IAFD/McCoy Report on Audrey Mestre's Death

Publisher Note: This report is reproduced "as-is" from the official IAFD report that is due to be published later today. We are providing no commentary in this piece and this does not include any of the appendices or attachments described in the document. Editorial commentary and analysis will follow in due course.

On October 12, 2002, Audrey Mestre, the well-known free-diver, tragically perished during her No Limits World Record attempt. Since that time, her husband, Francisco "Pipin" Ferreras and the International Association of Free Divers (IAFD) have attempted to understand what went wrong with the dive, leading to Ms. Mestre’s tragic death. This Report, and attachments, summarize that which is known and some conclusions about the tragic event.

In developing this Report, a number of sources have been consulted. Perhaps the most important is the detailed report of Kim McCoy. McCoy is a physical oceanographer and diver with decades of experience on and under the water and was present for Mestre’s final dive. He is affiliated with Ocean Sensors, which specializes in the design and manufacture of oceanographic and water-quality instrumentation. McCoy has observed and certified numerous world record free-dive attempts, and provides high resolution electronic data to measure the precise depth, velocity and duration of dives. The computer (OS500 D manufactured by Ocean Sensors), worn by Ms. Mestre during her final dive, provided ‘flight-recorder-like’ information. Neither McCoy nor Ocean Sensors are compensated in any way by the IAFD or Pipin Ferreras. McCoy is involved in academic and environmental research and is the Marine Technology Society Chair for Oceanographic Instrumentation.

McCoy gathered additional information from the Mares dive computer worn on Audrey’s wrist, the Aqualung dive computer on Audrey’s leg and several other Mares and UWATECs dive computers worn by various divers. Interviews with over a dozen people were conducted before and after the dive. Measurements were made of the descent and ascent fixtures including weights and volumes. Finally, mathematical tools were applied to help explain certain processes. Historical records, videos and photographs from the many previous training and record dives from 1995 through 2002 also were referenced.

An official autopsy was conducted in accordance with the laws of the Dominican Republic and concluded that the death was accidental. This report relies on that autopsy as well as Kim McCoy’s report, and on interviews and observations conducted directly by the IAFD team.

The conclusions reached by McCoy and the IAFD are very similar. McCoy’s report is reproduced in its entirety as an attachment to this report. In summary, the following conclusions were reached. There was no single cause for the tragedy that befell Audrey Mestre. Rather, many different factors, ranging from technology to ocean conditions, contributed to the events of October 12th. One factor was the new, thinner diameter cable used in this dive. This new cable was intended to minimize friction with the sled. However, due to the reduced weight of the new cable combined with a different type of weight at the bottom, the cable was free to move to the side, an effect that was noted by Kim McCoy. The full impact was not apparent during training dives because of excellent weather conditions. On the date of the world record attempt, however, the weather was stormy and much windier than during the preceding two weeks of training. While the waves on October 12th were not in and of themselves so problematic as to require a cancellation, they apparently exacerbated the movement of the cable, resulting in several starts and stops during the ascent. One of these "stops" occurred at 164 meters, at which time the sled did not move up or down for approximately 30 seconds.

It was also discovered later that a Teflon bushing installed to reduce friction between the cable and the sled appears to have been damaged during a training dive. An impact between the sled the support boat seems to have compromised the integrity of the Teflon bushing.

Equally important was the fact that there seems to have been some drift sideways by the bottom weight. Although a lead weight was used in most dives, on this occasion a concrete weight was used. In addition, wings were fitted on the camera attached to the sled in order to keep the sled from rotating during descent. The wings apparently created a lateral force, which continued as the sled descended further down the cable. The combination of the lateral force created by the wings, the different density of the bottom weight, and the lower drag than usual cable combined to cause inadequate tension on the cable and additional drift of the bottom weight. As a result, neither the descent nor the ascent was completely vertical, and the sled was forced to advance on a slight incline. Mr. McCoy found that this incline caused some "hydrodynamic instabilities at low speeds", which in turn led to a "braking" of the ascent bag on the slightly slack cable, exacerbated by surface wave heave movement.

The autopsy was conducted by Dr. Danyd Moquete Mendez and Dr. Ana Falete Mercedes on October 13, 2002. It is a matter of record in the Dominican Republic. The autopsy report indicates that blood and urine samples were examined and that neither alcohol nor drugs were detected. It notes that the lungs were found augmented or increased in size (which would not be uncommon for a well-trained free diver). The conclusions are reproduced in their entirety in the original Spanish, with a translation, which follows. "Conlusion: el deceso de la nacional francesa Audrey Anne Marie Mestre Ferreras, se debio a asifixia por submersion. MUERTE ACCIDENTAL." "Conclusion: the death of the French national Audrey Anne Marie Mestre Ferreras, occurred because of asphyxia by submersion. ACCIDENTAL DEATH."

