In this issue:
- Concerns over FAA's plan for ADS-B
- Inspector General on general aviation's ATC usage
- The staffed "virtual tower"
- Controller hiring—trouble with numbers
- "Rebaselining" hides ongoing procurement problems
- Quotable Quotes
Just about everyone in aviation thinks that ADS-B--a precise position-location system based on GPS--should replace ground-based radar, to provide much more accurate information about where planes are and where they are going. It's a key building block of next-generation air traffic management, being embraced by Europe and Asia, as well.
But the FAA's notice of proposed rule-making (NPRM) on plans for aircraft operators to equip for the first phase (called ADS-B/Out) produced a lot of negative feedback. Leading industry groups making critical comments included ATA, IATA, ACI-NA, NBAA, and AOPA. The general thrust of their comments is this: having to equip all planes for ADS-B/Out by 2020 will mean mostly costs and only the promise of future benefits for aircraft operators; the primary benefits will accrue to the FAA, as it begins to cut back on the number of costly radars.
It's true that potentially large benefits will only materialize once the whole fleet (in controlled airspace) is equipped for ADS-B/In as well as Out. The "Out" capability simply requires each aircraft to broadcast its GPS-derived position; that helps controllers and FAA, but does nothing for the aircraft operator except require it to add a costly avionics box. The "In" capability requires another on-board box, to display traffic (and possibly weather) information, enabling automation-aided "self-separation" of planes in flight, thereby enabling reductions in separation between planes, thereby expanding airspace capacity. But FAA's NPRM deals only with a deadline to equip planes with ADS-B/Out by far-off 2020. FAA has not defined any of the post-2020 requirements for ADS-B/In, let alone an implementation schedule. So airlines and airports that want the increased performance and lower costs from NextGen have only a vague promise that these things will occur sometime in the decade(s) after 2020, but have no idea what getting there will cost, or how soon any of it will actually happen.
AOPA, representing numerous private pilots, has long supported ADS-B and has been a big booster of the ongoing large-scale ADS-B demonstration in Alaska, called Capstone. AOPA is especially concerned about the equipage costs, but thinks the 2020 deadline is far too early (in contrast to the views of airport and airline groups). But it, too, sees the real benefits coming from the still vaguely defined ADS-B/In. AOPA and NBAA especially want "an affordable transition," meaning some kind of subsidized help with the cost of equipage and a guarantee that the traffic and weather information provided by ADS-B/In will be available at no charge.
These concerns are understandable, but it's very frustrating to see that while interest groups squabble, other countries are moving swiftly ahead. Europe's SESAR consortium seems to be better at integrated decision-making for implementing its version of NextGen. Airbus earlier this month announced that an ADS-B/In system from Honeywell will be offered on future A320, A340, and A380 aircraft. Australia is well along in implementing ADS-B nationwide, and its commercialized air navigation service provider, Air Services Australia, is at the forefront of assisting other parts of the Asia-Pacific region (such as Indonesia) to do likewise. And Nav Canada is busily equipping our northern neighbor.
The current FAA NPRM imbroglio is a foretaste of many similar battles still to come as NextGen moves from concept to implementation plans. The problem is one of coordinating the capital investments of the FAA's Air Traffic Organization and its numerous customers who operate aircraft. The huge uncertainties as to which NextGen capabilities will be available when, and whether they will be worth what they cost, creates a great reluctance by aircraft operators to invest in the near-term (high costs now, zero benefits now). And decades-long implementation periods pose a real risk that early adopters will be stuck with technologically obsolete equipment by the time the larger system exists in which they hope to make use of it.
To me, this says the organizational and decision-making structure is all wrong. The customers need to be at the table all the way along, helping to decide on the ATO's investment priorities and setting what become mandatory equipage deadlines. If those customers (stakeholders) were members of an ATO board of directors, they would have far greater incentives to devise implementation plans that produce benefits commensurate with costs. And they might just figure out a way to create affordable ADS-B boxes for the general aviation fleet, as well.
One of the main sticking points in the endless battle over FAA reauthorization is the useful Senate Commerce Committee proposal for a bondable $25 per flight segment ATC fee for all turbine (jet or turboprop) planes flying IFR, to pay for $5 billion worth of NextGen capital equipment. Those high-performance aircraft are the principal consumers of ATC services (other than control tower services). Yet after two years of debate, and detailed cost-allocation studies, Ed Bolen of the National Business Aviation Association could still tell Dow Jones this month that FAA "cooked up their own methodology" to arrive at the claim that general aviation uses 16% of ATC services but contributes only 3% of the cost via aviation user taxes. NBAA still maintains that GA is a "marginal user" of ATC.
