Harpy Eagle Training: Exploring the Potential of Positive Reinforcement
Barbara Heidenreich; Animal Training and Consulting Services, Good Bird Inc, Erik Corredor; Dallas Zoo, Nathan Compton; Dallas Zoo
First presented at The International Association of Avian Trainers and Educators Conference 2010
The Dallas Zoo is home to 0.1 Harpy Eagles (Harpia harpyja). The mature female hatched in 2002 has been on exhibit since 2004. In September 2007 Keepers initiated a training program to facilitate optimal care of this individual. Initial training included the use of targets to teach hands off behaviors such as moving from point A to point B, shifting into a holding area, stepping onto a scale and flying from perch to perch on cue. Eventually keepers were able to work in the enclosure with the eagle and incorporated these behaviors into educational presentations. As the bird continued to progress, it was decided to train behaviors that would allow the eagle to make appearances outside of the enclosure. This included teaching the bird to wear falconry equipment and sit comfortably on a glove in public areas. Because of the bird’s role as an exhibit bird it was important to the facility that equipment could be removed completely for when the bird was on display. This challenged keepers to train the eagle to allow daily manipulation of feet, extensive touching to the tarsus while anklets and jesses were attached and removed, and learning to step onto a glove voluntarily. By adhering to a strictly positive reinforcement based training approach staff members were able to successfully achieve their goals and more. This paper will share the triumphs and challenges of using positive reinforcement to train one of the largest and most powerful species of eagles in the world.
The Wings of Wonder section of the Dallas Zoo is home to a variety of raptor and new world vulture species. The recently refurbished exhibit is primarily designed to give zoo guests the opportunity to view these impressive birds of prey in their large, attractive and naturalistic enclosures. There is an additional goal of obtaining reproductive success with a number of the species on display. To date a pair of King Vultures (Sarcoramphus papa) has successfully reproduced while housed at the refurbished Wings of Wonder.
In the past caring for exhibit and birds in reproduction programs has usually involved a “less is more approach.” The idea being that the fewer disturbances a keeper makes in a bird’s life, the more likely the bird is to show behaviors indicating comfort and therefore potentially breed. However with the addition of an animal training consultant and zoo keeping staff members with experience and interest in training it was decided a different approach could prove successful.
In 2007 it was decided a proactive approach to training would be implemented for the birds in Wings of Wonder. The goal was to train as many birds as possible for behaviors that would facilitate husbandry and daily care. Initially several of the vulture species were identified as excellent training candidates. The criteria used to determine this was the level of comfort in close proximity to keepers and their willingness to accept food from keepers. A female Andean Condor (Vultur gryphus) was involved in the training program until her transfer to another facility. The King Vultures were successfully trained to target, shift, enter a kennel, present a foot, and allow tactile to the chest. They also bred and raised offspring and participated in training throughout the process. The offspring also learned to target and stand on a scale. 1
As training proved to produce positive results it was decided to consider adding other candidates to the training roster. Another bird that demonstrated a high level of comfort in the presence of the keepers was a female Harpy Eagle named Killa (pronounced kee ya).
The Training Candidate
The Harpy Eagle is one of the largest species of eagle. They can weigh from ten to twenty pounds and have a six and a half foot wingspan. Their talons are comparable in size to the claws of a grizzly bear. They feed primarily on animals that live in the trees, like sloths, monkeys, opossums, and some reptiles and birds. Harpy Eagles are highly maneuverable fliers and strike their prey after a rapid pursuit through the trees. 2
Killa was hatched in 2002 at the Avian Propagation Center at the San Diego Zoo. Hatched in an incubator, she was then hand raised using techniques to avoid imprinting on humans. While in the brooder she was covered with a towel, set behind a curtain, and a CD of rainforest sounds was used to drown out the sounds of human voices. During feeding keepers placed a sheet over their heads and used a Harpy Eagle puppet to deliver food. 3
Because of her large size and power, birds such as Killa have traditionally been regarded as potentially dangerous. For safety purposes the Dallas zoo policy requires that a keeper is only allowed to enter her enclosure if another keeper is present. However what keepers observed was a bird that appeared quite comfortable in the presence of people. She had never been observed to present aggressive behavior. Nor did she demonstrate a fear response. This in part may be the result of her experiences as a chick with humans at the Avian Propagation Center. Other individual birds in Wings of Wonder such as the much smaller Ornate Hawk Eagle (Spizaetus ornatus) would show overt aggressive behavior, flying directly at keepers. The Spectacled Owls (Pulsatrix perspicillata) showed extreme fear responses and escape behavior when keepers were present. Compared to these individuals Killa appeared to be a training candidate with excellent potential.
