Our modeling approaches indicate that, in Tapacurá Reservoir, caiman detectability is influenced by combinations of abiotic factors, habitats and fishing practices. This study represents the longest dataset of spotlight counts for a broad-snouted caiman population. Most studies for the species are based on estimates of population structure and distribution are limited to surveys spanning less than two years (Carvalho and Verás-Batista 2013; Fusco-Costa et al 2008; Mascarenhas-Júnior et al 2020; Mourão and Campos 1995; Passos et al 2014). Our present evaluation stands out for its comprehensive approach and long-term perspective. Long-term monitoring is fundamental for revealing reliable population and demographic trends and patterns in wild crocodylians (Woodward and Moore 1993). Crocodylians are long-lived and slow-growing animals, meaning that it could take several years to unveil how changes in landscape, abiotic conditions and human disturbance influence their population parameters and distribution (Charruau et al 2022; Fujisaki et al 2011; Ortiz et al 2020). For instance, broad-snouted caimans reach sexual maturity at five years in captivity (Verdade et al 2003), but competition and resources availability in the wild can delay the onset of breeding until ca. 10 years (Moulton et al 1999; Simoncini et al 2013). To our knowledge, our assessment is the first broad-scale study allowing a better understanding of how broad-snouted caimans respond to environmental and human factors and how such responses influence their autecology and demography.
The average encounter rate of caimans we found in Tapacurá was about one caiman per kilometer, within the range of encounter rates observed for broad-snouted caimans in other regions of Brazil, which vary between 0.07 and 11.3 caimans per kilometer (Marques et al 2016; Mourão and Campos 1995). However, most population assessments revealed encounter rates less than one individual per kilometer (Fusco-Costa et al 2008; Carvalho and Batista, 2013; Passos et al 2014; Mascarenhas-Junior et al 2020). In comparison to other species of the genus, the broad-snouted caiman has a relatively lower encounter rate, likely as a result of lower population abundance. For example, estimates based on populations of the Yacaré (Caiman yacare), resulted in more than 1,000 individuals per kilometer in some central-South America wetland areas (Campos, et al 2020a), and studies of wild populations often present encounter rates higher than 10 individuals per kilometer (Aguilera et al 2008; Coutinho and Campos 1996; Mourão et al 2000). Demographic estimates of the Spectacled Caiman (Caiman Crocodilus) have also reported encounter rates higher than five individuals per kilometer (Balaguera-Reina et al 2021; Cartagena-Otálvaro et al 2020; Parra-Torres et al 2020), even in disturbed areas (Hernández et al 2021; Pereira et al 2022). The low detection probability observed during our study, combined with the fact that approximately one-third of caimans submerged upon boat approach, indicates a notable level of wariness within the population (Brazaitis et al 1996; Rebelo and Lugli 2001), which may also affect encounter rates. Crocodylians exhibit elusive behavior in areas experiencing human pressures including boat traffic, illegal hunting, or fishing activities (Grant and Lewis 2010; Tellez et al 2016). In Tapacurá Reservoir, high levels of intense fishing practice and hunting pressure are suspected to increase caiman wariness (Mascarenhas-Júnior et al 2020; Rodrigues et al 2021).
The population size of broad-snouted caimans in Tapacurá has not changed significantly since the beginning of monitoring (2015). Such stability could be attributed to stable birth and survival rates (Barboza et al 2021; Briggs-Gonzalez et al 2017), or to high site fidelity and poor dispersal patterns, which are not uncommon in broad-snouted caimans (Borges et al 2018; Verdade et al 2002; Zucoloto et al 2021). Moreover, the reservoir has abundant food supply, including both native and nonnative fish (El-Deir et al 2012), so the prevalence of agonistic interactions may be minimal and death rates by cannibalism(Campos and Mourão 2021; Drews 1990) may be low. Notably, data from interviews suggest that local human communities in Tapacurá perceive caiman population as stable in recent years (Bezerra, et al 2019a).
Although we did not find trends in population size over the years surveyed, we documented potential shifts in population structure. Juveniles and subadults make up most of the population, but the increasing number of adults is noteworthy. The relatively high number of immature caimans could be an indication of potential impacts of illegal hunting (Balaguera-Reina and González-Maya 2009; Pereira et al 2022), as larger caimans are often targeted for meat in areas where subsistence activities are prevalent (Cook et al 2022). Furthermore, they may be killed by fishermen, who perceive them as competitors for fish resources or as retaliation for damage caused to fishing gears (Hilevski and Velasco 2020). An additional possibility is that adult caimans are harder to count, as individual wariness is often positively correlated with caiman size, as result of past negative interactions with humans (Pacheco 1996b). This may influence size-class estimation during surveys, as large caimans tend to dive in advance to observer approaching, resulting in many records unassigned to any size-class, and biasing counts towards larger numbers of small individuals(Aguilera et al 2008; Briggs-Gonzalez et al 2017; Flores-Escalona et al 2021). However, the number of adults spotted in Tapacurá has increased with time, with major changes between 2021–2022. Detecting a greater proportion of larger caimans, even with no significant reduction on counts in smaller size classes, could be an indicative of high survival and growth rates (Campos, et al 2020b), and these changes could be attributed to recent conservation efforts targeting caiman populations (Borteiro et al 2008; Joanen et al 2021; Piña et al 2010). For instance, since 2018, multiple educational initiatives were implemented in Tapacurá, mostly trying to engage fishermen and the local community in conservation efforts and reduce hunting and poaching pressure (Bezerra, et al 2019b). Effects on population size from these programs may only be apparent after many years, but changes to size-class structure may be more immediate and detectable.
