BREEDING ECOLOGY, NESTING HABITAT AND THREATS TO A BLACK-AND-CHESTNUT EAGLE SPIZAETUS ISIDORI POPULATION IN THE MONTANE FORESTS OF CENTRAL PERU

· The Black-and-chestnut Eagle Spizaetus isidori is an endangered and li(le-known top predator of South American montane forests. To be(er understand the breeding ecology and threats of this eagle, we studied several pairs between 2017 and 2022 in the central Andes of Peru. We recorded 62 adults, one immature, and 36 juveniles in 36 territories. These territories were in mountainous areas (al?tudinal range 690–3,810 m a.s.l.), widely covered by montane forests (43.8–99.7% cover), and secondarily by open land (0.3–56.2% cover), far from urban se(lements (1.3–27.1 km). Nes?ng sites (N = 15) were at medium al?tudes (1,330–2,330 m a.s.l.) in steep hillsides or ravines (15–55°), having no preferen?al cardinal orienta?on, and rela?vely close to permanent water courses (20–800 m), open areas (30–930 m) and sites with human ac?vity (120–2,200 m). Nests (N = 15) were placed at the top of tall (28–40 m) and thick-stemmed (DBH range 0.53–1.52 m) emergent trees of nine genera, with Ficus and Juglans being the most used. Incuba?on and brood-rearing occurred during the dry season (Mar–Nov). The wooded slopes where the eagles nested are being replaced by crops and livestock pastures, causing an es?mated loss of 218.2 km ² of forest cover in our study area (5,056 km ² ) during the last 20 years. We detected human persecu?on in 55.5% of the territories, resul?ng in 26 juveniles, four adults, two immatures, and nine unaged eagles killed. Preven?ng the local ex?nc?ons of these eagles will require long-term popula?on monitoring, improving knowledge of its ecology, mi?ga?on of human-eagle conﬂict, development of educa?on programs, and strengthening of land use inspec?on.


INTRODUCTION
Raptors (birds of prey) play a role in determining ecosystem balance, structuring biological communi?es, promo?ng biodiversity, and serving as environmental barometers (Sergio et al. 2008, Donázar et al. 2016. Despite their importance, human ac?vi?es have caused the ex?nc?on of several species and the popula?on decline of many others on a global scale (Sarasola et al. 2018). Large raptors are especially vulnerable to ex?nc?on because they need vast areas of suitable habitat to meet their ecological requirements (Newton 1979). Apart from scavengers, raptor species most likely to be threatened are larger species that live in forests, have small distribu?onal ranges, and long genera?on ?mes. These species are considered a high priority for research and conserva?on (McClure et al. 2018, Buechley et al. 2019. Thus, accurate knowledge of their breeding ecology and documen?ng their main threats play a central role in effec?ve conserva?on of large forestdwelling raptors (McClure et al. 2018). For most Neotropical species, these parameters are poorly known, mostly because many live in structurally complex habitats in developing countries, which represents a great challenge for their study due to difficulty accessing the remote areas where they live (Thiollay 1989, Buechley et al. 2019).
The Black-and-chestnut Eagle Spizaetus isidori Des Murs 1845, is a large top avian predator (63-74 cm; 2-3.5 kg) endemic to South American montane cloud forests, distributed along wooded slopes of the Andes and nearby mountain ranges (1,000-3,500 m a.s.l.) of Venezuela, Colombia, Ecuador, Peru, Bolivia and northern Argen?na (11°N-28°S; Fjeldså & Krabbe 1990, Thiollay 1994, Ferguson-Less & Chris?e 2001, Rivas-Fuenzalida et al. 2022a. Despite its wide yet narrow con?nental distribu?on, this species is one of the more poorly known Neotropical raptors (Monsalvo et al. 2018). It has been affected by intense deforesta?on throughout much of its range; it is es?mated that current habitat loss in countries such as Colombia could reach up to 60% . Another threat in several countries is human persecu?on in retalia?on to poultry preda?on (Lehmann 1959, Márquez & Delgado 2010, Zuluaga & Echeverry-Galvis 2016, Aráoz et al. 2017, Restrepo-Cardona et al. 2019, Zuluaga et al. 2022. Due to these anthropogenic pressures, it is currently es?mated that the global popula?on of Black-and-chestnut Eagles is less than 1,000 mature individuals, and thus are classified as Endangered by the IUCN (BirdLife Interna?onal 2022). The be(er-known aspect of the species' biology is its feeding ecology, along with anecdotal reports of nests in Colombia, Ecuador, Bolivia, and Argen?na (Lehmann 1959, Hilty & Brown 1986, Strewe 1999, Márquez & Rengifo 2002, Zuluaga & Echeverry-Galvis 2016, Aráoz et al. 2017, Zuluaga et al. 2018a, Restrepo-Cardona et al. 2019. In Peru, the Black-and-chestnut Eagle occupies a forest strip of more than 1,600 km long between 900-3,500 m a.s.l. (Valdez & Osborn 2004, Walker et al. 2006, Schulenberg et al. 2010. Despite this, very li(le is known about the species in the country, and there are only a few behavioral observa?ons of two pairs of eagles (Valdez & Osborn 2004). Here we provide new informa?on on the Black-and-chestnut Eagle based on observa?ons of several nes?ng territories found in the Andean montane forests of central Peru. Our general objec?ves were to describe its breeding ecology, nes?ng habitat, and main threats.

