Microburst windshear research in Colorado in 1992. The photo was taken from a tail-mounted camera that photographed the aircraft's upper fuselage and the horizon.

The TCV 737 made numerous contributions in many different areas over its 20 years at the Langley Research Center. But no effort involving the 737 was more successful, or had a greater impact, than the airborne windshear detection and avoidance program.

Forward flight deck of the NASA 737 at the time the aircraft entered a microburst windshear cell in 1992.

Members of the Aeronautics Advisory Committee that oversees the research efforts of NASA called the windshear program "NASA at its best."[Ref 5-1] The windshear program was, indeed, a classic research success story, and its achievements were even more extraordinary because of the significant organizational and technical challenges the program involved. The windshear research effort required an unusual number of cooperative relationships, not only among several different directorates at Langley, but between NASA and both the FAA and industry manufacturers. The research was subjected to a high degree of public pressure and scrutiny, with a deadline by which results had to be achieved. The technical obstacles were substantial, and many knowledgeable sources doubted they could be overcome at all. If technical solutions could be found, the flight testing would still be unusually challenging, and potentially dangerous if not planned and executed extremely well.

Yet the NASA and industry researchers and technicians involved with the program rose to the challenge. In a remarkable sevenyear effort, the windshear research team at Langley developed, demonstrated, and successfully transferred the technology to tame an aviation weather hazard that had caused 26 airline accidents and claimed more than 500 lives since 1964. [Ref 5-2]

The immediate catalyst for the NASA/FAA windshear program was a tragic event at the DallasFt. Worth (DFW) airport on the afternoon of August 2, 1985. Thunderstorms were in the area of the Texas airport as Delta Flight 191, a L1011 jumbo jet with 163 passengers and crew on board, approached runway 17L for landing. There was a rain shaft and scattered lightning coming from a thunderstorm cell in the airliner's final approach path, but the pilots decided the weather was passable and continued the approach. Fifteen to 30 seconds after the L1011 entered the weather, however, the rain and lightning intensified, and the airplane was buffeted by a violent series of up and down drafts. The headwind increased rapidly to 26 knots, and then, just as suddenly, switched to a 46 knot tailwind, resulting in an abrupt loss of 72 knots of airspeed. The jet was only 800 feet above the ground when it encountered the severe weather, leaving the pilots little room to maneuver when the airplane began to lose airspeed and altitude at the same time. Thirtyeight seconds later, Delta Flight 191 crashed into the terrain short of the runway, killing all but 26 of those on board. [Ref 5-3]

Windshear display used on the NASA 737 research aircraft; displays the location and intensity of the weather hazard. The aircraft's position is at the bottom of the cone; the dark red areas indicate teh greatest shear intensity. The red dot shows the F-Factor.

The crash of Flight 191 jarred the nation. Just three years earlier, a Pan American World Airways Boeing 727 departing the New Orleans International Airport had also encountered a severe windshear and crashed, killing 145 people on the airplane and eight more on the ground.[Ref 5-4] Two such accidents in three years, with the loss of almost 300 lives, was too much. With the wide publicity the Dallas crash received, "windshear" suddenly became a household term and the safety of air travel began to be questioned. Members of Congress were deluged with phone calls from constituents asking that something be done to prevent any more accidents.[Ref 5-5]

The 737 crew conducting windshear research from the aft flight deck. Clockwise, from lower left, Roland Bowes, program manager; Lee PErson, research pilot; Michael Lewis, deputy program manager; experimenters Emedio Bracalente and David A. Hinton.

Actually, the FAA, NASA and other organizations had been working on the problem of low level windshear for some time. As early as 1949 researchers had begun to explore some of the potentially dangerous characteristics of thunderstorms, and in 1971 the FAA started a joint program with several research organizations to work on improving the forecasting and detection of windshear. The issue received renewed emphasis, however, when an Eastern Airlines Boeing 727 crashed on approach to runway 22L at the John F. Kennedy International Airport in New York on June 24, 1975, killing 112 people. [Ref 5-6]

Information from meteorological measurements and another aircraft in the area indicated that the airliner had encountered a strong downdraft of air and an abrupt switch in wind direction just prior to the crash. Researchers analyzing the accident looked more closely at the thunderstorm conditions the airliner had encountered and concluded that a particularly violent, deceptive, and deadly kind of windshear, which one researcher named a "microburst," had caused the Eastern Airlines jet to crash. [Ref 5-7]

Windshear, defined as a sudden change in wind velocity and/or direction, was not always dangerous. Aircraft often experience a shift or drop in winds in cruise flight, but at that point they have enough speed and altitude to compensate for the loss of airflow over their wings. If an airplane encountered windshear during a takeoff or landing it was more serious, because the plane was close to the ground and had very little extra speed. A microburst windshear, however, was a unique phenomenon that posed a particularly deadly threat to airplanes.

A microburst occurred when the precipitation in a column of rising air evaporated, cooling the air very quickly. Since air became more dense as it cooled, the column of air would fall rapidly, spreading out in all directions with a great deal of force as it neared the ground. The phenomenon could occur in a variety of conditions, and not all microbursts had any rainfall associated with them. However, the strong, convective air currents in thunderstorms and towering cumulus cloud buildups were particularly conducive to the formation of microbursts.

Meterologist Dr. Fred H. Proctor on the ground directing the NASA 737 toward microbursts using a Terminal Doppler Weather Radar.

