneider, 1801) have small home ranges that are contained within the home range of at least one male, and Campos et al. (2003) reported coordinated terrestrial movement among Caiman yacare (Daudin, 1802) individuals. The present study aims to determine the home range and movement patterns of broad-snouted caimans in a modified silvicultural landscape of São Paulo state, southeastern Brazil, which is essential for understanding the strategies used by the species to survive in these anthropogenically modified environments.

MATERIALS AND METHODS

Analytical methodology

Home range was estimated by two methods: Minimum Convex Polygon (MCP; convex polygon formed by the union of external recorded positions; Mohr, 1947; Fuller et al., 2006) and a 95% fixed Kernel Density Estimate (KDE; probability density contours based on kernels; Worton, 1987, 1989). MCP is unable to distinguish between areas of high and low use by the animal, so KDE has been considered more accurate than MCP. KDE uses nonparametric statistics to calculate the probability of finding an individual in a particular location. The number and size of activity centers (Hayne, 1949) were determined by 50% fixed KDE. Spatial analyses were performed in Animal Movement (Hooge and Eichenlaub, 2000) and Hawth's Analysis Tools extensions of the ArcView 3.2 and ArcGIS 9 (ESRI, 2011), respectively.

Table 1.

Frequency VHF (Hz), sex (F: female, M: male), snout–vent length (SVL, cm), body mass (kg), number of position recorded, home range size estimated by Minimum Convex Polygon (MPC) and 95% fixed Kernel Density Estimate (KDE), number of activity centers (N°AC), area of activity centers (Area AC), and sampling period of each individual of Caiman latirostris monitored in silvicultural landscape, southeastern Brazil.

Normality and homoscedasticity of data (home range size and daily movements) were previously tested to guide the choice of parametric or nonparametric tests. Student's t-test (Zar, 1996) was used to examine possible differences in BM and SVL between sexes.

Home range analyses were carried out in two different ways: (a) including all individuals and (b) excluding the outlier individual (individual 148.180; see Table 1). The Whitney U Test and t-test (Zar, 1996) were used to examine possible sex differences in home range (MCP and K95) considering all individuals and without the outlier, respectively. Linear regression (Zar, 1996) was applied to investigate the relationship between home range size and BM and between home range size and body condition (body mass/SVL). In addition, linear regression was used to evaluate the relationship between home range and number of recorded locations. Results are presented as mean ± SD (min–max). Movement pattern was determined by the daily movement (m/day) of each individual during the sample period. The locations of each animal were transformed into a vector using ArcView 3.2 (ESRI, 2011) to construct a georeferenced image to estimate movement.

Daily movement was calculated by the linear distance (m) between subsequent locations. These daily movements were calculated on the basis of the probability of record one position per day independent of the period (spring/ summer and autumn/winter). Therefore, one position per day was selected at random for the movement estimates in spring/summer. The Whitney U Test was used to test for possible differences in mean daily movement between males and females. In addition, the Whitney U Test was used to detect possible differences in daily movement between the reproductive/hot period (October–February; Vac et al., 1992; Verdade, 1995) and the non-reproductive/ cold period (March–September). All statistical tests were performed in Minitab 16 (Arend, 1993).

RESULTS

Caiman mean BM and SVL were 16.50 ± 6.58 kg (8.75–27.00 kg) and 82.25 ± 9.13 cm (72.00–98.00 cm), respectively. There was no difference in BM [t = -2.00, df = 7, P = 0.105; males: 20.50 ± 6.95 kg (14.00–27.00 kg); females: 12.50 ± 3.18 kg (8.75–15.50 kg)] and SVL [t = -2.06, df = 7, P = 0.108; males: 87.75 ± 9.18 cm (80.00–98.00 cm); females: 76.75 ± 5.42 cm (72.00–84.50 cm)] between sexes.

Mean home range size for all individuals was 96.6 ± 183.9 ha (7.2–548.5 ha) and 43.2 ± 78.6 ha (2.8–236.2 ha) estimated by MCP and KDE, respectively (Table 1). There was no difference in home range sizes of males and females neither for MCP (Whitney U Test, P = 0.665) nor for KDE estimates (Whitney U Test, P = 0.885; Table 2; Figs. 1 and 2).

