Skip Navigation

July-August 2015

ResizeResize Text: Original Large XLarge

How Meteorologists Help Airlines Beat the Weather Odds

Most regular airline fliers have probably experienced something like this: Forty minutes before your flight is due to depart, an airplane pulls into the gate and passengers stream out through the waiting area. A few minutes after the last passenger walks off the jet way and you're lining up to board, the gate agent announces that the flight will be delayed. She says the allowed time on duty for the crew that just landed the airplane has expired, and the crew scheduled for your flight is trapped in Providence, Rhode Island, because a snowstorm has closed the airport. In other words, bad weather far from your airport has delayed your flight until a new crew, which the airline has called in, arrives after driving from their homes or catching rides on other flights.

This is a good example of why Warren Qualley, who retired as manager of weather services for American Airlines in 2003 after 25 years with the airline, says, “An airline's flights are like a big puzzle,” with many moving pieces including the airplanes, the flight crews, and the cabin crews.

Other considerations include a storm's effects, such as heavy snow or ice that would keep employees from getting to work and how well equipped an airport is to plow runways. Even if bad weather doesn't close an airport, poor visibility can slow down the number of airplanes that can land or take off each hour. “When weather causes flight cancellations or delays, different parts of this puzzle might end up at the wrong airport,” Qualley says. The flight crew trapped in Providence is a good example of this.

Qualley has more than 30 years of aviation meteorology experience, including 25 years at American Airlines, where he retired as the manager of weather services. He's now the Senior Weather Expert in the Harris Corporation's Mission Critical Networks group, based in Washington, D.C. Since he first started as a forecaster with American in 1978, Qualley has seen the focus of airline meteorology shift from predicting the weather at airports and along routes to helping airlines solve their complicated bad-weather puzzles at the least cost.

“American was one of the first airlines to proactively cancel flights in the face of impending bad weather, finding that it was a better experience for the customer and that the airline recovered more quickly after the weather abated,” Qualley says. “Today it's common practice, with meteorologists supplying the weather odds that the managers can use to make the best estimate of the costs of various actions.” These managers, in an airline's flight operations, are an airline meteorologist's main customers, but many others depend on weather information, including those who have to notify passengers of canceled flights and rebook them.

Delays and cancelations are expensive for both passengers and airlines. For example, Tulinda Larsen, president of masFlight, which is an aviation operations analytics firm, says that for the winter of 2014–2015 through March 6, 2015, her company calculated that U.S. airlines canceled 75,000 flights because of weather, which affected 5.4 million passengers.

masFlight estimates that the costs were a total of $6 billion for passengers and $382 million for airlines. The estimated costs for passengers include expenses such as hotels, taxis, and meals (40%) and lost productivity (60%). While pilots need good weather forecasts to operate safely, airlines also need reliable forecasts to save money for themselves and their passengers.

Qualley says that when he started at American in 1978, the airline's forecasters issued forecasts for every airport American flew to and every alternate airport its airplanes would use if they couldn't land at their destinations. Meteorologists saw their main “customers” as flight dispatchers and pilots. An airline's dispatchers, who must pass a Federal Aviation Administration test to work for a U.S. airline, have joint responsibility with pilots for planning flights and staying in contact with the pilots during a flight, which includes advising them of weather changes that might affect the flight.

“We used to think that we were with the airlines for safety,” Qualley says of himself and other airline forecasters. “Certainly, that was the case, but what it came down to was cost-savings from getting out ahead of the weather.”

Today airlines and their pilots rely mostly on the NWS, or foreign weather services on flights out of the United States for the forecasts of the airports they'll be landing at and weather along the way. In day-to-day operations, an airline's forecasters help save fuel by finding the best routes to avoid headwinds, especially 100-mph-plus jet stream winds at cruising altitudes above 20,000 feet. On the other hand, forecasters help airliners save both time and fuel by telling pilots where they can hitch a ride on jet stream winds blowing in the direction they want to go.