Low altitude encounters with microbursts were especially hazardous to aircraft because the first effect a pilot would notice was a performanceenhancing headwind (as the plane first encountered the outflow of the burst). If an airplane was on a landing approach, the pilot would typically respond by reducing the engine power to maintain the proper glide path angle and speed. As the plane progressed into the center of the microburst, however, the plane would be hit with severe downdrafts. Then, as it passed into the far side of the burst, the headwind would be replaced with a strong tailwind, causing a sudden loss of performance and airspeed. If the pilot had reduced power during the first stage of the microburst, this loss in performance would be intensified even further. Turbine engines take several seconds to spool up to provide additional power, and by the time a pilot realized the nature of the problem, it was often too late. [Ref 5-8]

Another reason microbursts posed such a danger to pilots was that they were extremely difficult to detect. A microburst was usually less than 2.5 miles in diameter and lasted only a few minutes. To help give pilots better warning of potentially dangerous shear conditions, the FAA developed a Low Level Windshear Alert System (LLWAS) in 1976. LLWAS consisted of an array of wind velocity measuring instruments that were installed at various locations around an airport. The LLWAS compared the wind direction and velocity readings from the different sensors and, if a 15 knot or greater difference existed, transmitted an alert to the air traffic controllers, who could then notify pilots in the area. The system had a number of limitations, however. The instruments could not measure winds above the ground sensors and could not record vertical wind forces. An extremely localized microburst on a final approach path might not even be recorded by the sensors. Even alerts that were recorded took a couple of minutes to reach the controllers.

Nevertheless, the LLWAS was an improvement over the existing detection methods, which consisted of weather forecasts and pilot reports of weather conditions they had just encountered. By 1983, the FAA had installed LLWAS at 59 major airports and had plans to install the system at an additional 51 locations.[Ref 5-9] The FAA also issued special Advisory Circulars in 1976 and 1979 that contained information for pilots on windshear recognition and recovery techniques. In 1977, the FAA amended Part 121 of the Federal Aviation Regulations to require air carriers to adopt an approved system for obtaining forecasts and reports of adverse weather conditions, including low altitude windshear. The agency even issued an advanced notice of proposed rulemaking (NPRM) in 1979 that proposed making airborne windshear detection equipment mandatory for scheduled airlines, but the regulation was not enacted. [Ref 5-10]

In truth, the technology for airborne windshear detectors and the level of knowledge about the phenomenon itself were still very limited, although researchers were working on the problem. For example, engineers and meteorologists at the Langley Research Center began to explore windshear and microburst behavior in more detail in the early 1980s as part of a simulation technology program. In order to simulate the weather hazard accurately, the engineers had to first understand how it operated and how it affected aircraft performance. Understanding windshear behavior was a difficult challenge, however. The NASA researchers first had to study the basic meteorology and atmospheric physics associated with windshear and then, using extremely high fidelity fluid dynamics models run by super computers, try to model the phenomenon's characteristics. [Ref 5-11]

In another effort known as the Joint Airport Weather Studies (JAWS), the National Center for Atmospheric Research (NCAR) conducted windshear experiments in the Denver, Colorado, area during the summer of 1982. Researchers concluded that some of the microbursts recorded during the JAWS program created windshear too severe for landing or departing airliners to survive if they encountered it less than 300500 feet above the ground. [Ref 5-12]

Yet even as the JAWS researchers were collecting microburst data in Denver, Pan American World Airways Flight 759 crashed in New Orleans, Louisiana. The airport's LLWAS did trigger an alert, but not in time. The tower controller's first broadcast warning of possible windshear came two seconds after the Pan Am Boeing 727 hit the trees off the end of the runway.[Ref 5-13] The accident prompted Congress to pass Public Law 97369, mandating the FAA to contract with the National Academy of Sciences (NAS) to "study the state of knowledge, alternative approaches and the consequences of windshear alert and severe weather condition standards relating to take off and landing clearances for commercial and general aviation aircraft."[Ref 5-14]

The NAS report was issued in May 1983 and concluded that low altitude windshear presented an "infrequent but highly significant hazard to aircraft while landing or taking off." Among the report's recommendations was continued research into airborne windshear detection systems. [Ref 5-15] That November, the FAA cleared the way for certification of airborne windshear detectors by issuing Advisory Circular No. 12041, "Criteria for Operational Approval of Airborne Windshear Alerting and Flight Guidance Systems."[Ref 5-16]

Researchers aboard the NASA 737 during windshear detection flight tests.

The crash of Delta's L1011 in Dallas in August 1985, however, provided a dramatic catalyst that suddenly turned the search for an effective weapon against the microburst into a top priority, organized program. In addition to the fact that it was the second windshear accident with a large number of fatalities in three years, the Delta crash highlighted just how inadequate the detection and warning technology still was. The DallasFt. Worth airport was equipped with a Low Level Windshear Alerting System, but none of the system's five sensors issued an alert until several minutes after the Delta airplane hit the ground. [Ref 5-17]

Among the Congressmen who were swamped with concerned calls from constituents following the Dallas crash was Representative George Brown from California. A couple of weeks after the accident, Congressman Brown visited the Langley Research Center and asked for a presentation on windshear. An FAA manager named George "Cliff" Hay and Dr. Roland L. Bowles, a Langley research engineer, had been working on a plan for an airborne windshear detection research program for six months prior to the Dallas crash, but the only concrete research that was actually underway at the Langley Research Center at the time was the work on microburst and windshear modelling. Dr. Jeremiah F. Creedon, head of the new Langley Flight Systems Directorate, put together a quick briefing on the windshear problem and potential technologies that might be able to combat it. Congressman Brown asked how much money it would take to develop a solution. Nobody at Langley had thought that far down the road, but Dr. Creedon gave the Congressman a rough estimate of at least several million dollars. Brown reportedly commented that the amount of money Creedon had quoted was nothing. "It's a lot of money if you don't have it," Creedon replied. The Senator whispered to an aide, and as the contingent left the briefing, the aide told one of the Langley managers that NASA had just gotten itself a windshear program. [Ref 5-18]

The hazard that NASA tackled: aircraft approaching a runway are espcially at risk from sudden and intense downdrafts brought on by abrupt emperature changes. Developing a reliable warning system was the only answer.

