The Influence of Intrinsic Factors and Release Trials on Site Fidelity and Dispersal of Reintroduced Golden Parakeets (Guaruba guarouba, Aves, Psittacidae)

Abstract

Reintroducing threatened fauna has been established as a valuable conservation practice despite the obstacles that such projects still face. In avian reintroductions, elevated dispersal of released individuals is one of the main factors preventing the establishment of sizeable populations in target areas. We started reintroducing Golden Parakeets in a protected area in the Amazon Forest, and we faced problems regarding low site fidelity. To address this issue, we tested a new methodological approach named “release trials”, consisting of repeatedly releasing and recapturing individual birds to avoid early dispersion. We address the results of applying this method and the influence that intrinsic factors had on the site fidelity of released individuals. We released seven groups of parakeets with an average first-month site fidelity of 46%. Individuals who underwent more release trials before the group release were more likely to present site fidelity. The level of aggression suffered by resident parakeets was the main factor leading to the dispersion of individuals. Older birds were more likely to suffer higher aggression, but age alone did not explain the result of dispersion. Individuals less aversive to humans and previously paired before release were less likely to disperse. Our results show that every bit of methodological care during pre-release training and individual selection may increase the chance of establishing a group with site fidelity.

1. Introduction

Wildlife reintroductions are becoming an increasingly important conservation practice aimed at restoring populations of organisms to parts of their historic range where they have been extirpated [1,2,3]. The success of these projects depends on various factors related to the species’ natural history, the origin of the individuals, and the environment [4,5]. Despite their popularity and the establishment of extensive guidelines to direct these initiatives, reintroductions still encounter numerous challenges that often lead to their failure [6,7,8].

Among the main challenges facing animal reintroductions are high predation rates, low flock cohesion, and, for avian reintroductions in particular, increased post-release dispersion due to flight [9,10]. High dispersion of birds during releases can reduce individual survival because their fitness declines during the critical post-release phase when they are adjusting to the wild [11,12]. Flocking becomes challenging due to increased dispersion, which significantly affects species that rely on flock size to evade predation and find food sources [13]. Furthermore, dispersion pushes individuals away from the target area, which ideally possesses the best conditions for survival and where the original causes of species extinction are addressed [7,8]. As a result, dispersion may act as a barrier that prevents the target population from achieving a minimum sustainable size, ultimately resulting in the failure of the reintroduction project [4,14].

Site fidelity is a crucial factor in determining the movement patterns of individuals or populations in relation to their release site [11,15]. Individuals that demonstrate higher site fidelity are more likely to stay in the ideal target area, which is easier to monitor, supplement, and protect. Therefore, this trait is advantageous for reintroduced birds in conservation projects [9,16].

Management techniques are designed to enhance site fidelity and minimize the dispersal of released birds, such as providing attractants (supplementary food, artificial nests, and roosting sites), eliminating threats (predators, competitors), and ultimately conditioning the animals to stay in a specific area [9,17,18]. However, most of these techniques focus on post-release actions and pre-release training, while the actual release process and traditional soft release methods do not primarily aim to increase site fidelity but rather to give the animals an opportunity to return if they choose.

The intrinsic characteristics of individuals selected for reintroduction are also crucial for the process’s success [19,20]. Using captive-bred individuals in reintroductions rather than those captured from the wild is regarded as a beneficial strategy for protecting existing populations; however, it presents greater challenges due to the lack of experience and learned skills that captive-born individuals require for immediate survival in the wild [21,22,23,24]. Management techniques during the pre-release period can help address these challenges, including flight training, food recognition training, predator recognition, and human aversion training [17,25,26]. However, individual factors such as age, sex, breeding status, and established behaviors may still affect a bird’s performance once released into the wild [27].

