Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Tropical deforestation induces thresholds of reproductive viability and habitat suitability in Earth’s largest eagles

Abstract

Apex predators are threatened globally, and their local extinctions are often driven by failures in sustaining prey acquisition under contexts of severe prey scarcity. The harpy eagle Harpia harpyja is Earth’s largest eagle and the apex aerial predator of Amazonian forests, but no previous study has examined the impact of forest loss on their feeding ecology. We monitored 16 active harpy eagle nests embedded within landscapes that had experienced 0 to 85% of forest loss, and identified 306 captured prey items. Harpy eagles could not switch to open-habitat prey in deforested habitats, and retained a diet based on canopy vertebrates even in deforested landscapes. Feeding rates decreased with forest loss, with three fledged individuals dying of starvation in landscapes that succumbed to 50–70% deforestation. Because landscapes deforested by > 70% supported no nests, and eaglets could not be provisioned to independence within landscapes > 50% forest loss, we established a 50% forest cover threshold for the reproductive viability of harpy eagle pairs. Our scaling-up estimate indicates that 35% of the entire 428,800-km2 Amazonian ‘Arc of Deforestation’ study region cannot support breeding harpy eagle populations. Our results suggest that restoring harpy eagle population viability within highly fragmented forest landscapes critically depends on decisive forest conservation action.

Introduction

Securing an adequate food supply is a biological imperative that drives fundamental elements of a species biology, including mating1, risk assessment2, survivorship3 and geographic range boundaries4. Consequently, feeding ecology studies have become central to population ecology3, and have shown how consumer density is linked to food abundance5 and why some species can persist in human-dominated landscapes6,7,8, while others cannot9,10,11. Apex predators are particularly sensitive to changes in the spectrum and availability of food because of their high metabolic requirements12. They depend on stable and secure prey populations to fuel their relatively high daily survival and breeding requirements13.

Anthropogenic habitat degradation and prey scarcity induced by poaching forces predators to adapt to feeding on alternative prey species to meet their basic metabolic needs. There are several examples of prey switching as a consequence, including jaguars (Panthera onca) that feed mostly on armadillos (Cingulata) in semi-defaunated habitats14, and golden eagles (Aquila chrysaetus) that feed extensively on mesopredators at sites lacking their usual lagomorph prey15. Some natural predators may switch to domestic livestock8,16,17,18, but once exacerbated by habitat loss, these human-wildlife conflicts often drive the landscape-scale extirpation of apex predators9,10. Predators may adjust to these changes and thrive in anthropogenic landscapes dominated by an agricultural matrix as suitable prey remains abundant, and humans show tolerance19,20,21. Consequently, the threshold of minimum food availability—which can be used to predict the population viability and persistence of apex predators and their coexistence with humans—has been extensively investigated in degraded landscapes22,23,24.

The harpy eagle (Harpia harpyja; Fig. 1) is an apex predator that feeds primarily on forest canopy vertebrates25,26. As one of the Earth’s largest eagles (males and females averaging 5.9 and 7.3 kg, respectively), they have stringent ecological requirements27 that include ~ 800 g of food per day for adults28 and emergent nest trees taller than 40–45 m29. Harpy eagles have a 30–36-month-long breeding cycle, lay two eggs but fledge only a single eaglet30,31,32,33. Harpy eagles are typically very long-lived; a 54 years-old wild-caught adult individual was still alive with the publication of the latest studbook34. They usually show strict nest site fidelity, breeding in the same specific T-shaped emergent nest tree for several consecutive decades35,36. These trees are typically of commercial interest to the logging industry29,37. Studies of their breeding biology are still considered a high research priority33. Their global distribution has contracted by 41% since the nineteenth century, and currently, 93% of their distribution range is within Amazonian Forests, their last stronghold10. Leading causes explaining their rapid decline are habitat loss10, and shooting by settlers and, to a lesser extent, reprisal for killing livestock38,39.

Figure 1
figure1

Harpy eagle (Harpia harpyja) arriving at a nest with a brown capuchin monkey (photo: Jiang Chunsheng).

Harpy eagles exert strong top-down control on their prey populations28,40. The literature widely describes sloths as the main prey taxon, followed by mid-sized primates26. Large reptiles (notably green iguanas, Iguana iguana) and large birds (such as guans and curassows; Cracidae) are less important prey species for harpy eagles26. Terrestrial ungulates such as brocket-deer (Mazama spp.) and peccaries (Tayassuidae) are even less frequent in their diet25. Studying how landscape degradation affects the feeding ecology of harpy eagles can shed light on how they adjust their diet with habitat loss. While dietary profiles are the main aspect studied in harpy eagle ecology, with over 1000 prey records available in the literature25, there is no information on the impacts of landscape-scale primary habitat conversion on the feeding ecology of harpy eagles. Yet ecological limitations such as habitat loss and prey scarcity are key to understanding their extirpation thresholds in human-modified landscapes.

To understand the impacts of forest loss on harpy eagle feeding ecology, our aims here were fourfold. Firstly, we examined the effects of harpy eagle sexual dimorphism on their feeding ecology. We predicted that larger-bodied females would prey on larger-bodied and less diverse prey species than smaller males. Secondly, we tested the effects of landscape degradation on predation rates of non-forest and disturbance-tolerant prey species. We predicted that pairs nesting in highly deforested sites would prey on these species, and sustain a more diverse diet. Thirdly, we investigated the effects of landscape degradation on numerical and biomass rates of prey delivery to eaglets. We predicted that eaglets would face lower prey delivery rates at longer intervals in degraded landscapes, resulting in lower prey biomass acquisition. Lastly, we scaled up our estimates of the deforestation tolerance threshold of harpy eagles and associated prey scarcity to a much larger human-modified region of the Amazon.

By exploring the above, we were able to estimate the consequences of extensive tropical deforestation on the persistence of breeding pairs of harpy eagles. Finding and capturing sufficient prey is the key issue for the persistence of such an apex predator. Therefore, their population viability in human-modified landscapes depends on their tolerance threshold to food stress, which is presently unknown.

Results

Prey profiles

We recognised a total of 306 prey items from harpy eagle nests, 253 of which from camera-traps and 53 from bone material. From a total of 279 prey deliveries recorded by camera-traps, 253 (89.7%) were identified to the level of species or genus across 16 different harpy eagle nests. We complemented camera-trapping data with analysis of skeletal detritus collected both within and underneath nests. We found bones of 76 different individual prey, of which 53 (69.7%) could be identified. We recorded 92 and 75 prey deliveries to nests by female and male eagles, respectively (♀♀ 34.5 vs. ♂♂ 27.2%). An additional 74 (26.9%) prey items observed on nests could not be unambiguously linked to delivery by an adult eagle. These additional prey of unknown origin either resulted from the eaglet hunting its own prey, or camera traps failing to record the adult’s arrival. For the remaining 34 prey deliveries (12.3%), we could not identify the sex of the adult because its talons were not visible in the camera trap images. All prey deliveries, combined with bone identification analysis, yielded 306 prey items representing 37 vertebrate species (Table 1). These were dominated by arboreal mammals, with three top-ranking species representing 49.7% of all deliveries and 50.0% of the overall prey biomass: two-toed sloths (Choloepus didactylus, 23.9%), brown capuchin monkeys (Sapajus apella, 18.3%) and grey woolly monkeys (Lagothrix cana, 7.5%).

Table 1 Species composition of harpy eagle prey items recorded in this study from both skeletal material and camera-trapping data.