Actually, it was not quite that simple. Congressman Brown, who was a ranking member of the House Science and Technology Committee, certainly had a lot of influence with both the FAA and NASA. But the Delta crash had created a tremendous amount of public and political focus on the windshear problem. The media coverage following the Dallas accident, public concern, and the interest of highranking Representatives and Senators all helped garner support for a substantial, coordinated interagency research effort to address the windshear problem.

In April 1986 the FAA announced the formation of a National Integrated Windshear Plan. The plan was an umbrella program that incorporated numerous independent research efforts, some of which were already in progress.[Ref 5-19] Even before the Dallas crash, for example, the FAA had begun working on a Windshear Training Aid for the airlines. The project was a joint effort by all three major commercial transport airframe manufacturers (Boeing, Lockheed, and McDonnell Douglas), and the training program and techniques they developed to help pilots handle severe windshear situations proved extremely effective. [Ref 5-20]

In addition to the training aid, the FAA National Plan incorporated LLWAS, research into an improved ground detection method called Terminal Doppler Weather Radar (TDWR), plans to improve terminal area communications of windshear threats, more detailed characterization of the windshear threat, and an airborne windshear detection research program.[Ref 5-21] This airborne system research program would be conducted by the Langley Research Center. In addition to its aeronautics expertise and the microburst modelling work that had already been done there, Langley had one other asset that made it a logical choice for the windshear research: a fully instrumented air transport test plane, with advanced displays, that could take the sensor technology development through a test flight stage. On July 24, 1986, NASA and the FAA signed a Memorandum of Agreement formally authorizing the start of a joint Airborne Windshear Detection and Avoidance Program. A Windshear Program office was created in the Flight Systems directorate at the Langley Research Center, headed by Dr. Bowles. [Ref 5-22]

The FAA/NASA airborne windshear research program had three major goals. The first was to find a way to characterize the windshear threat in a manner that related to the hazard level it presented for aircraft. The second was to develop airborne remotesensor technology to provide accurate, forwardlooking windshear detection. The third was to design flight management concepts and systems to transfer that information to pilots in such a way that they could respond effectively to a windshear threat. [Ref 5-23] The program also had to pursue these goals under unusually tight time constraints.

Part of the pressure came from the fact that both Congress and the general public were demanding a solution to the windshear threat as soon as possible. An even greater motivating factor, however, was a proposed FAA regulation that would greatly aid the transfer of any new technology developed in the NASA program if the research could be completed quickly enough. Since 1979, the FAA had contemplated requiring air carriers (operating under Part 121 of the Federal Aviation Regulations) to install airborne windshear detectors in their airplanes, and the proposed regulation was brought up again after the Dallas accident. It was finally enacted in September 1988 and required Part 121 air carriers to install airborne detectors in all their aircraft no later than December 1993. The minimum requirement was only for a reactive system, but Northwest Airlines, American Airlines, and Continental Airlines obtained exemptions that allowed them two more years to explore forwardlooking detection systems. [Ref 5-24]

In order for the technology being researched by the NASA/FAA windshear program to have a real impact, the airlines, who were the ultimate customers, would have to support forwardlooking detection systems. Three major airlines had already expressed an interest in forwardlooking systems, but if NASA's research was not completed until after the airlines had outfitted their entire fleets with reactive systems, airline support for the new technology would be greatly reduced.

The emphasis on forwardlooking systems reflected a growing realization of how severe the microburst threat could be. The JAWS research in Denver in 1982 discovered microbursts too strong for an airplane to fly through safely, no matter what kind of immediate recovery techniques the pilots used. [Ref 5-25] In its official report on the 1982 New Orleans Pan Am crash, the National Transportation Safety Board (NTSB) noted that although reactive systems could help improve pilot performance, "programs must be pressed to develop airborne and ground systems with greater lead time predictive capabilities." [Ref 5-26] The FAA Windshear Training Aid also warned that the maximum windshear capability of jet transports in some situations was 40 to 50 knots wind speed change, and that "some windshears cannot be escaped successfully (once they are actually entered)." [Ref 5-27]

Advance warning would give pilots the ability to increase the engine power and, if necessary, level the airplane before entering the microburst, so the airplane would have more energy, altitude, and speed with which to combat the effects of the windshear. Or, if traffic permitted and the shear was strong enough, the pilots could elect to maneuver around the microburst altogether. One of the first questions the Langley windshear researchers had to answer, however, was exactly how much advance warning was required to prevent an airplane from getting into a shear situation that it could not fly through safely. The Langley simulation technology program had already yielded significant insights about the characteristics of windshear and microbursts. The researchers integrated this information with analyses of aircraft energy states, tested their results in piloted simulations of microburst recovery techniques, and concluded that even 1520 seconds of advance warning was enough for an airliner to avert or survive an encounter with a microburst windshear. In fact, simulations indicated that aircraft with 1020 seconds of advance warning could fly through even relatively strong windshears without losing any altitude. [Ref 5-28]

A flat plate airborne radar used in the microburst windshear detection research program was housed in the nose of the NASA 737 aircraft.

In order to allow the pilots to take appropriate corrective action, however, the warning had to convey how much of a hazard the impending shear posed for the airplane. Some windshears could be penetrated safely with only small power additions; others were more dangerous and required immediate full thrust or evasive maneuvers. The solution devised by the Roland Bowles, manager of the Langley windshear program, was a hazard index that he called the "F-Factor."

