Big cats like water: occupancy patterns of jaguar in a unique and insular Brazilian Amazon ecosystem

doi:10.21203/rs.3.rs-2600459/v1

Patterns of detectability and occupation of the jaguar (Panthera onca) vary throughout its distribution, being determined primarily by vegetation cover, prey availability, and anthropogenic factors. However, there is still a large knowledge gap regarding what determines jaguar occupancy, especially in the Amazon. This knowledge gap is even more pronounced for oceanic islands, which represent unique and very sensitive ecosystems, such as the Maracá-Jipioca Islands of the Northeastern Brazilian Amazon. Our study aimed to establish the spatial ecology of jaguars in this insular ecosystem and to provide information to facilitate sustainable management of the population there. We assessed how different factors (vegetation type and prey availability) potentially influence detectability and occupancy patterns of the jaguars of the Maracá-Jipioca Islands. We found that greater wetland (beach-sea and lagoon-channel) cover was the main driver of jaguar detectability and occupancy. We reveal factors driving the population and spatial ecology of jaguars in an insular system. Despite imminent threats to the region, the knowledge we present can inform the sustainable management of jaguars to ensure that the fundamental and unique ecosystem services provided by this top predator are maintained.

Maracá-Jipioca

Panthera onca

prey availability

wetland conservation

The jaguar (Panthera onca) is one of the most widely distributed apex predators worldwide and the biggest cat inhabiting the Western hemisphere (Hunter 2020). Its extensive distribution in the Americas have resulted in its colonization of very distinct landscapes, ranging from rainforests, savannas, and dry tropical forests to sub-tropical scrublands, some of which are seasonally flooded (Rabinowitz and Zeller 2010). Such a wide range of habitats and its conservation status as Near Threatened (Quigley et al. 2017) have prompted much research into occupancy patterns of this felid, revealing several factors including, vegetation cover and food availability, shape those patterns throughout most of its range (Villalva and Palomares 2022).

Although environmental contexts may differ dramatically between areas occupied by jaguars (ranging from Northern Mexico to Northern Argentina), three main drivers seem to determine occupancy patterns. First, prey availability has a positive influence on jaguar occupancy (Anile et al. 2020; Arroyo-Arce et al. 2014; Rabelo et al. 2019). Second, broad native vegetation cover and greater net primary production is positively influential (Jędrzejewski et al. 2018). Finally, and in contrast to the latter two drivers, human activities exert a negative influence on occupancy (Jędrzejewski et al. 2017).

Nevertheless, jaguars seem to tolerate habitat disturbance to varying degrees, as they can move through deforested areas and those frequented by humans (Balbuena-Serrano et al. 2021). However, the functional role of such areas is incompletely understood, i.e., whether they represent dispersal routes or potentially colonized areas (de Azevedo and Murray 2007; Foster et al. 2010; Amit et al. 2013; Carral-García et al. 2021). Although occupancy studies cover most of the jaguar´s range and landscape contexts, few studies have targeted an insular population of this large felid, where environmental conditions and notably the marked isolation can pose additional challenges for individual survival.

Maracá-Jipioca Ecological Station (ESEC-MJ), located north of the mouth of the Amazon on the Atlantic coast, is one such system of islands. It is characterized by extensive beaches, mangroves, flooded grasslands, lagoons, and wetlands, creating a humid Neotropical region displaying unique characteristics (Ferreira et al. 2017). Strongly influenced by the Amazon River delta and located ~ 6 km from the mainland, a significant proportion of the island system experiences daily tides of 10-m amplitude that flood its extensive beach areas (Anthony et al. 2014, 2021; Santos et al. 2016; Ferreira et al. 2017). Covering ~ 60,000 ha, the two main islands harbor high concentrations of jaguar prey (e.g., fish, crocodilians, turtles, and wetland birds) (Ferreira et al. 2017) and, consequently, host an abundant population of jaguars (6.67 individuals/100 km²), whose occupancy patterns are probably influenced by the hydrodynamics of the islands and its influence on water-dependent prey (Duarte et al. 2022). Despite the spatial limitations of the islands, collectively they provide a wide array of food resources, both aquatic and terrestrial, which can be exploited by opportunistic predators such as jaguars (Jędrzejewski et al. 2018; Duarte et al. 2022; Friedeberg-Gutiérrez et al. 2022). The unique conditions of ESEC-MJ likely shape the ecology and behavioral patterns of the jaguars inhabiting this system, which display a strong dependency on both aquatic and terrestrial resources (Sanderson et al. 2002; Morato et al. 2016; Ferreira et al. 2017; Duarte et al. 2022). This dynamic relationship to both types of ecosystems may be fundamental to how this large felid provides ecosystem services and influences the wildlife communities of ESEC-MJ (Arias-Alzate et al. 2017; Craighead and Yacelga 2021).

Therefore, to better understand how jaguars have adapted to this unique insular ecosystem and to provide information supporting a sustainable management plan for ESEC-MJ, it is fundamental to understand the processes driving habitat use by the jaguar and to assess their occupancy patterns. Our study aims to fill this knowledge gap by assessing if habitats at the interface of land and water, namely those along lagoons, channels, and beaches, together with the occurrence of the most consumed prey species, potentially influence the detectability and occupancy patterns of jaguars in ESEC-MJ. We hypothesized that jaguars would be more easily detected in habitats with reduced vegetation cover, such as beaches, and those dominated by perennial aquatic environments (e.g., lagoons) (Duarte et al. 2022). Moreover, we predicted that jaguar occupancy would be promoted by a higher probability of prey occurrence (Eriksson et al. 2022). We tested these hypotheses by adopting a camera-trapping approach, supplemented by analyzing the frequency of prey in jaguar diets, to assess how habitat type and prey occurrence drive the spatial ecology of this unique insular population of jaguars.