Study area.
We worked in the montane regions (1,000-3,000 m a.s.l.) of Junín and Pasco Departments in the eastern Andes of central Peru. The general study area covers approximately 5,056 km ² and is located in the interior of the Selva Central (10°09'S -11°24'S; 75°16'O -75°30'O, Figure 1), within the Yunga vegeta?on zone, which is divided into three ecosystems located at different al?tudinal ranges: 1) The Yunga basimontane forest (between 600-1,800 m a.s.l.), 2) The Yunga montane forest (1,800-2,500 m a.s.l.), and 3) the Yunga high montane (pluvial) forest (2,500-3,800 m a.s.l.). The Yunga basimontane forest ecosystem is a transi?onal complex with botanical species belonging to both the low Amazon basin and Yunga, with tall emergent trees reaching 35 m in height. With increasing al?tude, the Yunga montane forest emerges. This ecosystem is consistently covered by mist and hosts abundant epiphytes, lichens, bromeliads and orchids, emergent trees, and tree ferns 10-20 m height. Finally, above 2,500 m a.s.l. the Yunga high montane forest reaches the top of the mountains, has low trees, and a great abundance of epiphytes. Above the high montane forest lies the páramo grasslands and dwarf forest, represented by plants of the families Ericaceae, Solanaceae, Asteraceae, Polemoniaceae, Rosaceae, among others (MINAM 2018). The two main protected areas within the study area are the Yanachaga-Chemillén Na?onal Park and the Pui Pui Protected Forest, together harboring c. 180,000 ha of montane forests and páramo.
Extensive human-caused forest degrada?on has occurred during the last 60 years in Peru's humid eastern montane forest due to ?mber extrac?on, oÑen followed by conversion to low-quality rangeland. Most of this impact occurs through several kilometers of roads associated with recent colonists or illegal loggers, which extract resources from state-owned lands (Young & León 1999). In the study area, deforesta?on pressure is high outside of protected areas. However, there are s?ll rela?vely large fragments or relicts of intact to nearly intact forest (c. 3,840 km ² ) above 1,200 m a.s.l., resul?ng in lower levels of anthropogenic impact as the distance from populated centers increases (Antón & Reynel 2004 Habitat around neseng territories. A nes?ng territory was defined as "an area that contains, or historically contained, one or more nests within the home range of a mated pair" or "a confined locality where nests were found, usually in successive years, and where no more than one pair was known to have bred at one ?me" (Steenhof & Newton 2007). For our study, we also considered nes?ng territories, the sites where we did not find a nest, but the pair ac?vely occupied the area. In loca?ons where we observed adults or dependent juveniles (<1 year of age, vocally reques?ng food) in suitable habitat for nes?ng more than twice throughout the breeding season, we considered the center of the post-fledging family area (PFA, defined as "areas occupied by the family group, from the ?me the fledgling leaves the nest un?l it stops depending on adults to eat"; Tapia et al. 2007) as the nest site. On the other hand, considering that territorial adults have predictable daily ac?vity pa(erns (e.g., usually soaring with mid-morning thermals over nes?ng sites; Valdez & Osborn 2004, Rivas-Fuenzalida pers. observ.), we assume that solitary adults repeatedly soaring at these hours in sites with suitable habitat were within their nes?ng territory, even though we were unable to find their nests due to difficult access.
To characterize the habitat around each breeding territory, we established circular buffers of 4 km in radius (c. 5,000 ha, the approximate home range of a nes?ng pair in the study area; Rivas-Fuenzalida et al. 2022a) around nests and nest sites loca?ons (N = 36). We then measured the surface-area of two habitat types: 1) forest areas and 2) open areas (including grasslands, cul?vated fields, paramo, landslides, and water bodies) using ArcGIS 10.3.