The F-Factor was a dimensionless number that interpreted the vertical and horizontal strength of a windshear in terms of the amount of climb performance it would take away from the airplane. Or, to put it another way, the F-Factor of an windshear would indicate how much excess power an airplane would have to have to fly through it without losing airspeed or altitude. A typical twin engine jet transport plane, for example, might have engines capable of producing .17 excess thrust on the F-Factor scale. So if a windshear registered higher than .17, the airplane would not be able to fly through it, even at full power, without losing airspeed or altitude. Information from past windshear accidents indicated that the warning threshold for most jet transports in landing or takeoff configurations should be an F-Factor of .1. Consequently, a cockpit warning display could be preset to only show alerts for windshears with an F-Factor of .1 or more to eliminate nuisance or unnecessary warnings. [Ref 5-29]

The invention of the "F-Factor" was an important step in the development of windshear detection systems, because it provided a way for information from any kind of sensor to be presented to the pilot in a relevant and easily understood form. But the hazard index was only one of five basic requirements the windshear research team at Langley had identified for an effective forwardlooking detection system. The technology also needed to identify hazards while rejecting nonthreatening information, locate the position and track the movement of a potentially dangerous air mass, and annunciate the hazard to the flight crew. In addition to an "FFactor" type of hazard index, the display also needed to provide information on the proximity and volume of the windshear. [Ref 5-30]

By 1986, Boeing and the Sperry Corporation were already in the process of developing reactive airborne windshear systems that would alert pilots to windshear once it was actually encountered. But technology with the capability the Langley researchers wanted, especially in a forwardlooking system, did not yet exist. [Ref 5-31] In fact, although the NASA/FAA windshear program listed the design of forwardlooking windshear detection systems as one of its three primary goals, it was still not known whether such technology was even possible. The 1983 National Academy of Sciences (NAS) report on LowAltitude Windshear and Its Hazard to Aviation discussed three different potential techniques for remote sensing of windshear, but found problems with all three that made them impractical. [Ref 5-32] Even the 1986 memorandum of agreement that initiated the FAA/NASA windshear program noted that "there is no assurance that a practical airborne forwardlooking system capable of detecting both wet and dry severe windshear 'microbursts' can be achieved." [Ref 5-33]

Close-up of the side-mounted infrared sensor used in the windshear detection program.

Despite their drawbacks, the forwardlooking technologies that seemed to have the most potential were the microwave Doppler radar, Doppler light detecting and ranging (LIDAR), and passive infrared radiometry systems discussed in the 1983 NAS report. Microwave Doppler radar operated by transmitting radio waves of uniform frequency ahead of the airplane. The waves would be reflected back to the plane when they hit water particles, and the frequency shift of the return signal would indicate the direction and velocity of the raindrops and, therefore, the wind. Just as the whistle of a train would be a high pitched noise as it approached and would become a lower and lower pitched sound as it went further away, the frequency of the radar return signal differed in direct proportion to the speed and direction of the water particles. If they were coming toward the airplane, as in the case of a headwind, the frequency would be high. If the particles were going away from the plane, the frequency would be proportionately lower. A variation of the frequency, and therefore the velocity and direction of the water particles, would indicate the presence of a windshear.

Close-up of the Light Detection and Ranging (LIDAR) radar mounted on the lower fuselage.

The Terminal Doppler Weather Radar (TDWR), which was a microwave Doppler radar system, was proving very effective in ground detection of windshear, but there were several problems with trying to adapt the technology for an airborne system. First, the ground Doppler radars were much larger and more powerful than the size equipment that would fit on an airplane, so an airborne version might not be as effective. The radar would also be looking not only ahead, but also down toward the ground as the airplane descended. As a result, the radar would be reflected off objects on the ground, creating "clutter" in the return signal. In addition, not all microbursts contained rain, and microwave signals typically received few returns from dry air.

Doppler LIDAR operated in much the same manner as Doppler radar, except that it used a laser light beam instead of radio waves. The LIDAR system also used reflections off tiny dust particles in the air instead of raindrops to determine wind direction and velocity. The size of existing LIDAR equipment posed a potential problem for its use in an aircraft environment, but the biggest concern about LIDAR was that its signal tended to be absorbed by raindrops. Consequently, its signal was weak, or "attenuated," in the presence of the kind of heavy rain that was often found in thunderstorm microbursts.

During the microburst windshear detection research laser beams, shown here, were used to align the optical hardware of the infrared and LIDAR systems.

Passive infrared radiometry was based on the premise that since microbursts were formed by a column of cooler, rapidly descending air, their presence would be marked by a sharp temperature shift in the air ahead of the airplane. The technology was simpler, less expensive and lighter than the other systems, but it had several potential problems. First, there was no firm evidence that microbursts or gust fronts were the only weather phenomena containing temperature shifts, so there was a potential problem of nuisance alerts with an infrared system. Even if the nuisance alert problem was solved, there would have to be a direct relationship between the amount of temperature change and the severity of the windshear in order for the system to be an effective warning device. [Ref 5-34]

To try to overcome the technical obstacles to the different types of forwardlooking airborne sensors, NASA enlisted the help of several industry manufacturers. Lockheed Missiles and Space was given a contract to develop an airborne forwardlooking LIDAR detection system, and funds from NASA's Small Business Innovative Research (SBIR) program were awarded to a Boulder, Colorado, company named Turbulence Prediction Systems to develop an infrared radiometry sensor. [Ref 5-35]

The radar windshear detection technology was developed by the researchers at Langley, but the system's basic component was a specially modified Model 708 Xband weather radar built by Rockwell Collins, Inc. The NASA engineers had developed a radar and ground clutter simulation model at the beginning of the windshear program, which they had been using to explore various radar design and signal processing methods. Their work indicated that by making some design modifications to the radar, using filtering in the signal and data processing, and managing the tilt of the radar antenna, the ground clutter problem could be eliminated without diminishing the radar's ability to detect windshear. [Ref 5-36]

The NASA researchers also designed two additional warning systems that did not use forwardlooking airborne technology. The first was an improved "in situ" reactive system that used airspeed, accelerometer, angle of attack, groundspeed, and other data from aircraft sensors to verify when windshear was actually encountered. Although reactive systems were already being developed by commercial manufacturers, the NASA version had more comprehensive, threedimensional capabilities. A comprehensive and precise reactive system was critical to the windshear research because the "In Situ" detector was the "truth" measurement against which the accuracy of the forwardlooking systems would be judged.