There was also a suggestion that the lift bag was inadequately inflated or leaked. It appears that there was inadequate lift provided by the lift bag. One of the divers, in response to this observation, did successfully add some gas to the bag at the beginning of the ascent. It is possible that the compressed air bottle used for lift bag inflation was not fully filled, but this factor alone would not have been responsible for all the difficulties encountered. It is clear that there was some air in the compressed air bottle. The lift bag, upon inspection after the accident, did show a section of wear that could have been the cause of some leaking. However, the amount of air that would have leaked would not have been sufficient to have significantly affected the ultimate performance of the sled. The compressed air bottle did not show any signs of damage or leaking after the dive.

In summary, no single reason can be determined for the tragic death of Audrey Mestre. Many factors contributed to it and it may be the case that if any one had been different, the dive would not have resulted in the tragic ending which occurred. Since the dive, IAFD and Pipin Ferreras have tried to determine exactly what went wrong in order to learn for the future, to insure that Audrey’s death will not have been in vain. Several changes will be made. Among them, lead bottom weights will be used instead of concrete, and the previous, heavier cable will be utilized until further testing can determine how the lightweight cable impacted the October dive. The wing design will be changed to minimize lateral force. Acceptable oceanographic conditions under which to conduct a dive will be redefined based on what has been learned. Measures will be taken to insure that the descent and ascent is as vertical as possible. New protocols and checklists regarding dive-related equipment and activities will be implemented to insure adequate dive preparations.

The tragic death of Audrey Mestre has devastated her family, friends and colleagues, especially her husband and her parents. Free diving enthusiasts everywhere have been saddened by her loss. As a result of this investigation, it can now be said that the death of Audrey Mestre was a tragic but unforeseeable accident. Even in death, the lessons learned from her final dive, as well as her courage, dedication and passion, will be an inspiration to all free divers, and indeed, to all of those involved in testing the limits of human endurance.

Kim McCoy Report

"A man gazing at the stars is proverbially at the mercy of the puddles in the road."

The descent was only slightly abnormal, but abnormal. A ‘wing’ had been fitted to the camera on the sled to keep the sled from rotating. The wing created a lateral (sideways) force*. As the sled descended, the continual lateral force of the wing pushed sideways along its entire cable’s length. Below, the concrete bottom weight (15 kg in water) was swung to the side. The cable extending from the boat to the concrete weight was inclined (not vertical). There was minimal ocean current. The data indicates an unusual increase in speed during the last 10 meters of the descent.

*Calculated sideways force ~10 kg at a descent velocity of 1.6 m/sec.

The ascent was unusual compared to other dives I have observed. The return to the surface was delayed several times. The cumulative delay time was more than 60 seconds. The causes for delay are associated with decreased buoyancy and increased drag. The increase in drag was a result of the ascent-bag sliding along the inclined cable. During mid ascent (164 to 120 meters depth) there were several short periods (~2 secs) during which the ascent almost stopped. At 120 meters, 3:50 into the dive, the ascent gracefully slows and Audrey, like a ‘leaf in autumn’ begins to descend. Blackout. She descended for only 15 seconds before Pascal Bernabe intercepted and reversed her descent. Audrey was transferred to Pipin and returned to the surface at 8:38.

Time depth plot of 12 OCT 2002 dive (divide samples by 4 to get seconds) (Training dives)

IAFD/McCoy Report on Audrey Mestre's Death freediving competition records  report pipin mestre mccoy kim iafd ferraras death audrey

My opinion is based upon the following sources of information:

Instrumentation: (used during Audrey’s dive into history)

  • OS500 D
    (on Audrey, oceanographic instrument 4 measurements per second e.g. 4 Hz)
    (Accuracy +/- 0.05 meters, useable resolution 0.005 meters)
    (Maximum data acquisition rate 1000 samples per second e.g. 1000 Hz)
    (Corrections made for water density and local gravity)
  • Mares dive computer
    (time and depth computer on Audrey’s wrist)
  • Aqualung dive computer
    (max depth recorder and elapsed time, on Audrey’s leg)
    (3 worn by Pascal Bernabe on his wrist)
  • Dive computer
    (1 worn on Wiki’s wrist, Mares)

People: (all personally interviewed by Kim McCoy)

  • Audrey Mestre (I have know Audrey and her parents for 6 years)
  • Pascal Bernabe (Mixed gas diver 170 meter)
  • Eduardo ‘Wiky’ Orjales (air diver 90 meters)
  • Denis Bourret (air diver 60 meters)
  • Matt Briseno (surface support diver 0-30 meters)
  • Orlando ‘Tata’ Lanza (surface support diver 0-30 meters)
  • Carlos Serra (IAFD)
  • Francisco ‘Pipin’ Ferreras (IAFD)
  • Bill Stromberg (AIDA observer, surface observer, in water)
  • Robert Margaillan (photographer and journalist, longtime friend of Audrey)
  • Eddie Matos (first aid, EMT, on catamaran)


