Martinez was working at Loro Parque in the Canary Islands when the 29-year-old was brutally killed by Keto, a 6,600-pound orca on December 24, 2009. His death barely made the press, and wouldn't until Dawn Brancheau was killed by Tilikum at SeaWorld Orlando two months later.
Martinez' fiancee, Estefanía Luis Rodriguez, was initially told after the incident that he was okay. According to Tim Zimmerman's, Blood in the Water
, Rodriguez received a call from Orca Ocean supervisor Miguel Diaz:
He told Rodriguez that Martinez had been involved in an incident with a killer whale but that he would be fine, that he was being taken to the University Hospital in San Cristóbal de La Laguna, about 20 miles away.
The incident report however, revealed that Martinez was anything but okay. The corporate brief, marked confidential, is believed by insiders to have been filed by SeaWorld, not Loro Parque. Former trainers emphasized that the format is similar to that of SeaWorld's own, and Loro Parque is not a corporation.
The report obtained by Digital Journal, details how Keto pushed Martinez to the bottom of the pool before returning to the surface without him.
Initial efforts to get Keto under control were unsuccessful. The killer whale headed for the trainer again and returned, "with Alexis lying motionless on his rostrum." By the time Keto had been funneled away from Martinez, almost three minutes had passed since the incident onset. A SeaWorld supervisor named Brian Rokeach, along with another trainer, entered pool A to, "retrieve Alexis from the bottom."
The preliminary autopsy report
confirmed that Martinez endured a "violent death." The immediate cause of death was cited as "pulmonary edema", with the fundamental cause of death given as, "mechanical asphyxiation due to compression and crushing of the thoracic abdomen with injuries to the vital organs".
An email from someone claiming to be a SeaWorld employee indicated that "Rokeach was distraught." Spanish trainers also alleged "that Alex had blood coming out from every orifice."
When Miss Rodriguez arrived at the hospital with the Martinez family, Zimmerman writes that they were met by, "Wolfgang Kiessling, Loro Parque’s president ... along with legal representation."
Orca issues at Loro Parque
After Martinez was killed, sporadic news reports indicated that the orcas at Loro Parque had behavioral issues. This media piece by Laopinion.es
, errantly reported:
Informed sources have suggested that the orca responsible for the attack is the same animal that starred in another incident with one of the caregivers two years, which he produced serious injuries ago [translated].
In this instance however, Keto it seems, was mistaken for Tekoa, who attacked trainer Claudia Vollhardt in October 2007. All of the orcas at Loro Parque are SeaWorld owned. Keto was one of a group of four who were sent to the park on a breeding loan in 2006.
Digital Journal took the report for analysis to former SeaWorld trainers Jeffrey Ventre, John Jett, Samantha Berg, and Carol Ray. They highlighted several areas of concern that occurred prior to the incident.
Ventre, Berg, Jett and Ray are current members of the group: Voice of the Orcas
, and all of them appeared in the documentary, Blackfish
After the reading the corporate incident report, Ventre told DJ, "that one element of the attack reminded him of the famous 2004 Kyuquot attack on trainer Steve Aibel at SeaWorld of Texas."
The incident seen in this NBC segment
, was captured and uploaded by SeaWorld visitor Michael Gibson
Ventre emphasized, "the orcas determine when and if a waterwork trainer is allowed to exit the water. In Aibel's case, Ky eventually let him out. Keto made a different decision."
To understand the report further, requires some understanding of trainer terminology.
The term 'primary' is shorthand for "primary reinforcement." It is used to reinforce the animal ( via food). This is not seen as a conditioned reinforcer.
"Secondary reinforcement," is a conditioned reinforcer, and it is something that gains its reinforcing value by being paired with a primary, or events that are already reinforcing.
A bridge is used to indicate the exact moment when an animal performs a correct behavior. A bridge is trained by being paired with an established reinforcer, for example when working with a naive animal like a new calf, the whistle would be paired with fish.