The final detection system was simply a VHF radio data link that would allow information from a ground Terminal Doppler Weather Radar (TDWR) to be transmitted to the airplane directly. The standard TDWR design transmitted the information from the radar to a tower controller, who then had to transmit a verbal caution to flight crews. Clearly, a data link would improve the timeliness of windshear warnings to any airplane landing at an airport equipped with a TDWR. But the NASA researchers also planned to use TDWR information, uplinked to the TSRV 737 airplane and processed to indicate the "FFactor" hazard level of any indicated windshear, to help them locate microbursts for flight testing the forwardlooking sensor systems. [Ref 5-37]

Technicians service the tail-mounted reciever on teh NASA 737 used in the windshear detection research project.

Developing the various sensor technologies took several years. In the process, the researchers made extensive use of computer models and piloted simulations developed by engineers at the Langley Research Center. Potential versions of the different detection systems were "flown" against computer models of past windshear accidents many times to determine how accurately the hazards were detected and measured. Yet especially with tricky, unstable phenomena like windshear and microbursts, computer simulations could only go so far. In order to get a true sense of the accuracy and performance of the sensors, they had to be tested in actual windshear conditions. In May 1990, a second Memorandum of Agreement was signed between the FAA and NASA to support a flight test program for demonstrating and validating the advanced wind shear detection systems. [Ref 5-38]

Flight testing this particular technology raised some significant safety considerations, however. The microburst windshear that was so lethal to aircraft occurred only close to the ground. To flight test the windshear sensors, NASA's 737 test plane would have to intentionally fly into microburst conditions at low altitude. At best, the tests would be intense, turbulent encounters with an extremely severe weather environment. But if the sensors underestimated the severity of a microburst, there would be much less margin than usual to correct for the error.

To minimize the risks, the researchers at Langley conducted an unusually thorough examination of the flight test plan before it was approved. First, research pilots "flew" possible flight test scenarios in the Transport Systems Research Vehicle (TSRV) fixed base piloted simulator to establish operating procedures with adequate safety margins. After experimenting with numerous parameters and procedures, the researchers drew up a list of guidelines to ensure flight safety during the tests. The first was to minimize weather exposure. The team would work with isolated thunderstorm cells, but would avoid cells embedded in frontal systems. Furthermore, although the penetrations would be flown by the safety pilots in the forward flight deck, who could see the actual conditions the airplane was encountering, they would be guided by the pilots monitoring the electronic windshear and navigation displays in the 737's experimental aft flight deck (AFD). Before each microburst was entered, the AFD pilots would give the safety pilots an escape vector in case the cell was stronger than expected. Ground obstacles that might be a factor at low altitudes would be programmed into the moving map display in the aft cockpit, as well, so the crew there could make sure the airplane stayed clear of them. [Ref 5-39]

Some weather risks could not be completely eliminated, however. It was going to be impossible to conduct microburst experiments, for example, without running some risk of a lightning strike. The biggest danger posed by lightning was that a spark might ignite the airplane's fuel, triggering a catastrophic explosion. So before the flight tests began, NASA called in specialists to thoroughly inspect and seal the 737's fuel tanks and system. Lightning could also severely damage the research equipment on board the airplane, so Langley technicians tried to harden the equipment against strikes as much as possible. While lightning still posed a threat to some of the equipment, the researchers felt the risks had been reduced to an acceptable level. [Ref 5-40]

Second, the researchers set firm operating limits for microburst penetrations. For cells with an "Ffactor" greater than .1, they set a minimum flight altitude of 750 feet above the ground and a minimum indicated airspeed of 210 knots. Microbursts with an Ffactor greater than .15 would be avoided. The plane would also stay clear of weather cells with extremely high levels of "reflectivity," or dense precipitation, to minimize the risk of hail damage. Even in less dense rain, however, the igniters in the 737's JT8D jet engines would be left on to minimize the chance of a flame out due to water ingestion. The researchers also planned a phased approach to the microburst experiments, starting with mild shear conditions and working up to the more severe storm cells.

Third, an extensive training program was developed for pilots and researchers who would be taking part in the flight tests. Flight crews were put through literally hundreds of simulated windshear penetrations so they could practice appropriate responses. The flight and research crew then underwent two weeks of rehearsal flights in the vicinity of the Langley Research Center and the Wallops Island Flight Test Facility before the team deployed for the actual microburst experiments.

Schematic of microburst windshear and its threat to aircraft. How the microwave detection system works.

The 737 was outfitted with some additional equipment for the windshear flight tests. The radar sensor used the flat plate Doppler weather radar that was already in the nose of the airplane, but a whole pallet of processing equipment for the system had to be installed in the cabin. The LIDAR also required a pallet of equipment in the cabin, but the actual sensor was attached to the underside of the 737's fuselage. The infrared sensor was installed in a window on the left side of the plane. Changes were also made to the research flight deck to allow the windshear data to be displayed. The electronic map and primary flight displays in the left seat position of the aft cockpit were left alone, but three of the cathode ray tube (CRT) screens on the right side were modified to display windshear information, and two additional CRT displays were installed above the standard instrument panel. Each of the two monitors above the instrument panel could be configured to show a realtime forwardlooking image from the television camera in the nose of the plane, the radar sensor display or the LIDAR sensor display. Another CRT screen showed information from the infrared and in situ sensors, and two additional monitors displayed the data linked information from the TDWR.