With the permission of the curator Killa became the next training subject for the Wings of Wonder team.
Traditional raptor training practices often involve capturing and restraining a bird to get a bird on a scale to determine a “starting” weight for training. However the process of grabbing and restraining a bird against its will is contradictory to the goals of a positive reinforcement training program. Furthermore knowing the weight of an animal is not necessarily important to reaching training goals. While the information can be of interest for evaluating health, it is not required to train an animal.
Instead keepers tried to ascertain based on observations, the amount of food Killa typically ate in a day. They also observed her responsiveness to food when she was relaxed and comfortable. More or less food could be added or removed from the diet based on her response during a training session to maintain an adequate level of interest in training. This amount changed in correlation with the time of year. Colder temperatures meant an increase in diet. Warmer weather meant a decrease.
Because Killa showed a high level of comfort around people she readily took food from keepers in the first training session. Per zoo policy the food was offered from hemostats as opposed to hands. Training was initiated with keepers outside of the enclosure. The first behavior Killa learned was to target. The target was a blue plastic circle attached to a dog clip. The target could be hung on the outside of the cage near the eagle. At first food was held near the target. Killa learned to touch the blue target with her beak to earn the food reinforcer. In the early stages of training, the target became a useful tool to call Killa to the front of the cage for keeper talks. This offered zoo guests a great opportunity to see the impressive wingspan of a Harpy Eagle as she flew to the front of the enclosure.
The target was also used to call her to the back of the cage to work on training other behaviors. To initiate introduction of a kennel, the target was used as a means to pair reinforcers with the kennel. The kennel was placed outside of the enclosure and Killa was reinforced for targeting nearby.
Success Leads to More Training Goals
As hoped, Killa learned quickly and was an eager participant. In addition her continued calm demeanor helped increase the teams’ confidence that the presentation of aggressive behavior would be very unlikely. Because of this keepers moved from training outside of the enclosure to training inside.
Very quickly Killa learned to fly to perches when cued by the keepers. This made for an even more dynamic presentation for zoo guests.
Killa also learned to shift into a holding area. This was an important training goal. Colder weather often required birds to be placed in holding. Training Killa to voluntarily enter her holding area made this process stress free and more reliable.
Crate training progressed with the crate now being in the enclosure and keepers in there as well to help direct Killa into the kennel using the target and food reinforcers.
A scale was able to be brought into the enclosure to get a weight on the eagle. This was after months of training. Therefore although it was helpful to know her weight for record keeping purposes it was not integral to her training success.
A Small Setback
One of the challenges for the Harpy Eagle was her exhibit. On several occasions Killa was found caught in the netting/wire of her enclosure. This meant keepers had to restrain her and untangle her legs from the netting. Pairing her trainers with this aversive experience had the potential to affect the trusting relationship that had been nurtured via positive reinforcement. Unfortunately the problem occurred more than once before Killa was moved to another enclosure while the netting was replaced with something more suitable.
The entanglement incidents did cause Killa to show some hesitancy with her trainers, but because of a long positive reinforcement history she was able to return to her consistent performance during training rather quickly.
Training to Wear Falconry Equipment
The move to the new enclosure proved to be a blessing in disguise. Although the enclosure was smaller, it opened up the possibility for some new training goals. Because Killa had been doing so well, the idea of bringing her out of the enclosure for educational opportunities had been suggested. However her role as an exhibit (and potentially breeding) bird was also important. To bring Killa out of her enclosure would require the bird to be equipped with anklets and jesses. However it was agreed for her safety, based on her entanglement history, and for her role as an exhibit bird that she not wear equipment while in her enclosure. For these reasons it was decided to train her to allow keepers to apply anklets, jesses and any other equipment for excursions outside of the enclosure. These would then need to be removed when she returned to her enclosure.
At the time, to the team’s knowledge, training a Harpy Eagle to accept the level of manipulation required to apply and remove equipment daily had not been attempted before. A shaping plan was devised to achieve the training goal. The steps included the following:
Touch foot with hand
Touch foot with anklet
Drape anklet on foot
Wrap anklet around leg
Secure anklet around leg
Add swivel and leash
Step to the glove
Short hop to the glove
Walk with the bird on the glove
Training sessions were scheduled to occur once daily. However due to staffing challenges, training could not always take place every day. Despite this, the goal of securing the anklets comfortably around Killa’s legs was achieved in thirty three training sessions. It took nine more training sessions for the eagle to be comfortable with the jesses being threaded. Sessions to add the swivel and the leash followed. After eleven more training sessions Killa was stepping to the glove without hesitation.
Training Killa to accept touching on her feet also lead to some other behavior goals. These included training her to allow a keeper to wrap his hands around her legs and apply some pressure. The goal was to be able to approach and calmly secure Killa’s legs if she needed to be captured and restrained. This approach is far less stress inducing than traditional methods that require birds to be chased, netted and grabbed for restraint.
An additional bonus to the tactile training was the ability to evolve this into touching Killa’s sternum and pectoral muscles. This allowed keepers another means to evaluate her physical condition in a manner that was stress free for the bird.
Since Killa’s initial training it was discovered that nature cinematographer and falconer Neil Rettig has also had success training a Harpy eagle to accept the application of anklets and jesses and allow touching.4 With more than one Harpy eagle responding well to the use of positive reinforcement to train these behaviors, it is possible more trainers will be inspired to consider this as a training goal.
Addressing Aggressive Behavior and Fear Responses
Because Killa continued to be an excellent training subject it was an important goal to maintain her calm demeanor. Avoiding creating aggressive behavior or fear responses were high priority. If any activity elicited aggressive behavior it was immediately discontinued and smaller approximations were employed. The smallest demonstration of body language typically associated with aggressive behavior would cause trainers to reevaluate their strategy. This was especially relevant when introducing touching with hands or the anklets and when manipulating equipment, for example when threading the jesses.
Another factor that contributed to some aggressive behavior was using large pieces of food. To encourage Killa to hop to a glove and remain for a period of time, a large piece of food was used to lure her initially. However if Killa was particularly motivated for training she would show some body language indicative of aggressive behavior. For these reasons trainers quickly switched to using smaller pieces of food and waiting for presentation of the behavior paired with calm body language before delivering reinforcers.
Although some female eagles have been observed to present aggressive behavior at sexual maturity, this has not been observed with Killa. She has also not yet been provided with a nest site, nesting materials or a mate. Other birds at Wings of Wonder have successfully continued training during breeding and rearing of offspring. It will be interesting to observe if this can be repeated with Killa at some point in the future.
Keepers also did not want to create a fear response. This was especially important when the equipment was on Killa and she was on the glove. Keepers wanted to avoid creating a situation in which the eagle would be likely to fly or bait off the glove. When a bird baits it is often trying to escape a situation it perceives as uncomfortable. In other cases a bird may be trying to go towards something, for example a preferred perch. In Killa’s case her few instances of baiting appeared to be in line with wanting to go away from the situation. They were always associated with the keeper walking around the enclosure with Killa on the glove. Learning to adapt to the movement was a challenging behavior.
In many cases a bait can be avoided by observing a bird’s body language closely and reinforcing the bird for remaining calm or if possible removing the bird from the situation. However if the moment is missed and a bird does bait it can be very challenging for some birds to recover. Restricted movement is not something to which Killa was accustomed. When she did bait, the trainer needed to assist her to get back on the glove and often she was no longer interested in participating in the training session after a bait.
At this point in Killa’s training another major setback occurred. Killa was found disoriented and wobbly. Because of this she was sent to the veterinary hospital for medication and observation. Training was discontinued for approximately thirty days while she remained at the hospital. She was eventually returned to Wings of Wonder. Her training was resumed. However she was eased back into the process. Behaviors practiced initially included hopping on the scale, targeting around the enclosure and crate training.
Along the way other training challenges including finding a crate large enough to comfortably hold a large eagle. Another challenge was creating stable portable perches that could withstand the force of a Harpy Eagle landing and launching into flight. Other setbacks included lack of funds during hard economic times to purchase materials to facilitate training. Because zoo policy required two keepers to enter Killa’s enclosure training was sometimes delayed due to lack of staff availability. However despite these challenges training did progress.
To date Killa has successfully learned to target, fly from point A to point B on cue, shift into a holding area, fly to a scale, enter a kennel, allow falconry equipment to be applied/removed, hop to a glove, perch on a glove while a trainer walks around her enclosure, allow her feet to be manipulated, allow a trainer to wrap his hands around her legs and apply pressure, and allow touching of her chest and beak. Killa’s training is ongoing and it is the team’s hope to continue to add more behaviors to her repertoire.
As the potential of positive reinforcement is explored more, traditionally held beliefs about animal care are challenged. Exhibit birds can function as education program animals, falconry methods can be modified to ensure the most positive methods are utilized and coercion is reduced or eliminated, micromanaging diets and weights is not required to train, and potentially dangerous animals can be trained and trusted to be non threatening. Most importantly birds such as Killa can be an example of what is possible when you train with positive reinforcement.