Most detections were made near or among floating aquatic vegetation, which provide protection to caimans against predators and human disturbances (Mascarenhas-Júnior et al 2020), while harboring more prey items than open water (Borteiro et al 2009). However, excluding unknow size class, this pattern was observed only among juveniles. Due to their limited interactions with humans and potential predators, juvenile caimans tend to exhibit less wary behavior (Pacheco 1996b), and are less likely to dive or seek refuge within aquatic vegetation when boat approached. We suspect subadult and adult individuals tend to conceal themselves behind macrophyte banks (Portelinha et al 2022), decreasing detection and accurate size estimates.
Seasonality did not predict changes in caiman encounter rates. When surveyed areas are fixed, crocodylian counts often increase when water level and rainfall decrease, because the scarcity of available areas forces them to congregate in smaller territories (Lance et al 2011; Ouboter and Nanhoe 1988; Pantoja-Lima et al 2010). Still, in regions where floodplains are small or inaccessible, such as in Tapacurá, where the margins are relatively steep, it may be difficult to detect an effect of water level on counts (Wood et al 1985). As our survey routes were designed to be adaptative across all accessible areas within the reservoir it is not surprising that we did not detect seasonal differences in encounter rates. Furthermore, we anticipate that the absence of other major neighboring rivers or reservoirs downstream of Tapacurá will lead to a reduction in emigration and immigration processes. As a result, encounter rates are expected to remain relatively stable across seasons.
Water temperature had positive effects on total encounter rate, while daily rainfall had the opposite effect. In crocodylians, warmer conditions lead to higher metabolic rates, increasing foraging, reproduction, or territorial behavior (Hutton and Woolhouse 1989; Mazzotti et al 2019; Nifong and Silliman 2017). They also increase prey availability and abundance, influencing crocodyilian foraging tactics and movements (Rosenblatt and Heithaus 2011; Somaweera et al 2011). During our surveys, water temperature varied between 24.6–30.8oC, a range sufficient to increase movement rates in other crocodilian species (Fujisaki et al 2014; Goodwin and Marion 1979; Nifong and Silliman 2017). Interestingly, temperature was present in most of juvenile and subadult best-ranked models, but not in the adult models. As larger individuals have greater thermostability than smaller ones, they could be less affected by changes in temperature (Grigg and Kirshner 2015). Also, juvenile and subadult broad-snouted caimans have higher metabolic rates than mature individuals (Mascarenhas-Junior et al 2021a), and metabolism and movements are probably more sensitive to changes in temperature. Conversely, adverse weather conditions and rain may impact crocodylian behavior, for example, by limiting their opportunities for feeding or hindering emergence from shelters (Ahizi et al 2021; Strickland et al 2018). Heavy rainfall can also limit habitat availability for smaller individuals (Herrera et al 2015). However, recent studies on West African crocodile (Crocodylus suchus) suggest that precipitation can have positive effects on counts, possibly due to effects on prey availability, increasing crocodylians’ foraging movements (Velo-Antón et al 2014); Nifong and Silliman 2017).
Cloud coverage and air relative humidity were both significant factors in predicting caiman encounter rate. Cloud coverage in our study and others had a negative effect on counts, possibly because of its association to increased wind speed, which can limit caiman emergence and feeding (Pacheco 1996a; Sarkis-Gonçalves et al 2004; Strickland et al 2018). Overcast skies can reduce the amount of moon and starlight, which can disorient smaller individuals by visibility reduction (Murphy 1981), and prompt their movements towards open waters. Conversely, humidity positively affected counts, mainly because it is closely correlated to the decrease in air temperature. Due to specific heat properties, water tends to remain warmer for longer periods compared to the air. As a result, when the water temperature exceeds air temperature, evaporation processes occur, consequently increasing the air relative humidity. Crocodylians often bask on the land or logs, even during the night (Nordberg and McNight 2023) but retreat to the water when the air temperature drops, potentially increasing their detectability (Hutton and Woolhouse 1989). It is important to note that dense fog can limit spotlight reach and can negatively impact counts (Woodward and Moore 1990).