Study area. — Our study was conducted on the Maracá-Jipioca Islands (02°01′13″N, 50°30′20″W), located off the coast of the Brazilian Amazon (Fig. 1), which originated from erosional processes and deposits of tertiary sediments on consolidated mud (Allison et al. 1995; Nittrouer et al. 1995). ESEC-MJ is a federally protected area comprised mostly of estuarine habitats (Ferreira et al. 2017). With annual rainfall ranging from 2,300 to 2,800 mm (Am climate type, according to Köppen’s climate classification) and limited variation in elevation (mean altitude = 3 meters), the islands have extensive flooded areas and lagoons (Tavares 2014; Ferreira et al. 2017). The distance from the islands to the continental coast ranges from 6–10 kilometers (Ferreira et al. 2017; Duarte et al. 2022). The vegetation on the islands is primarily composed of mangroves, grasslands, and formations of mixed vegetation (locally called Tesos) consisting of mangrove trees (Rhizophora sp. and Avicennia sp.), bamboo (Guadua sp.), and shrubs (Ferreira et al. 2017). Together, the islands harbor at least 11 vertebrate orders that have been reported as jaguar prey, including tegu (a lizard; Salvator merianae), shorebirds, rodents, and ungulates. However, no other species of terrestrial carnivore inhabits the islands, probably due to spatial limitations and the competitive pressure imposed by jaguar (Duarte et al. 2022). Therefore, the functional role of the jaguar as a predator in this system is irreplaceable.

Field Survey. — Camera traps are a widely used non-invasive method to survey wild felids, as they provide reliable data for assessing space use and occurrence patterns of wildlife, enabling population-level inferences (Hebblewhite and Haydon 2010; Burton et al. 2015; Karanth et al. 2017; Carvalho et al. 2019). Our data on jaguar detectability was gathered from 25 camera-trap stations installed at an average spacing of 2.4 ± 1.1 km, which were active for 90 days, from 25 November 2018 to 22 February 2019 (Duarte et al. 2022). Each station (considered a spatial replicate) hosted two paired camera traps (Bushnell Trophy Cam HD), installed 30–60 cm above the ground. Individual jaguars were identified using the patterns of rosettes along their flanks and spots on other body parts. Due to logistical constraints with installing camera-traps across the islands, we selected a major continuous and accessible region, which was sufficiently representative of the entire island landscape. A minimum convex polygon encompassing our camera trap sites covered 25% of the total area of the islands (Duarte et al. 2022). The cameras functioned continuously and were programmed to take three sequential photos and a 40-s video when triggered. The total sampling effort was 2,250 station-trap days (i.e., the number of stations multiplied by the number of days that they were active; Srbek-Araujo and Chiarello 2007), with all camera traps used in our analyses working throughout the sampling period. We did not deploy baits to avoid biasing the detection probabilities of jaguar and their prey (Duarte et al. 2018).

Characterization of camera-trap stations. — Each camera-trap station was characterized based on a set of environmental variables that best represent the landscape context of the islands and that, according to other studies of jaguars, may influence the detectability and occupancy patterns of this species (Ferreira et al. 2017; Duarte et al. 2022). Accordingly, we determined the habitat composition of each camera-trap station based on four habitat types: wetland (beach and sea, lagoons and channels), grassland, mangrove, and Teso (Ferreira et al. 2017) (Table 1). We used raster information on the islands and the surrounding marine area available from the MapBiomas web platform (MAPBIOMAS 2022) to assess coverage (ha) of each habitat type (Fig. 1). These data were imported to QGis software to assess the proportion (ha) of each habitat type within a buffer of 1 km radius around each monitoring station (Allen et al. 2022; McManus et al. 2022; QGIS Development Team 2022). Buffer size was determined based on the distances between camera trap stations, i.e., ~ 50% of the average distance between each station. In addition, we also used the software Google Earth Engine (Gorelick et al. 2017) to estimate the distance from each monitoring station to the water (sea, lagoons, and channels) (Table 1).

Prey selection. — Data on mammals, reptiles, and birds recorded at more than five of our monitoring stations and previously described or recorded as jaguar prey (see González and Miller 2002; Duarte et al. 2022) were considered as possible drivers of jaguar occupancy patterns. Among all the mammals detected, we considered only four mammals as potential jaguar prey, i.e., agouti (Dasyprocta leporina), black-eared opossum (Didelphis marsupialis), giant anteater (Myrmecophaga tridactyla), and white-tailed deer (Odocoileus virginianus) (Table 1). Tegus were the only species of reptile that we included in our analysis (Table 1). We also included three orders of birds as potential jaguar prey, namely Anseriformes, Gruiformes, and Pelecaniformes (Table 1). We did not consider shorebirds at species level, since they are social and our records correspond to mixed flocks, which would result in super-abundance estimates (Miller 1984; Farmer and Durbian 2006; Lanctot et al. 2008). Despite jaguars being recognized as also consuming crocodilians, turtles, and fishes (Supplementary Information – Table S1), we could not estimate their abundance from our camera-trap data since camera-trapping is not an adequate method for sampling those species and so we have no detection data for those groups. Therefore, they were not considered as drivers in our occupancy model analysis. However, we acknowledge that these groups are present on the islands (personal observation) and in the diet of the islands’ jaguars (Supplementary Information – Table S1), so likely also influence the spatial ecology of this felid.