Mean home range size without the outlier (see Table 1, Individual 148.180) was 32.0 ± 23.8 ha (7.2–72.2 ha) and 15.6 ± 11.0 ha (2.8–35.7 ha) estimated by MCP and KDE, respectively. There was no difference in home range sizes of males and females neither for MCP (t = 0.04, df = 6, P = 0.973) nor for KDE estimates (t = 0.64, df = 6, P = 0.554; Table 2). Caiman home ranges encompassed between 1 and 3 activity centers depending on the individual (Table 1). Considering all individuals, the mean activity-center area was 12.9 ± 28.2 ha (0.4–82.5 ha), being 22 ± 40.3 ha (0.5–82.5 ha) for males and 3.8 ± 3.7 ha (0.4–9.1 ha) for females. However, the mean activity-center area without the outlier was 2.9 ± 3.0 ha (0.4–9.1 ha), being 1.7 ± 1.9 ha (0.5–4.0 ha) for males and 3.8 ± 3.7 ha (0.4–9.1 ha) for females.

Figure 1.

Home range estimated by Minimum Convex Polygon (MCP) and 95% fixed Kernel Density Estimate (KDE) for four female Caiman latirostris (each identified by a unique numerical code) in a silvicultural landscape, southeastern Brazil (Datum: SAD 69 - UTM). The numbers indicate the water bodies where the individuals were captured: 1, Açude Vermelho; 2, Lagoão; 3, Lagoa Suja; 4, Açude do Casemiro; and 5, Três Lagoas. X indicates capture location.

Figure 2.

Home range estimated by Minimum Convex Polygon (MCP) and 95% fixed Kernel Density Estimate (KDE) for four male Caiman latirostris (each identified by unique numerical code) in a silvicultural landscape, southeastern Brazil (Datum: SAD 69 - UTM). The numbers indicate the water bodies where the individuals were captured: 1, Açude Vermelho; 2, Lagoão; 3, Lagoa Suja; 4, Açude do Casemiro; and 5, Três Lagoas. X indicates capture location.

Table 2.

Mean home range size of males and females of Caiman latirostris estimated by Minimum Convex Polygon (MCP) and 95% fixed Kernel Density Estimate (KDE) in silvicultural landscape, southeastern Brazil. Home range analyses were carried out in two different ways: a) with home range sizes of all individuals and b) with home range sizes without the outlier (individual 148.180).

Table 3.

Regression equations between home range size, body measurements (body mass and body condition), and number of position of Caiman latirostris in silvicultural landscape, southeastern Brazil.

There was no relationship (P > 0.05) between home range size (MPC and K95) and BM nor body condition considering all individuals or excluding the outlier (Table 3). With the increase in the number of locations recorded, there was an increase in home range size when considering all individuals (MPC: P = 0.025; K95: P = 0.018); however, that relationship was not found when the outlier was excluded (Table 3).

Mean daily movement was 37.6 ± 18.6 m/day (1.2–238.0 m/day) for all individuals monitored during the study period. Males had higher mean daily movement than females (Whitney U Test, P = 0.030; females = 22.2 ± 4.8 m/ day; males = 53.0 ± 12.7 m/day). Mean daily movement was higher during the reproductive/hot period than the non-reproductive/cold period (Whitney U Test, P = 0.025; reproductive/hot period: 61.8 ± 33.9 m/day, non-reproductive/cold period: 27.3 ± 17.3 m/day; Fig. 3). Movements of up to 850 m were recorded for some locations inside Eucalyptus plantations.

DISCUSSION

The wide variation in home range size among individuals found in this study was similar to that of other crocodilian species. Crocodylus johnstoni Krefft, 1873 had home ranges of 2–137 ha in Lynd River, Queensland (Tucker et al., 1997). Alligator mississippiensis (Daudin, 1802) had home ranges of 2.6–16.5 ha in Rockefeller Refuge (Joanen and McNease, 1970), 0.6–256 ha in northern Louisiana (Taylor, 1984), and 6.1–165 ha in southwestern Louisiana (Rootes and Chabreck, 1993). Crocodylus porosus Schneider, 1801 had home ranges of 110–7,250 ha in the Gulf of Cambridge in Australia (Campbell et al., 2013) and 127–6,427 ha in Wenlock river in Australia (Kay, 2004).

The home range size and movement pattern of wild animals can be influenced by several factors, such as availability of food resources, breeding period, social hierarchy, anthropogenic activities, predator presence, and population density (Kenward, 2001; Millspaugh and Marzluff, 2001). Therefore, animal movement in the landscape is usually related to spatial-temporal distribution of resources or interactions with other individuals (Pough et al., 2003). Large home ranges may be the result of long-distance movements to obtain the necessary resources for the survival of individuals (Milan and Melvin, 2001). This seems to be the pattern found for broad-snouted caimans in this study.