Females and males preyed on proportionally similar arboreal prey (♀♀ 75.8 vs. ♂♂ 70.6%), whose difference is not statistically significant (null model, P = 0.756). The same applied to both scansorial (♀♀ 17.5 vs. ♂♂ 16.0%, null model, P = 0.611) and terrestrial prey (♀♀ 4.4 vs. ♂♂ 6.6%, null model P = 0.283). Females preyed primarily on adult prey, while the smaller males preferred juveniles and subadults (Fig. 2), which led to a significant difference in prey body mass between the sexes. Male prey averaged 0.74 ± 0.69 kg (geometric mean ± SD), while female prey was 57.3% heavier, averaging 1.29 ± 0.88 kg (null model, P < 0.001, Fig. 3). The largest individual prey brought to a nest by any female was a ~ 4.7-kg, lower body section of an adult spider monkey, while for any male, this was a ~ 3.2 kg lower body quarters of a howler monkey. One adult whose sex could not be identified simultaneously carried into a nest the lower body of a common opossum (Didelphis spp.) and the lower body of a juvenile spider monkey (totalling ~ 3.6 kg). Levin’s niche width analysis yielded similar values for the sexes: 7.20 for females and 7.42 for males (n = 84 ♀♀, 69 ♂♂). Following bootstrapping based on a null model, however, the niche width fell to 2.40 for females and 4.62 for males, which was significantly different (P < 0.001). Photographic problems induced by mismatched fields of view or malfunctioning camera traps rendered 29% of all cameras inoperative (see Supplementary Information Table S1).

Figure 2
figure2

Prey age classes consumed by either male and female harpy eagles. The 10 most important prey species (representing > 80% of the total) are ordered top to bottom by adult body mass. Prey age class (stacked horizontal bars) are indicated according to the legend. Male harpy eagles are less likely to pursue adult prey, particularly in large-bodied prey species.

Figure 3
figure3

Prey body mass brought into nests by either male (n = 75) or female (n = 92) harpy eagles. Male prey averaged 0.74 kg, while female prey was 57.3% heavier, averaging 1.29 kg (hatched lines). The largest individual prey items brought into a nest were a ~ 4.7 kg lower body of a spider monkey killed by a female, and a ~ 3.2 kg lower body of a howler monkey killed by a male.

Nest density and buffer definition

Using the maximum packed nest density method, the average distance to the nearest neighbouring harpy eagle nest within a nest cluster was 2.99 km (Cluster A, n = 6, SD = 1.52) and 5.90 km (Cluster B, n = 5, SD = 0.89). A total of 124 km2 of the 195 km2 previously forested area around Cluster A still retained forest cover, while forest cover spanned 254 km2 of the 633 km2 around Cluster B. These data translated into a nest density of 1.97–4.84 nests/100 km2 of forest habitat, or 0.79–3.07 nests/100 km2 if we also include deforested areas (Fig. 4). We thus used both the 3-km and 6-km radii as buffers for landscape analyses. A rough approximation between these (4 km, or 50 km2) was used as the breeding territory size (consistent with the literature41), and therefore used as the hexagonal cell size for the regional-scale analysis of habitat loss.

Figure 4
figure4

Examples of the spatial distribution of active breeding nests. Nest density was calculated by the Maximum Packed Nest Density (MPND) method. Distances to the nearest neighbouring nest were 2.99 km to 5.90 km depending on the nest cluster identity, which we used to calculate the 3-km and 6-km landscape buffers. The two southernmost nests in cluster A are the ones excluded for sitting in riparian forests (QGIS V.3.16; https://www.qgis.org).

Prey profile responses to deforestation

Contrary to our expectations, the level of deforestation did not significantly alter the prey species composition for harpy eagles (Fig. 5), nor did distance to cattle pastures (range: 0–6300 m, X̅ = 1159 ± 1703 m, n = 14 nests) or amount of forest cover within a 3-km radius (range: 14–100%, X̅ = 69 ± 28%; NMDS, 999 permutations, stress = 0.1335, P = 0.332 for forest loss and 0.079 for distance to pasture, n = 14 nests). Eagles continued to capture sloths and primates as their main prey, even in nests sited in severely deforested areas. The degree of forest loss did not affect Levin’s niche breadth index nor prey species richness (GLM, residual deviance: 152.59 on 10 DF, P = 0.23, n = 12 nests), showing that harpy eagles relied primarily on large arboreal prey and barely switched to open-country terrestrial vertebrates even in highly-deforested landscapes (Fig. 6).

Figure 5
figure5

Species composition of prey items delivered to active harpy eagle nests. Main prey species are ordered in terms of overall abundance. Each column represents a nest, ordered by the landscape context from the lowest to the highest amount of forest loss within a 3-km buffer area around each nest site. The harpy eagle prey base consists of arboreal forest vertebrate species even when nests are situated in highly deforested landscapes.

Figure 6
figure6

Effects of forest loss (3-km buffer) on the predation ecology of harpy eagles. Eagles nesting in highly deforested landscapes exhibited levels of prey diversity and dietary niche breadth that were similar to those in more relatively intact landscapes, with sloths and primates as the main prey.

Effects of deforestation on prey delivery and biomass

Data from 16 nest-phases (12 nests) and 189 prey delivery events showed that habitat loss induced significant food stress on nesting harpy eagles (Fig. 7). For these analyses, we excluded the two nests that were located in otherwise deforested riparian forest remnants, which are clear outliers because they experienced abnormally high prey biomass delivery rates, which was equivalent to breeding territories in entirely forested landscapes. Prey delivery intervals were on average 4.20 ± 1.97 days (n = 16 nest-phases). Prey biomass delivery rates averaged 0.37 ± 0.27 kg/day (range: 0.10–0.65 kg/day, n = 16 nest-phases). Details on prey body parts brought to the nests are shown in Table 2. The nesting phase had a negative effect on prey delivery rates, with longer time intervals recorded for older eaglets. Forest habitat loss resulted in an increase in the number of days between consecutive prey deliveries considering both the 3-km (GLMM, F2,11 = 102.06, P < 0.001 for forest cover; F2,11 = 795.01, P < 0.001 for nesting phase, n = 16 nest-phases) and the 6-km buffers (GLMM, F2,11 = 35.74, P < 0.001 for forest cover; F2,11 = 780.82, P < 0.001 for nesting phase, n = 16 nest-phases). Habitat loss resulted in lower delivery rates in terms of prey biomass, which was significant for both the 3-km (GLMM, F2,11 = 34.80, P < 0.001, n = 16 nest-phases) and 6-km buffers (GLMM, F2,11 = 27.29, P < 0.001, n = 16 nest-phases). For instance, prey delivery to nests within entirely forested landscape contexts was on average 0.69 kg/day, but this declined to only 0.11 kg/day for nests in areas deforested by 50% or more. There were no effects of the nesting phase (nestling, fledged and dependent juvenile) on the prey biomass delivered.

Figure 7
figure7

Effects of forest habitat loss on the feeding ecology of harpy eagles, as depicted by 16 nest-phases. Habitat loss reduced the frequency of prey delivery (right panels) as well as the prey biomass delivered to the nest (left panels). These slopes were significant in all cases.

Table 2 Prey body part brought into the nest for the 10 most common prey species.

Food stress threshold

Considering that we found no active nests in landscapes that had succumbed to > 70% forest loss (n = 34, excluding nests in riparian forest remnants), we determined an approximate threshold above which harpy eagles can no longer tolerate landscape-scale deforestation. Yet harpy eagle pairs nesting within landscapes containing 50–70% forest cover were unable to feed fledged eaglets up to the stage of dependent juveniles. We witnessed three fledged eaglets dying of starvation at three nests surrounded by 63.3%, 54.3% and 53.0% of deforestation at 3-km buffer. We, therefore, consider levels of forest habitat loss higher than 50% as prohibitive for successful nestling rearing in harpy eagles. Under this assumption—and considering the 50-km2 hexagons—we estimate that forest habitat loss in the southern Amazon has extirpated harpy eagle populations from 35% of their original distribution throughout northern Mato Grosso (Fig. 8). We further estimate that this represents a habitat-induced metapopulation decline of ~ 3256 breeding pairs of harpy eagles. This is particularly concerning given that 28.7% of this entire region has been deforested in the last 35 years, with only 64.9% of forest cover currently remaining.