The Golden Parakeet (Guaruba guarouba) is a psittacine species native to the Brazilian Amazon. Their striking appearance and behavior have made them prime targets for poachers and traffickers since colonial times [28,29]. Combined with the extensive deforestation of their habitat, this species is classified as vulnerable to extinction (VU) and has already been extirpated from a significant portion of its historical range [30,31,32]. Despite this, they have thrived in captivity worldwide, making them excellent candidates for reintroduction once protected areas are established in their native habitat [33]. Golden Parakeets are notably different from other parrots, exhibiting uncommon behaviors such as communal nest assistance and sentinel watching during feeding, while demonstrating a high level of sociability [34]. This underscores the importance of group cohesion and size for their survival. Although this species carries considerable cultural and symbolic significance in Brazil, Golden Parakeets have yet to be thoroughly studied, particularly regarding their behavior and ecology in the wild.

We began reintroducing captive-bred Golden Parakeets in 2018, and, as expected, one of the main challenges we faced was the rapid post-release dispersion of the birds. Since then, we have proposed an alternative method to the traditional soft-release approach, designed to enhance the birds’ site fidelity, ease their transition to the wild, and improve the monitoring process.

In this study, we aim to present our soft-release method based on release trials, test how efficient it is in reducing dispersion for Golden Parakeets, and determine the influence of each intrinsic individual factor on the outcome of the reintroduction.

2. Materials and Methods

2.1. Study Area

This reintroduction project occurred in Belém, the capital of the State of Pará, where the species has been locally extinct since the early 20th century [31]. The target area is Utinga State Park, a protected region founded in the 1990s. Covering 1393 hectares, the park connects with another protected area to the east and south while directly bordering urban areas to the north and west [35].

The park landscape mainly comprises primary rainforest (60%), secondary vegetation with altered landscapes (15%), and aquatic environments (25%), primarily represented by the two large lakes that supply water to the city [35].

The area immediately surrounding the aviary (30 m radius) is filled with various short plants, some of which form part of the diet of Golden Parakeets, along with several trees that are under 10 m tall. The nearby environment showcases a diverse range of habitats, primarily composed of continuous primary vegetation with tall trees to the north and west, roads utilized by park staff and visitors lined with numerous power lines, a research station with artificial lakes to the south, and a large water reservoir to the east, which connects to the park’s most pristine areas on the opposite side.

2.2. Project Description

The acclimatization aviary was built in 2017 and consists of two modules: a maintenance module measuring 6 × 2 × 2 m, connected to a larger one that measures 5 × 5 × 5 m. The maintenance module enables better control of the birds during management interventions or group separation, while the larger cube provides ample space for the birds to develop their flight abilities and undergo training. The larger module features a sliding door that connects to the outside, allowing the captive birds to be released at specific times or to return inside as they wish. Both aviaries are furnished with native vegetation components, preferably species known to be consumed by the parakeets, along with nest boxes designed similarly to those the parakeets were already familiar with at the Lymington Foundation, where they originated.

We received seven groups of Golden Parakeets between 2017 and 2024. The birds were raised in captivity at the Lymington Foundation (LF), a conservation-focused non-governmental organization that runs a captive breeding facility for endangered psittacines. Each year, LF prepared groups averaging 10 Golden Parakeets by selecting individuals with strong flight skills and positive social interactions. The birds were screened for common psittacine diseases (herpesvirus, bornavirus, circovirus, mycoplasma, chlamydia) in accordance with local animal translocation regulations two to three times before the flight to Belém. Individuals showing physical, feathering, or behavioral issues were excluded from reintroduction. The groups were sent from LF to Belém by plane and then transported by car to the acclimatization aviary.

Once a group arrived, they were kept in the aviaries for at least five months before their release. During this acclimatization period, they were exposed to potential predators both directly, by introducing live boa snakes into the enclosure to startle them, and indirectly, as birds of prey and mammals naturally present in the area were occasionally seen hunting nearby. They also received daily training to improve their flight strength by promoting movement between enclosures and through active pursuits. Their diet gradually transitioned from commercial fruits and rations to fruits native to the Amazon region. Native items were attached to branches and whole plants as part of the foraging training [36].

After the acclimatization and training period, we reintroduced the golden parakeets while keeping the larger aviary open. This allowed them to stay outside or return if they preferred. Immediately after releasing one group, we placed the next group in the maintenance aviary to attract the others and provide vocalizations to guide them back to the site, enhancing their site fidelity.