Characterisecs of nest sites.
To determine eleva?on at the nest sites, we used loca?ons taken under nest trees by GPS (GARMIN 64S). We measured the distances from nest trees to water courses, open areas, and inhabited places using ArgGIS. The slope of every nest site was calculated from a digital eleva?on model (DEM) in ArcGIS.

Characterisecs of nest trees and nests.
We used a clinometer to es?mate tree height and nest height above the ground. Using a measuring tape, we measured the Diameter at Breath Height (DBH). To iden?fy tree genera and species, we relied on our knowledge. In some cases, when possible, we took photos and samples of leaves and flowers, for iden?fica-?on by Peruvian botanists. We recorded the number of branches suppor?ng the nest, whether the nest was built on epiphytes, nest loca?on within the tree crown, nest shape (es-?mated by direct measurement or drone photos), and composi?on (e.g., branches, green leaves, prey remains). We climbed the tree with arborist equipment to accurately measure the nest and used a measuring tape.
Producevity and breeding phenology. To es?mate breeding phenology, we made observa?ons of courtship behavior, eggs, and morphological development stages of chicks at 10 nests, visi?ng each nest at least four ?mes per season and considering: 1) a courtship period of three months, 2) incuba-?on period of 51 days (Juan Manuel Grande pers. com.), 3) chick-rearing period of 60 days (un?l the age at which the young abandon the nest, but remain on the nest tree), 4) the first flight out of the nest tree at 92 days aÑer hatching, and 5) post-fledging period of 4.5 months un?l juvenile dispersal (Aráoz et al. 2017, Zuluaga et al. 2018b, Rivas-Fuenzalida et al. 2022a. We es?mated produc?vity as a successful a(empt being a breeding pair producing a fledging. Threats: human persecueon and habitat loss. To obtain informa?on on the aàtudes of local people towards the eagles, we conducted informal interviews with villagers familiar with the species, located ≤ 4 km from a nes?ng site. We interviewed every poultry owner who agreed to talk with us (N = 22), asking them two ques?ons: 1) Do you have a conflict with the species due to poultry preda?on? and 2) How many eagles have you killed to defend your poultry? In many cases, we only obtained informa?on about killed eagles by the villager´s descrip?ons. However, in a few cases, we also obtain physical evidence (e.g., talons, heads, feathers, embalmed birds, en?re bodies) or photographic evidence of kills. To determine the age classes of killed eagles in cases with no physical or photographic evidence, we showed photos of juvenile, immature, and adult eagles to the villagers. We assume that each killed eagle belonged to the nearest nes?ng site. To evaluate habitat loss, we also measured the forest cover change derived from Global Forest Watch (Hansen et al. 2013) between 2001 and 2020 as a proxy for habitat conversion in an area of 5,056 km ² , corresponding to a polygon that enclosed all the 4 km buffers around the territories.

RESULTS
We monitored 36 breeding territories, 14 in the Pasco region and 22 in the Junín region (Figure 1). Three territories were found in 2017, 9 in 2018, 1 in 2019, 8 in 2020, 14 in 2021, and 1 in 2022. We found nests in 15 territories, while in the remaining 21, we assumed the existence of nests. In these territories, we observed 62 adults, one immature, and 36 juveniles ( Figure 2). The 16 territories observed for more than one season were considered as independent breeding events.
Habitat around neseng territories. The territories (N = 36, Table 1) were located in areas of very rugged terrain, with numerous ravines and steep slopes. Within the area limited by a radius of 4 km around the nest loca?ons, the minimum al?tude ranged between 690-2,080 m a.s.l. (mean ± SD = 1,329.4 ± 397.7 m a.s.l.), and the maximum al?tude varied between 1,710-3,810 m a.s.l. (mean ± SD = 2,915.6 ± 530 m a.s.l.). In this same area, the propor?on of forested area within breeding territories ranged from 43.8-99.7% (mean ± SD = 90.4 ± 11.5%), while open areas ranged from 0.3-56.2% (mean ± SD = 9.6 ± 11.5%). The linear distance from the nest loca?on to the edge of the nearest urban area ranged from 1.3-27.1 km (mean ± SD = 9.1 ± 6.7 km). (range = 26-38 m, mean ± SD = 31.9 ± 3.8 m) of tall emergent (height range = 28-40 m, mean ± SD = 34.5 ± 4.2 m), and thick-stemmed trees (DBH range = 0.53-1.32 m, mean ± SD = 1.01 ± 0.27 m), belonging to nine genera, with Ficus and Juglans being the most frequently used (N = 3). The number of branches that supported the nests varied between two and four, and several nests were built on epiphy?c plants. Nine nests were on the side of a treetop, three in the middle of the crown and three in the main top. Three nests had oval, and 10 had circular shapes. The only nest measured (Huatziroqui) had the following dimensions: 130 cm wide, 140 cm long, and 50 cm high. However, some large branches in the base of the nest reached up to 2 m. Before egg laying, the nest cup was approximately 15 cm deep. As the chick-rearing period progressed, the female filled the cup with twigs un?l it was prac?cally a flat plaçorm. Unmeasured nests had similar dimensions (c. 1.4 m diameter) as es?mated by comparison with the size of adults perched. The nests were built with dry branches, twigs with green leaves, lichens, and bark, and contained pellets and prey remains such as feathers, hair, and bones.