Figure 8
figure8

Comparison of cells in northern Mato Grosso containing less than 50% of forest (purple cells indicate unsuitable habitat) along the time series examined in this study, from 1985 (left) to 2019 (right) in a region that was once entirely covered by Amazonian forest. No harpy eagle nests has ever been found in landscapes with less than 30% of forest cover, and pairs attending nests in areas surrounded by 30–50% forest cover where unable to feed nestlings/fledglings into the dependent juvenile phase (see “Results” section). We therefore consider 50% to be the harpy eagle habitat loss threshold in terms of reproductive viability, at least in terms of completing a nesting cycle. We further estimate that 35% of northern Mato Grosso can no longer support reproductively viable populations of harpy eagles (QGIS V.3.16; https://www.qgis.org).

Discussion

Our data and analyses showed that forest loss is associated with severe reductions in prey delivery rates and prey biomass by adult harpy eagles to their eaglets; nestlings reared in highly fragmented areas received less food with longer intervals between consecutive feeding bouts. On the other hand, prey species composition changed relatively little in areas that had succumbed to varying amounts of deforestation. This provides clear evidence that harpy eagles depend heavily on forest canopy vertebrate prey. In fact, they exhibited limited hunting ability in non-forest areas, including cattle pastures, targeting few terrestrial vertebrates. These traits render them unable to both use any nest surrounded by > 70% forest loss, or feed young until they reach the dependent juvenile phase in nests located within areas deforested by 50–70%. These findings provide the basis for understanding the persistence thresholds of reproductively viable harpy eagle populations within fragmented forest landscapes.

The dietary profile of harpy eagles in northern Mato Grosso shows that sloths comprised only a quarter of all prey items, which is much lower than studies elsewhere in which sloths account for at least two-thirds of all prey26. Importantly, three-toed sloths (Bradypus spp.), which are normally the most important components of harpy eagle diets26,42, were largely absent in our study region. The absence of three-toed sloths in this region of southern Amazonia is confirmed by several reports of arboreal wildlife rescue operations in hydroelectric dams43. Differences between the age and size of prey taken by male and female harpy eagles were similar to those in previous studies. However, in terms of prey habits (terrestrial or scansorial), the differences were smaller than typically observed in unpaired harpy eagles44. Another surprising pattern emerging from our study area was the conspicuous absence of small livestock in harpy eagle diets. Other apex predators in the same landscape, such as jaguars, frequently prey on bovine cattle18. We conducted prior research in the region to understand the prevalence of retaliatory killings of harpy eagles. While the reported prevalence of livestock predation was relatively low (< 20% of 184 interviewees38), we expected eagles to kill small livestock, especially those nesting near homesteads. Perhaps livestock predation is primarily associated with floaters and therefore was undetected in nests where breeding pairs did virtually all the hunting.

In particular, we found that harpy eagles took a large number of relatively large-bodied prehensile-tailed (Atelinae) primates (both spider monkeys Ateles chamek and woolly monkeys Lagothrix cana). It has been suggested that, as an advantage of their large size, these relatively large-bodied neotropical primates are virtually exempt from aerial predation45. The general absence of previous records of harpy eagles preying on large Atelinae primates has led primatologists to consider these predation events as extremely rare. We suspect that this could be partly explained by the fact that our study area does not experience human hunting of Atelinae primates, and consequently boasts high population densities of large-bodied monkeys43,46. In northern Mato Grosso, hunting of canopy vertebrates is mostly limited to gamebirds such as guans and curassows47. This is largely attributed to cultural taboos against eating monkeys and sloths48,49,50, given that hunters in this area are typically descendants of recent migrant settlers from southern Brazil51. Grey woolly monkeys were frequently represented in harpy diets, particularly considering that they only occur in forests west of the Juruena River52, or roughly half of our study area. However, high Ateles and Lagothrix population abundance is a rare phenomenon in most parts of Amazonia52,53,54, where fragmentation and hunting often have a synergistic effect on prey density. Viable breeding populations of harpy eagles are indeed unlikely to persist in areas that are both overhunted and largely deforested, given the “double-whammy” effects of these combined stressors53. Our conservative estimates of landscape-scale deforestation thresholds that can be tolerated by nesting harpy eagles can therefore be considered as optimistic in most of the Amazon.

It remains unclear from our study why and how harpy eagles in other highly fragmented forest landscapes can rely so heavily on terrestrial prey such as armadillos (A. Blanco, C. Tuyama, pers. comm.). At one much more southerly harpy nest that we excluded from this study, the prey profile of adults provisioning their nest consisted almost entirely of armadillos. The two eagles that fledged on the southern outlier nest were proficient at processing armadillo carcasses. In contrast, the few armadillos recorded in this study were delivered belly-down, and the fledglings failed to turn the armadillos, which subsequently went uneaten. Prey handling skills and hunting profiles may therefore be culturally inherited and dependent on previous experience and prey availability. Another possibility is that the local abundance of armadillo prey, or the species of armadillos available, causes those differences, since naked-tailed armadillos (Cabassous spp.) are the vast majority of individual killed in the aforementioned nest. Predation by harpy eagles on other non-forest, open-country prey species was at best infrequent, as evidenced by the very few records of opossums, macaws, mesocarnivores and other non-forest generalist prey in their diet; harpy eagles were, therefore, unable to switch to non-forest prey species in our study area.

The Levin’s index of trophic niche breadth was much greater than the previous estimates42. We believe this higher trophic breadth index is associated with nest monitoring using camera-traps. Specifically, our camera-traps added 14 new species to the previous total of 102 known prey species taken by harpy eagles anywhere in their range25. Many prey species would have gone undetected had we relied exclusively on bones to identify prey species (Table 1), which would also induce a methodologically biased level of prey importance for sloths. Eagles carry whole small-bodied prey under specific orientation that reduces aerodynamic drag (Fig. 1). For larger prey, butchering carried out by the eagles to reduce flight drag leaves few easily identifiable body parts for most prey, as shown in Table 2. Further evidence of the importance of camera-trapping comes from our prey species composition. Had this been based exclusively on skeletal remains recovered within and underneath nests, then this would have shown only “the usual suspects” of harpy eagle prey species, with the notable exception of Atelinae primates. Therefore, we highly recommend the use of camera-traps in further feeding ecology studies of rainforest raptors. Our 89.7% prey identification rate was also higher than the 47.2% obtained using surveillance cameras54. However, given the several camera-trapping problems we encountered, further safeguards are necessary to reduce camera-trap failure to a bare minimum (see Supplementary Information Table S1).

The harpy eagle prey base consisted of arboreal forest vertebrate species even when nests were situated in highly deforested landscapes. As central-place foragers, breeding harpy eagles incur much greater travel costs to forage in faraway forest fragments in heavily deforested landscapes. This is probably the critical inflection point that precludes nesting in landscapes that had been deforested by > 70%, as well as failure in raising chicks in landscapes experiencing 50–70% of forest loss.

Harpy eagle prey delivery rates in terms of frequency and biomass clearly declined with forest habitat loss. On three occasions involving recently-fledged eaglets, delivery rates were so low that the fledglings died of starvation. One of those nests was monitored by camera-traps, where prey delivery intervals were consistently longer than 15 days, whereas the typical interval required to sustain recently-fledged eaglets is ~ 2.5 days. In another case in which the fledgling survived, the parents were clearly food-stressed and attempted to fall back on a diet of forest birds. That pair was responsible for all the predation events reported here on blue-and-gold macaws (Ara ararauna) and crested curassows (Crax fasciolata). Their dependent juvenile male eagle, however, quickly learned how to hunt black vultures (Coragyps atratus) and accounted for 9 of our 10 records of harpy predation on vultures. Hunting by recently-fledged harpy eaglets is not unheard of55 and is especially common in fledged males56,57. We emphasise that a harpy eagle requires ~ 800 g of prey each day58, an unfeasible target if they are forced to rely on a mixed diet of forest birds. In all other nests, adults continued to prey on sloths and primates. The impacts of sustained depletion of adjacent prey populations by a strict central-place predator hunter within ever smaller forest remnants deserves further study, as harpy eagles clearly exert strong top-down control on some prey species28,40.