In an effort to get another third-party assessment of these claims, the House Transportation & Infrastructure Committee asked the DOT Office of Inspector General to assess who uses what degree of ATC services, and to look into whether a funding system based solely on per-gallon fuel taxes would accurately reflect ATC usage. I commend to you the OIG's report, "Use of the National Airspace System," Report No. CR-2008-028, March 3, 2008 (available at www.oig.dot.gov).
In short, the Inspector General "found that air carriers and non-air carriers, including general aviation and business jet operators, all make sufficient use of the NAS [National Airspace System] so as to materially contribute to FAA's costs and congestion in general." In fact, the OIG found that non-air carriers (general and business aviation) constitute 59% of tower operations, 49% of TRACON operations, and 17% of en-route miles flown.
A separate section breaks out business jets, pointing out that under today's bizarre user tax system, the same Gulfstream or Citation can be charged a significantly different amount, depending on whether it is operated by a corporation for its executives or by a fractional-jet company serving similar executives. Because FAA does not track "business jets" as a category, OIG derived their data by taking total jet operations and subtracting out airline jets. By this reckoning, business jets alone account for 12% of tower services, 13% of TRACON services, and 11% of en-route operations. (Terminal-area services-towers and TRACONs-account for more than half of all ATC costs, according to the FAA's latest cost allocation study.)
Next, the researchers looked into whether non-air carrier planes contribute to congestion. They verified that such planes make very little use of large, congested hub airports. But when it comes to the congested metro-area TRACONs, the picture changes. "We found that both air carriers and non-air carriers contributed to congestion at the terminal area radar facilities [TRACONs] we examined." For example, non-air carriers accounted for 20 to 30% of instrument approach operations handled by the New York TRACON during peak periods. In general, non-air carrier TRACON operations peak at the same times as airline operations, so they definitely contribute to congestion there. And the same phenomenon was observed for en-route operations. At the very busy Cleveland center, non-air carriers averaged 17% of operations-but 18-23% during peaks. Likewise at Atlanta center, general aviation averaged 21%--but as much a s 28% during peaks.
How about obtaining better cost responsibility by having all NAS users pay a fuel tax rather than today's complex set of aviation excise taxes? That would make a slight improvement, according to the OIG analysis. Using a sample of nonstop markets, the current system has air carriers paying 98.8% of the aviation taxes while operating 92.3% of the flights. Under a fuel tax system, the air carriers would pay 96.9%. Non-air carriers, with 7.7% of all flights, would go from paying 1.2% to 3.1%.
The one other thing the House committee asked was whether there's a more useful way for FAA to organize its data about various categories of aircraft operators. The historic approach divides aircraft operators by trip purpose-commercial vs. non-commercial. That's the origin of the bizarre disparity in what business jets pay, depending on who is operating them (where "commercial" is defined by whether someone paying to use the plane; hence the executive who charters a jet or belongs to a fractional program makes it a "commercial" flight, but the jet owned and flown by Exxon-Mobil or Pepsico is "non-commercial"). More meaningful categories, says the OIG report, would be air carrier vs. non-air carrier and piston vs. turbine power. The type of propulsion dramatically affects the flight regime and hence the use of ATC services. And the non-air carrier grouping would eliminate the artificial distinctions among differ ent uses of business jets, all of which need exactly the same ATC services and should pay the same, regardless of who is operating them.
I hope the FAA sees fit to make these useful changes in how it keeps track of the data.
Several years ago, when I first began reading about what was then called NGATS (now NextGen), one of he most intriguing components was something called a "virtual tower." Instead of each airport having to have a costly several-hundred-foot tall structure located on its premises, with system-wide information available everywhere in real time you could provide the tower functions in a remote location. After all, FAA increasingly uses tower simulators to train new controllers. Although most are located at specific airports, a number are located at the FAA Academy in Oklahoma City.
Why might we want to do this? To get more bang for the aviation customer's buck, and by reducing the cost of tower functions, possibly to expand tower services to a larger fraction of airports. Many smaller airport towers are not staffed at night, due to the cost of keeping controllers on duty to serve a handful of flights. Using virtual towers, a number of low-activity airports could get tower functions from a single facility, providing enough total workload to justify controller staffing.
