TSB featured article

This article was published by Ship & Boat International in their March/April issue.

This article was also published 23 March in SWZ magazine.

The knockdown and capsize of Concordia:
Lessons to learn

By Jonathan Seymour, FICS, MNI

On 17 February 2010, at approximately 1423 local time, the sail training barquentine Concordia was knocked down and rapidly capsized after encountering a squall off the coast of Brazil. All 64 crew, faculty, and students successfully abandoned the vessel into life rafts. They were rescued two days later by two merchant vessels and taken to Rio de Janeiro.

That everyone survived is remarkable. It was not an easy abandonment. Several life rafts could not be launched, and a distress call could not be made. Much of the credit goes to safety preparedness: there was twice the number of life rafts on board than required by regulation, the life jackets had recently been moved to deck lockers to improve accessibility, and regular safety drills had taken place. Youth and agility also played a significant role, though conditions in the life rafts proved difficult. The EPIRB (Emergency Position Indicating Radio Beacon) functioned as designed and its position was established within an hour of the knockdown. However, the contact number for the EPIRB was long out of date, a false alarm was suspected, and it took many hours before a comprehensive search and rescue attempt was initiated.

Concordia was registered in Barbados and the owner was a Bahamian company. However, its ties to Canada were substantial.

Photo of the sail training barquentine Concordia
Photo 1. Sail training barquentine Concordia (Credit: Matt Jacques Photography)

The charterer was a Canadian school; management was located in Canada; and Lunenburg, Nova Scotia, was its home port. Because of this, and the fact that many witnesses were located in Canada, the Transportation Safety Board of Canada (TSB) investigated the accident. The Board's report (M10F0003) was published on 29 September 2011.

Initial reports on the capsize all seemed to agree that the knockdown was caused by a vicious microburst associated with a thunderstorm and that there was nothing that could have been done to avoid it. This quickly became “common wisdom,” and the focus of media comment moved on to the lengthy delay in recovering the survivors.

But the evidence available to the TSB, which included video and photographs, along with statements from those on deck at the time, suggested that the winds never exceeded Beaufort force 7–weather conditions that Concordia had encountered many times in its 20 years of ocean voyaging. Satellite imagery was obtained and weather experts were consulted. In comparison to thunderstorms known to have produced microbursts, it was clear that this one was not sufficiently formed to do so.

So what happened? To find out, it was quickly decided that a thorough assessment of Concordia's stability was required.

We were fortunate to have good data to work from: the ship's plans, a copy of the stability booklet, on board observations, plus the video and photographs. All this proved sufficient to allow one of our naval architects to develop a computer model of the vessel and analyze its stability. (A full report on the vessel's stability assessment can be found on the TSB's website.) The model was then used to generate the righting arm curve and compare it to the theoretical wind heeling arm curve for the sail plan in use at the time and a range of wind speeds. These calculations were verified against the known conditions and events of the day to ensure reasonable accuracy. The results were intriguing.

The righting arm curve and wind heeling arm curves for various steady wind speeds are shown in diagram 1 for Concordia on the day of the occurrence.

Righting arm curve and wind heeling arm curves for various steady wind speeds
Diagram 1. Righting arm curve and wind heeling arm curves for various steady wind speeds

The steady heel angle is where they intersect. Up to about 27 knots of wind, the response is as one would expect: more wind induces more heel which in turn induces more righting force, resulting in a reasonable and steady angle of heel. But somewhere between 27 and 37 knots, the two curves substantially coincide over a significant range of heel, starting at about 38°. In strengthening near–gale conditions, the angle of heel rapidly increases to a partial knockdown.

As the heel approaches about 70°, the righting arm increases as the deck houses start to submerge and provide additional buoyancy (provided that the hull and deck houses are watertight). Indeed, the model suggests that Concordia would have retained positive buoyancy after being knockdown onto its beam ends and would, therefore, likely have recovered once the wind abated.

But the hull and deckhouses were not secured. Not only were the leeward doors open, but so were the engine room skylight and numerous vents. To reflect this, the model was modified to remove the buoyancy of each of the deck houses once water reached the first openings. The deck houses still provide some residual buoyancy but over a more limited range, as shown in diagram 2.