The display formats were created primarily by Langley engineers. Sensor data from the infrared and in situ detection systems was displayed in simple bar graphs indicating the "Ffactor" level of the hazard. The radar and LIDAR sensors were capable of detecting a much more comprehensive picture of the airspace ahead of the airplane, however, so the information from those sensors was processed into graphic, multicolor displays, coded to indicate levels of windshear severity. If the sensors detected a windshear over a .1 threshold, both displays would highlight the threat area with either a box, in the case of the radar display, or a diamond in the case of the LIDAR. The highlighted area would also display the numerical "FFactor" of the shear, and a large "ALERT" signal would appear on the screen.

The TDWR displays, which were the primary navigation aids in finding microbursts, integrated all the windshear detection information. The basic TDWR display would indicate microburst hazards with oval, race trackshaped icons. The icons were color coded, depending on the severity of the microburst, and a numerical indication of the Ffactor of each particular microburst accompanied the icons. When the radar or LIDAR forwardlooking sensors detected the shear, the warning box or diamond, with the Ffactor, would be replicated on the TDWR screen. Warnings from the infrared sensor were displayed in numerical format in a corner of the TDWR screens. When the airplane actually entered the microburst and the in situ sensor detected an Ffactor hazard over a .1 threshold, a red dot appeared over the microburst icon on the TDWR displays. The exact accuracy of the forwardlooking sensors could be verified by comparing the numerical data recorded from all five of the sensor systems. [Ref 5-41]

The initial flight tests were conducted during the summer of 1991, in Orlando, Florida, and Denver, Colorado. The locations were picked for two reasons. First, Orlando and Denver each had an TDWR system in the vicinity of the airport. Second, microbursts in Orlando tended to be "wet" cells, while a majority of those in Denver were "dry." By flight testing the sensors in both places, researchers could find out how well each system performed across a full range of microburst conditions.

The LIDAR equipment was not yet ready for flight testing in 1991, however, so the initial experiments involved only the radar and infrared systems. The first tests took place around the Orlando International Airport from June 10th through June 20th. A team of more than 50 researchers, technicians and meteorologists was involved, including researchers from the Massachusetts Institute of Technology's Lincoln Laboratory, who operated the Orlando TDWR equipment.

On a typical flight day, the research team received its first weather briefing in the late morning, since microbursts typically did not develop until the afternoon or evening hours. After the briefing, the researchers often had to endure hours of waiting before the right weather conditions materialized. When the meteorologists saw promising signs of microburst activity, the flight crew of nearly 30 researchers and pilots boarded the 737 and took off.

The flight operations themselves were complicated, because in order to use the TDWR to help locate microbursts, the NASA plane had to stay in the immediate vicinity of the airport. The researchers discussed the flight test plans and procedures with the airport air traffic controllers in advance, but the flight crew still had to stay in constant communication with controllers throughout the flight to avoid conflicts with other airport traffic. The crew also had to be in constant contact with the TDWR ground personnel for microburst and other weather information. Added to these communications were numerous other conversations over the airplane's internal intercom between the two flight decks and between the different research stations and the crew viewing the displays in the aft flight deck. Communication was actually one of the toughest challenges of the flight tests, as the work often required as many as five different voice channels to be active at the same time.

A test run began when the ground TDWR personnel notified the flight crew of a potential microburst target. Lee H. Person, the primary aft flight deck pilot for the 737 throughout its 20 years at the Langley Research Center, usually flew the airplane onto a final approach path to the microburst, because he had the advantage of the CRT navigation and TDWR displays. The team had to maneuver quickly, because microbursts were a shortlived phenomenon, lasting only five to 10 minutes. On approximately a two mile final approach to the microburst, Person would turn control of the airplane over to research pilot Kenneth R. "Dick" Yenni in the forward cockpit. Yenni, like Person, had been with the airplane since it arrived at NASA and was the primary safety pilot on the 737. Test runs were flown in level flight at altitudes between 750 and 1,000 feet above ground level, and before entering a severe microburst, the 737's airspeed was stabilized at 210 knots. The procedure was to enter each microburst with an indicated airspeed of 210 knots and then, as the plane encountered the initial headwinds, add power as necessary to maintain at least a 210 knot groundspeed. That way, the airplane would have enough power and airspeed to safely transit the microburst, no matter how great the windshear.

As the 737 approached the microburst, the activity in the airplane became highly focused and intense. The forward flight deck pilots were watching for air traffic, obstacles and unacceptable conditions, such as severe lightning. At the same time, they were keeping up a running conversation with the air traffic controllers, and listening for course corrections, hazard and escape vector information, and groundspeed reports from the pilots in the aft flight deck. In addition to the constant reports they were giving to the safety pilots, the crew in the aft flight deck was talking continually with with the ground TDWR personnel to make sure the windshear and reflectivity levels in the target microburst were still within acceptable safety limits and no other hazards had arisen. On two different intercom channels, researchers were giving status reports or positioning requests for the different sensors. Meanwhile, the weather around the airplane would be deteriorating rapidly as the 737 entered the kind of rainfilled thunderstorm cell typical of Florida microbursts. The plane would be buffeted by turbulence and rain would start to pound at the windows. The world outside the cabin became dark and threatening, lit up only by lightning flashes, which were sometimes so close the researchers could "hear thunder inside the airplane." The first few times out there were reportedly some wide eyes throughout the cabin, but after a while, the flights became almost routine to the research crew. Even during the first few flight tests, however, the sensor monitoring, status reports, and communications among crew members continued uninterrupted through even the most turbulent storm encounters.