Before the release, we installed three suspended feeders around the release aviary to provide daily food that promotes a gradual transition to wild feeding. Additionally, we placed nest boxes in the taller aviary and nearby trees to offer a safe place for resting and to facilitate breeding attempts (Figure 1).

After the release, we monitored the birds daily through active searches using playback and VHF radio tracking, following established trails within the park that covered most of the area. Individuals were identified by a combination of factors, including the emitted radio signal, metal tags, and visual and behavioral traits. We recorded which individuals remained in the vicinity of the aviary (within 30 m) and returned to feed in the suspended feeders at least once a day, as well as those that did not return or were never seen after the release. When individuals were found in the urban environment and could not feed, we recaptured them to prevent poaching by locals. They were released again after a recovery period in the aviary.

2.3. Release Trials

To prevent immediate post-release dispersal of the groups, we adopted a training method designed to decrease the likelihood of individuals straying from the site immediately after leaving the aviary. This approach will be referred to as a release trial. The trials consist of daily training sessions held weeks before the group release, during which we isolate one individual from the rest of the group in the larger module of the aviary and open the trap door, allowing it to exit. Additionally, we place attractants like Byrsonima crassifolia and Euterpe oleracea branches filled with fruit near the aviary (less than 10 m away). After the bird explores the area for 15 min, we entice it back inside the aviary by offering sunflower seeds near the service door, shutting it immediately once it enters (Figure 2). This encourages the bird not to venture too far during the exploratory period and rewards it for staying close to the familiar aviary. Subsequent trials for the same individual should be extended to allow it to explore further each time before calling it back, ensuring a gradual experience while still minimizing the risk of going out of reach.

The release trials were conducted only on individuals already deemed fit for release. During these trials, there is always a chance that, similar to a standard release, the individual might fly away due to various uncontrollable factors and become lost.

Ideally, each individual should have undergone multiple release trials prior to the group release. However, due to time constraints, individual issues, and operational obstacles, we were unable to implement the method thoroughly and uniformly for all seven groups. Therefore, in this paper, we will assess how varying levels of training for this technique may impact the outcome of group site fidelity.

2.4. Intrinsic Factors

To analyze individual post-release site fidelity, we considered two potential outcomes for the individuals: either they demonstrated site fidelity for at least one month after release (SF) or they stopped returning to the release site, dispersing before one month after release (D). We used the first month as a key indicator of the most critical period for the birds following release, during which site fidelity with supplementation may be crucial for their adaptation. The six main individual factors we assessed were individual sex, individual age at release, individual level of suffered aggression, individual human aversion, individual pairing, and the number of release trials per individual. The level of aggression experienced was determined by how the previously released individuals interacted with the captive individual in question. We categorized each individual into three groups: low (when the experienced individuals did not persecute the individual), medium (when the individual faced moderate agonistic interactions that did not lead to severe fights), and high (when the individual was actively persecuted by released individuals, with aggression ceasing only when they were separated). This factor was not applicable for two groups (the first and third) since there were no established parakeets in the area during the release.

The human aversion factor was assessed by observing how individuals reacted when approached by a human. Those who displayed immediate fear of humans and would not allow a person to come closer than 1 m inside the aviary without flying away were categorized as Aversive to Humans (A). In contrast, individuals that allowed humans to get within more than 1 m and did not fly away until a person actively attempted to grab them were classified as Non-Aversive (NA). Individuals who exhibited no fear of humans and permitted themselves to be held did not participate in any release and were returned to LF for reproduction. Pairings were established through ad libitum observations, where we searched for signs of breeding behavior such as allopreening and food regurgitation in pairs. When individuals displayed both behaviors frequently, they were recorded as couples.