  • Historical
    (data from training and record dives from 1995 to present)
  • Videos
    (from training and record dives 1995 to present, sled and mid-water)
  • Still photographs
    (from training and record dives 1995 to present, sled and mid-water)
    (Film and digital images with embedded time codes)
  • Mathematical tools
    (algorithms used in hydrodynamics)
  • Tape measures
    (correlation of sled cable length on land and of depth gauges)
  • Weights
    (estimated weights of the components of the sled, weights, cable, etc.)
  • Observed similar events
    (Pipin’s Cozumel blackout dive, Pipin’s Cabo San Lucas aborted 130 meter dive, Audrey’s Canary Island extended time at depth, Pipin’s aborted dives. All events instrumented by McCoy

Descent: 1 min 42 secs count down of -5 to zero minutes (completed by Carlos Serra)

Zero: (Audrey delayed a few seconds after zero was reached, commensurate with prior dives)
(Audrey was focused; no indications of excessive stress or discomfort)

Time: 0:00 (Zero all references to time referenced to Audrey’s complete immersion below surface)

Time: 0:04 (V=1.50 m/sec)

Time: 0:31 (depth 50 meters attains maximum V=1.90 m/sec)

Time: 1:00 (depth 100 meters V= 1.75 m/sec)

Time: 1:34 (depth 159 meters V=1.62 m/sec Pascal bangs on his tank as signal of approach)
(Cable ‘impulse’ of unknown origin; initially believed (falsely) to be impact with diver)
(Slight increase in V ~+0.30 m/sec, perhaps from change of sled or body positions)

Time 1:42 (large cable ‘impulse’ supported by all sources as sled reaching bottom of cable)

Ascent: from 1min 43 to 8 min 38 secs.
(Visual observations of Audrey indicated normal bodily actions, no distress)

Time 1:59 (depth 169 meters, upward increased V~0.30 m/sec normal for initial ascent from depth)
(Surface waves and boat motions visible in pressure record T=~ 7 seconds)
(Audrey does not request assistance)

Time 2:12 (depth 165, V=essentially 0, waves clearly visible in record indicating cable interaction)
(Ascent is impeded; Pascal attempts to add gas to ascent-bag)
(Audrey does not request assistance, appears calm)

Time 2:42 (depth 164, V=0.6 begins to rise uniformly)
(Pascal observes Audrey ascending above him for 2-3 secs)
(Pascal resumes his ascent; Audrey has ascended above him with ascent-bag)

Time 3:00 (depth 153, abruptly slows to V=0.0 for ~2 secs, then resumes V=0.8)
(‘Breaking’ of ascent-bag on cable, several other locations during ascent)

Time 3:30 (depth 136, V=0.8 fairly constant velocity during this rise portion

Time 3:50 (depth 120, V=0.0 ascent slows and begins to descend)

Time 4:05 (depth 124, V=-0.3 descending when Pascal reaches Audrey 15 secs after unconscious)

Time 6:00 (depth 91 V=0.0 no longer safe for Pascal to ascend, prepares to transfer)

Time 7:03 (depth 89, V=0.5 ascending Audrey transferred to Pipin for final ascent)

Time 8:38 (depth 0, V=0.0 surface reached)

Time 9:39 (Audrey removed from water and placed on port side of catamaran)

Estimated small boat transfer time to beach 5 to 6 min. (based on my stopwatch earlier that day)

Audrey was a well-trained athlete. I have come to the following conclusions about the accident.

1) The ascent bag provided inadequate lift (much less than normal) at bottom

2) There was inadequate tension on the cable (bottom weight ~15 kgs + cable weight ~23 kgs)

3) Wings (400 cm2) on the sled camera caused an unexpected force (~10kg sideways) on the cable.

4) The descent has an increase in speed just prior to bottom arrival, associated with the relative location of the bottom weight on the cable as the sled approached the bottom. Observed cable ‘impulse’ could be associated with the sled-cable-weight interactions.

5) Ascent bag upper bearing insert was marginal (slight damage) which increased drag.

6) An attempt to add gas to the ascent bag was made by Pascal Bernabe

7) The ascent was impeded by hydrodynamic instabilities at low speeds. The ascent was non-vertical, which caused ‘braking’ of the ascent bag on slack cable (at regular surface wave intervals).

Publisher Note: This report is reproduced "as-is" from the official IAFD report that is due to be published later today. We are providing no commentary in this piece and this does not include any of the appendices or attachments described in the document. Editorial commentary and analysis will follow in due course.