In all the flight testing in Orlando, there were no "safety of flight" issues, but there were more than a few tense moments. Once, the weather conditions around the airplane suddenly closed in, and the 737 crew found the airplane headed up a blind canyon of thunderstorm cells that were all beyond the safety limits set for the flight experiments. Fortunately, the safety pilots found a slightly lighter section of clouds between two cells and were able to steer the plane through the hole without incident. On several other occasions, the plane came close to being hit by lightning, but it never actually took a lightning strike. It did, however, run into hail in one of the more severe microbursts it penetrated.

To obtain the data required, NASA windshear researchers intentionally flew in conditions most pilots try to avoid.

Following the Orlando experiments, the plane and its research crew travelled to Denver Colorado for two more weeks of flight testing on July 8 July 24, 1991. The flight tests were staged out of Denver Stapleton Airport and used the NCAR "Mile High" Doppler ground radar in nearby Boulder to help locate the microbursts. The procedure was similar to that followed in the Orlando tests, although the conditions were expected to be somewhat different. The researchers were hoping to encounter "dry" microbursts in Denver, which would present severe windshears without the heavy rainfall of the Florida storm cells. Unfortunately, the weather refused to cooperate. The only windshears the researchers found in the two weeks at Denver were generated by gust fronts. The gust front shears were still quite strong, however, and they caused some unique flight conditions. Several of the fronts threw up dust clouds from the ground, and some of the 737's lowlevel test flights were at or below the tops of the dust clouds. As a result, the research crew found that they had discovered a new and unusual way to get the airplane dirty.

The results of the 1991 flight experiments showed that the forwardlooking radar detection system successfully identified and tracked high hazard areas in flight, although the low moisture microbursts that were expected to present the greatest challenge to radar were not tested. The in situ warnings also correlated well with the other detection systems. The infrared sensor did not fare anywhere near as well, but another round of flight experiments was scheduled for the summer of 1992, when the LIDAR equipment would also be tested. [Ref 5-42]

The Langley crew boarding the Boeing 737 research aircraft in preparation for a windshear detection flight at Orlando, Florida.

The 1992 flight tests very nearly did not happen, however. The plan was to repeat the basic 1991 schedule, with flight tests first in Orlando and then in Denver. Everything was going according to schedule until two months before the Florida tests, when several problems arose that should have cancelled the entire summer's test flights. First, both of the 737's engines developed problems and had to be overhauled or replaced. Then, even before the engine work had begun, maintenance technicians discovered some debonding in certain lap joints of the 737's fuselage skin. This same problem had caused an Aloha Airlines 737 to lose a section of its fuselage in flight, so the NASA airplane was grounded until all the affected areas could be riveted together. This was a major maintenance project, and under ordinary circumstances, the airplane would have been down for several months. [Ref 5-43] But this was not an ordinary airplane or program.

Throughout the history of the TCV/ATOPS program at Langley, the 737 and its research projects had attracted a unique type of person. Taking research all the way to flight required much more detail work and presented a lot of unknowns and obstacles that were never encountered in laboratory tests. Projects that included flight testing also forced researchers to focus on real world, practical applications of their ideas. It took an enthusiastic, pragmatic and resourceful person to enjoy working on flight test research programs, and the people who worked with the TSRV 737 were often characterized as having an energetic, "can do" attitude about their projects and the airplane. But the power of their resourcefulness and dedication was never more evident than in the 1992 windshear test program.

When the 737 technicians and researchers got the bad news about the airplane, most people assumed the summer windshear experiments would have to be cancelled. That would set the research back a full year, since microbursts occurred primarily during the summer months. In addition to being frustrating, a year's delay could make it more difficult for airlines to use forwardlooking detection systems, because it would leave little time for industry to incorporate the technology before the FAA deadline for installing windshear detectors.

Some of the people involved with the program refused to give up so easily on the flight tests, however. Artie D. Jessup, an electronics technician who worked on the plane's research systems, came up with the idea of leaving the research pallets and wiring in the airplane during the airframe repair. There was some risk that the wiring could be damaged during the riveting process, but if it worked, it would eliminate weeks of work. Jessup argued that the risk of damage could be minimized, and the repair technicians could probably work around most of the experimental equipment. So the equipment was left in place. The determination of the Langley crew also seemed to rub off on the repair facility in Birmingham, Alabama, because it put extra workers on the project, worked round the clock shifts, and returned the airplane a week ahead of schedule.

Even then, it appeared that, at best, the airplane would only be able to complete one of the windshear flight tests before the end of the summer. Although the program managers thought they might be able to gain a little extra time by conducting the Denver flight tests first, since the Florida microburst season tended to last a little longer, they were still doubtful that the airplane and its equipment could be ready in time. But then Michael Basnett, the airplane's crew chief, took up the challenge. He told the program managers that if they could get the research equipment ready, he and his crew would do whatever it took to make sure the airplane was also ready in time for both deployments. The positive attitudes exhibited by Jessup and Basnett began to spread. After all, if the crew chief was willing to make that kind of commitment, with all the maintenance problems the airplane was having, the other team members decided they could do no less. A growing conviction began to emerge from the windshear team that they might just be able to get the job done, after all. That conviction was backed by the managers of the Flight Systems directorate, who authorized the funding and overtime the effort required.

Over the next few weeks the researchers and technicians worked long hours, including weekends and holidays, to get the airplane and the research equipment ready. As Dr. Bowles described it, "Nobody wanted to be pointed to as the guy holding up the program." Problems that arose were handled quickly and quietly, with as little disturbance to other work areas as possible. A strong team spirit developed among those working on the project, and morale remained extremely high throughout the flight test program. The team even printed up colorful "Burstbusters" patches that they wore with pride alongside the NASA insignia on their flight suits. Finally, the long hours, dedication, and effort paid off on July 13th. Two months to the day after the airplane arrived back at Langley from the Birmingham repair facility, the windshear team departed for Denver. [Ref 5-44]

The Denver flight tests were held on July 14 18, 1992. This time, the researchers found the dry microbursts they were seeking and were able to get solid data on the performance of all three forwardlooking sensors. Two weeks after returning to Virginia, the team flew to Orlando for a second round of flight tests in "wet" microburst conditions. The flight tests ran from August 11 25, 1992. Aside from a twoday hiatus when the airplane had to be flown back to Langley to avoid Hurricane Andrew, the research went extremely well. Weather conditions cooperated, and the researchers were able to get excellent data on the radar, LIDAR and infrared systems.