2.5. Data Analysis

We performed a Multiple Correspondence Analysis (MCA) and a logistical regression to assess the impact of four intrinsic factors on each individual (Sex, Aggression, Coupling, Human Aversion) regarding release outcomes (Site fidelity or Dispersion). The “Age” factor was excluded from the multiple analysis since we could not verify the exact age of some individuals due to their diverse origins. To evaluate the potential influence of each factor separately, we conducted chi-square, t-test, and Pearson correlation, using BioEstat 5.3, while the MCA and regression were conducted in R 4.4.1 [37,38].

3. Results

Over seven years, we released seven groups of Golden Parakeets into the wild, reaching a total of 57 individuals. The immediate dispersal of individuals was common. Typically, a released parakeet would exit the aviary by climbing the walls or hopping small distances to nearby trees and then take a long flight after a few minutes outside, during which it would quickly gain altitude, soaring over the tallest treetops. Subsequently, the bird would circle while vocalizing with the other individuals that remained enclosed until it chose a direction and attempted to perch in the nearest tree. If the parakeet ascended too high, it would land too far away, making it difficult to pinpoint its exact location. When the individual perched near the aviary, it was easier to monitor them and exchange vocalizations with the rest of the group.

After the first flight, nearly all individuals remained silent and still in their perches for a period of one to eight hours. Following that, they would start flying and vocalizing again, seemingly searching for a way to return to the group. We noted that all individuals struggled to control their flight altitude once they tried to return. They often flew over the aviary at a significant height but found it difficult to descend when needed. This crucial period lasted from one to two days for each individual. The parakeets that learned to descend and return to the aviary displayed site fidelity, while those that did not ended up flying long distances in random directions until they vanished from sight, at which point they were deemed dispersed.

We observed that individuals who underwent more release trials before the final group release could perform flight maneuvers more quickly and found it easier to return to the release site. Those accompanied and guided by previously established birds also had an easier time returning to the site by following the other parakeets and learning their movements and flight techniques. However, the presence of established birds could also be detrimental if a released individual is not accepted by the group or by a specific individual. In such cases, they might chase the released bird and attack it until it lands or flies beyond the defended territory.

Site fidelity during the first month was generally moderate at 46% (SD 23%). Group site fidelity did not vary significantly in relation to the average number of release trials per individual (Pearson correlation R² = 0.4119; p = 0.12). However, the three groups that received the most release trials per individual demonstrated at least 50% site fidelity, while the groups that averaged fewer than one trial per individual (i.e., not every individual in the group underwent a release trial) had the lowest site fidelity values (less than 50%).

At the individual level, the birds that dispersed underwent fewer release trials on average (1 per individual; SD = 1.58), while individuals with site fidelity averaged 2.5 trials (SD = 2.02) (t-test t = −3.11; p < 0.01; df = 54) (Figure 3).

The dispersion of each individual (i.e., the day when the individual was last seen near the release site) primarily took place on the first day following the release. Some individuals, mainly from the first group, dispersed within the first week. Only four individuals left the release site one year after the release (Figure 4).

There was a trend such that individuals who dispersed were, on average, older than 7 years, than those who stayed at the release site (t-test t = −1.507; p = 0.07; df = 33), with only one individual over 8 years (18 years) showing site fidelity. Half of the individuals demonstrating site fidelity were at most 2 years old, while only three dispersed individuals were also 2 years old (Figure 5).

The level of aggression displayed by resident Golden Parakeets toward new individuals was a determining factor in determining site fidelity or dispersion (chi-square = 12.93; p < 0.01). No individual observed experiencing high levels of aggression from the residents was able to establish site fidelity, and most individuals who encountered moderate aggression were also more likely to disperse. Conversely, individuals who faced little to no aggression exhibited higher rates of site fidelity (Figure 6).

The aggression exhibited by resident parakeets was generally higher toward older individuals than younger ones (LowxHigh t-test t = −1.97; p = 0.04; df = 9). On average, individuals who experienced low aggression were 4 years old, those who faced moderate aggression were 6.5 years old, and those who endured the most aggression were 9.3 years old (Figure 7).