The calf would receive a whistle bridge as it receives the fish. Over time, the calf will associate the whistle bridge with that primary reinforcer and a correct response. Bridges can also be "tactile" and usually involve a double tap on the body of the animal.
An LRS is an acronym for "least reinforcing stimulus" or "least reinforcing scenario." It's a technique used to extinguish a behavior or communicate that the animal did not do what was expected by the trainer. During an LRS, a trainer performs a 2-3 second neutral or non-response to an animal that has given an incorrect response. It's given in place of reinforcement and is also used to calm the animal and get their attention back onto the trainer and the job at hand.
The LRS is itself a trained behavior, animals have to be taught to accept it. To make the LRS less negative for the animals, animals can be rewarded with primary or secondary reinforcers for calmly and attentively accepting the LRS.
According to Samantha Berg:
Animals at SeaWorld are conditioned to understand and "accept" a variable reinforcement schedule. This means they learn to receive a variety of secondary reinforcers in addition to primary, which allows the trainers to not have to feed the animals after each behavior. it would be impractical (for the trainers) and boring (for the audience) to have a show where the reward schedule was 1:1 - that is 1 fish per behavior/trick. But in reality, this is how most animals are trained to start out. It's only over time and with regular training that they learn to understand that they will not be rewarded with "primary" for every behavior.
SeaWorld also addressed aggression issues in killer whales and other captive animals at an Animal Behavior Management Alliance Enrichment Talk in 2004. Thad Lacinak of Busch Entertainment Group (SeaWorld) and Gary Priest of the San Diego Zoo and Wild Animal Park, co-authored this presentation
, which states that trainers can control aggression in animals with operant conditioning.
Keto and Alexis' training session
There appeared to have been a couple of issues with Keto's training session that may have contributed to the eventual outcome. Ventre believes that Keto was frustrated after the lack of a bridge-reward from the safety spotter on stage, which in this case was Brian Rokeach, the SeaWorld trainer supervising the session.
Alexis was subordinate to the "control trainer" on stage in every way. Rokeach was an American SeaWorld trainer with full autonomy (from SeaWorld & Loro Parque) to orchestrate and direct training sessions. They were working toward a new show. Alexis was dependent on this "expert" to protect him and make prudent decisions.
Instead, from what I read, Keto gave a high energy attempt on a stand on spy hop, and didn't receive a reward because the control trainer determined that it wasn't good enough. So they repeated, and the same thing happened: High energy ... no reward. Unfortunately, in this case, Alexis received the brunt of Keto's frustration, leading to his death.
Still, out of all of Loro Parque's trainers, insiders noted that Martinez had the greatest affinity for working with killer whales, even though he had only worked there for three short years.
According to Zimmerman's article, Martinez, who had begun his work with killer whales in 2006, had expressed concerns to Rodriguez about the situation at Loro Parque:
For months, Martínez had been telling her that all was not well at Orca Ocean, that there was a lot of aggression between the killer whales and that they sometimes refused to obey commands, disrupting training and the shows ... As he gained experience, according to Rodriguez, he began to fret about safety, and he twice contemplated leaving the job.
His concerns appear to have been well-founded, but Ventre believes that Brian Rokeach's subjective decision to play "hard ball" with Keto during the training session, also did not help.
"Training is based upon operant conditioning," Ventre explained, "but when you're dealing with highly evolved predators, it's also a very subjective art, with consequences, including the possibility of human death." The former trainer added:
One camp, the "hard liners" believes that the animal doesn't get rewarded even if the mistake is made by the trainer. For example, if the animal gives incredible energy on a stand on, is pretty straight, but the trainer loses balance, the animal is still held accountable. I'm not sure if Alexis slipped or if Keto slightly twisted, but based upon what I read, he certainly gave 100% effort on both attempts, and was given a neutral response, which he didn't appreciate.
Berg agrees, and explained:
It's always the trainer's discretion to decide how much primary reinforcement to deliver during a given session — and there's always the risk of misreading the animals and not delivering enough primary reinforcement.