Diagrm 1. Stability model modified to remove the buoyancy of each of the deck houses once water reached the first openings
Diagram 2. Stability model modified to remove the buoyancy of each of the deck houses once water reached the first openings

This would not last for long, however, because of downflooding into the hull through other unsecured openings.

But that still may not account fully for the rapid increase in heel between 70° and 90°. So was another factor potentially in play?

Although a microburst did not occur, squalls associated with thunderstorms can contain some downdraft element that reaches the ocean surface (see figure 1).

This diagram shows a typical thunderstorm front.
Figure 1. Typical thunderstorm front (Credit: NTSB Report MAR-87/01)

So, what happens if the winds are inclined? Although not immediately intuitive, this can be shown by shifting the heeling arm curve for a horizontal wind to the right on the x–axis. For example, to look at the effect of winds inclined to 30° from the horizontal, the horizontal wind heeling arm curve can be shifted 30° to the right. As can be seen in diagram 3, when this is done, any residual buoyancy from the deck houses is totally overcome, and the vessel goes over on to its beam–ends.

Diagram 3. Effect of horizontal wind heeling arm curve shifted 30° to the right

So what we have from the Concordia model is:

  • a vulnerability to a rapid increase in heel angle at a relatively modest wind speed,
  • a loss of buoyancy due to downflooding through open doors, windows, and vents; and
  • a possible inclination of the wind that overcomes any remaining residual buoyancy provided by the deckhouses.

In combination, these factors would result in the rapid knockdown and capsize.

All well and good–as explanations go–but Concordia was no spring chicken; rather, it was a seasoned veteran of many ocean voyages. While we do not know if the vessel had been knocked down before and recovered, we do know that, by reputation, the vessel had always been sailed conservatively, with minimized heel angles so as not to adversely affect classes.

Here's a final element to consider: Concordia was built in Poland and was originally flagged in the Bahamas. The Bahamas followed the UK rules and required that the vessel have ”squall curves“ included in its stability book (maximum steady heel angle to prevent downflooding in gusts and squalls). These rules were introduced in the UK following an inquiry into the loss of the Marques in 1987. When consulted by a deck officer who is knowledgeable in their use, these curves show when the margin of safety is being eroded and, therefore, when mitigating action–such as shortening sail or altering course�is required.

So what happened on Concordia? The Master handed over the con at 1200 to the officer of the watch (OOW) with a shortened sail plan that was ”good to 40 knots“ and instructions to bear off if the wind increased. The vessel was on a broad reach and making about 5.5 knots in 15 to 20 knot winds. The Master's standing orders required that he be called if the vessel was at risk. There was no discussion of the squall curves at the handover. This is key. In fact, the OOW, who was appropriately certificated for his position, was not aware they were on board, did not consult them, and was not trained in their use.

Had the stability book been consulted, it would have shown that, in order to provide a margin of safety in typical gusting conditions, the heel angle should be limited to 24°. Instead, as the squall approached, the angle of heel increased from around 10° to about 23° in wind speeds of about 23 knots. At this point, the margin of safety may have just been sufficient for a gust, but certainly was not sufficient for the encounter with the squall when it came. The OOW, not perceiving the risk to the vessel, was not concerned. But the heel angle then quickly increased and, within a very short time, Concordia was on its beam ends with the deckhouses flooded. The OOW's attempt to change course was too late.

TSB recommendations
  • The (Canadian) Department of Transport ensure those officers to whom it issues sailing vessel endorsements are trained to use the stability guidance information that it requires to be on board sailing vessels.
  • The (Canadian) Department of Transport undertake initiatives leading to the adoption of international standards for sail-training vessels on the provision of stability guidance to assist officers in assessing the risk of a knockdown and capsize, and for the training of officers in the use of this information.

Ultimately, no single factor caused this accident. Rather, it was a combination of the inherent limits of the Concordia; a lack of knowledge of these limits on the part of the OOW; the associated lack of awareness of the developing risks; and, consequently, the absence of mitigating action to reduce sail, change course, or secure the vessel watertight.

As a result of its investigation, the Board made two recommendations, which–if adopted–should help ensure that a similar accident does not occur in the future. In essence, every sail training vessel should have the necessary information on board to define its individual vulnerability to specific weather conditions, and its deck officers should be trained to use that information.

Jonathan Seymour FICS, MNI, was a Member of the Transportation Safety Board of Canada from 1999 until he retired from the Board in 2011