We observed that individuals who were more afraid of humans (A) were more difficult to manage through our method, while the less fearful could be guided back to the aviary more easily. However, no significant differences were revealed in site fidelity or dispersion (chi-square = 2.48, p = 0.11). The three individuals who dispersed, even though they were familiar with humans, belonged to the group that experienced high levels of aggression from the parakeets already established in the area, which may have influenced the outcome (Figure 8).

Considering their availability at LF, we released 58% males and 42% females. The sex ratio did not vary significantly for site fidelity or dispersion and for levels of aggression.

Individuals displaying breeding behaviors and forming pairs before release were less likely to disperse into the wild than those who did not. Although 62.5% of paired individuals demonstrated site fidelity, only 34.1% of birds that did not exhibit breeding behaviors remained on site (chi-square = 3.79; p = 0.05) (Figure 9).

The MCA biplot illustrates two dimensions that account for at least 52% of the variance in results, leading to either a positive outcome of site fidelity or a negative outcome of dispersion (Figure 10). The main contributing factors were the heightened aggression experienced by residents, the reduced aversion to humans, and the pairing of individuals. Specifically, individuals suffering high aggression had significantly lower odds compared to those suffering low aggression (Estimate = −0.701, p = 0.05, SE = −2.11). Individuals that showed high aversion to humans had marginally lower site fidelity (Estimate = −0.573, p = 0.06, SE = 0.28). Although the pairing of individuals was significant in the biplot representation and in the individual analysis (Figure 9), the regression results were not significant (Estimate = −0.349, p = 0.18, SE = 0.25).

4. Discussion

Reintroducing gregarious birds is generally seen as easier when there are established conspecifics in the target area [9]. The newcomers are expected to follow experienced individuals, learn from observing their behavior, and benefit from group living [16]. We have observed mixed outcomes for Golden Parakeets in this context. The first released group had no prior conspecifics to act as guides at the site and recorded the second-lowest site fidelity among all groups. A significant issue noted during their release was that individuals often flew past the range where their vocalizations could be heard. Their chances of returning to the release site could have been notably better during these instances. The presence of established individuals did not ensure success for a release cohort, as the territorial response was largely unpredictable. Nevertheless, subsequent groups showed higher site fidelity, likely due to experienced individuals finding and bringing back many lost members. The exception was the fourth group, where only one released individual demonstrated site fidelity despite the presence of conspecifics. This group had the fewest release trials per individual before their release, which may have been an important factor.

The release trials produced mixed outcomes. The main issue with the method was preventing individuals from getting lost during the trial, as some wandered off despite careful strategies designed to avoid this. This is evidenced by the fact that most lost individuals went out of sight on their first day in the wild, where they had little to no experience. These incidents were exacerbated by unpredictable events, such as the parakeets being startled by random animals and sounds, which led them to fly off abruptly during previously successful release trials. The challenge of navigating and controlling flight altitude during the return phase also posed a significant obstacle that could not be addressed during training due to the aviary’s limited size.

Overall, the method improved site fidelity for individuals when implemented more frequently. From a behavior-based management perspective, the technique is effective because it conditions birds to return to the release site with their preferred food available. This approach encourages a longer acclimatization period and a slower release process, which reduces the likelihood of immediate fly-offs.

Alternative techniques, such as employing free-flight-trained birds as guides (as suggested by Woodman et al. [26] and implemented by Brightsmith et al. [9]), may be used alongside the release trials to help prevent the erratic flights that cannot be avoided. However, these options might be limited to smaller groups of younger individuals who successfully pair with guides and do not exhibit intense antagonistic interactions—something that may be difficult to achieve in territorial species.

Fewer individuals dispersed after showing site fidelity for months or years at the release site, and their chances of mortality post-dispersion can be considered lower than those who dispersed immediately after release. However, the reasons that led them to separate from the main group remain uncertain. The decision to disperse may depend on habitat quality, individual preferences, and social interactions within the group [11,15]. Given that experienced individuals were unlikely to get lost and received consistent food and nesting boxes at the release site, social factors may provide the best explanation for their separation. One individual that remained on-site for one year dispersed from the group at the start of the breeding period after suddenly experiencing aggression from another individual that had previously coexisted with it. Three individuals that dispersed for no apparent reason were later located, with two forming a pair and one integrating into a mixed-species group (Psittacara leucophthalmus and Aratinga jandaya). Therefore, competition leading to aggression may be the decisive factor in late dispersal events.

