It has been more than five years since the engines of Southern Airways Flight 242 were literally drowned by a vicious thunderstorm that spawned at least six tornadoes and hail so heavy it shattered the airplane's windshield.
When pilot William W. McKenzie attempted a crash landing on the highway that bisects New Hope, Ga., he and 72 others died, but he left behind a stunning indictment of lethal flaws in how weather information reaches pilots.
National Weather Service radar had pinpointed the thunderstorm and identified it as intense at least 20 minutes before Flight 242 left Huntsville, Ala., but the Federal Aviation Administration's air traffic control system or the crew did not know that information until it was too late.
The crash jolted the federal bureaucracy into action for a time, but budget and jurisdictional issues between the weather service and the FAA have slowed installation of technology that could discover more timely information and speed its way to the cockpit.
Federal officials agree that many flaws remain in the reporting of weather information to aviators; those responsible for aviation safety are equally concerned, however, that pilots do not heed warnings adequately when they are given.
These issues merged in the crash July 9 of Pan Am Flight 759 in the New Orleans suburb of Kenner, La. The National Transportation Safety Board will examine that accident at hearings beginning tomorrow in Kenner.
Enormous progress has been made in the last decade in forecasting and tracking weather systems; at the same time, more has been learned about flying jet airplanes, which have proven remarkably reliable and safe. There are about 1,400 scheduled airline flights a day in the United States, and, before the Air Florida crash here Jan. 13, there had not been a fatal major airline crash in the nation since May, 1979.
The situation spawned complacency.
Maj. Gen. John E. Ralph, a retired Air Force pilot and now senior vice president for operations of the Air Transport Association of America, said, "I think we thought that most modern jet aircraft had the ability to fly through" severe weather conditions. New research, Ralph said, particularly into short-lived weather phenomena such as wind shear, "has cast doubt."
The dangers, even the existence, of wind shear have not been understood long. The safety board lists nine major airline accidents since 1973 in which wind shear was either a cause or major factor. The worst was in June, 1975, when 113 people were killed as Eastern Airlines Flight 66 from New Orleans crashed while trying to land at Kennedy Airport in New York.
A wind shear occurs when bands of winds traveling in different directions and at different velocities meet. Shears can accompany gust fronts moving through an area or, as is increasingly suspected in the New Orleans probe, can result from a small but powerful downdraft known as a microburst.
If a plane is flying at 25 knots above stall speed -- the speed at which it would stop flying -- and flies out of a 20-knot head wind into a 20-knot tail wind, it has encountered a 40-knot wind shear and is theoretically 15 knots below stall speed. In other words, it is about to plunge and crash.
It is generally agreed that big jetliners can pass safely through most wind shears if the planes are 500 feet to 1,000 feet above ground, because there is time and room to make corrections. Closer to the ground, however, the pilot risks destruction if he encounters wind shear.
That fact, some pilots and researchers believe, is not adequately stressed in pilot training or in flight scenarios created by simulators used in training.
"If you're going to show a guy wind-shear training, show him a wind shear he can't fly through," said William W. Melvin, a member of a panel of senior airline pilots provided by the Air Line Pilots Association for interviews.
Melvin also believes that safety board reports, in explaining what a pilot might have done to save the plane after encountering wind shear, lead "pilots down the garden path," encouraging them to think they can survive the condition "if they just have the Right Stuff."
Patrick Clyne, another member of the ALPA panel, put it another way:
"Operating in wind shear is a very common occurrence. It happens in probably a majority of flights you operate. . . . Relatively large-scale shear events . . . very seldom approach the magnitude where the airplane doesn't have enough performance reserve to get through them.
"The real question you're asking is why, on these rare occasions, don't crews recognize the imminent danger that they are in? In retrospect, I think the record's pretty clear. In almost every instance I'm aware of where we have a catastrophic wind-shear encounter, not only was the crew absolutely unaware they were approaching this, oftentimes they weren't aware even when they were in it how bad it was. . . . That was true certainly in New Orleans."
Clyne's assessment will be debated at the Kenner hearing, because there is no question that low-level wind-shear alerts were issued before Flight 759 took off and that its crew members heard them. The question appears to be one of risk assessment: did the crew members pay attention to what they heard, and, if not, why not?
John J. McCarthy directs a program called JAWS, for Joint Airport Weather Studies, a project of the National Center for Atmospheric Research in Boulder, Colo. JAWS studies wind-shear detection and forecasting, and a network of sensors and three state-of-the-art weather radar units around Stapleton International Airport in Denver have been used to find and track wind-shear events.
McCarthy and his team have confirmed that microburst wind-shear events occur regularly; they have discovered the remarkable fact that downbursts capable of creating airplane-killing shears can develop and dissipate within seven minutes but go undetected by current warning systems. Further, microbursts are born in both innocent-looking cumulus clouds as well as thunderstorms.
The FAA, reacting to the Eastern crash in New York, developed a low-level wind-shear alert system consisting of a series of wind gauges scattered around an airport. If the direction and velocity of winds at one of the outlying gauges differs from the wind at midfield by more than 15 knots, an alarm sounds and flashes in the control tower. The tower is then supposed to relay the information to flight crews, giving precise wind differentials.
That kind of alert was given in New Orleans, although the warning did not contain precise differentials, and crew members of Flight 759 did not ask for them.
"The bottom line is that pilots and controllers don't believe the alert . Pilots know damn well they've got to fly in wind shear; if they stopped flying every time there was a wind-shear alert, the whole system would be fouled up. So pilots and controllers say the alert is not accurate and timely," McCarthy said.
