1995-2 Embargoed for Release until: January 16-20, 1995 Contact: Joan Vandiver Frisch Manager, NCAR Media Relations Boulder, CO 80307-3000 Tel. 303-497-8607; Fax: 303-497-8610 E-mail: jfrisch@ncar.ucar.edu Snowfall Rate, Not Visibility, Makes the Difference in Aircraft Icing, NCAR Scientists Report DALLAS--In five major U.S. airline jet crashes of the past 20 years caused by icing just before takeoff, the weather conditions were remarkably similar except for an eightfold variation in visibility. This finding emerged from a study by researchers Roy Rasmussen, Jeffrey Cole, and Kevin Knight (National Center for Atmospheric Research, or NCAR), along with R. K. Moore and Murray Kuperman (United Airlines). The study implicates snowfall rate (specifically the amount of water in the snow) as a major variable in aircraft icing and cautions pilots not to rely on visibility as an indicator of icing hazard. The results, already being used by major passenger airlines to assess icing risk, are being presented at the Sixth Conference on Aviation Weather Systems during the American Meteorological Society's 75th Annual Meeting in Dallas, Texas, on Wednesday, January 18. The study was conducted at NCAR's Research Applications Program and sponsored by the National Science Foundation through an interagency agreement with the Federal Aviation Administration. The accidents studied took place at New York's LaGuardia Airport on March 23, 1992; Denver's Stapleton International Airport on November 15, 1987; Washington (D.C.) National Airport on January 13, 1982; Boston's Logan International Airport on February 18, 1980; and Newark International Airport on November 27, 1978. Four of the five crashes resulted in fatalities, ranging in number from 7 to 69. In all five cases, temperatures were between 25 and 31 degrees Fahrenheit, winds were between 10 and 15 miles per hour, and snow was falling at a liquid- equivalent rate of between 0.08 and 0.10 inches per hour, as measured in the hour preceding each crash. But visibility varied from 1/4 mile in the Washington case to two miles in the Boston case. Because standard techniques of reporting snowfall intensity are based entirely on visibility, the snowfall in the five cases was reported as anywhere from "light" to "heavy," even though roughly the same amount of water equivalent was being deposited in each case. The difference in visibility is due to the highly varied structure of snowflakes, according to the NCAR/United researchers. A fluffy snowflake can obscure visibility up to ten times more than a small, dense flake holding the same amount of water. Thus, the different visibilities in the five cases can be explained largely by the type of snow falling in each case. Pilots and air traffic controllers may be misled by relatively good visibility when a snowfall consists of dense, compact flakes that are actually depositing substantial amounts of frozen water onto aircraft, according to Rasmussen. It is this deposition that can cause ice accumulation on wings after deicing but prior to takeoff. "An accumulation of as little as 0.03 inch of ice on the upper wing surface can result in a 25% loss of lift and increase in drag during takeoff rotation," says Rasmussen. The fluids used in deicing can only protect aircraft from reicing for a limited time, and delays caused by poor weather can prolong the time planes must taxi before takeoff. The NCAR/United team also uncovered a secondary problem caused by the geometry of an aircraft wing surface. Planes typically taxi downwind to the takeoff runway in order to take off into the wind. When rain or snow is falling, any wind during the taxi to takeoff will blow the particles more directly on to the upper surface of the wing, which is typically angled at about 10 degrees from true horizontal. "For the wind speeds observed during the aircraft accidents studied, the enhancement factor ranges from 1.75 to 2.0," says Rasmussen. Thus, the accumulation rate of snow on a wing can nearly double if a plane is pointed in the same direction as a 10-15 mph wind while taxiing to the takeoff runway. Rasmussen and colleagues have relayed their findings to major airlines, noting that the liquid-equivalent snowfall rate of 0.08-0.10 inches per hour seems to be a critical window. The team also recommends that visibility not be relied upon as an indicator of snowfall rate. USAir and United Airlines have implemented these results into their training materials for winter weather operations. This research is sponsored by the National Science Foundation through an Interagency Agreement in response to requirements and funding by the Federal Aviation Administration's Aviation Weather Development Program. NCAR is operated by the University Corporation for Atmospheric Research under sponsorship of the National Science Foundation. This table shows the location of each of the five icing-related airline accidents studied by RAP, the date and time (local standard) of each accident, the average liquid-equivalent precipitation rate during the hour of the accident, and the most recent visibility report preceding each accident. A more complete table is included in the paper by Rasmussen et al., "Meteorological Conditions Associated with Ground De-Icing Accidents and the Effects of Wind on Snowfall Accumulation on Aircraft," which appears in the Preprints of the Sixth Conference on Aviation Weather Systems to be held January 15, 1995, in Dallas during the 75th Anniversary meeting of the American Meteorological Society. Table of Five Icing-Related Airline Accidents Studied by RAP between 1978 and 1992 Location Date, time Precipitation rate Visibility Newark, NJ 11/27/78 0.095 inches/hour 0.5 mile (Newark) 11:50 a.m. Boston, MA 2/18/80 0.08 in/hr 2.0 miles (Logan) 2:08 p.m. Washington, DC 1/13/82 0.09 in/hr 0.25 mile (National) 4:00 p.m. Denver, CO 11/15/87 0.10 in/hr 0.5 mile (Stapleton) 2:15 p.m. Flushing, NY 3/22/92 0.10 in/hr 0.75 mile (LaGuardia) 9:35 p.m. Writer: Bob Henson, UCAR Communications, (303) 497-8605