Some traits of the individuals selected for reintroduction were shown to be beneficial for forming a site-faithful group, while others proved to be harmful. Aggressive interactions with established individuals were the most significant barrier to integrating new individuals into the target area. Some individuals who initially exhibited attachment to the site were quickly expelled after being harassed by parakeets from previous cohorts, eliminating any possibility of them establishing themselves in the area. Considering the first few days in the wild to be the most critical period [11], conspecific aggression is expected to significantly reduce the chances of survival for new individuals. Thus, establishing multiple release sites, with considerable distances between them, may be essential for maximizing the success of released territorial species and countering the effect of increasing their density.

The age of released birds is relevant to their expected behavior in the wild, and we observed a lower average age among the parakeets that successfully established themselves in the area. In a systematic review, Miller et al. [10] also reported that for avian reintroductions, younger birds perform better than older ones. In parrots, younger individuals are considered more behaviorally flexible and innovative, while older ones typically display more neophobic behavior, which may be disadvantageous in a reintroduction scenario and may explain this trend [27,39,40]. Additionally, we observed that younger individuals experienced less aggression from residents, while many of the older individuals faced more intense persecution. One 18-year-old individual showed site fidelity and successfully transitioned to the wild; however, this bird was part of the first release group, and no resident individuals were present in the area to interact with. This further reinforces that the increased aggression faced by older individuals is a major factor, and that age alone should not hinder success, despite the overall lower success rates of older birds.

The aversion that released animals have towards humans is often seen as a beneficial trait for reducing the likelihood of recapture [41]. This is particularly crucial for parrots, regarded as some of the most charismatic bird groups, which makes them prime targets for illegal capture [3,14]. While it is preferable to release parakeets that are unlikely to approach humans and become trapped, we have noted that less fearful individuals were more receptive to training, handling, and recapture during our release trial efforts. The ease of management and the birds’ tendency to approach staff while being monitored may have significantly contributed to higher site fidelity in the less fearful animals compared to those that were intolerant of human proximity. Although there are concerns that these birds might experience problems related to human interaction, given that providing food at the release site may reinforce this behavior [41], we have also observed that this tendency was notably less apparent when they were outside the release area, along with the rest of the group. Therefore, we believe that parrots with manageable levels of human attachment should not be dismissed as candidates for release, and their presence could benefit the release and monitoring processes.

Individual sex, regarded as a significant factor in influencing bird behavior during reintroductions [20,42], was not identified as a determinant at any level of our analysis. Reintroductions should consider the sex ratio of a given species to achieve long-term population consolidation [43,44]. However, the average sex ratio in wild populations of Golden Parakeets is unknown, necessitating further studies on native populations.

Conversely, forming couples before the release was crucial for establishing individuals in the area. Paired individuals were less likely to disperse, though we noted cases where some individuals became lost after being separated from their partner following the first flight. We can infer that having multiple attractants at the release site, such as a mate still in captivity, may help prevent individuals from straying too far, as also described by White et al. [16]. Therefore, releasing individuals separated from their partners could be an additional strategy to minimize their immediate dispersion.

While our release method has not yet been standardized and may not suit every reintroduction scenario, we have shown that gradually acclimating an individual to the wild and repeating the process lessens its immediate impact. This approach should enhance control over a released population and aid in monitoring by improving site fidelity. Although “soft release” is the conventional method to prevent dispersion, its effectiveness varies across different animal translocations [45]. Therefore, additional techniques like these may bolster the chances of success.

It is recommended that individuals be carefully selected based on favorable traits to improve a reintroduction’s success. However, with endangered species, we often rely on a limited stock of individuals, which reduces the potential for choosing the ideal candidates. In these situations, every bit of methodological care to increase each individual’s chance of success is valuable, and we hope these results will serve as a basis for further improving bird release techniques.