Producevity and breeding phenology. In all the nests where we found eggs (N = 7), only one egg was laid per pair/ season. In the sites where we corroborated successful reproduc?on (N = 28), a single juvenile was fledged per pair/season. We confirmed that at least six pairs produced one juvenile every year, and three pairs produced only one fledging every 1.5 to 2 years. The incuba?on and chick-rearing periods were concentrated in the driest months of the year. The earliest laying occurred in mid-March, but most occurred between late April and mid-July (Figure 3). The earliest hatching occurred in early May and from mid-June to early September in the remaining nests. Fledglings made their first jumps out of the nest between the end of July and the end of November, while the first flights out of the nest tree occurred between the beginning of August and the beginning of December. The postfledging period extended between August and May, mainly during the rainy season. However, at one site, we observed a juvenile s?ll feeding at the nest (although it was already flying within 500 m of the nest) in early May, which probably dispersed in June or July.
Threats: human persecueon and habitat loss. In 55.5% of the nes?ng territories (N = 36), we recorded human persecu-?on towards Black-and-chestnut Eagles in retalia?on to poultry preda?on, resul?ng in 41 individuals killed between 1999 Table 1. Characteris?cs of the breeding territories (N = 36) of the Black-and-chestnut Eagle Spizaetus isidori in the Andean montane forests of central Peru. All measurements, except distance to the urban area, were calculated within a 4 km radius (5,000 ha) around the nest loca?on.   and 2019, according to the informa?on of local villagers. Deaths corresponded to four adults, two immatures ( Figure  4A), 26 juveniles ( Figure 4B), and nine individuals of undetermined age.
In our study area (5,056 km ² ), the loss of forest cover around nes?ng territories reached an es?mated 21,822 ha between 2001 and 2020 (Figure 1), averaging 1,091 ha per year. However, forest loss rates have been increasing, especially during the last five years, in which 9,794 ha were lost, averaging 1,959 ha per year.

DISCUSSION
Given that dependent juveniles we observed in Peru and by others observed in Colombia (Zuluaga et al. 2018b) concentrate their movements within a 500 m radius around the nest, we believe that our loca?on es?mates of unfound nests were accurate enough. In general, the characteris?cs of the nes?ng territories, nest sites, nest trees, and nests found in Peru coincided with those reported in other countries such as Venezuela (Hilty & Brown 1986), Colombia (Lehmann 1959, Strewe 1999, Márquez & Delgado 2010, Zuluaga & Echeverry-Galvis 2016, Ecuador (Zuluaga et al. 2018a), Bolivia (Hilty & Brown 1986) and Argen?na (Aráoz et al. 2017). The observa-?on of a nest in the Yanachaga-Chemillén Na?onal Park (Huampal) is the lowest al?tude record for the species (1,330 m a.s.l.). Based on observa?ons of a pair with an unlocated nest but confirmed reproduc?on, we es?mate that some nests can be located above 2,500 m a.s.l.
Black-and-chestnut Eagles showed lower produc?vity than most raptors (Newton 1979), with a maximum of one juvenile produced per season. However, compared with other Spizaetus, including the Ornate Hawk-Eagle S. ornatus, the Blackand-white Hawk-Eagle S. melanoleucus, and the Black Hawk-Eagle S. tyrannus, which produce only one juvenile every 1.5 or 2 years (Canuto et al. 2012;Whitacre et al. 2012aWhitacre et al. , 2012b, Black-and-chestnut Eagles showed higher produc?vity. This is likely related to the rela?vely short dura?on of the post-fledging period (4.5 months; Zuluaga et al. 2018b).