By establishing the amount of deforestation that harpy eagles can tolerate to meet their basic ecological requirements, we can estimate the harpy eagle population size lost to cattle pastures and other non-forest land uses in the Amazonian portion of Mato Grosso. Currently, 64.9% of northern Mato Grosso still supports forest cover, which may sound encouraging in retaining harpy eagles and other forest vertebrates. Much of the remaining forest habitat, however, lies within indigenous territories59. The vast majority of native Amazonians kill any number of harpy eagles whenever possible to harvest the primary wing and tail feathers for headdresses and fletching arrows60,61,62, trade headdresses commercially (a practice that is illegal for tribal peoples Brazil63) and capture harpy eagle nestlings to keep as caged pets in Indian villages60 or sell them on the black market (EBPM, pers. obs.). Proactive wildlife co-management64,65 with indigenous peoples should be a priority, including a no-take policy for highly vulnerable species to prevent the extirpation of harpy eagles and other low-fecundity species from Indigenous Lands66. We support indigenous cultural practices and trust that a better wildlife policy from government bureaus and NGOs working with indigenous people will stimulate sustainable use while reinforcing the legal framework that prevents these issues.

We observed that for two harpy eagle nests sited in riparian forest remnants, prey delivery rates were high despite deforestation levels exceeding 70%. These observations are encouraging but at odds with some of our predictions since some hexagonal-cells in our maps may represent riparian corridors that remain well connected to larger forest fragments. Riparian forest remnants in northern Mato Grosso are usually too narrow, poorly connected or highly degraded to support high bird and mammal diversity67. They should also be linked to large fragments, or continuous forests if they are to successfully support harpy eagle reproduction and juvenile dispersal.

Over half of all riparian forest corridors in the study region are degraded to some extent67,68,69, which inhibits movements of arboreal mammals such as woolly monkeys, spider monkeys and sloths. Furthermore, selective logging is widespread in those riparian forests68 and removes the same set of large canopy trees that harpies require for nesting29,68. Additionally, the principal nest tree species used by harpy eagles29—the Brazil nut tree Bertholletia excelsa—rarely occurs within riparian sites, and is largely restricted to higher terrain70. One strategy that can generate resources to fund activities for greater forest connectivity, engage local communities towards conservation through economic incentives and encourage governments to create and manage corridors is ecotourism71,72. As nature conservation is mostly an unprofitable activity73, tourism is one of the few lucrative activities that can generate financing for conservation74,75,76.

Our results contribute to a greater understanding of how harpy eagle feeding ecology is impacted by anthropogenic land-use change, with prey availability frequently as the main constraint for the persistence of any apex predator. Determining the deforestation thresholds that a large forest canopy raptor can tolerate is of prime interest for conservation actions such as adding value to forest habitat through ecotourism and carrying out eagle reintroductions or translocations, which can boost populations of this species in their forest environments. Finally, it is likely that nesting harpy eagles in hyper-fragmented Neotropical forest landscapes, such as the Atlantic Forest and the rapidly expanding Amazonian Arc of Deforestation, are become increasingly isolated in relatively small fragments. These stranded eagles will likely depend on decisive “hands-on” population management interventions, which could include translocation of juveniles and food supplementation to eaglets, if they are to persist.

Methods

Study area

Amazonian Forests have been degraded, particularly along a large deforestation frontier known as the Arc of Deforestation51,77. This study was conducted in the southern portion of the Arc of Deforestation in the northern state of Mato Grosso, Brazil (Fig. 9). The main agricultural land uses in this region are cattle ranching and soybean farming78,79,80. Brazilian forestry law requires that a minimum proportion of any private landholding (varying per ecoregion, but 80% in Amazonia), as well as riparian forests along rivers and streams, should be spared from deforestation81. Human occupation thus generates a hyper-fragmented landscape mosaic with varying levels of habitat loss and structural connectivity68,82. Koeppen83 classifies the region’s climate as “tropical wet climate” or Amazonian (tropical monsoon) climate. Annual rainfall averages 2350 mm, and annual ambient temperature averages 24.5 °C84. The wet season occurs from October through to March, amounting to ~ 80% of the annual precipitation, whereas the dry season spans from April through to September. Arboreal folivores such as howler monkeys (Alouatta spp.) and green iguanas are locally rare in forest areas43,46,85, while the three-toed sloth (Bradypus spp.) is largely absent43. These prey species are important elsewhere, as reported by most previous harpy eagle dietary studies42. This portion of the southern Amazonia was originally inhabited by several indigenous groups60, but they have been largely displaced since the 1970s by southern Brazilian settlers who joined a government-sponsored transmigration program51. Our study region is currently dominated by cattle-ranches, with the largest forest areas remaining within National Parks, State Parks and Indigenous Lands, and most smaller forest remnants set aside as forest reserves within private landholdings81.

Figure 9
figure9

The 903,357-km2 state of Mato Grosso in Brazil, showing the location of harpy eagle nests monitored in this study (yellow circles). Colour-coded areas represent the Amazonian biome portion of the state. Primary forest and deforested areas (mostly cattle pastures) are shown in green and purple, respectively. Blank areas indicate non-forest vegetation. Mapping data are based on 2019 Landsat imagery geoprocessed as part of the MapBiomas project (mapbiomas.org collection 5; year 2019; QGIS V.3.16; https://www.qgis.org).

Nest finding

We openly offered a reward amounting to ~ USD100 or BRL500 (about 50% of the Brazilian minimum monthly wage) for geographical positional information for any active harpy eagle nest reported to us. The reward was widely-publicised in posters and pamphlets that were disseminated among key groups of rural workers, indigenous people, timber industry personnel and especially Brazil-nut collectors who spend the most time in the forest. This reward payment allowed us to ‘discover’ harpy eagle nest locations unconstrained by biases related to forest location, as Brazil-nut collectors and indigenous people range widely within forest areas, while other informants were mostly urban dwellers. Once we identified a nest location, we proceeded to reach contractual agreements with the landowner. These contracts included formal agreements to ensure physical access to scientific research and low-disturbance ecotourism86.

Harpy eagle tourism originated as a cooperative venture between us and a private tour company that install near harpy eagle nests observation towers. Our initiative required that the company pay ~ USD20 (BRL100) per tourist per day to the landowner. In exchange, each landowner signed a legal contract stipulating several measures that prevent forest degradation86. Local inhabitants earned money from the project by transporting and building the towers, as well as trail-cleaning, driving, cooking for tourists and staff, and other associated logistical services86.

Climbing and camera-trapping protocols

We climbed harpy eagle nests using published, raptor-specific tree-climbing protocols87,88. We first determined the eaglet age class using binoculars. We did not climb trees with nests containing either eggs or nestlings younger than 15 days-old, as young eaglets depend on adults for thermoregulation89. We planned our climbing approach to avoid subjecting older nestlings to any rain or excessive heat. A team member ascended the nest by rope to install camera-traps (several models of Bushnell camera-traps). Camera-traps were installed 0.5–2 m from nests and fixed using 15-cm-long nails and malleable, 1.65-mm-thick wire. Camera-traps (containing ≥ 16 Gb memory cards and lithium batteries) were programmed to take one photograph every 10 min for 24 h a day. We used 2 or 3 camera-traps per nest to allow multiple angles to view the prey items delivered by adult eagles and minimise data loss from camera-trap failures. Two static camera-traps per nest allowed us to sex the adult eagles by comparing talon morphology, as female harpy eagles have much larger feet and thicker talons90. We removed camera-traps after they had photographed a nest for at least 90 days.