Mark Miller, Vice President and General Manager of Decision Support at WSI, which handles weather for several airlines, says when a big weather event such as a hurricane or a major winter storm threatens one of an airline's major hubs, “It can have a profound impact on the entire system—a cascading impact.” WSI begins issuing risk outlooks five days before a storm is expected to hit. “The attention is on its potential impact. As we get to within 48 hours, the airline will start making decisions about canceling flights,” he says.

Many of its WSI's meteorologists work at the company's headquarters in Andover, Massachusetts, and others at various airline operations centers. “Part of our job for an airline's operations staff is for them to be able to talk directly with a meteorologist. They are trying to get a sense of how confident we are about the forecast,” he says. “Questions include what are the chances it will start sooner or later than expected.”

“The operator wants to talk with the forecaster to see what the forecaster might feel,” Miller says. “We can express probabilities in terms of a graphic, but when a forecaster and an airline manager are talking about potentially serious weather, a lot of the communication is happening in things such as facial expression. An airline manager wants to look at someone when making a $20 million dollar decision: Do they want to cancel a flight or not?”

Miller describes an example of how a meteorologist focusing on the needs of a particular airline saved it money by avoiding a cancelation. A typhoon was approaching Hong Kong, and many airlines canceled flights due to arrive in Hong Kong at any time that day. “Our forecasters gave the customer precise times when the typhoon winds would be above a certain threshold,” Miller said. “Our customer shifted the flight by two hours and arrived safely.”

Larsen says that masFlight estimates that the cost to an airline of canceling an international flight can be as much as $40,000. Canceling a mainline U.S. domestic flight runs around $6,000, and a regional airline flight costs $1,050.

As Rick Curtis, the Chief Meteorologist at Southwest Airlines, says: “Dispatchers and crews worry about individual flights. My group is focused on keeping the machine running as efficiently as possible. We are trying to add value from the statistical point of view.”

Warren Qualley sums up the job of an airline meteorologist: “It's all about creating value-added products and services that address the particular customer, in this case commercial aviation. The last thing a dispatcher, pilot, operations manager, reservation manager, senior management, or airport personnel wants to hear about is ‘vorticity, upper level front, computer models, etc.’ They simply want to know what, when, where, and how bad.”

Some Airlines Supply Vital Weather Data

In addition to being major consumers of weather data and forecasts, some airlines are also supplying weather observations, which are helping to improve all weather forecasts. In fact, data from airliners has become an invaluable part of the global system of gathering upper atmospheric data for computer models, which is improving all forecasts, not just those for aviation.

Since the 1940s, upper atmospheric data have mostly been gathered from instrument packages lifted aloft by weather balloons launched from roughly 800 global sites, in many places twice a day. Compared with the automated data from airliners, the data from weather balloons is quite limited; only 100 U.S. NWS offices launch weather balloons.

As long ago as the 1970s, balloon data were not meeting the needs of the computer models that were becoming the bedrock all weather forecasting. Upper air data were coming from only 800 places around the world only twice a day; in fact many times less often, because many poor nations launch only one weather balloon a day, if that often. Fortunately, by then airlines were installing computerized flight management systems in their airplanes. These systems measure air temperature and calculate wind speed and direction using the airplane's speed, heading and its changing positions. Flight management systems also always “knew” the aircraft's altitude and latitude and longitude.

A few airlines, other aviation-related companies, and national weather services have established a system for airliners to automatically radio the location, altitude, wind, and temperature data to each airline's flight operations department, which sends it on to weather services around the world. Airliners usually transmit weather data from the time the wheels leave the runway on takeoff until the wheels touch the runway when landing. Since the program began in the 1970s, the weather reports have flowed through the Aircraft Communications Addressing and Reporting System (ACARS), which is operated by Rockwell Collins/ARINC with headquarters in Annapolis, Maryland. It was called “ARINC” until Rockwell Collins took over the company in 2013.

