Knowing when to doubt what you see
Posted on 01 November 2010
Written by Jim Flannery
Overreliance on electronics without other references is among the lessons from a maxi-yacht grounding
An internal inquiry by Australia's premier ocean racing club has found that GPS inaccuracies, a poor lookout and overreliance on electronic navigation contributed to a hard grounding in which a maxi-yacht broke apart on rocks at night, leaving two dead during a race last October off southeast Australia.
While sailing at about 15 knots, the Reichel Pugh 80 PricewaterhouseCoopers grounded on a ledge just 90 feet from visible rocks near the northeast tip of Flinders Islet. Winds were 15 to 20 knots and seas offshore were running 3 to 4 feet on top of 6- to 8-foot swells. The grounding occurred at 2:35 a.m. Oct. 10.
When the boat was about 218 yards from the islet, crewman Matt Pearce, who was at the bow, felt the vessel surge and begin to surf, and then he heard breaking waves. He looked forward and was shocked to see the boat heading directly for breaking water and a low rock shelf. He yelled aft, "Come away! Come away!" Skipper Andrew Short turned sharply to starboard to try to clear of the end of the islet, but it was too late.
"The boat stopped dead like a car crash," says a report from the Cruising Yacht Club of Australia that was released earlier this year. "People were thrown about on deck and gear was thrown about below decks."
In the shallows, the waves and swells combined into breakers about 8 to 10 feet high. They picked the yacht up and slammed it down on the ledge and then hurled it onto the rocks nearer to the shore. Crewman Peter Geddes ran below to start the engine, but a line fouled the prop, killing the engine in about 30 seconds. The mast broke and fell over the side. Then the 8.5-ton keel bulb snapped off, along with the rudder.
"The waves would lift the boat and cause it to roll to starboard and strike rocks, and then as waves receded the boat would be sucked about 6 meters [19 feet] clear of the rocks and roll back to port. ... The rolling motion was violent and the angle of heel to starboard at least 45 degrees. The crew were being swept around the deck; it was difficult to hang on and they were under water for a considerable time, having difficulty breathing," according to the report, which was based on testimony from the yacht's crew.
One series of waves sucked veteran racer Sally Gordon, 47, out of the cockpit and through the port lifelines. Two of the crew grabbed her tether and tried to pull her aboard, but she was unresponsive. Another wall of water threw them across the deck before they could further assist Gordon.
Meanwhile, skipper Short was standing at the port wheel. A wave slammed him against it and swept both him and the wheel overboard. A crewmember heard the 48-year-old skipper cry for help, but by this time everyone was just holding on for dear life. Short was not wearing a PFD, a safety harness or a tether.
Another wave swept the skipper's son Nicholas, 19, over the side between the boat and the rocks, sucked him under PricewaterhouseCoopers and carried him clear of the islet. Once he was away from the yacht, he inflated his PFD and was later rescued.
Fortunately, a wave lifted PricewaterhouseCoopers far enough onto the islet so that the remaining crew could scramble onto the rock ledge - 15 of them in 30 seconds - even with the boat bucking in the surf. They then made their way to higher ground. "The total time from the initial grounding to all being off the yacht was only about four to five minutes," the report says.
Responding to two flares that the survivors set off on the island, the crew of the yacht Ragamuffin discovered Gordon floating face down in the water. She had suffered a head injury and, although she had been wearing a safety harness at the time of the grounding, she was not wearing it when she was found. She was not wearing a PFD. The report surmised that Gordon, a veteran of 15 Rolex Sydney Hobart Yacht Races and the recipient of the Cruising Yacht Club's 2000 ocean racing crew of the year award, was ripped out of her safety harness when the boat broke apart.
The crew of the yacht Quest found skipper Short, also a veteran of 15 Sydney Hobart races, floating in the water. A police boat rescued Short's son Nicholas. Another of his sons, Mitch, 14, was among those who were able to abandon the boat and make it to shore. Police divers found the yacht's rig north of the grounding site, the severed keel bulb, a small piece of the bow and a "great deal of wreckage" strewn along the shore - much of it broken up into very small pieces.
"The inquiry could not identify a single reason or cause of the accident, but we have identified a series of errors in judgment, which by themselves could have been inconsequential, but combined are contributory factors to the tragic grounding," the report concludes.
Among the key contributory factors cited in the report was the lax navigation of PricewaterhouseCoopers, including the failure to keep a proper lookout and to leave enough sea room while rounding the north end of the islet. In addition, the report cited an overreliance on the GPS/chart plotter. The report identified these factors as the "most significant" in the grounding, but it went on to say that these factors probably were compounded by the "inaccuracy of the GPS at the critical time the boat approached" the island.
The inquiry team consisted of Chris Oxenbould, a retired rear admiral in the Australian navy with experience as a commander and navigator; David Kellett, a former Cruising Yacht Club commodore, veteran of 35 Sydney Hobart races and member of the executive board of the International Sailing Federation; and John Brooks, a racer, military and commercial pilot for 40 years, and vice president of the World Sailing Speed Record Council.
In the report, they say that "statements from the crew suggest that the chart plotter was the sole input into the boat's safe navigation and it was being used to determine a safe distance to pass to the north of the islet and clear of dangers. What is difficult to understand is why a visual reference was not made to the islet."
The crew reported seeing the high ground on the islet's south end from four miles away. At three miles out, skipper Short set a course for roughly the northern edge of Flinders. Crew on deck could see the whole islet, including the low ledge on the north end, at one mile away.
"The inquiry cannot understand why the course was not adjusted to pass clear to the north and why, with 18 people on board, somebody did not draw attention to the fact the boat was heading straight for the rocks," the report says.
At the wheel, Short - consulting the plotter - "showed no concern about the navigation situation and was not seeking any confirmation from the bow or any other crewmember sitting on the windward rail," the report says. "The conclusion is that the chart plotter must have been indicating the boat was safe and would pass the island at a safe distance. This may have been the case with the GPS inaccuracies being experienced at the time."
The inquiry team notes that Short's organizational style was "relaxed." In addition, Short elected not to delegate some key responsibilities to members of the crew. "The brunt of the workload fell on Andrew Short as the skipper, navigator and helmsman, and that evening he was on the helm for all but a few minutes - a total time of about seven hours," the inquiry team writes. On other boats that the crew had raced on, helm changes occurred every 90 minutes.
The race had started at 8 p.m. in Sydney and Short had spent a full day at work. He then spent seven hours at the helm. The team thought that Short might have been fatigued, although the crew said they saw no evidence of that.
While acknowledging the prerogative of any skipper or owner to lead or manage a boat based on personal preferences and styles, the inquiry team says it prefers a more regimented, less relaxed style of organization. The team also says it prefers a clear division of labor and crew assignments. "It has the advantage of being safer, more flexible [and] able to deal with and monitor more things at the same time. Furthermore, a more structured organization balances the workload and mitigates the very real issue ... of fatigue."
Bad timing, bad luck
The "GPS inaccuracies" were the result of an unusual confluence of events related to the satellites that the chart plotter's GPS used to determine the boat's position. The inquiry team interviewed Greg Halls, a hydrographic surveyor and yacht navigator, who says he had been working on a tunneling project under Botany Bay that required extremely accurate GPS positions and 24-hour-a-day monitoring of the GPS system.
"In the early morning of 10 Oct., the survey had to be stopped between 2:15 a.m. and 4:15 a.m. because the GPS accuracy was outside the strict parameters set for the project," Halls told the inquiry team.
PricewaterhouseCoopers ran aground at 2:35 a.m.
The yacht carried two Garmin chart plotters: a model 5012 on deck that Short was working with and a model 5008 at the nav station below. Both units shared a Garmin BlueChart g2 Vision SD card with advanced cartography, three-dimensional maps and satellite images. However, neither unit was equipped to receive differential GPS, which could have provided greater accuracy. They were only capable of autonomous positioning - the least precise type of GPS, with no corrections. According to Garmin, DGPS would have provided an accurate position to within 10 to 16 feet, on average, compared with about 50 feet without DGPS. Both models were equipped with alarms that would have alerted the skipper to the degraded accuracy. The report does not say whether the alarms were set.
Based on detailed satellite data from Oct. 10, the dilution of precision - a measure of the quality of a GPS position - was very degraded between 2 and 3:30 a.m. and could have resulted in an "error possibly in excess of 100 meters [328 feet]," the report says.