It is just a matter of time, McCarthy said, before wind shear causes another accident.
"The thing is," Clyne said, "the low-level wind-shear alert system is meant to be one of the things you consider in a decision to take off. . . . The existence of an alert in and of itself is not sufficient reason to suspend operations. . . .
"What we've been trying to say is that the weather information we're getting is incomplete in many cases; it can be enhanced by giving us better sensors. . . . Much of the stuff we're given is outdated, is incomplete, is misleading and there's absolutely nothing state of the art that it takes to solve that. . . .
"There's lots and lots of weather information out there that gets collected but never gets beyond the pad of the person that makes the observation, and that's the problem. . . ."
Ramon Alvarez is deputy director of the FAA's air traffic service. "I think the National Weather Service has got to go on record that they need more resources for aviation," he said. " . . . I would like them to emphasize aviation weather as one of their top priorities."
However, weather reports to pilots are not the top priority of air traffic control; in fact, with the exception of a "significant meteorological condition," weather information is listed in the air traffic control manual as an "additional service." The first priority of air traffic control is to keep airplanes from hitting each other.
Richard E. Hallgren is director of the National Weather Service, part of the Commerce Department's National Oceanic and Atmospheric Administration, and he serves many masters. The weather service is responsible not only for aviation, but for all U.S. weather forecasting in general and such specifics as tornado and hurricane alerts, severe storm warnings, flood predictions and information for farmers.
"I desperately want to do a better job of getting weather information flowing, especially in the terminal airport areas," Hallgren said in an interview. Aviation weather information, he said, "has got to be better, the detection system has to be improved."
But forecasting for a terminal area, or a specific runway, is a difficult, expensive proposition. "You're trying to define weather in an area that is three miles long and one-tenth of a mile wide," he said. Because significant weather events at airports are "small-scale and short-term, we handle forecasting on a decentralized basis" through regional centers that produce forecasts for many localities.
The FAA is one recipient of those more generalized forecasts, and it passes the information to pilots and airline dispatchers. Information for a local area is frequently two to three hours old when handed to a pilot, even though a microburst can form and dissipate within seven minutes and a severe thunderstorm can become powerful and move across a vast area within three hours.
That happened to Southern Flight 242. After that crash, National Weather Service meteorologists were assigned to each of the 20 regional air traffic control centers operated by the FAA. These centers handle air traffic in several states and coordinate departures and approaches at major airports within their areas. The meteorologists are supposed to give the controllers, who talk to airports and pilots within their sectors, more accurate and immediate weather information.
The first Center Weather Service Unit, as these are called, was set up in Atlanta, the regional center where the Southern crash occured. That unit was given a radar terminal with which the meteorologist could dial almost any of the several hundred weather radar antennas in the United States, assess the situation and inform air traffic controllers of problems in their sectors.
That was four years ago, and everybody who has seen it says the Atlanta radar is wonderful. The other 19 regional air traffic control centers will receive their radars "by next summer," according to the weather service.
It has taken time to write the specifications and acquire the equipment under competitive bids with constrained budgets, according to the FAA and weather service personnel. The FAA is buying the equipment and paying the salaries of the meteorologists.
Fred Foss is chief of the Center Weather Service Unit at the regional air traffic center in Longmont, Colo., which covers parts of several states and has Stapleton as its biggest airport. Foss has been the beneficiary of more data than most centers receive because he is getting an experimental feed from many of the instruments being used in JAWS and other projects centered around Stapleton.
"We have found," Foss said, "that once we got sharp in using the equipment, we were having a direct impact on" operations at Stapleton. In other words, once the weather information became believable for controllers and pilots, they began to pay attention to it.
With the benefit of sensitive sensors, Foss could predict with great certainty at one point that a huge thunderstorm would roll across the airport in 15 minutes, and it did. The information was relayed to the tower, and takeoffs were delayed.
Just as importantly, because of that prediction, controllers in the regional center where Foss works were able to delay incoming planes some distance from Denver and avoid the multiplane holding patterns that are despised by pilots and controllers and that encourage poor decision-making in cockpit and control room.
The radar that can spot wind shears and other types of wind shifts is called a Doppler radar. The Doppler effect says that if a sound such as an automobile horn moves toward the listener, it will have a different pitch and tone than it will have as it moves away. The same effect applies to weather fronts as they move toward or away from a Doppler radar antenna.
Technology has reached the point at which a color radar display can indicate the speed of moving air. If there is a wind shear, it will be shown on the radar screen, because winds would move sharply in one direction, then calm, then move sharply in another direction.
If the Doppler radar were placed to look down a runway, it could accurately spot a wind shear. However, reading the Doppler accurately requires training, and it takes time to interpret results and relay them to the cockpit.
The Doppler radar will be the next nationwide weather radar system and is called NEXRAD, which is being jointly funded by the Commerce, Transportation and Defense departments and is supposed to be in place by 1990. It will not cover all airports' runways, though, Hallgren said.
Meanwhile, JAWS' McCarthy said, his team has learned much about the kinds of conditions in which wind shear can occur that can be detected by a pilot looking out the window before takeoff.
McCarthy is trying to find funds to make a film showing conditions that presage wind shear, notably dust blowing along the ground and a virga shaft -- wisps of precipitation that evaporate before they reach the ground and show as an arrowhead descending from a larger cloud.
In a heavy thunderstorm, he said, "you have to say wind shear may be present, so I'm not going.
"Civil airliners should not take off in a high rain core -- and they all do that."