Video display of a windshear research flight undertaken by the NASA 737 aircraft in 1991. Although the aircraft was not hit by lightening, it came very close.

The flight tests showed that the radar system accurately and reliably detected and tracked hazards in both wet and dry microburst conditions. The advance warning given by the radar was impressive, sometimes alerting the pilots to a hazard as much as a minute in advance. The LIDAR detected the dry microbursts well in Denver, but was less effective in the heavy rain conditions encountered in Florida. The LIDAR was the least developed of the technologies, however, and was still considered to hold some promise for the future. The researchers finally concluded that the infrared technology, on the other hand, was unable to reliably detect windshear activity. [Ref 5-45]

Throughout the course of the FAA/NASA windshear program, researchers had kept in close contact with potential commercial manufacturers of the technology. The Langley Research Center held yearly conferences to update industry on the progress in the program and made all the computer modelling and simulation technology developed by the NASA researchers during the program available to any interested company. The Westinghouse Electric Company, the Bendix/King Division of Allied Signal Aerospace and the Collins Air Transport division of Rockwell International were all acutely interested in forwardlooking windshear detection technology, especially since the FAA was going to require some kind of detectors in all future airliners. Representatives from these three companies visited Langley often and talked to researchers almost weekly over the telephone. Consequently, the manufacturers knew as much about the technology as the NASA researchers by the time the test flights were completed. In perhaps the most telling measurement of the program's success, all three companies had commercial radarbased forwardlooking windshear detection systems almost ready for FAA certification less than a year after the NASA windshear research was completed. [Ref 5-46]

Ground radome at Boulder, Colorado, providing data for winshear detection research project from the Center for Atmospheric Research.

In fact, the NASA/FAA windshear program was an almost classic example of a successful government/industry research effort from the very start of the project. Like many of the early National Advisory Committee for Aeronautics (NACA) research projects, it was a very focused, missionoriented program that took a proven, significant threat to aviation and air transportation and developed new technology that could defeat it. At the same time, the windshear program illustrated how complex a technology development effort could be. The program structure was a joint venture between NASA and the FAA, and the airborne detection program was only one component of an even bigger National Integrated Windshear Plan. Within NASA itself, the program involved numerous different research directorates, including electronics, engineering, aeronautics, and flight systems. In addition, the program involved researchers from industry and academia, as well as agencies like the National Center for Atmospheric Research (NCAR).

The windshear research itself was also complex. It incorporated physics, meteorology and engineering and consisted of theory, analysis, simulation, and flight testing. The work also had to be completed in a relatively short period of time. In addition, the flight tests were unusually demanding, and in order to complete them successfully, the researchers had to pull together as a team to overcome some significant obstacles. Yet despite the difficulties, the research was highly successful. There were several scientific and technological breakthroughs made during the windshear program, including the invention of the "Ffactor" and the development of a Doppler radarbased ,forwardlooking windshear detector. As a result, the program was able to develop new technology to finally tame the microburst windshear threat. Furthermore, since the NASA engineers had worked with potential commercial manufacturers of the technology from the start of the program, the transfer of the technology occurred smoothly, quickly, and effectively.

Of course, there were other factors that influenced the success of the windshear program, as well. The research might not have been done at all, and certainly would not have been completed in such an expeditious manner, if it had not been for the airline accidents in New Orleans and Dallas and the public and political pressure that followed them. By the same token, no matter how impressive or worthwhile the research results were, the information might not have been applied quite as quickly by industry manufacturers if it had not been for the FAA regulation requiring airliners to install windshear detectors. The FAA regulation created an automatic market for windshear detection technology, which made commercial companies more willing to develop forwardlooking detection systems. The manufacturers still had to compete with reactive systems, but they did not have to argue the basic costbenefits of windshear detectors.

The NASA 737 also played an important role in gaining acceptance for the forwardlooking detection systems among manufacturers and airlines. The fact that the windshear research included flight tests in the 737 meant that the technology was developed further than if the program had been limited to computer simulations. Consequently, there was a much smaller gap for the manufacturers to close between the NASA research effort and the commercial applications of the technology. In addition, the NASA test data, collected in a transport airplane and in realistic microburst conditions, presented unassailable proof that an airborne Doppler radar could reliably and accurately detect windshear 40 seconds or more before the airplane entered it. This conclusive evidence gave both the manufacturers and their customers a critical boost of confidence in the new technology.

The 737 flight tests were only the final, visible step in a complex, sevenyear research effort, however. Without the wide diversity of research talent both at the Langley Research Center and in the companies, universities, and other government agencies that participated in the program, the support of managers at NASA and the FAA, and the extraordinary dedication of all the research team members, the flight tests would not have happened. In recognition of the extraordinary effort put forth by those involved and the significance of what they achieved, the NASA/FAA airborne windshear research program was nominated in 1992 for a Collier Trophy the highest honor an aviation research effort could receive. Industry evaluations of the program by NASA's Aeronautics Advisory Committee were full of praise, describing the research effort as a "perfect role for NASA in support of national needs" and "NASA at its best." [Ref 5-47] The windshear research team at NASA agreed. "This was the best we can do," Dr. Creedon said. "We might get that good again, but we can't get any better." [Ref 5-48]