Lisa Nolan of Rockwell Collins says the system includes more than 2,000 airplanes that supply 200,000 observations a day, which accounts for roughly 20% of the air-to-ground radio network's data. Airliners also use the system for a variety of other purposes, including automated reports to airline maintenance departments, dispatcher-pilot communications, and the information for the connecting gates announcements that flight attendants make before an airliner lands at a hub. The official name for the weather reporting system is AMDAR (Aircraft Meteorological Data Relay), but meteorologists often refer to it as ACARS data in their technical discussions of forecasts and in scientific papers.

U.S. airlines that supply data from some—but not all—of their airplanes are American, Delta-Northwest, Southwest, United, and two package carriers, FedEx Express and UPS Airlines. Since the package airplanes fly mostly overnight and into the early morning, their automated reports fill gaps when few passenger airliners are in the air.

Water Vapor, the Missing Element

While the automated temperature, atmospheric pressure, and wind data were useful for both computer models and aviation meteorologists from the beginning, one extremely important kind of information that balloons collected was missing: how much water vapor is in the air? Knowing the amounts and locations of water vapor and how it is moving is vital for not only the obvious forecasts, such as where and when rain for snow might fall and where and when the sky will be cloudy or clear, but also how much energy will be available to power thunderstorms.

Randy Baker, who is Chief Meteorologist for UPS Airlines, said his company became involved in the first tests of devices that would measure and report the amount of water vapor in the air because morning fog and low clouds are big problems at some of its hubs, including the main UPS Worldport hub at the Louisville, Kentucky, International Airport. Knowing the amount of water vapor in air moving toward an airport at various altitudes is a key to forecasting when fog is likely to form at the airport and determining how thick it will be.

For example, in the Seattle area, UPS operates out of Boeing Field, which has an elevation 400 feet lower than the Seattle-Tacoma region's main airport, Sea-Tac, and is in a valley, which makes morning fog more likely there. If the lowest clouds at Boeing Field are less than 300 feet above the runway, UPS airplanes can't land there and will land at Sea-Tac instead. “We make sure we always have a water-vapor sensor equipped 757-8 go to Boeing Field every evening,” Baker says.

The UPS forecasters also receive data, including water vapor data, from Southwest Airlines 737s taking off and landing at Sea-Tac. Rick Curtis, Southwest's Chief Meteorologist, says the airline was a good addition to the system of automated weather reports: “We have lots of ups and downs.” He was referring to the 3,400 flights a day the airline to makes to 93 destinations across the United States and five additional countries.

The fact that Baker and his forecasters can access data from another airline's aircraft illustrates the advantage of supplying automated data. Airlines that participate can access data from all airlines, not just their own. Otherwise, only NWS forecasters have real-time access to it.

Developing a Humidity Sensor

In the late 1990s, Baker and UPS Airlines joined researchers, including those at the University Center for Atmospheric Research in Boulder, Colorado, and other institutions searching for a humidity sensor that would work on airplanes. The first experiments used humidity sensors much like those used on weather-balloon radiosondes. They didn't work well. One reason was that “everyday airport gunk,” such as de-icing fluid and runway dirt, contaminated them. Radiosondes are mostly use-it-once-and-forget-it devices. The rare radiosonde that someone finds after it drifts down to earth under a tiny parachute after the weather balloon carrying it bursts has a notice on it directing the finder to drop it in a mailbox to return it to the NWS to be refurbished.

The basic water vapor sensing technology currently used in aircraft was developed and manufactured by SpectraSensors of Houston, Texas, and was originally used to measure the water vapor concentration in a natural gas pipeline. “This is done to ensure that moisture remains low and does not cause a potential for corrosion,” says Bryce Ford, Vice President of Atmospheric Programs for SpectraSensors. During 2008 and 2009, SpectraSensors and NWS engineers conducted tests to see whether a device based to some extent on those used for pipelines, known as the WVSS-II, would work under the extremely demanding conditions encountered on each flight. It was also tested on UPS 757s, with the final tests being conducted in 2009–2010. It passed the tests, and by early March 2015 a total of 120 U.S. airliners were using the WVSS-11 water vapor sensor.