The home range size of the broad-snouted caiman varied greatly among the individuals monitored. The male 148.180 had a much larger home range size than other males (see Table 1). Campbell et al. (2013) reported two behavioral tactics adopted by males of Crocodylus porosus to increase their reproductive success: (1) site-fidelity: dominant males (larger body) patrol their territory, which includes the home ranges of several females; and (2) nomadic: subordinate males (smaller body) must travel long distances to find unprotected females. However, these dominance/subordination dynamics do not seem clear among males of the broad-snouted caiman in this study.

Subordinate males are smaller and have larger home ranges because they must travel long distances to find females. This was not observed in the study population, as the two males with the largest home ranges (individuals 148.180 and 148.260; Fig. 1) were not the smallest individuals (i.e., they were probably not subordinate males). Moreover, other factors probably influenced male home range size in the present study, such as anthropogenic disturbances occurring in the study area. Caiman yacare can move on land to avoid anthropogenic disturbances (e.g., hunting; Campos et al., 2006) or move to search and consume food resources (Campos, 2003). Furthermore, an individual pattern of space use may be occurring in the study area, as previously described in captivity (Verdade et al., 2006).

Figure 3.

Mean daily movement (m/d) of individual Caiman latirostris (each identified by unique numerical code) in a silvicultural landscape, southeastern Brazil (M: male; F: female).

Broad-snouted caiman daily movement was higher during the reproductive period (October–February) than the non-reproductive period (March–September). Crocodilians are ectothermic animals that undergo considerable changes in their physiology and behavior in relation to environmental temperature (Lang, 1987b; Pough et al., 2003). Increased movement during the warm seasons is common in this taxon (Cott, 1961; Chabreck, 1965; Goodwin and Marion, 1979; Smith, 1979; Kay, 2004) and could be related to the search for food (Verdade et al., 2006) or females. Seasonal variation in temperature can affect metabolism and, hence, appetite and food consumption in crocodilians (Coulson and Hernandez, 1983; Larriera et al., 1990), and broad-snouted caimans increase food consumption, food conversion rate, and growth rates when exposed to higher temperatures (Larriera et al., 1990; Parachú et al., 2009).

During the reproductive season of the broad-snouted caiman, breeding occurs from October to December and nesting from January to February (Vac et al., 1992; Verdade, 1995; Portelinha et al., 2015). Males are hypothesized to increase their movement in order to find a higher number of females for reproduction, thus maximizing their fitness. On one hand, intra-sexual hierarchy can result in this process of partner search and promote agonistic interactions (Verdade, 1992). On the other hand, it can occasionally fail to prevent subordinate males to successfully breed, with consequent multiple paternity in some clutches (Amavet et al., 2008).

Females can also increase their movements in order to find nesting habitats (Zug, 1993; Campbell et al., 2013), which is possibly related to temperature-dependent sex determination (Campos, 1993; Piña et al., 2007) and fitness of the hatchlings (Verdade, 2001). According to Larriera (1995), female broad-snouted caimans can nest in savannah (sites with low slope that overflow in periods of heavy rain), floating vegetation (type of vegetation that floats when the water level increases), and forest. After nesting, females generally remain close to the nest area during the incubation period (Lang, 1987a). We did not find nests during this study period; however, the permanence of female 148.240 in vegetation during the reproductive period suggests that she was caring for her nest.

The use of different water bodies or wetlands by aquatic reptiles might be needed in areas where these environments are spatially dispersed (Joyal et al., 2001; Roe et al., 2003; Roe and Georges, 2007). As such, landscape characteristics (e.g., habitat area and connectivity) can directly influence movement (Roe and Georges, 2007), and the presence of activity centers in different water bodies within the home range of monitored individuals underscores the importance of matrix permeability to the movements of species. The present study suggests that Eucalyptus plantations can be considered permeable to broad-snouted caiman movements. Permeability for specialist species as well as habitat quality for generalist species are crucial for biodiversity conservation in agricultural landscapes (Verdade et al., 2014).

The present results suggest that the wide variation in the home range sizes of the broad-snouted caiman might result from the combination of social behavior and seasonal variation in resource availability. Future studies on crocodilians in agricultural landscapes should prioritize other dimensions possibly related to space use, such as agricultural practices of the matrix, water contamination by agrochemicals, and hunting pressures.