Their breeding phenology was similar to that observed in other countries, with incuba?on and chick-rearing periods concentrated during the dry season (Aráoz et al. 2017;Zuluaga et al. 2018aZuluaga et al. , 2018b. This permits tropical raptors to minimize the adverse effects of storms on their eggs and chicks, and increases the chances of juvenile survival during the rainy season when prey is more abundant (Whitacre & Burnham 2012).
The human-eagle conflict has been known along virtually the en?re Black-and-chestnut Eagle's distribu?onal range for a long ?me. In Colombia, several studies revealed the occurrence of human persecu?on since the early 20 èê century (Lehmann 1959, Córdoba-Córdoba et al. 2008, Zuluaga & Echeverry-Galvis 2016Restrepo-Cardona et al. 2019, 2020, and several shot individuals have been received in rescue centers of Colombia, Ecuador, and Peru (Alex Ospina, Joep Hendrix, Ernesto Arbelaez, and José Antonio Otero pers. com.). Some dietary studies found that poultry consump?on can occasionally be greater than the consump?on of natural prey (Márquez & Delgado 2010, Zuluaga & Echeverry-Galvis 2016, Restrepo-Cardona et al. 2019. In Argen?na, the presence of domes?c fowl in these eagles' diet presumes that conflict with humans can occur (Aráoz et al. 2017).
At a global scale, human-raptor conflict derived from domes?c animal preda?on by large raptors has caused significant popula?on declines due to direct human persecu?on, being one of the main threats for many species (Newton 1979). This conflict has been recorded for several medium to largesized raptors which inhabit forest habitats in South America, including the three lowland Spizaetus: Black-and-white Hawk-Eagle (TRF pers. observ., Alex Ospina pers. com.), Ornate Hawk-Eagle (Ffrench 1973, Friedmann & Smith 1955 and Black Hawk-Eagle (Whitacre et al. 2012b, TRF pers. observ.). In the austral temperate forest ecoregion (southern Chile and Argen?na), the Rufous-tailed Hawk Buteo ventralis -an ecological equivalent of the Black-and-chestnut Eagle in southern South America-also suffers intense persecu?on by villagers in retalia?on to poultry preda?on; together with habitat loss, such persecu?on is responsible for local ex-?nc?ons within some areas of its distribu?on (Rivas-Fuenzali- Montane forest loss and degrada?on represent the main threat to the Black-and-chestnut Eagle in the long-term (Thiollay 1991, Echeverry-Galvis et al. 2014), since structure and func?on recupera?on of old-growth forests may take centuries (Newton 1979). If we es?mate a loss rate of 1,959 ha per year, in 55.5 years -or three genera?ons-of the Blackand-chestnut Eagle (genera?on length = 18.5 years; BirdLife Interna?onal 2022) the total forest cover loss will be 108,724 ha (i.e., the area that covers the needs of c. 22 breeding pairs, 61% of the currently known breeding pairs in the area). However, with increasing rates every year, this amount of forest loss and its impact on Black-and-chestnut Eagle popula?ons could be much higher.
At a finer scale, deforesta?on may reduce the reproduc?ve viability and habitat suitability for large eagles due to decreasing prey availability (Miranda et al. 2021). In the case of Blackand-chestnut Eagle, this makes them more prone to hunt domes?c fowl, increasing the conflict with humans (Restrepo-Cardona et al. 2020). In Chile, it has been observed that forest loss can increase the vulnerability of Rufous-tailed Hawks to human persecu?on since they are more visually exposed, and even the fledglings are killed by domes?c dogs when they fall to the ground during their first flights (Rivas-Fuenzalida et al. 2011, Rivas-Fuenzalida & Figueroa 2019

Recommendaeons.
To halt the local ex?nc?on of this species, we recommend to: 1) carry out long-term monitoring of breeding popula?ons, 2) improve knowledge of its natural history (e.g., diet, habitat use, home range, juvenile dispersal, popula?on densi?es), 3) promote projects that mi?gate the human-eagle conflict from different approaches and adjust to local reali?es and opportuni?es (see Rivas-Fuenzalida 2019), 4) develop permanent environmental educa?on programs, 5) promote be(er regula?on and supervision of land use by government en??es, and 6) promote the crea?on of new protected areas priori?zing the conserva?on of large tracts of primary montane forests.
?ons, commentaries, and sugges?ons of Carlos Bosque (Editor-in-Chief), and two anonymous reviewers significantly improved this manuscript's quality. We also appreciate the improvements in English usage made by Daniel Brooks through the Associa?on of Field Ornithologists' program of editorial assistance.