Ford describes the conditions the sensor faces: “On the ground, you might have conditions like air pressure of 1,013 millibars and air temperature of 77°F and 75% relative humidity. In those conditions, that is a water vapor concentration of 24,100 parts per million by volume. At 40,000 feet, the plane is typically experiencing outside air pressures around 186 millibars, air temperatures of around −70°F, and a water vapor concentration of 16 parts per million by volume. That's a temperature change of 140°F, a pressure change of more than 800 millibars, and a water vapor change from 24,100 parts per million by volume to 16 parts per million by volume, all in a matter of minutes—twice on each flight as the aircraft goes from the ground to high altitude and back down. Doing that while traveling at over 500 mph with the data quality necessary for meteorological purposes is a tough order to fill.”

By early March 2015, a total of 120 U.S. airliners were using the WVSS-11 water vapor sensors. These sensors consist of an air sampler that attaches to the outside of an airplane's fuselage without affecting aircraft performance. A hose carries air from the sampler to the a metal box inside the airplane, where the sample flows past a tunable diode laser that accurately measures the amount of water vapor in the air. Another hose carries the air back to the sampler and back into the atmosphere. The company says the details of the unit are proprietary, but the company does say that “the laser is specifically selected to be at a wavelength that corresponds to the absorption wavelength of water. Therefore the absorption of the laser light is proportional to the amount of water in the sampled air, and can be measured by the detector in the analyzer” and radioed back with other weather data the airplane is collecting.

Rick Curtis of Southwest notes that getting the approval of the Federal Aviation Administration (FAA) and similar bodies in other nations to add the humidity sensors is complicated. Approval requires showing that any new system “won't mess anything up,” he says. For example, the FAA has to be convinced that the required hoses and wiring will be secure, and that the system won't interfere with the flight management system or anything else inside the airplane. Separate FAA certification is needed for each type of airplane using the sensor.

Curtis compares adding humidity data to the other airplane data as “going from single photographs to movies” of an event. He says, “Think of it, in 2015 we're still sending up balloons when we have thousands of airplanes that could be supplying data.”

The Value of AMDAR Reports

A 2014 World Meteorological Organization report said that AMDAR is enhancing the upper-air observing programs of national meteorological and hydrological services with the “provision of very high quality upper-air meteorological data at a lower cost relative to that derived from conventional radiosonde programs. The data that are derived from this low cost, high quality, high temporal and spatial resolution observing system are particularly valuable in areas or regions that do not reliably or routinely provide radiosonde profiles.”

The tiny laser sensor that measures water vapor in the air. The penny illustrates how small it is.

“While the radiosonde provides data to a higher altitude and incorporates a measurement of humidity as standard, the addition of a water vapor sensor can make an AMDAR program even more cost-efficient, useful, and increase the positive impact of the data,” the report says. AMDAR “can produce the needed data at about one-third the cost of radiosondes.”

The data from airliners are improving the predictions of computer forecasting models for all kinds of weather. “The predictive ability and accuracy of numerical weather prediction models relies heavily on the quality and quantity of data that are assimilated into the ‘initial state field,’” the report says. “Studies and experiments have shown that AMDAR and other aircraft-based observations generally provide an improvement in forecasting ability through a reduction in Numerican Weather Predication forecast error of up to 20%.”

The aftermath of the September 11, 2001, attack on the World Trade Center and the Pentagon illustrated how much difference aircraft reports made even then. From September 11 to September 13, the FAA grounded aircraft across the United States. Without the benefit of aircraft observations, the three-hour forecasts produced by NWS models became only as accurate as 12-hour forecasts normally were.

JACK WILLIAMS was the founding weather editor of USA TODAY and is now a freelance writer. He's the author or co-author of seven books, five of which are about weather.       

In this Issue

On this Topic

© 2017 Taylor & Francis Group · 530 Walnut Street, Suite 850, Philadelphia, PA · 19106