One of the few witnesses to the sinking of a sealing vessel last March called an investigation into the fatal accident a "joke" Thursday after learning that simulations of the incident were done in ice-free, open waters.

Sealer Wayne Dickson scrambled to help the crewmen of L'Acadien II after it overturned while under tow by a coast guard icebreaker in ice-clogged waters off Cape Breton, leaving three Quebec fishermen dead and one presumed missing. Two other crew members survived.

He was mystified Thursday when the Transportation Safety Board of Canada said recreations of the incident were conducted in water that had none of the large ice cakes thought to have caused the accident, and were being considered a key part of the investigation.

"It's just a joke," Dickson, whose boat was trailing behind L'Acadien II the night of the accident, said from his home on the Îles de la Madeleine, Que.

TSB 'confident' board clear on what happened

"How do you simulate something when the conditions are entirely different?"

His comments came the same day the TSB announced that investigators had completed their draft report into the accident after conducting interviews with crew aboard all vessels involved, collecting data and reviewing communications from the ships.

Lead investigator Donald Eaves said he was "very confident" the board had a clear understanding of the events leading up to the capsizing, especially after monitoring the two separate sea trials carried out by the coast guard.

The simulations took place near Halifax harbour in June, with another coast guard icebreaker towing a boat the same size as the 12-metre L'Acadien II.

The second set was done last month in the Baie des Chaleurs, N.B., using another fishing boat and a sister ship to the Sir William Alexander, the coast guard icebreaker involved in the accident.

Too difficult to replicate large floes

Eaves said the trials were invaluable and meant to help them "understand the dynamics of what would cause the accident the way it did."

But when asked why they weren't done with ice or similar obstructions, he said "we didn't think that was as important because when the accident happened, both vessels had entered a pool of open water."

He also said it would have been too difficult to replicate the large ice floes for the trials, even though it's widely accepted that the much smaller sealing boat was hauled up onto a large ice cake by the icebreaker and went over on its side before being pulled into the frigid waters.

Dickson and two survivors aboard L'Acadien II said the Alexander began towing the boat through the ice after the smaller vessel lost its rudder and issued a call for help.

What appeared to be an uneventful tow ended when the big ship swerved to dodge a massive chunk of ice in the otherwise relatively clear waters.

The ice cake ended up in the direct path of L'Acadien II and flipped it almost immediately, but the icebreaker continued to tow it and apparently dragged it into the ocean.

Water rushed into the boat while three sealers — including the captain — slept in their bunks down below. Their bodies were later recovered from the vessel, while a fourth man who was on watch is missing and presumed to have drowned.

Dickson, who was captaining the nearby Madelinot War Lord, and the survivors insist no one was monitoring the tow from the stern of the Alexander, as is recommended practice in the coast guard's manual.

But Eaves said they have electronic data and anecdotal evidence to prove that crew members were observing the procedure, refusing to elaborate on the data.

"We are convinced that there were lookouts on the quarter deck of the Sir William Alexander at the time of accident," he said in an interview.

Dickson also alleges that the icebreaker's lights suddenly went off as they were scanning the dark waters for survivors, and that no one on the icebreaker's deck responded to repeated, frantic calls to stop when the boat overturned.

Eaves said he couldn't comment on the claims while the draft report is undergoing a review by interested parties for the next 30 days.

He said those issues will be addressed in the final report, due late this fall, which might also include safety recommendations in the leadup to the sealing season this winter.