Effects of forest loss on the feeding ecology

Addressing complex interactions between landscape ecology and feeding ecology requires a multifaceted approach. The flow chart represented in Fig. 10 shows a step-by-step presentation of our framework, which is also described in detail below.

Figure 10
figure10

Clockwise flow chart summarizing the steps taken to estimate the effects of habitat loss on harpy eagle population viability. Inter-nest distances are used to define nest-centered landscape buffer sizes. Buffer area is used to obtain data on forest habitat loss. Camera traps provide evidence of tolerable food stress levels. Combining information on tolerable habitat loss and nesting territory size, we can scale up the habitat viability to much larger degraded landscapes (Author: Paula Viana, Corel Draw X8; https://www.coreldraw.com/br).

To define the landscape buffers around each nest, we estimated the nest density using the maximum packaged nest density method (MNPD). The MNPD is based on demarcating a circle around each nest, for which the radius is defined as half the mean distance to the nearest neighbouring nest in a cluster of nests91. A constant is then added to fill the interstitial spaces between circles using the equation A = π × r2 × 1.158. We used two nest clusters, with six and five nests each, respectively, in two study subregions in which we had comprehensively surveyed the forest (using the reward system) over many years86, so we assumed that all harpy eagle nests in those two areas had been detected.

We estimated the percentage of forest area lost within each of those buffers using land cover data available from the MapBiomas Project (collection 5, year 2019; mapbiomas.org). MapBiomas-classified rasters have a georeferenced, 30-m pixel resolution and a general classification accuracy of 97.3% for the Amazon biome. We used Google Earth’s Engine cloud computing platform (code.earthengine.google.com) to access the MapBiomas 2019 collection and extract all pixels corresponding to natural forest formation (ID code 3). We further assessed and added data on recent deforestation that was revealed in post-2019 Google Earth images.

Within the buffers described above, we analysed the effects of habitat loss on prey species composition. In addition to the amount of forest cover, we added distance in meters to the nearest pasture areas as a covariate because even pairs of harpy eagles nesting within forest territories could access open-habitat prey in nearby pastures. We identified prey species from photographs using reference information in the zoological literature92,93, and reference collections at the Federal University of Mato Grosso (UFMT), Brazil. We also identified prey skeletal material collected both underneath and inside our study nests.

We defined prey biomass delivery rates by estimating the prey biomass delivered to each nest (using prey data exclusively from camera-traps). Harpy eagles frequently prey on subadult individuals and reduce the weight of large-bodied carcasses to reduce drag during flight42. Eagles reduce flight loads (i.e. prey mass) by consuming parts of the carcass. We, therefore, used the following estimated reductions: subadult prey were estimated to represent 66% of adult body mass; for very young prey (because ungulates taken by the eagles are almost exclusively newborns26), we estimated that they represented 20% of adult body mass; sloths received a further 33% reduction, because of the large amount of gut foliage content, which represents ~ 30% of the body mass of living sloths94. After killing sloths, eagles discard the digestive tract. For dismembered prey, we used the following approximate body mass reductions: 10% (head or viscera), 20% (reduced per arm/leg missing or added per single-member delivered), 50% (lower or upper body missing), and 90% (prehensile tails of Atelinae primates from the genera Ateles, Lagothrix and Alouatta; tails of porcupines, Rodentia: Erethizontidae; tails of lesser anteaters, Tamandua tetradactyla) of the total carcass mass7. Body masses of these prey species were obtained from the literature92,95. When calculating the overall body mass reduction for analyses related to the feeding ecology (rather than biomass delivered to nests), we only reduced whole body mass to account for subadults, rather than including body part reductions. This allowed us to account for the mass of prey parts consumed by adults before carcass delivery at the nest.

We used the forest habitat amount remaining around each nest to determine how much forest loss harpy eagles could tolerate while still exhibiting clear evidence of successful breeding. We defined a hexagonal cell area representing the home range of a harpy eagle pair using the mean distance to the nearest neighbouring nest. Deforestation levels around known active nests surrounded by the highest levels of forest loss—but still successfully raising eaglets—were used to identify hex-cells containing sufficient amounts of forest cover assumed to be suitable for harpy eagles. We defined those below this threshold as unsuitable habitat. We then extended this rationale for the entire 428,800-km2 area of the northern portion of the State of Mato Grosso, and calculated the amount of regional-scale habitat suitability both for the first (1985) and last year (2019) of our land-use time series.

Statistical analyses

We compared prey profiles between the two sexes for several traits such as prey size and Levin’s niche breadth index, using null models96. Levin’s index is defined as Bsta = B − 1/(n − 1), where B is Levin’s index (B = 1/Σpj2), pj is the frequency of occurrence of each prey species (or prey item), and n is the total number of prey species97. Our null model consisted of the following steps: (1) bootstrapping samples of 25 prey records for both males and females; (2) calculating niche breadth for each sex using a niche breadth measure; (3) creating a pairwise difference in niche breadth between males and females; and (4) determining whether the difference in niche width found between breeders and floaters was larger than expected by chance by comparing differences between two randomly labelled bootstrapped samples (n = 25). We carried out 1000 iterations to calculate niche width differences between sexes, and examined to what extent this departed from chance alone. We selected a sample size of 25 prey items, given that previous work showed this sample to be sufficiently large to adequately represent all prey species brought to active nests at frequencies exceeding 5%26. Alpha levels were defined as 5%.

We tested whether the amount of forest loss within buffer areas around each nest affected prey species composition using non-metric multidimensional scaling (NMDS) ordination. We selected the NMDS with the goal of collapsing information from multiple dimensions (in our case, a nest by prey species matrix) into a few axes so that they could be interpreted. To explain the effects of forest loss on harpy eagle feeding ecology, we used the amount of deforestation and the linear distance to the nearest pasture area (in meters). Deforestation likely limits harpy eagle access to canopy prey such as sloths and monkeys, while exotic pastures increase access to ground-dwelling alternative prey such as armadillos.

We also tested the effect of forest loss within nest neighbourhoods on nest-specific prey species richness and Levin’s niche breadth index. We then used the amount of forest loss and distance to nearest pasture as covariates explaining Levin’s niche breadth and prey species richness based on a Generalized Linear Model (GLM). We used a Gaussian error structure and a logit link function98.

To determine the impact of forest loss on the feeding ecology of harpy eagles, we used GLMMs with a Gaussian error to explain prey delivery rates both in terms of individual prey items and prey biomass (defined as response variables). We defined feeding rates as the time interval (in days) between consecutive prey deliveries by adults, for which 0 represents two prey items delivered on the same day. We further added the nesting phase (nestling, fledgling and dependent juvenile) as ordinal covariates since these provision intervals slowly increase from 1.8 to 5.12 days per prey item from the nestling to dependent juvenile phases, respectively99. We, therefore, used each nest phase per nest to be an independent replica after adding each nest as a random factor in the GLMMs. All analyses were performed using the R coding environment, version 4.0.2100.

References

  1. 1.

    McQueen, A. et al. Evolutionary drivers of seasonal plumage colours: colour change by moult correlates with sexual selection, predation risk and seasonality across passerines. Ecol. Lett. 22, 1838–1849 (2019).

    PubMed  Google Scholar 

  2. 2.

    Menezes, J. F., Kotler, B. P. & Dixon, A. K. Risk pump in Gerbillus pyramidum: quality of poor habitats increases with more conspecifics. Ethol. Ecol. Evol. 31, 140–154 (2019).