Until recently, the idea had never been put to a real test. But the Winter 2008 issue of The Journal of Air Traffic Control carries a fascinating report about the first serious demonstration of the idea, at the FAA's research facility in Atlantic City. The prototype system made use of DBRITE displays of terminal area traffic and ASDE-X displays of ground traffic. Customized large, medium, and small display screens were developed specifically for this project, which included both Local and Ground Control positions. The equipment was installed in the Atlantic City facility's tower cab simulator, which also includes a 360 degree out-the-window (OTW) view simulation that would be used to simulate normal tower operations, with the virtual tower using the new displays instead.
For the simulation, the experimenters recruited three retired controllers and one former FAA Academy instructor, all with tower and TRACON experience and previous use of DBRITE and ASDE-X. To compare staffed virtual tower (SVT) operations with conventional OTW operations, six different conditions were modeled: staffed virtual tower VFR, SVT IFR, out-the-window VFR day, OTW VFR night, OTW IFR 6,000 ft. visibility, OTW IFR 1,000 ft. visibility. Each scenario was given the same set of arrivals and departures to deal with, and the primary metric used to compare virtual tower operations with out-the-window operations was the departure rate-how many departures can take place in between the same set of arrivals.
The results were striking. For VFR daytime operations, the departure rate was 52/hour with OTW and 55/hour with SVT. But at night, the OTW rate dropped to 37/hour, while the SVT rate stayed at 55/hour. For IFR operations with 6,000' visibility, the departure rates were practically the same: 45 and 46.5 per hour. But under low-visibility (1,000') conditions, the OTW rate plunged to 29/hour while, again, the SVT rate remained at 45.
But wait, it gets better. While the controllers were more productive in the SVT operation, their perceived workload was less, since the new displays made their job easier. During VFR day conditions, subjective workload was perceived to be about the same in SVT and OTW operations. But workload was perceived to be significantly higher for OTW operations at night. For IFR operations, there was no difference in perceived workload with 6,000' visibility, but under 1,000' visibility conditions, the ground control position's perceived OTW workload nearly tripled. On a post-simulation questionnaire, both controllers strongly preferred the SVT operation to the OTW operation.
Being cautious people, despite their enthusiasm for the SVT the participating controllers did express concerns about how SVT would perform under emergency situations, suggesting the need for further simulations beyond this initial exercise. But overall, this first-ever test of the virtual tower concept "indicates that it is feasible to provide ATC services at an airport using an SVT system." Not only did it deliver comparable performance under clear-weather conditions, but it clearly demonstrated "superior performance when visibility deteriorates."
Carelessness with numbers is one of my pet peeves. A striking case in point was recent reporting on the FAA's ongoing effort to hire and train new controllers to replace the many post-strike controllers who are retiring this decade.
If you turned to Congress Daily (Feb 4th) reporting on the FAA's FY09 budget proposal, you would have read, "The spending plan calls for hiring 306 air traffic controllers . . . ." Traffic World two weeks later: "[T]heir budget calls for hiring 306 more controllers." And even our trusted Aviation Daily (Feb. 5) said the budget would "hire and train more than 300 air traffic controllers."
What's wrong with this picture? You don't have to rely just on hyperventilating news releases from controllers union NATCA to know that controller retirements are running at record levels. Over this 10-year period, the FAA will be hiring more than 16,000 controllers; that averages over 1,600 per year. In FY07, the agency hired 1,815 and aims to hire 1,877 in FY08. So didn't anybody think to question a FY09 number as low as 306?
I got those hiring numbers right out of the February 6th hearing summary paper prepared by the House Aviation Subcommittee for the budget hearing at which FAA's CFO, Ramesh Punwani, testified. That background paper did not give the actual number of controllers FAA plans to hire in the FY09 budget. But it clearly showed the agency's general M.O. of hiring slightly more trainees each year than the number of retirements (leading to a net increase of 256 in FY08), and the report implied that a similar new-hire number in the 1,800 range was in the FY09 budget. Unfortunately, at least in his written testimony, Punwani only gave the expected net increase that would result from this year's hiring: 306 (compared with the 256 net increase in FY08, based on gross hiring of 1,877). So that's apparently where the 306 number came from.
Perhaps realizing the misleading impression that the 306 number had created, the FAA put out a news release on Feb. 15th, headlined "FAA Speeds Up Process to Hire Air Traffic Controllers." There, in the fifth paragraph, was the statement that "The FAA will hire more than 1,800 new controllers this year and increase total controller staffing to more than 15,000."
Too bad we couldn't depend on transportation reporters to do their homework and tell us that.