Data Analysis. — The detectability and occupancy patterns of jaguars and their main prey were assessed using an occupancy modeling approach developed by MacKenzie et al. (2017). This modeling procedure estimates the probability of a site being occupied/used (ψ) by jaguar and/or its prey when detectability is imperfect (a detection probability of < 1, i.e., the species is present but not detected). We structured our model based on 25 sites (camera-trap stations) divided into six groups of 15 days. We also scaled the environmental variables (ha) and prey probability (%) into maximum likelihood estimates (Arnold 2010). Prior to conducting the modeling procedure, we tested for correlations among variables by estimating Spearman correlation coefficients between all independent variables and considered that no significant correlation existed if the coefficient value was between − 0.7 and 0.7 (Dormann et al. 2013).

The modeling procedure was implemented in two phases according to the following approach. First, we tested which factors were more influential in determining target species detectability while maintaining occupancy as constant (i.e., ψ = 1). We tested four environmental variables as drivers of jaguar detectability and for each prey type (Table 1 – Environmental variables). We built models corresponding to all combinations of these four environmental variables and ranked them according to the Akaike Information Criterion (AIC) (Burnham et al. 2011). We also calculated overdispersion factors (c-hat) to assess overdispersion of our best models (MacKenzie et al. 2017). The model with the lowest AIC and ΔAIC = 0 (the difference between the AIC of the model and the lowest AIC value; MacKenzie et al. (2017)) was identified as the best model and the variables included in this best model were considered the most influential variables of the detectability component of all subsequent models (Castro et al. 2022). Secondly, we used the variables from the best detectability model and included them in all models built subsequently to assess the drivers of occupancy. We followed the same approach described above for model building and selection (Castro et al. 2022) to conduct occupancy modeling. However, in this case, we tested three different hypotheses: H1 - jaguar occupancy is mostly determined by environmental variables (i.e., habitat) (Table 1); H2 – jaguar occupancy is more dependent on occupancy patterns of its potential prey (Table 1) (i.e., prey); H3 – jaguar occupancy is determined by the concurrent action of both habitat characteristics and prey occupancy (Table 1) (i.e., habitat + prey). Therefore, we built three sets of models. For H2 (prey), first we built occupancy models for each prey type individually and then identified the best model for each species (using the same procedure as applied for jaguars). We used the predicted occupancy from each of those prey models to estimate the probability of occupancy of each prey type per camera-trap station. Then, we used the percentage probabilities to determine approximate prey availabilities as covariates in H2 and H3. We identified the hypothesis with strongest support as that which presented the lowest AIC value for the best occupancy model (i.e., overall ΔAIC = 0). We identified informative variables by highlighting those whose 95% confidence interval of the coefficient did not include 0 (Arnold 2010), i.e., we were confident that their influence on jaguar occupancy was either positive or negative.

We used the software ‘R’ (R Core Team 2022) and the packages ‘camtrapR’ (Niedballa et al. 2016), ‘unmarked’ (Fiske and Chandler 2011), ‘wiqid’ (Meredith 2020), ‘AICcmodavg’ (Mazerolle and Mazerolle 2017), and ‘MuMIn’ (Barton 2009a; b) for these analyses.

Table 1

—Explanatory variables used (see text for details of habitat classification) to test each of our formulated ecological scenarios about jaguar detectability and occupancy patterns in ESEC-MJ.
Hypothesis	Variable	Data Type	Description
Environmental variables (H1)	Grassland	Numerical continuous	Proportion in ha of Grassland within a 1 km radius
	Mangrove and teso	Numerical continuous	Proportion in ha of Mangrove and teso within a 1 km radius
	Wetland	Numerical continuous	Proportion in ha of Wetland within a 1 km radius
	Distance to water	Numerical continuous	Distance in m of camera traps to the water (sea, channels, and lagoons)
Prey variables (H2)	Probability occurrence Agouti	Numerical continuous	Probability of occurrence of Agouti (0–99%)
	Probability occurrence White-tailed deer	Numerical continuous	Probability of occurrence of White-tailed deer (0–99%)
	Probability occurrence Giant anteater	Numerical continuous	Probability of occurrence of Giant anteater (0–99%)
	Probability occurrence Black-eared opossum	Numerical continuous	Probability of occurrence of Black-eared opossum (0–99%)
	Probability occurrence Tegu	Numerical continuous	Probability of occurrence of Tegu (0–99%)
	Probability occurrence Anseriformes	Numerical continuous	Probability of occurrence of Anseriformes (0–99%)
	Probability occurrence Gruiformes	Numerical continuous	Probability of occurrence of Gruiformes (0–99%)
	Probability occurrence Pelecaniformes	Numerical continuous	Probability of occurrence of Pelecaniformes (0–99%)
Environmental and prey variables (H3)	Environmental variables + Occupancy probability of prey	Numerical continuous	Environmental variables (H1) together with the occupancy probability of each prey (H2)

We obtained 32 independent detections of jaguars at 16 stations and with at least one record in each of the six 15-day sessions. Mangrove and teso were negatively correlated (r= -0.728) with grassland and positively correlated with Gruiformes occupancy (r = 0.857). Grassland presented a negative correlation with prey species occupancy for D. leporina, D. marsupialis and S. merianae (r=-0.826; r=-0.816; r=-0.769; respectively). Myrmecophaga tridactyla was the only species that displayed a positive correlation with distance to water (r = 0.701). The Gruiformes group was also correlated positively with S. merianae (r = 0.880), as was D. leporina with D. marsupialis (r = 0.843) (Supplementary Information – Table S2). Accordingly, we tested two different scenarios for the detectability models and for hypothesis H1; one with Mangrove and teso and another with Grassland. Four scenarios/models were considered for hypothesis H2: D. leporina and S. merianae; D. marsupialis and S. merianae; D. leporina and Gruiformes; and D. marsupialis and S. merianae. Finally, for hypothesis H3, we considered the variables of the best model for all scenarios, excluding those displaying correlations (Supplementary Information – Table S4).

From our jaguar models, we identified a greater proportion of wetland and distance to water as being the most influential drivers promoting detectability (i.e., ΔAIC = 0; Table 2). However, as a predictor of detectability in our occupancy models, we considered only wetland coverage as a variable due its significance (p < 0.01) and to the absence of a correlation with any of the other variables (Supplementary Information – Table S2). Although we examined several covariates and have relatively few sites, our best models did not present any significant overdispersion, with estimated values of c-hat for H1 = 1.31, for H2 = 1.34, and for H3 = 1.29. Accordingly, we assume that our models present no overdispersion or lack of fit, since that would only be the case if c-hat is ~ 1 (MacKenzie et al., 2017). The occupancy hypothesis with the strongest support among the tested scenarios, i.e., comprising the model with the lowest AIC and significance in relation to the tested variables, was the one that considered only habitat characteristics as drivers of jaguar occupancy patterns (i.e., H1; Table 3).

Table 2

— Detection probability models, ranked by ΔAIC [showing degrees of freedom (df), AIC and ΔAIC; p – detection probability; Ψ (1) – constant occupancy probability. Only the three models with the highest fit (lower AIC) and weight > 0.1 per scenario are presented] and a summary of the best overall detectability model [Estimated variable coefficients; Coef – Coefficients; SE – Standard Error; P-value; CI 95% − 95% Coefficient Confidence Interval)].
Model			df	AIC		∆AIC		Weight
p (Wetland + Distance to water), Ψ (1)			4	136.5		0.0		0.600
p (Wetland + Distance to water + Grassland), Ψ (1)			5	138.5		1.95		0.226
p (Wetland), Ψ (1)			3	140.1		3.57		0.101
Best model characteristics
Estimated variable	Coef	SE	Z value		P		CI 95%
Intercept	-2.1	0.38	-5.5		< 0.001		-2.2/-1.03
Wetland	0.68	0.24	2.8		0.005		0.28 / 1.09
Distance to water	-1.02	0.55	-1.8		0.06		-1.92 / -0.11

Table 3

—Top occupancy models per hypothesis and respective scenarios (H1: habitat, H2: prey, H3: habitat + prey). The model best supported by our data (i.e., overall ΔAIC = 0, AIC = 138.9 and weight 0.640) is the top model for the habitat hypothesis (Supplementary Information – Table S4). P – detection probability; Ψ – occupancy probability; df - degrees of freedom; AIC - Akaike Information Criterion; ΔAIC – the difference in AIC between the top model within each scenario and weight.
Model	df	AIC	Weight
p (Wetland), Ψ (Distance to water) (H1)	4	138.9	0.640
p (Wetland), Ψ (Black-eared opossum + Gruiformes + Giant anteater) (H2)	6	137.8	0.689
p (Wetland), Ψ (Distance to water) (H3)	4	138.9	0.412

According to the overall best model, jaguar detectability is significantly higher (i.e., the 95% confidence interval of the variable coefficient did not include 0) in environments with higher proportions of wetland and with access to the sea (Table 4). The occupancy pattern of jaguar was significantly higher near aquatic environments, i.e., sea, lagoons, and channels (Table 4). Furthermore, some of the evaluated prey items seem to influence jaguar occupancy patterns, as the probability of occurrence of Gruiformes, D. marsupialis and M. tridactyla were all included in the best prey models (H2) (Table 4 and Supplementary Information – Table S4). Although all these prey items seem to influence jaguar occurrence, individually they do not present a significant relationship with jaguar occupancy (Table 4).

Table 4

—Estimated variable coefficients for the overall best models from H1 and H2 to explain occupancy and used to explain jaguar detection and occupancy patterns (Coef – Coefficient; SE – Standard Error; P – p value; CI 95% − 95% Coefficient Confidence Interval).
Best H1 model p (Wetland), Ψ (Distance to water)	Detection Estimated variable	Coef	SE	Z value	P	CI 95%
	Intercept	-1.11	0.28	3.89	< 0.001	-1.58 / -0.64
	Wetland	0.75	0.23	3.15	0.001	0.35 / 1.14
	Occupancy Estimated variable	Coef	SE	Z value	P	CI 95%
	Intercept	1.49	1.25	1.19	0.23	-0.57 / 3.56
	Distance to water	-2.12	1.28	1.65	0.09	-4.24 / -0.009
Best H2 model p (Wetland), Ψ (Black-eared opossum + Gruiformes + Giant anteater)	Detection Estimated variable	Coef	SE	Z value	P	CI 95%
	Intercept	-1.12	0.23	4.83	< 0.001	-1.50/-0.74
	Wetland	0.77	0.23	3.33	< 0.001	0.39/1.15
	Occupancy Estimated variable	Coef	SE	Z value	P	CI 95%
	Intercept	42.3	38.7	1.09	0.27	-56.9/152.5
	Black-eared opossum	11.0	16.5	0.67	0.50	-50.59/76.0
	Gruiformes	27.3	26.1	1.04	0.29	-17.8/90.0
	Giant anteater	-43.1	41.7	-1.03	0.3	-153.1/53.8

The detection probabilities of jaguars proved higher in wetlands and water-associated habitats, specifically on the beaches of the Maraca-Jipiocá Islands. Beaches are characterized by open areas lacking in or with reduced vegetation, allowing greater visibility and consequently a higher probability of detecting prey (Ferreira et al. 2017; MacKenzie et al. 2017; O’Connor et al. 2017). Moreover, their openness also facilitates jaguar movement for hunting and territory patrols, enhancing jaguar detectability (O’Connell et al. 2011; Kolowski et al. 2021). Thus, it is unsurprising that these environments play an important role in the spatial ecology of jaguars in ESEC-MJ.

The detection probability of jaguar represents the capacity to detect an individual in a monitored site, so a detection record indicates that an individual used the monitoring site, but the absence of detection is not indicative of avoidance (MacKenzie et al. 2002). Moreover, the assumption of “use” can be misleading since, for example, an animal in transit could be detected but not necessarily occupy that location (MacKenzie et al. 2017; O’Connor et al. 2017). However, in our study the pattern of detectability is corroborated by the positive influence of habitats near lagoons, channels, and the sea on the occupancy probability for jaguar. These results highlight that detection records in areas displaying such characteristics indicate that the jaguars are not simply moving through these sites but reflect frequent occupancy in these areas as part of their daily activities (e.g., hunting) and territory. Apart from the advantages of easier movement through open habitats, the occurrence of jaguar in such habitats is also related to prey access. Many vertebrates congregate near water, enhancing the probability of capturing prey such as Black-eared opossum and Gruiformes (Duarte et al. 2022; Eriksson et al. 2022). Notably, the Maracá-Jipioca Islands are subjected to daily tides of 10-m amplitude, which extensively flood beach areas (up to 2 km inland) and channels, creating the flooded habitats that favor jaguar predation (Santos et al. 2016; Ferreira et al. 2017; Duarte et al. 2022).

The best model of hypothesis H2 (prey) revealed that jaguars in ESEC-MJ preferred areas displaying higher occupancy for some target prey. However, the results did not allow us to determine any preference for a specific group of prey. In fact, the model encompassed a diverse set of prey - Gruiformes, black-eared opossum (D. marsupialis), and giant anteater (M. tridactyla) - highlighting that apart from aquatic prey, many of the tested prey groups exert some influence on the probability of jaguar occupancy. Jaguars are opportunistic predators with a varied diet and high tolerance to some degree of disturbance (de Azevedo and Murray 2007; Foster et al. 2010; Amit et al. 2013; Carral-García et al. 2021). Although our models for prey were not statistically significant, the presence of black-eared opossum and Gruiformes remains in the fecal samples of jaguar further support that both these prey groups are contributory factors in the occupancy patterns of ESEC-MJ’s jaguars.

Their diverse predatory behaviors enable jaguars to exploit prey from both terrestrial and aquatic ecosystems, resulting in diets that vary according to localities (Jędrzejewski et al. 2017; Eriksson et al. 2022; Friedeberg-Gutiérrez et al. 2022). Islands are spatially restricted environments typically displaying resource limitations, so insular populations of a large predator would likely face challenging survival scenarios (MacArthur and Wilson 2016). However, the high frequency of fishes, crocodilians, and other detected prey in our jaguar fecal samples (Supplementary Information – Table S1) show the extensive dietary breadth of the jaguar population on the Maracá-Jipioca Islands. Thus, these jaguars consume prey not only available from the islands’ terrestrial and lagoon environments, but also from the surrounding sea. This capacity to exploit niches that are not limited by the islands’ terrestrial boundaries might represent an ecological strategy enabling this insular population to overcome the adversities imposed by limited terrestrial area and, consequently, of terrestrial prey availability.

We acknowledge that including data on these other important prey categories in our occupancy modeling would likely have increased the robustness of our models. However, we could not consider occupancy of fishes or crocodilians in our models because camera-trapping is an ineffective approach for sampling both these prey groups. Such aquatic prey play important roles in the ecology and behavior of jaguars in other regions, such as in the Pantanal and Venezuelan Llanos (Jędrzejewski et al. 2017; Eriksson et al. 2022). Further studies on the Maracá-Jipiocá Islands are needed to investigate the relationships between aquatic prey and jaguars and assess their ecological importance and how they may influence the spatial ecology of the predators.

Our study sheds light on the factors that influence detectability and occupancy of an insular population of jaguars. Although further assessments are necessary to reveal the ecological adaptations of this unique population, the environmental context currently shaping its ecology will likely change under forecasted climate change scenarios that will affect the entire Amazon coastal region. Introductions of exotic species (e.g., Bubalus bubalis), freshwater salinization, increased poaching, and habitat reductions due to increased sea levels will all challenge jaguar survival along the Amazon coast (Anthony et al. 2014, 2021; Dos Santos et al. 2018; Fernandez et al. 2019; Morcatty et al. 2020). Thus, it is extremely urgent to understand the dynamics between jaguars and aquatic and terrestrial ecosystems (including their prey) in this region, how they affect their spatial-temporal ecology. Despite imminent threats to the region, the knowledge generated in this study can support conservation strategies and inform sustainable management plans, thereby preventing population declines or local extinction of this unique population and consequently maintaining its fundamental and exclusive ecosystem services as a top predator.