    Google Scholar 

  3. 3.

    Stephens, D., Brown, J. & Ydenberg, R. Foraging: Behavior and Ecology. (University of Chicago Press, 2007).

  4. 4.

    Schweiger, A., Fünfstück, H.-J. & Beierkuhnlein, C. Availability of optimal-sized prey affects global distribution patterns of the golden eagle Aquila chrysaetos. J. Avian Biol. 46, 81–88 (2015).

    Google Scholar 

  5. 5.

    Carbone, C. & Gittleman, J. L. A common rule for the scaling of carnivore density. Science (80-.) 295, 2273–2276 (2002).

    ADS  CAS  Google Scholar 

  6. 6.

    Athreya, V., Odden, M. & Linnell, J. A cat among the dogs: leopard Panthera pardus diet in a human-dominated landscape in western Maharashtra, India. Oryx https://doi.org/10.1017/s0030605314000106 (2014).

    Article  Google Scholar 

  7. 7.

    Van der Meer, T., McPherson, S. & Downs, C. Temporal changes in prey composition and biomass delivery to African Crowned Eagle nestlings in urban areas of KwaZulu-Natal, South Africa. Ostrich 83, 241–250 (2018).

    Google Scholar 

  8. 8.

    Miranda, E. B. P., Ribeiro-Jr., R. P. & Strüssmann, C. The ecology of human-anaconda conflict: a study using internet videos. Trop. Conserv. Sci. 9, 26–60 (2016).

    Google Scholar 

  9. 9.

    Paviolo, A. et al. A biodiversity hotspot losing its top predator: the challenge of jaguar conservation in the Atlantic Forest of South America. Sci. Rep. 6, 37147 (2016).

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

  10. 10.

    Miranda, E. B. P., Menezes, J. F. S., Farias, C. C., Munn, C. & Peres, C. A. Species distribution modeling reveals strongholds and potential reintroduction areas for the world’s largest eagle. PLoS ONE 14, e0216323 (2019).

    PubMed  PubMed Central  Google Scholar 

  11. 11.

    Marshall, B. M. et al. Hits close to home: repeated persecution of King Cobras (Ophiophagus hannah) in Northeastern Thailand. Trop. Conserv. Sci. 11, 1–14 (2018).

    Google Scholar 

  12. 12.

    Carbone, C., Pettorelli, N. & Stephens, P. A. The bigger they come, the harder they fall: body size and prey abundance influence predator-prey ratios. Biol. Lett. 7, 312–315 (2011).

    PubMed  Google Scholar 

  13. 13.

    Garcia-Heras, M. S., Mougeot, F., Simmons, R. E. & Arroyo, B. Regional and temporal variation in diet and provisioning rates suggest weather limits prey availability for an endangered raptor. Ibis (Lond. 1859) 159, 567–579 (2017).

    Google Scholar 

  14. 14.

    Miranda, E. B., Jácomo, A. T. D. A., Tôrres, N. M., Alves, G. B. & Silveira, L. What are jaguars eating in a half-empty forest? Insights from diet in an overhunted Caatinga reserve. J. Mammal. 99, 724–731 (2018).

    Google Scholar 

  15. 15.

    Ellis, D. H. & Gombobaatar, S. Ecology of the Golden Eagle in Mongolia, part 2: prey. J. Raptor Res. 54, 30–37 (2020).

    Google Scholar 

  16. 16.

    Zuluaga, S. & Echeverry-Galvis, M. Á. Domestic fowl in the diet of the Black-and-chestnut Eagle (Spizaetus isidori) in the Eastern Andes of Colombia: a potential conflict with humans. Ornitol. Neotrop. 27, 113–120 (2016).

    Google Scholar 

  17. 17.

    McPherson, S. C. & Brown, M. Downs CT (2015) Diet of the crowned eagle (Stephanoaetus coronatus) in an urban landscape: potential for human-wildlife conflict?. Urban Ecosyst. https://doi.org/10.1007/s11252-015-0500-6 (2015).

    Article  Google Scholar 

  18. 18.

    Michalski, F., Boulhosa, R. L. P., Faria, A. & Peres, C. A. Human-wildlife conflicts in a fragmented Amazonian forest landscape: determinants of large felid depredation on livestock. Anim. Conserv. 9, 179–188 (2006).

    Google Scholar 

  19. 19.

    Lamichhane, B. R. et al. Rapid recovery of tigers Panthera tigris in Parsa Wildlife Reserve, Nepal. Oryx 52, 16–24 (2018).

    Google Scholar 

  20. 20.

    Tortato, F. R., Izzo, T. J., Hoogesteijn, R. & Peres, C. A. The numbers of the beast: valuation of jaguar (Panthera onca) tourism and cattle depredation in the Brazilian Pantanal. Glob. Ecol. Conserv. 11, 106–114 (2017).

    Google Scholar 

  21. 21.

    Macdonald, C. et al. Conservation potential of apex predator tourism. Biol. Conserv. 215, 132–141 (2017).

    Google Scholar 

  22. 22.

    Karanth, K. U., Kumar, N. S., Nichols, J. D., Link, W. A. & Hines, J. E. Tigers and their prey: predicting carnivore densities from prey abundance. Proc. Natl. Acad. Sci. USA 101, 4854–4858 (2004).

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

  23. 23.

    Nelson, A. A. et al. Native prey distribution and migration mediates wolf (Canis lupus) predation on domestic livestock in the Greater Yellowstone Ecosystem 94(4). Can. J. Zool. 94, 291–299 (2016).

    Google Scholar 

  24. 24.

    Terborgh, J. & Estes, J. Trophic Cascades: Predators, Prey, and the Changing Dynamics of Nature (Island Press, 2013).

  25. 25.

    Miranda, E. B. P. Prey composition of harpy eagles (Harpia harpyja) in Raleighvallen, Suriname. Trop. Conserv. Sci. 11, 1–8 (2018).

    Google Scholar 

  26. 26.

    Miranda, E. B. P. Conservation implications of harpy eagle Harpia harpyja predation patterns. Endanger. Species Res. 29, 69–79 (2015).

    Google Scholar 

  27. 27.

    Vargas González, J. D. J. et al. Breeding habitat suitability index for the harpy eagle in Panama: conservation implications. J. Raptor Res. Press, (2020).

  28. 28.

    Touchton, J., Hsu, Y. & Palleroni, A. Foraging ecology of reintroduced captive-bred subadult harpy eagles (Harpia harpyja) on Barro Colorado Island, Panama. Ornitol. Neotrop. 13, 365–379 (2002).

    Google Scholar 

  29. 29.

    Miranda, E. B. P., Peres, C. A., Marini, M. Â. & Downs, C. T. Harpy Eagle (Harpia harpyja) nest tree selection: logging in Amazonian forests threatens Earth’s largest eagle. Biol. Conserv. 250, 108754 (2020).

    Google Scholar 

  30. 30.

    Muñiz-López, R. et al. Movements of Harpy Eagles Harpia harpyja during their first two years after hatching Movements of Harpy Eagles Harpia harpyja during their first two years after hatching. Bird Study 3657, 509–514 (2016).

    Google Scholar 

  31. 31.

    Muñiz-López, R. Harpy Eagle (Harpia harpyja) mortality in Ecuador. Stud. Neotrop. Fauna Environ. 30, 1–5 (2017).

    Google Scholar 

  32. 32.

    Urios, V., Muñiz-López, R. & Vidal-Mateo, J. Juvenile Dispersal of Harpy Eagles (Harpia harpyja) in Ecuador. J. Raptor Res. 51, 439–445 (2017).

    Google Scholar 

  33. 33.

    Monsalvo, J. A. B., Heming, N. M. & Marini, M. Â. Breeding biology of neotropical accipitriformes: current knowledge and research priorities. Rev. Bras. Ornitol. 26, 151–186 (2018).