The FAA has a long, troubled history of major procurements ending up years late and significantly over budget. Understandably, that history makes its aviation customers very cautious when the agency tells them it can implement NextGen on-time and on-budget, delivering real value for their user-tax money.
According to former Administrator Marion Blakey, those concerns should be laid to rest. "It stings when I listen to criticisms about the FAA that are based on something that happened 10 or 20 or 30 years ago," she said in a speech last spring in Hartford. In the last few years, we have achieved enormous management efficiencies, and at the end of fiscal year 2006, 97% of our major capital projects were on time and on budget." The next month, before a Senate subcommittee, Blakey upped the ante: "As we speak, 100% of our major capital programs are on schedule and on budget."
That sounds way too good to be true, and it isn't. As DOT Inspector General Calvin Scovel testified at that very same hearing, what FAA did during Blakey's tenure was to systematically move the goalposts on major projects. It's called "rebaselining," which means after a project goes badly off track, you define a new completion date and a new total cost-to-complete, and then measure accomplishments against those new baselines.
"The rebaselining process explains why the Wide Area Augmentation System, according to the FAA's logic, is still on budget, even though its costs have grown from $892 million to over $3 billion since 1998," Scovel testified. His written testimony elaborated:
"Of the 37 major procurements that the FAA evaluates in claiming its near-perfect procurement record, more than 25% of them have had their schedules and budget revised by the agency since the initial project contracts were signed . . . . Through the rebaselining process, just since 2001, the FAA has added 296 months, or over 25%, to the original schedules for these programs. The FAA has added almost $1.7 billion, or 26%, to the total costs of those programs over the same time period. When you look to the agency's record going back to the inception of each of these programs, the FAA has added almost 400 months to their schedules, an increase of about 46% . . . [and] has added more than $5 billion to their costs. That represents cost growth of 109%."
This is not just a matter of interpretation, for this chronic problem has major consequences. For example, as Scovel's March 8, 2008 testimony before the House Aviation Subcommittee points out, the STARS program to replace controllers' displays was so over budget that in 2004 the FAA reduced it from the original 172 sites (at an original $940 million) to just 47 sites at $1.46 billion. The next year it came up with a lower-cost modernization for some additional sites, but "this leaves over 100 sites still in need of modernization," which Scovel estimates will require another billion dollars.
Another example is the FAA's highly touted ATOP program to modernize its oceanic ATC systems. Back in 2003, the FAA's Charlie Keegan and then-Inspector General Ken Mead emphasized to the House Aviation Subcommittee that this was a fixed-price contract in which any cost overruns would be absorbed by contractor Lockheed Martin. ATOP started out as an off-the-shelf procurement at a fixed price of $217 million. For some reason, FAA and Lockheed decided the New Zealand off-the-shelf system wasn't quite good enough and needed software customization. Today, thanks to software development problems, that "fixed-price" contract has grown to $376.5 million-and the system still isn't working right.
The fact that major procurements continue to experience cost growth and schedule slips suggests underlying problems that no degree of rebaselining can paper over. It would be a big mistake to embark on implementing the $20 billion NextGen transformation of air traffic control without fundamental reform of program management.
"One thing that has changed since 1990 and that is public outrage. With delays at all-time highs, and levels of service at all-time lows, the traveling public is no longer willing to put up with an outdated and inefficient air traffic management system. In responding to its constituents, Congress has two choices. The first is to create a professionally managed, independently funded, non-political organization to operate the nation's ATC system. Failing this, the alternative is to impose such draconian capacity constraints favoring public transportation [airlines] which will all but eliminate business and general aviation. Put simply, unless the aviation community can come together and urge Congress to adopt the first choice, matters will only get worse, and general aviation will be the first to feel the heat."
--Jonathan Howe, former president, NBAA, letter to Professional Pilot, February 2008.
"I think that anyone that is in a safety position has a responsibility to not obviously use inflammatory language and maintain their comments on a factual basis."
--Jim Hall, former chairman of the National Transportation Safety Board, quoted in "Air Traffic Controllers' Labor Tactics Raise Concern," Los Angeles Times, Jan. 17, 2008.
"Today's paradigm of how air traffic control works has about run out of steam and has reached the limits of its scalability. A human-controlled, voice-communicated, radar-surveillance-based system by one person talking to one pilot at a time inherently can only take so much growth. A controller can handle a dozen or 15 or so airplanes at a time; if you get more than that you have to start subdividing the sectors. And at a certain point, there's a diminishing return on subdividing the airspace."
--Mike Lewis, Boeing ATM, Airport Business, February 2008