Acknowledgments

We thank Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) for funding of the doctoral scholarship to the author HOBD. WDC was supported by 'Ayudas Maria Zambrano' (CA3/RSUE/2021-00197), funded by the Spanish Ministry of Universities. Additional training support was provided by Pós-graduação em Biodiversidade Tropical (PPGBio), Universidade Federal do Amapá (UNIFAP), funded by the Programa Nacional de Cooperação Acadêmica da Amazônia (PROCAD-AM/CAPES, no. 88887.200472/ 2018-00). We thank Programa Nacional de Cooperação Acadêmica na Amazônica (PROCADAmazônia/CAPES) for funding of an international exchange during the PhD of HOBD, which was realized in collaboration with cE3c - Center for Ecology, Evolution and Environmental Change & Change - Global Change and Sustainability Institute, Faculdade de Ciências, Universidade de Lisboa, Portugal, which was funded by the FCT/MCTES through national funds and co-funding by FEDER within the PT2020 Partnership Agreement and Compete 2020 (UIDB/00329/2020). Finally, we are extremely grateful to the World Wide Fund for Nature WWF-Brazil and the Chico Mendes Institute for Biodiversity Conservation (ICMBio) for their funding and operational logistics support for field research, and especially to all the firefighters and field assistants for all their hard work in the field.

Ethical Statement

No direct funding was obtained for the present work, but the data presented herein were collected with funding and operational logistics support for field research from the World Wide Fund for Nature (WWF-Brazil) and the Chico Mendes Institute for Biodiversity Conservation (ICMBio). Therefore, there is no conflict of financial interests. The study was authorized by ICMBio. The methods applied were not invasive, and we did not use animal experimentation, so there was no need for specific approval from an appropriate ethics committee for research involving animals.

Allen, M. L., A. M. Green, and R. J. Moll. 2022. Modelling the distribution and intraguild associations of an understudied mesocarnivore across the contiguous United States. Diversity and Distributions 28:1022–1033. https://doi.org/10.1111/ddi.13502
Allison, M. A., C. A. Nittrouer, and L. E. do C. Faria Jr. 1995. Rates and mechanisms of shoreface progradation and retreat downdrift of the Amazon River mouth. Marine Geology 125:373–392. https://doi.org/10.1016/0025-3227(95)00020-Y
Amit, R., E. J. Gordillo-Chávez, and R. Bone. 2013. Jaguar and puma attacks on livestock in Costa Rica. Human-Wildlife Interactions 7:77–84.
Ancrenaz, M., A. J. Hearn, J. Ross, R. Sollmann, and A. Wilting. 2012. Handbook for wildlife monitoring using camera-traps. BBEC II Secretariat. Natural Resources Office Chief Minister’s Department, Kota Kinabalu, Sabah, Malaysia.
Anile, S., E. Greenspan, and C. K. Nielsen. 2020. Determinants of jaguar occupancy at the northern range edge. Mammal Research 65:667–677. https://doi.org/10.1007/s13364-020-00511-0
Anthony, E. J., E. S. Brondizio, V. F. Dos Santos, A. Gardel, and M. Besset. 2021. Sustainable Management, Conservation, and Restoration of the Amazon River Delta and Amazon-Influenced Guianas Coast: A Review. Water 13:1371. https://doi.org/10.3390/w13101371
Anthony, E. J., A. Gardel, and N. Gratiot. 2014. Fluvial sediment supply, mud banks, cheniers and the morphodynamics of the coast of South America between the Amazon and Orinoco river mouths. Geological Society, London, Special Publications 388:533–560. https://doi.org/10.1144/SP388.
Arias-Alzate, A., J. F. González-Maya, J. Arroyo-Cabrales, and E. Martínez-Meyer. 2017. Wild felid range shift due to climatic constraints in the Americas: a bottleneck explanation for extinct felids? Journal of Mammalian Evolution 24:427–438. https://doi.org/10.1007/s10914-016-9350-0
ARNOLD, T. W. 2010. Uninformative parameters and model selection using Akaike's Information Criterion. The Journal of Wildlife Management, 74(6), 1175-1178. https://doi.org/10.1111/j.1937-2817.2010.tb01236.x
Arroyo-Arce, S., J. Guilder, and R. Salom-Pérez. 2014. Habitat features influencing jaguar Panthera onca (Carnivora: Felidae) occupancy in Tortuguero National Park, Costa Rica. Revista de Biología Tropical 62:1449–1458.
de Azevedo, F. C. C., and D. L. Murray. 2007. Evaluation of potential factors predisposing livestock to predation by jaguars. The journal of wildlife management 71:2379–2386. https://doi.org/10.2193/2006-520
Balbuena‐Serrano, Á., M. M. Zarco‐González, O. Monroy‐Vilchis, R. G Morato, and R. C De Paula. 2021. Hotspots of livestock depredation by pumas and jaguars in Brazil: a biome‐scale analysis. Animal Conservation 24:181–193. https://doi.org/10.1111/acv.12619
Barton, K. 2009a. MuMIn: multi-model inference. http://r-forge. r-project. org/projects/mumin/.
Barton, K. 2009b. MuMIn: multi-model inference. R package version 1. 0. 0. http://r-forge. r-project. org/projects/mumin/.
Burnham, K. P., D. R. Anderson, and K. P. Huyvaert. 2011. AIC model selection and multimodel inference in behavioral ecology: some background, observations, and comparisons. Behavioral Ecology and Sociobiology 65:23–35. https://doi.org/10.1007/s00265-010-1029-6
Burton, A. C. et al. 2015. Wildlife camera trapping: a review and recommendations for linking surveys to ecological processes. Journal of Applied Ecology 52:675–685. https://doi.org/10.1111/1365-2664.12432
Carral-García, M., I. Buenrostro, H. Weissenberger, V. Rosales, and J. Pérez-Flores. 2021. Dog predation by jaguars in a tourist town on the Mexican Caribbean. Neotropical Biology & Conservation 16. https://doi.org/10.3897/neotropical.16.e68973
Carvalho, W. D., L. M. Rosalino, M. S. M. Godoy, M. F. Giorgete, C. H. Adania, and C. E. L. Esbérard. 2019. Temporal activity of rural free-ranging dogs: implications for the predator and prey species in the Brazilian Atlantic Forest. NeoBiota 45:55. https://doi.org/10.3897/neobiota.45.30645
Castro, G. et al. 2022. Drivers of occupancy patterns for the red fox, Vulpes vulpes, in Mediterranean Eucalyptus plantations. Forest Ecology and Management 519:120293. https://doi.org/10.1016/j.foreco.2022.120293
Craighead, K. A., and M. Yacelga. 2021. Indigenous peoples’ displacement and jaguar survival in a warming planet. Global Sustainability 4. https://doi.org/10.1017/sus.2021.6
Dormann, C. F. et al. 2013. Collinearity: a review of methods to deal with it and a simulation study evaluating their performance. Ecography 36:27–46. https://doi.org/10.1111/j.1600-0587.2012.07348.x
Duarte, H. O. B., V. Boron, W. D. Carvalho, and J. J. de Toledo. 2022. Amazon islands as predator refugia: jaguar density and temporal activity in Maracá-Jipioca. Journal of Mammalogy 103:440–446. https://doi.org/10.1093/jmammal/gyab142
Duarte, H. O. B., D. Norris, and F. Michalski. 2018. Assessment of Attractants for Neotropical Mammals. Tropical Conservation Science 11:1940082918800665. https://doi.org/10.1177/194008291880066
Eriksson, C. E. et al. 2022. Extensive aquatic subsidies lead to territorial breakdown and high density of an apex predator. Wiley Online Library. https://doi.org/10.1002/ecy.3543
Farmer, A., and F. Durbian. 2006. Estimating Shorebird Numbers at Migration Stopover Sites. The Condor 108:792–807. https://doi.org/10.1093/condor/108.4.792
Fernandez, G. B. et al. 2019. Natural landscapes along Brazilian coastline. Pp. 199–218 in The Physical Geography of Brazil. Springer. https://doi.org/10.1007/978-3-030-04333-9_10
Ferreira, L. M., V. Theulen, I. S. Coutinho, C. G. Oliveira, B. T. Gomes, and and L. F. Jr. 2017. Plano de Manejo da Estação Ecológica de Maracá-Jipioca (EEMJ). Pp. 1–314 in. Brasil.
Fiske, I., and R. Chandler. 2011. Unmarked: an R package for fitting hierarchical models of wildlife occurrence and abundance. Journal of statistical software 43:1–23. https://doi.org/10.18637/jss.v043.i10
Foster, R. J., B. J. Harmsen, B. Valdes, C. Pomilla, and C. P. Doncaster. 2010. Food habits of sympatric jaguars and pumas across a gradient of human disturbance. Journal of Zoology 280:309–318. https://doi.org/10.1111/j.1469-7998.2009.00663.x
Friedeberg-Gutiérrez, D. B. et al. 2022. Landscape patterns in the occupancy of jaguars (Panthera onca) and their primary prey species in a disturbed region of the Selva Maya in Mexico. Mammalia. https://doi.org/10.1515/mammalia-2021-0149
González, C. A. L., and B. J. Miller. 2002. Do jaguars (Panthera onca) depend on large prey? Western North American Naturalist:218–222.
Gorelick, N., M. Hancher, M. Dixon, S. Ilyushchenko, D. Thau, and R. Moore. 2017. Google Earth Engine: Planetary-scale geospatial analysis for everyone. Remote sensing of Environment 202:18–27.
Hebblewhite, M., and D. T. Haydon. 2010. Distinguishing technology from biology: a critical review of the use of GPS telemetry data in ecology. Philosophical Transactions of the Royal Society B: Biological Sciences 365:2303–2312. https://doi.org/10.1098/rstb.2010.0087
Hunter, L. 2020. Field guide to carnivores of the world. Bloomsbury Publishing.
Jędrzejewski, W. et al. 2017. Predicting carnivore distribution and extirpation rate based on human impacts and productivity factors; assessment of the state of jaguar (Panthera onca) in Venezuela. Biological Conservation 206:132–142. https://doi.org/10.1016/j.biocon.2016.09.027
Jędrzejewski, W. et al. 2018. Estimating large carnivore populations at global scale based on spatial predictions of density and distribution–Application to the jaguar (Panthera onca). PloS one 13:e0194719. https://doi.org/10.1371/journal.pone.0194719
Karanth, K. U., A. Srivathsa, D. Vasudev, M. Puri, R. Parameshwaran, and N. S. Kumar. 2017. Spatio-temporal interactions facilitate large carnivore sympatry across a resource gradient. Proceedings of the Royal Society B: Biological Sciences 284:20161860. https://doi.org/10.1098/rspb.2016.1860
Kolowski, J. M., J. Oley, and W. J. McShea. 2021. High‐density camera trap grid reveals lack of consistency in detection and capture rates across space and time. Ecosphere 12:e03350. https://doi.org/10.1002/ecs2.3350
Lanctot, R. B., A. Hartman, L. W. Oring, and R. I. Guy Morrison. 2008. Response to Farmer (2008): Limitations of Statistically Derived Population Estimates, and Suggestions for Deriving National Population Estimates for Shorebirds. The Auk 125:983–985. https://doi.org/10.1525/auk.2008.21008.2
MacArthur, R. H., and E. O. Wilson. 2016. The theory of island biogeography. The Theory of Island Biogeography. Princeton university press.
MacKenzie, D. I., J. D. Nichols, G. B. Lachman, S. Droege, J. Andrew Royle, and C. A. Langtimm. 2002. Estimating site occupancy rates when detection probabilities are less than one. Ecology 83:2248–2255.
MacKenzie, D. I., J. D. Nichols, J. A. Royle, K. H. Pollock, L. L. Bailey, and J. E. Hines. 2017. Occupancy estimation and modeling: inferring patterns and dynamics of species occurrence. Elsevier.
MAPBIOMAS. 2022. No Title. The Project [s. l.]. V. 6. <https://mapbiomas.org/> (30 July 2022).
Mazerolle, M. J., and M. M. J. Mazerolle. 2017. Package ‘AICcmodavg.’ R package 281.
McManus, J., M. P. E. Schurch, S. Goets, L. Faraut, V. Couldridge, and B. Smuts. 2022. Delineating Functional Corridors Linking Leopard Habitat in the Eastern and Western Cape, South Africa. Conservation 2:99–121. https://doi.org/10.3390/conservation2010009
Meredith, M. 2020. wiqid: Quick and Dirty Estimates for Wildlife Populations. R package version 0.3. 0.
Miller, E. H. 1984. Communication in Breeding Shorebirds BT - Shorebirds: Breeding Behavior and Populations. Pp. 169–241 in (J. Burger & B. L. Olla, eds.). Springer US, Boston, MA.
Morato, R. G. et al. 2016. Space use and movement of a neotropical top predator: the endangered jaguar. PloS one 11:e0168176. https://doi.org/10.1371/journal.pone.0168176
Morcatty, T. Q. et al. 2020. Illegal trade in wild cats and its link to Chinese‐led development in Central and South America. Conservation Biology 34:1525–1535. https://doi.org/10.1111/cobi.13498
Niedballa, J., R. Sollmann, A. Courtiol, and A. Wilting. 2016. camtrapR: an R package for efficient camera trap data management. Methods in Ecology and Evolution 7:1457–1462. https://doi.org/10.1111/2041-210X.12600
Nittrouer, C. A., S. A. Kuehl, R. W. Sternberg, A. G. Figueiredo Jr, and L. E. C. Faria. 1995. An introduction to the geological significance of sediment transport and accumulation on the Amazon continental shelf. Marine Geology 125:177–192. https://doi.org/10.1016/0025-3227(95)00075-A
O’Connell, A. F., J. D. Nichols, and K. U. Karanth. 2011. Camera traps in animal ecology: methods and analyses. Springer.
O’Connor, K. M., L. R. Nathan, M. R. Liberati, M. W. Tingley, J. C. Vokoun, and T. A. G. Rittenhouse. 2017. Camera trap arrays improve detection probability of wildlife: Investigating study design considerations using an empirical dataset. PLoS One 12:e0175684. https://doi.org/10.1016/j.scitotenv.2017.12.211
QGIS Development Team, J. 2022. QGIS geographic information system. Open source geospatial foundation project.
Quigley, H., R. Foster, L. Petracca, E. Payan, R. Salom, and B. Harmsen. 2017. Panthera onca (errata version published in 2018). The IUCN Red List of Threatened Species. 2017: e. T15953A123791436.
R Core Team. 2022. R: A language and environment for statistical computing. Vienna, Austria.
Rabelo, R. M., S. Aragón, and J. C. Bicca-Marques. 2019. Prey abundance drives habitat occupancy by jaguars in Amazonian floodplain river islands. Acta Oecologica 97:28–33. https://doi.org/10.1016/j.actao.2019.04.004
Rabinowitz, A., and K. A. Zeller. 2010. A range-wide model of landscape connectivity and conservation for the jaguar, Panthera onca. Biological conservation 143:939–945. https://doi.org/10.1016/j.biocon.2010.01.002
Sanderson, E. W. et al. 2002. Planning to Save a Species: The Jaguar as a Model. Conservation Biology 16:58–72.
Dos Santos, E. S. et al. 2018. The impact of channel capture on estuarine hydro-morphodynamics and water quality in the Amazon delta. Science of the total environment 624:887–899. https://doi.org/10.1016/j.scitotenv.2017.12.211
Santos, V. F. dos, A. D. Short, and A. C. Mendes. 2016. Beaches of the Amazon Coast: Amapá and West Pará. Pp. 67–93 in Brazilian beach systems. Springer.
Srbek-Araujo A.C., A. G. Chiarello. (2007) Armadilhas fotográficas na amostragem de mamíferos: considerações metodológias e comparação de equipamentos. Rev Bras Zool 24:647–656. https://doi.org/10.1590/S0101-81752007000300016
Tavares, J. P. N. 2014. Características da climatologia de Macapá-AP. Caminhos de geografia 15.
Tobler, M. W., S. E. Carrillo-Percastegui, A. Z. Hartley, and G. V. N. Powell. 2013. High jaguar densities and large population sizes in the core habitat of the southwestern Amazon. Biological Conservation 159:375–381. https://doi.org/10.1016/j.biocon.2012.12.012
Villalva, P., and F. Palomares. 2022. A continental approach to jaguar extirpation: A tradeoff between anthropic and intrinsic causes. Journal for Nature Conservation 66:126145. https://doi.org/10.1016/j.jnc.2022.126145

SupplementaryInformation.docx

Big cats like water: occupancy patterns of jaguar in a unique and insular Brazilian Amazon ecosystem

Status:

Version 1

Abstract

Figures

Introduction

Methodology

Results

Discussion

Declarations

References

Supplementary Files

Status:

Version 1