    Google Scholar 

  34. 34.

    Hall, C. Harpy Eagle Studbook Harpia harpyja North American Regional. (2011).

  35. 35.

    Alvarez-Cordero, E. Biology and conservation of the harpy eagle in Venezuela and Panama. DSc Thesis. (University of Florida, Florida, USA, 1996).

  36. 36.

    Rettig, N. Breeding behavior of the harpy eagle (<i>Harpia harpyja<i/>). Auk 95, 629–643 (1978).

    Google Scholar 

  37. 37.

    Giudice, R., Piana, R. & Williams, M. Tree architecture as a determinant factor in nest-tree selection by Harpy Eagles. In Neotropical Raptors (eds. Bildstein, K. L., Barber, D. R. & Zimmerman, A.) 14–22 (Hawk Mountain Sanctuary, 2007).

  38. 38.

    Miranda, E. B. P. de, Peres, C. A. & Downs, C. T. Perceptions of livestock predation (or the lack of it) drive the persecution of Earth’s largest eagle. Anim. Conserv. Press (2020).

  39. 39.

    Giraldo-Amaya, M. A. T. E. O., Aguiar-Silva, F. H., Aparício, K. M. & Zuluaga, S. Human persecution on the harpy eagle: a widespread threat?. J. Raptor Res. 55, 1–6 (2020).

    Google Scholar 

  40. 40.

    Terborgh, J. et al. Ecological meltdown in predator-free forest fragments. Science (80-). 294, 1923–1926 (2001).

    ADS  CAS  Google Scholar 

  41. 41.

    Aguiar-Silva, H. Uso e seleção de recursos por harpia em múltiplas escalas espaciais: persistência e vulnerabilidade (INPA, 2016).

  42. 42.

    Aguiar-Silva, F., Sanaiotti, T. & Luz, B. Food habits of the Harpy Eagle, a top predator from the Amazonian rainforest canopy. J. Raptor Res. 48, 24–45 (2014).

    Google Scholar 

  43. 43.

    Silva, D. A. Comunidade de mamíferos de médio e grande porte em fragmentos florestais da amazônia meridional (Unemat - Nova Xavantina, 2016).

  44. 44.

    Miranda, E. B. P., Campbell-Thompson, E., Muela, A. & Vargas, F. H. Sex and breeding status affect prey composition of Harpy Eagles Harpia harpyja. J. Ornithol. 159, 141–150 (2017).

    Google Scholar 

  45. 45.

    Terborgh, J. Five New World Primates: A Study in Comparative Ecology (Princeton University Press, 2014).

  46. 46.

    Oliveira, A. T. M. et al. Primate and ungulate responses to teak agroforestry in a southern Amazonian landscape. Mamm. Biol. 96, 45–52 (2019).

    Google Scholar 

  47. 47.

    Michalski, F. & Peres, C. A. Gamebird responses to anthropogenic forest fragmentation and degradation in a southern Amazonian landscape. PeerJ 5, e3442 (2017).

    PubMed  PubMed Central  Google Scholar 

  48. 48.

    Barbosa, H. Estrutura de comunidades de mamíferos de médio e grande porte em fragmentos florestais da Amazônia Meridional (Unemat - Cáceres, 2012).

  49. 49.

    Michalski, F. & Peres, C. A. Anthropogenic determinants of primate and carnivore local extinctions in a fragmented forest landscape of southern Amazonia. Biol. Conserv. 124, 383–396 (2005).

    Google Scholar 

  50. 50.

    Trinca, C. T. & Ferrari, S. F. Caça em assentamento rural na amazônia matogrossense. Diálogos em ambiente e sociedade no Brasil (2006).

  51. 51.

    Schneider, M. & Peres, C. A. Environmental costs of government-sponsored agrarian settlements in Brazilian Amazonia. PLoS ONE 10, e0134016 (2015).

    PubMed  PubMed Central  Google Scholar 

  52. 52.

    Cavalcante, T. et al. Niche overlap between two sympatric frugivorous Neotropical primates: improving ecological niche models using closely-related taxa. Biodivers. Conserv. 29, 2749–2763 (2020).

    Google Scholar 

  53. 53.

    Peres, C. A. Synergistic effects of subsistence hunting and habitat fragmentation on Amazonian forest vertebrates. Conserv. Biol. 15, 1490–1505 (2001).

    Google Scholar 

  54. 54.

    Bowler, M. et al. Harpy eagles (Harpia harpyja) nesting at Refugio Amazonas, Tambopata, Peru feed on abundant disturbance-tolerant species. Food Webs 24, e00154 (2020).

    Google Scholar 

  55. 55.

    Cavalcante, T., Tuyama, C. A. & Mourthe, I. Insights into the development of a juvenile harpy eagle’s hunting skills. Acta Amaz 49, 114–117 (2019).

    Google Scholar 

  56. 56.

    Campbell-Thompson, E., Vargas, F. H., Watson, R. T., Muela, A. & Cáceres, N. C. Effect of sex and age at release on the independence of hacked harpy eagles. J. Raptor Res. 46, 158–167 (2012).

    Google Scholar 

  57. 57.

    Watson, R. T., McClure, C. J. W., Vargas, F. H. & Jenny, J. P. Trial restoration of the harpy eagle, a large, long-lived, tropical forest raptor panama and belize. J. Raptor Res. 50, 3–22 (2016).

    Google Scholar 

  58. 58.

    Touchton, J. The Harpy Eagle. In The eagle watchers: Observing and conserving raptors around the world (eds. Tingay, R. & Katzner, T.) 264 (Cornell University Press, 2010).

  59. 59.

    Crisostomo, A. C., Alencar, A., Mesquita, I., Silva, I. & Dourado, M. Terras Indígenas Na Amazônia Brasileira: reservas de carbono e barreiras ao desmatamento (2015).

  60. 60.

    Villas Boas, O. & Villas Boas, C. A marcha para o oeste: a epopéia da expedição Roncador-Xingu (Globo, 1994).

  61. 61.

    Tufiño, P. Cunsi Pindo: The Mistress of the Monkeys (Simbioe, 2007).

  62. 62.

    Reina, R. E. & Kensinger, K. M. The Gift of Birds: Featherworking of Native South American Peoples. (University Museum of Archaeology & Anthropology, 1991).

  63. 63.

    Anonymous. Lei de Proteção à Fauna, Lei 5.197, de 03 de janeiro de 1967. (1967).

  64. 64.

    Campos-Silva, J. V. & Peres, C. A. Community-based management induces rapid recovery of a high-value tropical freshwater fishery. Sci. Rep. 6, 34745 (2016).

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

  65. 65.

    Antunes, A. P. et al. Subsistence hunting rights in the Brazilian Amazon. Land Use Policy 84, 1–11 (2019).

    Google Scholar 

  66. 66.

    Aleixo, A. & Galetti, M. The conservation of the avifauna in a lowland Atlantic forest in south-east Brazil. Bird Conserv. Int. 7, 235–261 (1997).

    Google Scholar 

  67. 67.

    Lees, A. C. & Peres, C. A. Conservation value of remnant riparian forest corridors of varying quality for Amazonian birds and mammals. Conserv. Biol. 22, 439–449 (2008).

    PubMed  Google Scholar 

  68. 68.

    Zimbres, B., Machado, R. B. & Peres, C. A. Anthropogenic drivers of headwater and riparian forest loss and degradation in a highly fragmented southern Amazonian landscape. Land Use Policy 72, 354–363 (2018).

    Google Scholar 

  69. 69.

    Michalski, F., Metzger, J. P. & Peres, C. A. Rural property size drives patterns of upland and riparian forest retention in a tropical deforestation frontier. Glob. Environ. Change 20, 705–712 (2010).