Our study and others have failed to find an association with moonlight and water depth (Caut et al 2019; Da Silveira and Magnusson 1999; Villamarín et al 2017), though the effect of water depth may be marginally significant, and some authors have identified its significant influence on crocodylians counts (Nifong and Silliman 2017; Flores-Escalona et al 2021). Shallow waters in the Tapacurá Reservoir are found mostly near the margins, creating an ecotone between terrestrial and aquatic habitats. This ecological interface generally has a high species richness, influencing prey abundance (e.g., invertebrates and fishes) and thereby caiman detection (Caut et al 2019; Da Silveira and Magnusson 1999; Villamarín et al 2017). Moreover, these shallow areas often have extensive coverage of aquatic vegetation on the water surface (Mascarenhas-Júnior et al 2020), providing a protective habitat for caimans. Effects of moonlight are controversial between studies and can be site dependent (Da Silveira et al 2008). During dark nights, invertebrates can increase their activity due to lower predation risk, possibly increasing prey consumption by crocodylians (Eversole et al 2015; Perry and Fisher 2006). Nevertheless, intense moonlight can negatively affect observer visibility by dimming eyeshine reflection, reducing detection rate (Sarkis-Gonçalves et al, 2004).
All three reservoir habitat sectors (i.e., dam, river, and forest) were suitable for caiman distribution, as indicated by occupancy model predictions. This suggests that there are no discernible habitat conditions that restrict caiman distribution in Tapacurá Reservoir. The broad-snouted caiman is a highly adaptable species capable of inhabiting a diverse range of habitats, including human-altered areas such as agricultural lands, urban areas, and dams (Borteiro et al 2008; Marques et al 2016; Mascarenhas-Junior et al 2021b). However, we did find that occupancy is higher in the forest and river sectors compared to the dam. Forested areas are the most protected habitats with the lowest levels of boating traffic and illegal hunting. Forested fragments also provide ideal nesting sites, offering protection from high solar radiation and human interference, while providing organic material for mounts due to dense vegetation cover (Banon et al 2019; Cintra 1988; Rodrigues et al 2021). In Tapacurá, nests are located away from human settlements and only found within the forest (Barboza et al 2021). Additionally, forested areas are expected to harbor a greater abundance of prey items due to habitats complexity and nutrient cycling (Arantes et al 2018; Harper et al 1997; Lo et al 2021). The river sector likely has high prey availability from nutrient influences and continuous water flow, compared to stagnant water close to the dam (Maavara et al 2020). The decreased occurrence of caimans in the dam sector could be associated to a reduced availability of prey, due to hydrological alterations. These changes have the potential to disrupt habitat functioning by reducing spawning sites and creating barriers for species dispersion, reducing prey richness and abundance (Gehrke et al 2002). While occupancy is lower in habitats close to the dam, the average occupancy within the sector remains relatively high at over 87%, indicating that resources are still available in this area.
Fishing activity was positively correlated with an increase in caiman occupancy. The overlap between gillnets and caiman distributions has two potentially linked explanations: 1) Both caimans and fishers choose areas with higher fish abundance and compete for resources, and/or 2) caimans may be attracted to fish entangled in gillnets given that they are opportunistic and generalist feeders (Borteiro et al 2009). In both scenarios, bycatch and entanglement can be detrimental to caimans as they may sustain injuries, such as body compression and limb amputation (Mascarenhas-Junior et al 2018). Furthermore, entangled crocodylians can damage gillnets impacting subsistence activity (Amarasinghe et al 2015; Cook et al 2022). The presence of caimans in fishing areas can result in negative perceptions among local communities, potentially increasing in intentional killing of entangled individuals for retaliation (Pooley et al 2021). Although fishing activity does not appear to impact the overall population size of caimans in Tapacurá so far, the potential increase in negative interactions highlights a conservation concern.
Abiotic conditions, habitat features, and human disturbance determine caiman abundance and distribution. Enhancing our comprehension of how alterations in habitats and environmental factors impact caiman populations is of utmost importance in formulating effectives strategies for species conservation and management in Atlantic Forest. Our study yields several novel insights into the ecology of the broad-snouted caiman with implications to the species’ conservation: 1) The population size of caimans in the Tapacurá Reservoir remained relatively stable during our study, but observations include an upward trend in the number of adult individuals recorded in recent years. This observation potentially indicates a positive recovery trend following a period of significant human pressures; 2) Weather conditions are significant covariates impacting the detectability of caimans. These abiotic factors have the potential to influence crocodylians ecological dynamics, physiological stress, and the accuracy of observer detection. Nevertheless, effects vary across different size classes and interact with habitat characteristics, individual wariness, and human disturbance; 3) While all reservoir general habitat sectors were found to be suitable for caimans, river and forested areas exhibited greater occupancy compared to the dam. The availability of resources and level of protection may be influential factors that shape the preference of caimans; 4) A compelling correlation exists between fishing activity and the distribution of caimans. This relationship can be attributed to both selecting areas with abundant fish resources and/or caimans being attracted by fishes entangled in fishing nets. We are confident that our findings and the continued monitoring of broad-snouted caiman in the Atlantic Forest of Brazil will not only be useful for regional managers, conservationists, and decision-makers, but will also provide insights into the ecology and population dynamics of an important predator in a fluctuating environment. We expect our results to also be applied to other less studied and perhaps more threatened crocodylians.