This leaves all sessions open to a lot of interpretation about what was the right call or what should have been done in hindsight. But in the moment, trainers are always making decisions "on the fly" about the best way to reinforce as well as communicate clearly to their animals. Unfortunately if an undesired outcome results, trainers are often blamed by management (trainer error) for making the wrong decision and therefore inadvertently causing the animals to attack or injure other trainers or animals.
Carol Ray noted that the training session appeared to start "off okay, and then Alexis lost him." [Lost the orca's attention]. "Keto never got anything for the call back response," she said, adding that the orca should have been reinforced for coming back to stage.
In hindsight, Ray acknowledged, "it can be a tough call, but it sounds like Keto did give him better control — a good response to hand target and Alexis is cued to get out — but Keto still didn't get anything for it."
Ray thinks that if Keto had received some food, it might have increased the likelihood of him staying, but she quickly added, "in no way would it have necessarily prevented this accident. Keto was making the decisions," she said, "and he was going to do what he was going to do."
The nature of the report, indicates that after Keto's initial attack on Martinez, the trainer never surfaced again, meaning that the orca's attack must have been swift and devastating. The report confirms this after just 40 seconds into the incident — Alexis appeared "motionless on Keto's rostrum."
Samantha Berg told DJ that the moment before things went desperately wrong is also worthy of mentioning. "Keto was noted as being 'calm'," she said, "until he wasn't." The indication that something was wrong Berg said, "was when Alexis surfaced with Keto and Keto positioned himself between Alexis and the stage."
After Keto had split towards Alexis the report stated, "Brian observed the intensity of the situation." Berg estimated, "my guess, is that Brian actually observed Keto ramming Alexis, not just taking him down to the bottom."
Keto was also reported as being "big-eyed" after he had positioned himself between the Alexis and the stage and just prior to the attack. Although it could indicate focus, according to SeaWorld's training protocols, it could have been a precursor. See pg. 7
of OSHA versus SeaWorld of Florida, which states:
A precursor is a behavior that indicates a predictable behavior will happen next (Tr. 142). Precursors that indicate killer whales may engage in aggressive behavior include putting their heads down, avoiding eye contact with their trainers, opening their eyes wider, vocalizing.
Digital Journal also ran the report by a current marine mammal trainer who, for obvious reasons, chooses to remain anonymous. The trainer, who has more than two decades of experience with marine mammals, evaluated the report as follows:
During the session Keto was sent on a perimeter swim for desense and was received by Alexis where he was reinforced with secondary (tactile). From that behavior Alexis went into the stand on that was not bridged due to criteria.
Keto was sent to stage and received secondary reinforcement (ice), then attempted the stand on again. The behavior was not bridged again due to criteria. An LRS was given and Keto was recalled to stage where he was reinforced with primary.
In this case the primary was given because Keto responded to the recall and returned to stage when asked. Looking at the notes Keto received primary reinforcement at the start of the session when pointed out to A pool, but once there he was sent on a perimeter swim, then performed a behavior that was not bridged.
It seems to set up a cycle of non-primary reinforcement where Keto went through several behaviors (some not bridged) before receiving any sort of primary reinforcement.
Moving forward through the session notes, Keto continues with this progression of incorrect behavior during the approach to the stage and the haul out. Keto moves out of position and Alexis moved his hand to target Keto, which appears to terminate the behavior.
Keto then surfaces but positions himself between Alexis and stage. He was recalled and responded but was "uncommitted" to staying with the trainer on stage. It's not clear in the report when or if Keto received primary for responding to the recall. I will say that anyone with experience working cetaceans can tell when they are going to "break" from station and it sounds like the situation here.
The trainer at stage felt strong enough about it not to leave Keto to retrieve primary reinforcement, so had another trainer bring the bucket. Possible delay here in primary but Keto was already "leaning" so it may be moot. When Alexis started swimming, Keto broke from control.
The trainer at stage who had "control" of Keto directed Alexis to the slide out which was the closest exit but this also took him close to Keto. It sounds like a mistake, but in that situation its hard to say, decisions are made in the water on a case-by-case basis and even though SOP (Safety Operating Procedures) exist, a trainer can make an independent decision based on their experience, it's a judgement call.
While interpretations of the report tendered diverse responses, one underlying theme could not be ignored. Despite SOP's and safety measures placed upon trainers by marine park executives, in a moment of crisis, trainers are on their own.
Tim Zimmermann touched upon this issue in the 2013 piece: Anatomy Of A Minor Killer Whale Incident
Zimmermann implied that killer whale personalities, which science states are highly complex and socially driven, always need to be considered. "Trainers must constantly make subjective judgments," he said, "about how to deal with the almost infinite challenges or wrinkles which can pop up at any time."
Zimmermann then provides an example of an incident between Jeff Ventre and a young orca named Taku. "A small decision, or minor incident," Zimmermann wrote, "if not handled with the right judgement can easily become a major incident, with major consequences."
But Dawn Brancheau was certainly not inexperienced. She was one of the most exulted trainers that SeaWorld employed. Known for being notoriously safety conscious, not even her experience saved her when she attacked by Tilikum two months to the day after Martinez was killed.
During the OSHA versus SeaWorld trial that ensued after Brancheau's death, SeaWorld Florida animal-training curator Kelly Flaherty Clark told the court
[pg. 5] that prospective trainers, "won’t be the person poolside making decisions, behavioral decisions with the killer whale until you’ve been interacting with killer whales for more than three years."
Martinez' entire employment history at Loro Parque was three years in length. His work with killer whales was less than that.
As the trial progressed, OSHA revealed that between 1988 and 2009:
SeaWorld generated 100 incident reports, twelve of which documented injuries (or, in A. M.’s case, death) to trainers:
Not every event of undesirable behavior by a killer whale resulted in an incident report. Chuck Tompkins is the corporate curator for zoological operations for SeaWorld Parks & Entertainment (Tr. 352). He acknowledged that SeaWorld failed to document several known events of undesirable behavior by killer whales when working with trainers: "[W]e missed a few." (Pg. 24).
Flaherty Clark also testified under oath that SeaWorld was not associated with Loro Parque. OSHA's report argued differently:
Sea World Parks & Entertainment sent several of its personnel, including SeaWorld of California's supervisor of animal training, Brian Rokeach, to Loro Parque to demonstrate its use of operant conditioning and to help implement its training program.
SeaWorld's killer whales are still housed at Loro Parque, but since Martinez' death, waterwork has been suspended. In America, trainers returned to the water just two days
after he was killed. Two months later in Orlando, Florida, Brancheau was killed by Tilikum.
Brancheau's death led to an OSHA investigation and a final court ruling that at last offered SeaWorld trainers some protection. During showtimes, a federal judge ruled, trainers and orcas had to be separated by a physical barrier. It is a ruling that SeaWorld continues to try and overthrow.
One thing that all of the trainers appeared to agree on was that an apex predator who is also highly intelligent, can form a deadly combination. John Jett told DJ:
One thing we know is that wild orcas will block exits from their prey items. Keto and Tilikum and many others seem to demonstrate this when attacking trainers. They seem to eliminate or reduce escape routes. These animals have an amazing evolutionary history that isn't removed from life in captivity. They're wired for this kind of behavior. They don't make mistakes.
Berg agreed that this this is an issue often overlooked during killer whale training. She added:
The more intelligent an animal is, the more training becomes a subjective rather than objective science. And some animals are simply more predictable than others. But that does not negate the fact that every killer whale in SeaWorld's collection is capable of doing what Keto did — it's just that the majority of them choose not to do it. I wonder how many Seaworld trainers are truly aware that this is the case. In my experience, I believed that as long as I did the right thing, I wouldn't get hurt. Clearly, that's not the whole story.