    Google Scholar 

  70. 70.

    Mori, S. A. & Prance, G. T. Taxonomy, ecology, and economic botany of the Brazil nut (Bertholletia excelsa Humb. & Bonpl.: Lecythidaceae). Adv. Econ. Bot. 8, 130–150 (1990).

    Google Scholar 

  71. 71.

    Buckley, R. Conservation Tourism (CAB International, 2010).

  72. 72.

    Ribeiro, S. M. C. et al. Can multifunctional livelihoods including recreational ecosystem services (RES) and non timber forest products (NTFP) maintain biodiverse forests in the Brazilian Amazon?. Ecosyst. Serv. 31, 517–526 (2018).

    Google Scholar 

  73. 73.

    Strand, J. et al. Spatially explicit valuation of the Brazilian Amazon Forest’s Ecosystem Services. Nat. Sustain. 1, 657 (2018).

    Google Scholar 

  74. 74.

    Kirkby, C. A. et al. Closing the ecotourism-conservation loop in the Peruvian Amazon. Environ. Conserv. 38, 6–17 (2011).

    Google Scholar 

  75. 75.

    Kirkby, C. A. et al. The market triumph of ecotourism: an economic investigation of the private and social benefits of competing land uses in the Peruvian Amazon. PLoS ONE 5, e13015 (2010).

    ADS  PubMed  PubMed Central  Google Scholar 

  76. 76.

    Vianna, G. M. et al. Shark-diving tourism as a financing mechanism for shark conservation strategies in Malaysia. Mar. Policy 94, 220–226 (2018).

    Google Scholar 

  77. 77.

    Fearnside, P. M. Deforestation in Brazilian Amazonia: history, rates, and consequences. Conserv. Biol. 19, 680–688 (2005).

    Google Scholar 

  78. 78.

    Junior, C. S. & Lima, M. Soy Moratorium in Mato Grosso: deforestation undermines the agreement. Land Use Policy 71, 540–542 (2018).

    Google Scholar 

  79. 79.

    Lima, M. et al. The paradoxical situation of the white-lipped peccary (Tayassu pecari) in the state of Mato Grosso, Brazil. Perspect. Ecol. Conserv. 17, 36–39 (2019).

    Google Scholar 

  80. 80.

    Eri, M. et al. Capitalizing on opportunities provided by pasture sudden death to enhance livestock sustainable management in Brazilian Amazonia. Environ. Dev. 4, 100499 (2020).

    Google Scholar 

  81. 81.

    Anonymous. Novo Código Florestal, Lei 12.651 de 25 de maio de 2012, Dispõe sobre a proteção da vegetação nativa (Subchefia de assuntos jurídicos, 2012).

  82. 82.

    Zimbres, B., Peres, C. A. & Machado, R. B. Terrestrial mammal responses to habitat structure and quality of remnant riparian forests in an Amazonian cattle-ranching landscape. Biol. Conserv. 206, 283–292 (2017).

    Google Scholar 

  83. 83.

    Koeppen, W. Climatologia: con un estudio de los climas de la tierra (1948).

  84. 84.

    Radam-Brasil. Projeto Radam-Brasil: levantamento de recursos naturais (1983).

  85. 85.

    Ayres, J. M. Observações sobre a ecologia e o comportamento dos cuxiús (Chiropotes albinasus e Chiropotes satanas, Cebidae: Primates) (1981).

  86. 86.

    Miranda, E. B. P. de et al. Harpy Eagle nest activity patterns: Potential ecotourism and conservation opportunities in the Amazon Forest. Bird Conserv. Int. (in press) (2021).

  87. 87.

    Rosenfield, R. N., Grier, J. W. & Fyfe, R. W. reducing management and research disturbance. In Raptor Research and Management Techniques (ed. Bird, D. M.) 351–364 (Hancock House Publishers, 2007).

  88. 88.

    Pagel, J. E. & Thorstrom, R. K. Accessing nests. In Raptor Research and Management Techniques (ed. Bird, D. M.) 171–180 (Hancock House Publishers, 2007).

  89. 89.

    Ellis, D. H. & Schimitt, N. J. Behavior of the Golden Eagle: An Illustrated Ethogram. (Hancock House Publishers, 2017).

  90. 90.

    Ferguson-Lees, J. & Christie, D. Raptors of the World (Houghton Mifflin Harcourt, 2001).

  91. 91.

    Brown, D. A test of randomness of nest spacing. Wildfowl 26, 102–103 (1975).

    Google Scholar 

  92. 92.

    Emmons, L. & Feer, F. Neotropical Rainforest Mammals: A Field Guide (University of Chicago Press, 1997).

  93. 93.

    Sick, H. Ornitologia brasileira, uma introdução (Universidade de Brasília, 1984).

  94. 94.

    Goffart, M. Function and Form in the Sloth (Pergamon Press, 1971).

  95. 95.

    Dunning, J. Handbook of Avian Body Masses (CRC, 1993).

    Google Scholar 

  96. 96.

    Gotelli, N. & Aaron, M. A Primer of Ecological Statistics (Sinauer Associates, 2005).

  97. 97.

    Krebs, C. Ecological Methodology (Benjamin/Cummings, 1999).

  98. 98.

    Ashe, E., Noren, D. P. & Williams, R. Animal behaviour and marine protected areas: incorporating behavioural data into the selection of marine protected areas for an endangered killer whale population. Anim. Conserv. 13, 196–203 (2010).

    Google Scholar 

  99. 99.

    Miranda, E. B. P., Peres, C. A. & Downs, C. T. Changes in soil fertility mosaics in the Amazon Forest induced by an apex predator. Press (2020).

  100. 100.

    R Core. R: A Language and Environment for Statistical Computing. (2020).

Download references

Acknowledgements

This research would not be possible without the contributions of Roberto Stofel, Carlos Tuyama, Guilherme R. R. Brito, Guilherme S. T. Garbino, Caio Kenup, Lorena Castilho and Alexander Blanco. CTD is grateful to the University of KwaZulu-Natal (ZA) and the National Research Foundation (ZA, Grant 98404) for funding. We greatly appreciate the financial support of the following donors: SouthWild.com Conservation Travel System, Rainforest Biodiversity Group, Idea Wild, The Mamont Scholars Program of the Explorer’s Club Exploration Fund, Cleveland Metroparks Zoo, and the Rufford Small Grants Foundation (18743-1, 23022-2 and 31091-B). Logistical support was given by the Peugeot-ONF Carbon Sink Reforestation Project, based at the São Nicolau Farm in Cotriguaçu, Mato Grosso, Brazil. This project is a Peugeot initiative to fulfil some of the Kyoto Protocol directions and is run by the ONF-Brasil enterprise. Finally, we thank ICMBio for authorising the project activities under the permit SISBIO 58533.

Author information

Affiliations

Authors

Contributions

E.B.P.M. and C.A.P. formulated research ideas, goals and aims, performed the analyses, management activities to collect, curate and store research data, raised funds and wrote the first draft. V.C.R. produced the maps and G.I.S. analyses, and reviewed the first draft. B.V.M., N.L. and N.H. collected, curated and stored data, and reviewed the first draft. C.A.M. performed fundraising, language review and the critical review of the first draft. T.B.F.S., T.V.F., J.B.P., V.Q.P. identified prey species and performed the review of the first draft. C.A.P., M.Â.M. and C.T.D. provided useful comments on the final draft, its critical review and approval.

Corresponding author

Correspondence to Everton B. P. Miranda.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Publisher's note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Miranda, E.B.P., Peres, C.A., Carvalho-Rocha, V. et al. Tropical deforestation induces thresholds of reproductive viability and habitat suitability in Earth’s largest eagles. Sci Rep 11, 13048 (2021). https://doi.org/10.1038/s41598-021-92372-z

Download citation

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Search

Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing