In January 1982, two airline accidents within a matter of weeks suddenly focused national attention on the hazards of winter flying. On January 13, 1982, an Air Florida 737 crashed into the 14th Street Bridge after departing Washington National Airport, killing 78 people on board the airplane. Ten days later, a World Airways DC10 landing at Logan Airport in Boston, Massachusetts ran off the end of the icy runway and into the bay, killing two passengers and injuring four others. [Ref 8-10] Spurred by those two accidents, the House subcommittee on aviation recommended that the FAA conduct a focused research and development program in runway friction measurement in an effort to improve safety in one of the critical areas of winter flight operations. The recommendation led to an appropriations bill for the FAA and NASA to "study the correlation between aircraft stopping performance and runway friction measurements on wet and contaminated surfaces." The bill also specified that the research would focus on "determining if it is possible to predict aircraft stopping performance based on runway friction measurements using new technology friction measuring devices." [Ref 8-11]

The Boeing 737 testbed during runway friction tests at Brunswick Naval Air Station, Maine, in 1985.

The FAA, NASA and other groups had actually been studying runway friction and the effects of wet, snowcovered, icy, or otherwise "contaminated" runway surfaces on aircraft performance for some time. Engineers at the Langley Research Center, for example, had been investigating the impact of different runway surfaces on braking performance since 1968. Yet by 1982, pilots still had to rely solely on the subjective reports of other pilots landing or departing ahead of them for information on runway conditions. These reports were imprecise and often of limited value, because what was "adequate" for a Northwest Airlines pilot accustomed to flying in winter conditions might be considered "severe" by an Air Florida pilot. Groundbased frictionmeasuring vehicles were in existence in 1982, but they were only used to tell airport operators when repairs needed to be made to runway surfaces. The research recommended by the House subcommittee was an effort to provide pilots and airport operators with more precise and reliable information about runway conditions in bad weather. Pilots could then either modify their landing technique to accommodate the conditions or, if performance was degraded beyond a minimum safety level, operations could be suspended until the runway conditions improved. [Ref 8-12]

Researchers wanted data from more than one kind of transport aircraft, so both NASA's TSRV 737 and a Boeing 727 owned by the FAA were used for the tests, although the FAA airplane did not have the extensive research equipment or capabilities of the 737. NASA technicians installed a special instrumentation packet on the 727 to enable it to collect the necessary friction data for the research experiments, but the 737 was still able to collect more detailed information, such as how hard the pilots were braking during the test runs. The tests, which ran from June 1983 through March 1986, took place at NASA's Wallops Island Flight Facility and the Langley Air Force Base in Virginia, the FAA Technical Center in New Jersey, and at the Brunswick Naval Air Station in Maine. The tests were conducted at numerous sites so data could be collected in a variety of weather conditions, and researchers could evaluate the impact of different runway surfaces on braking friction as well as the quality of the ground vehicle measurements.

The Boeing 737 and ATOPS researchers working on the runway friction tests in 1989 at Wallops Flight Facility.

The research effort evaluated a number of different ground friction measuring devices. A diagonalbraked vehicle (DBV), for example, was a car with a specially modified braking system that allowed only two diagonally opposed tires to lock up when the brakes were applied sharply. The DBV measured the speed, acceleration and stopping distance from the point of brakedwheel lockup to determine the friction level of the runway. Another ground device called a MuMeter consisted of a 540 lb. trailer towed behind a truck that determined surface friction by measuring the sideforces imposed on the trailer wheels. A third ground vehicle, called a BV11 Skiddometer, measured the speed of the vehicle, the torque applied during braking, and the slip ratio of an instrumented wheel to determine the runway friction level. Over the course of the tests, half a dozen different ground vehicles were evaluated at different locations.

The tests were conducted by having a ground vehicle make a pass down the runway first, followed by one of the two test airplanes. The measurements of the ground device and the airplane would then be compared. Braking performance was measured over a range of speeds, ranging from 40 mph to over 100 mph. For the slower runs, the aircraft would accelerate to the required speed and then apply maximum braking, but for the faster runs, the aircraft would take off, land, and then test the braking performance as it slowed down to a stop. The research also looked at the impact of engine reversers on aircraft braking performance in contaminated runway conditions, and the effectiveness of different kinds of runway deicing substances. [Ref 8-13]

Measurements were taken on a wide variety of runway surfaces and conditions. Dry runway measurements were taken to establish the standard against which degraded performance could be measured. Tests were then conducted on wet, flooded, slushcovered, snow covered and icy runways. In some cases, natural rainfall allowed researchers to measure the impact of both surface water levels and the intensity of rainfall on runway friction levels. If it was not raining, wet runway conditions could be created by flooding surfaces with water trucks. Some of the winter conditions would have been harder to simulate but, fortunately, that was not necessary.

The TSRV 737 and its research crew travelled to the Brunswick Naval Air Station in March 1985 to test runway friction levels in snow and icy conditions. The researchers wanted to test snow levels up to eight inches, and immediately after the plane arrived, eight inches of snow fell. A 2,000 foot test section of snow was prepared, with a long cleared strip at either end to allow the airplane to stop and turn around. The first thing the researchers discovered was that eight inches of snow increased the friction level so much that although the 737 finally reached the 90 knot speed the test required, the snow caused so much drag that the aircraft could never have reached its takeoff speed. The snow was progressively reduced for additional tests with six, four and two inches of snow, by which time the conditions had deteriorated into the precise kind of slush conditions the researchers wanted to test next. After the snow measurements were completed, the temperature at New Brunswick fell, allowing the runway to be frozen for the icing tests. The lack of friction on icecovered runways became very clear when the 737 accelerated to 90 knots as it entered the icecovered section and, despite maximum braking efforts by the pilot, exited the section at 92 knots. [Ref 8-14]

The runway friction tests showed researchers that the ground vehicle measurements correlated extremely well with aircraft performance, meaning that they could be used to give pilots and airport operators reliable information about runway friction conditions in bad weather. The results also indicated that grooved runway surfaces were an extremely effective method of maintaining safe friction levels in poor weather conditions. Braking performance on transversely grooved runway surfaces was almost as good in wet and slippery conditions as it was on dry runways. These results played a significant role in expanding the use of grooved surfaces in the transportation industry. By 1993, over 800 commercial runways in the world were constructed with grooved surfaces. In addition, sections of highway in all 50 states within the U.S. were using grooved surfaces to improve traction in poor weather. While most of the highway applications used longitudinal grooving for ease of installation, transverse grooving was also being installed at an increasing number of intersections to improve braking performance. [Ref 8-15]

The use of ground vehicles for frictionmeasuring also increased dramatically as a result of the NASA/FAA tests. Before the runway friction research, fewer than a dozen airports had the machinery, and there were no guidelines to tell operators how to use the devices to judge friction levels in poor weather. By the early 1990s, between 80 and 100 airports had purchased ground frictionmeasuring devices. In addition, the FAA issued advisory circulars that spelled out what the minimum friction measurements should be for each different vehicle in various conditions to ensure safe aircraft braking performance. Airport operators now had a much more reliable method for judging the safety of runway conditions and knowing when operations should be suspended. [Ref 8-16]