POSTED 9 NOV 2000 One minute, the Edmund Fitzgerald was visible on radar. The next minute, the 729-foot ore carrier was gone. On Nov. 10, 1975, the giant ship mysteriously sank on a routine run from Superior, Wis., to Detroit, laden with 26,000 tons of iron ore.

The S.S. Edmund Fitzgerald on the Saint Mary River, May 1975. Photo by Bob Campbell.

There was no SOS call. There was no word why the ship, whose experienced captain had just radioed that Fitzgerald was "holding our own," abruptly vanished, killing all 29 aboard.

How did the star of the Great Lakes fleet disappear beneath the waves, with her lifeboats lashed to the deck and her radio silent?

Twenty-five years after the tragedy that sparked the "The Wreck of the Edmund Fitzgerald," by Gordon Lightfoot, we have some answers.

So don your life jacket and board your lifeboat. The Why Files is heading into the cold, stormy waters of Lake Superior to learn why the Fitzgerald went down -- and whether it could happen again.

Bad weather ahead
When the Fitzgerald headed east on Lake Superior from the port of Superior, Wis., a storm was already heading for the Great Lakes. Over the next day and a half, poor weather forecasts -- and poor ship maintenance -- proved fatal.

Here's what happened, with thanks to our friends at the Cooperative Institute for Meteorological Satellite Studies at the University of Wisconsin-Madison.

Nov. 9, 1975 Fitzgerald leaves Superior, Wis.

7 p.m. The National Weather Service issues a gale warning for Lake Superior, predicting winds of 34-40 knots (one knot equals one nautical mile per hour, or 1.15 regular mph, so a gale ranges from 39 to 46 mph). Winds were predicted from the northeast, shifting to the northwest by the next day.

Nov. 10, 2 a.m. The captains of the Fitzgerald and the Anderson, a nearby ship, decide to head for the Canadian coast, to seek shelter from the northeast wind, and the waves it will generate.

7 a.m. The storm starts across Lake Superior.

2:45 p.m. Winds are shifting to the northwest, depriving the ships of shelter. Winds have reached 43 knots with waves between 12 and 16 feet. The Fitzgerald reports minor damage. A tilt to one side indicates that water has entered her hull, probably due to damaged hatch covers on deck.

7:10 p.m. The Anderson is in radio and radar contact with the Fitzgerald, whose captain reports "We are holding our own." Within minutes, the Fitzgerald abruptly disappears from the Anderson's radar.

Why did the good ship go down?
Why did the Fitzgerald sink? Why couldn't anyone escape? The answer starts with waves. You probably know that, except for tsunamis, wave energy comes from the wind. In fact, wave height reflects three factors related to wind:

• Wind speed: Higher speeds make higher waves.

• Wind duration: The more hours a wind blows, the higher the waves.

• Wind fetch: The longer the stretch of open water affected by a wind, the higher the waves.

While mariners cannot control wind speed or duration, they can reduce fetch, which explains why the two ships headed toward Canada. Unfortunately, crude forecasts concerning wind speed and direction put the Fitzgerald in harm's way. According to Jonathan Martin, assistant professor of atmospheric and oceanic science at UW-Madison, the storm was far more intense than predicted. While the Weather Service forecast winds from 28 to 42 knots, they actually exceeded 70 knots -- hurricane force. Gusts reached 85 knots.

Combine that with the premature shift of wind direction -- and leaks aboard the Fitzgerald that caused her to ride low in the water -- and you have a recipe for disaster.

Educated guesses based on examinations of the wreck indicate that the ship plowed into a wave. Her cargo shifted forward, her stern reared out of the water, and she sank before a radio call could be issued -- let alone before anyone could get into a lifeboat.

Forecast: better forecasts
Forecasts have improved since the sinking, says Martin, particularly the short-range ones that concern conditions less than 48 hours away. Better data from weather satellites ranks high among the reasons for the improvement, but Martin points to two other factors:

1. Meteorologists have learned to estimate (or "paramaterize") factors they can't pinpoint. "We've spent 25 years developing better ways to approximate some processes where we can't get a perfect answer," Martin says.

2. You guessed it -- better computers allow forecasting models to be far more sophisticated. The inputs and outputs of these computer programs describe conditions at uniformly spaced grid points. When the Fitzgerald went down, Martin says, grid points were 160 miles apart; today, they are 10 to 15 miles apart.

A grid is like a sieve. Just as sand falls through a coarse sieve, small, intense storms can fall through a coarse meteorological grid. A finer grid, in contrast, "catches" more storms, and allows finer-grained forecasts.

These improvements were tested precisely 13 years after the Fitzgerald sinking. On Nov. 10, 1998, a Lake Superior storm that Martin describes as "at least as intense, it was almost an exact replica, it developed very rapidly over the exact same path."

This time, however, ships had adequate warning, and "there were no incidents at sea."

In a way, better forecasts are a self-fulfilling prophecy. In 1975, if you were in port and a storm was forecast for tomorrow, you might consider sailing because two-day forecasts were accurate only 50 percent of the time.

Today's two-day forecasts are much more accurate, Martin says. In the same situation, "a person would not put himself at risk."

-- David Tenenbaum

More shipwreck tales at The Great Lakes Shipwreck Museum.