At the time, six or seven satellites were visible to the GPS receiver at Flinders Islet, but only three or four were usable. The others were low on the horizon - under 30 degrees. The signals from those satellites were likely to be of poor quality because they had to travel farther through the atmosphere, and they were more likely to deflect off waves, rigging, deck hardware and crew than signals from overhead. This was especially so with the boat heeled and the antenna lower to the water in lumpy seas.
That wasn't the only problem, according to the report. "A close examination of the sky plot [of the satellites] indicates that three of the four usable satellites were on the same bearing or within 20 degrees of the reciprocal bearing, which produces a poor-quality fix," the report says. The most accurate GPS position readings are from satellites spread across the sky, not clumped together or in a line. On top of that, accuracy of the hydrographic charting of Flinders Islet used in the chart plotter reportedly was plus or minus about 54 yards in horizontal distance.
Healthy GPS skepticism
The report provides some sound advice for all boaters. "What is displayed on a chart plotter needs to be interpreted and not accepted as 100 percent accurate and without error. A safety margin needs to be added when using a GPS/chart plotter, but more importantly a second source of positional information should be used to check and verify the boat's position relative to the danger. This could be a visual reference if the danger can be seen, visual clearing bearings, depth soundings, radar clearing ranges, visual or radar fixes placed on a chart or any other means." A warning to that effect is given in the Garmin chart plotter owner's manual.
The inquiry team says it also is concerned about the "casual passage planning. ... At best, the passage plan is done by placing a few waypoints on, at times, a small screen with a cursory glance at the compressed detail available on the plotter. Close reference to the more detailed paper hydrographic charts, as directed by the chart plotter warning, is not being practiced."
The inquiry team notes that Short had voiced some concerns earlier in the race about a discrepancy between the displayed positions on the boat's two plotters, which the skipper apparently resolved to his satisfaction by rebooting both systems. The team also notes that the crew of PricewaterhouseCoopers said they had observed chart plotter errors in other races. In one race, the plotter showed the boat crossing reefs when the vessel clearly was in safe water. "This could have been a problem with the update interval in creating the past track or possibly a setup, installation or calibration problem on [the boat]," the report says.
In addition to critiquing the navigation on PricewaterhouseCoopers, the inquiry team has made the following recommendations:
- The Cruising Yacht Club should review its PFD and harness rules and consider making it mandatory for racing crews to wear them at night and at times of "heightened risk."
- Yachting Australia, the national sailing authority, should consider a requirement for a cutting or quick-release mechanism to separate a tether from a safety harness when the tether is under full strain.
- The yacht club should reiterate the need for all ocean-racing crews to carry personal locator beacons while on deck, and to encourage the use of PLBs fitted with GPS.
- Yachting Australia should consider a requirement to carry a grab bag with a hand-held VHF radio, two flares, and possibly an EPIRB inside that is readily accessible from the cockpit in the event of a catastrophic accident. This gear was not in the cockpit of PricewaterhouseCoopers. The boat lost all power in the earliest stages of the grounding, leaving the crew without a radio to send a mayday. The emergency gear was stowed in the pitch-black cabin, which was too dangerous to enter. Fortunately, one crewmember grabbed two flares - the only means of signaling their distress.
- The yacht club should remind owners of boats built before June 2001, which can carry life rafts below deck, that the raft must weigh less than 88 pounds, be stowed next to the companionway and be capable of launch in 15 seconds. The rafts aboard PricewaterhouseCoopers were stowed below, were not close to the companionway and could not be retrieved.
- The yacht club should recommend that boats be equipped with battery-powered emergency lights below and that each member of the crew carry a waterproof flashlight.
- Yachting Australia should consider requiring AIS transponders on racing boats as another way to lead rescuers directly to a boat in an emergency.
"Good seamanship, navigation and management on a boat significantly reduce the risks," the report concludes.
Launched in Sydney in 2000 as Neville Crichton's Shockwave V, PricewaterhouseCoopers also raced internationally as Morning Glory under the ownership of German Hasso Plattner. Short, owner of Andrew Short Marine, a Brunswick dealership with four locations in Australia, bought Pricewaterhouse
Coopers in the United States in 2008. Before taking the boat home, Short won line honors with her in the St. David's Lighthouse Division of the Newport Bermuda Race that year under the name Andrew Short Marine Shockwave V.
This article originally appeared in the November 2010 issue.