In any specific instance, an individual or organization would appear to have some rational basis for resisting change. In some cases, the rationale is pretty clear. A labor organization doesn’t want its members to lose jobs or Congressmen and their constituents don’t want to lose the revenue associated with a federal facility in their district. In other cases, people don’t want to deal with taking on a lot more work or headaches. They say that the contemplated change is not operationally feasible, there will be workload issues, or the technological changes are too complicated. Or, we can’t get all the stakeholders to agree, or one of the stakeholders will object and block the desired change. NextGen has been subjected to all of these kinds of resistance to change. Is it any wonder why it’s been so hard to get it rolling forward? Why changes so far have been either innocuous on one hand or pretty miraculous on the other?

I think it’s useful to reflect on an object lesson from a few years ago: the outsourcing of Automated Flight Service Stations. When this first occurred in 2003-2004, I used to quietly talk about this event as an object lesson. I feel that enough time has passed to be able to talk more openly about it now, especially since I think it’s relevant to today’s insidious resistance to change and its impact on NextGen.

Flight Service Stations (FSS) seem like they had existed since the dawn of aviation. When I first joined the FAA in 1980, there were 318 FSSs, most of which were located at airports and fixed base operator locations where general aviation pilots could readily access them to get preflight weather briefings and file their flight plans. They were typically set up with walk-up counters, where pilots could interact with FSS specialists. FSSs also provided in-flight advisories to general aviation pilots.

By the mid-1980s, technology had progressed to a point where pilots could access weather briefings and do their flight planning and filing first over the telephone and then later via home computer. The growing use of remote pilot access and other technology made it attractive to consolidate FSSs into 61 Automated Flight Service Stations (AFSS). While the business case for this change was clearly compelling, there was resistance from nearly every quadrant. The FAA’s program manager faced all of this resistance and did a masterful job dealing with it. In the end, manual FSSs for Alaska were retained, but the consolidation was allowed to occur throughout the rest of the system.

As we progressed into the 1990s, changes continued to affect the AFSS landscape. You could clearly see it in the data. The demand for human-based preflight briefings continued to decline significantly with each passing year, like a straight line moving in a downward direction. During that period, I visited an AFSS and noted how little activity seemed to be going on. I was told that the weather was beautiful that day, and that they would be swamped on a poor weather day. So, I waited for a period of poor weather and returned to observe the operation. During my visit that lasted all morning, the peak number of calls was four. I was in a camp that felt that all preflight briefings and flight filing could be done remotely by computer and only with an occasional need for human interaction that could be handled by exception; and in-flight assistance could easily be done out of en route centers. I felt that AFSSs largely could be decommissioned. Yet, we were outvoted by those who felt that it was politically too difficult. After the difficulties in pushing through AFSSs just 10 years earlier, elimination of AFSSs was viewed as a “third rail issue.” So, in the 1990s, despite overwhelming evidence to the contrary, the FAA made a major capital investment to modernize the automation systems in AFSSs. Funding is always in short supply, so this capital investment was particularly expensive in terms of opportunity cost.

By the early 2000s, less than 10 years later, demand had continued to decline. Other capital investments were up for consideration if we were to continue to operate AFSSs. Yet the supply and demand curves for this type of service continued to diverge. It happened that the Bush administration around this time had initiated a program requiring federal agencies to out-source federally provided functions to the private sector if it made sense to do so. The requirement and procedure for out-sourcing was nothing new. It had long existed as an Office of Management and Budget guidance document, known as A-76. But, the Bush administration set hard requirements for each agency to do a minimum specified amount of out-sourcing. This was new.

The gap between AFSS supply and demand had become so great that it could no longer be concealed or ignored. When the out-sourcing edict was issued by the White House, AFSSs jumped out as something that was ripe for out-sourcing. So, this is how AFSSs became subject to A-76. Those people and organizations that currently provide the service are allowed to complete with alternatives to out-source the service. If sufficient efficiencies can be extracted by the incumbents, they can beat the out-sourcing alternatives and win the opportunity to make those changes to continue providing the service. But, if one of the out-sourcing alternatives wins, then the service is out-sourced. Once something enters the A-76 process, it simply must run its course. There is little opportunity to exert external influence over the outcome. It’s essentially taken out of your hands.

In early 2005, the AFSS function was out-sourced via an FAA contract awarded to Lockheed Martin. The out-sourcing proposal called for consolidating nearly 60 AFSSs into three major hub facilities located in Virginia, Texas and Arizona. By this decision about 2,700 FAA employee positions were eliminated. Efforts were made to place FSS specialists in other parts of the FAA where job requirements and skills could be matched, but very few FAA reassignments were possible. A good many FSS specialists accepted positions with Lockheed Martin and were required to move to one of the three hub locations. Many chose to not move and had to look for new employment. The termination of all of these FAA employees was one of the most painful experiences in recent FAA history. It was a tragedy. It especially was a tragedy because it could have been averted.

The object lesson for me is this: When you continue to resist change in the face of a growing need for change, the magnitude of change grows. You reach a point where the need for change is so great and so obvious that the problem suddenly is taken out of your hands, and then change happens quickly, perhaps with a bang. No one is happy with the outcome. Many suffer huge amounts of pain because the change is so sudden and so great. Much of it could have been avoided if the need for change had been met with incremental change over time.

As we enter this New Year, let’s be more receptive to change. Let’s embrace the changes required by NextGen. We know that demand for air travel is largely driven by the health of the economy. It’s only a matter of time before the US economy turns around. When it does, delays could go through the roof. Ticket prices will too. The ills of the aviation system will return to page 1 of the daily newspapers. The gap between supply and demand will be so great that it won’t be tolerated. It will be taken out of the hands that currently control it. Change will then be sudden, and there will be a lot of pain. We can act now to avoid that kind of outcome—and avoid being run over by the need for change.

On the autobahn, a maximum speed of 130 km per hour (a little over 80 mph) is “suggested.” I can recall doing 130 km per hour and seeing a blur of high-end cars whizzing past at what must have been 120 or 140 mph. One of the major differences with the autobahn is the discipline exercised by drivers. Over the course of our 2700 km journey on the autobahn, we never saw anyone dawdling in the fast lane. We never saw any one cut off. Very cautious lane changes and always signaled. Rarely any passing on the right side. Move over to the fast lane to overtake, and then quickly back to the slower lane. When in doubt, wait. Discipline.

Our Interstates aren’t any safer just because we drive slower. The statistics show that despite the higher speeds on the autobahn, the rate of fatalities is about the same as on US Interstates (it was interesting to see the huge billboards on the autobahn with children’s Crayola drawings of auto accidents and kids crying over the deaths of parents and family members – rather direct, but seemingly without much effect).

Getting from point A to B seems a lot faster on the autobahn. It doesn’t seem quite as congested even during busy periods. Although, if all of the traffic is moving 50% faster, then in the same unit of time, you can move the same number of cars with only two thirds as many on the road at any given time. Is that the reason it seems less congested?

I’ve been told that the pavement on the autobahn is roughly twice as thick as the pavement of the US Interstate. This may explain why the surface is so much smoother and able to more easily support higher speeds. While the initial cost is significantly higher, the recurring maintenance cost is likely to be lower. There are a lot of trucks in Germany, but they seem to have many more tires and fatter tires than those in the US, probably resulting in a significant reduction of loading and deterioration of the pavement.

Two cultures share the same problem, but produce totally different solutions. Which one is better? Have we been thinking about culturally different solutions to our air transportation problem? Could there be a significantly better version of NextGen out there? What might it look like?

One idea from the autobahn that hasn’t gotten enough attention in NextGen is the importance of protecting our heaviest traffic flows. What can we do to avoid the analogous need to step on the brake or change lanes to avoid another car? Because whenever there is even a slight disruption in the flow, that disruption cascades back and multiplies. This is the reason there are a lot of delays even at the highest altitudes on a completely clear day in the system. The largest traffic flows in the system need to be protected from disturbances. Large converging flows need to be separated or synchronized to avert flow disturbances. If we require lower rate crossing traffic to deviate around these heaviest flows, we can expect the total number of deviations to go down, just because there will be fewer cascading disturbances in traffic flows. Fewer aircraft will be subjected to vectoring, holding, speed and altitude changes, which means more time, fuel and emissions savings.

Another forgotten idea: In the mid-1990s, the FAA and aviation community came together under RTCA Task Force 3 to work on Free Flight. In the beginning of these discussions, there was a great desire on the part of the community to give aircraft much more freedom and latitude to fly where they wanted to fly. It wasn’t about the unalienable right to fly freely, but the airlines in particular wanted to fly more efficient routes to save time and fuel. The most efficient routes moved around every day with the winds. What ensued was a debate about routes that moved around with the winds vs. fixed, structured, predictable routes, the latter of which is important to air traffic controllers. At the time, no win-win solution could be found. So, we continued and largely still continue to fly the same fixed jet routes today, irrespective of what the winds are doing. We are in a different place today.  Oil was $19 a barrel in 1995 (in terms of 2010 dollars). It’s increased fivefold since then. We have more technologies available to us on the aircraft and on the ground. We have an opportunity in NextGen to find that win-win solution.

As we design tomorrow’s system, let’s remember the analogy of the autobahn vs. our Interstate system. Two societies had the same problem, yet developed very different solutions. We need to ask which is better. Can we have the best of both? Perhaps most importantly, let’s stay focused on the fundamental problem we are solving.

Recently, ICAO held the Global Air Navigation Industry Symposium (GANIS) in Montreal. The focus of GANIS was driven by ICAO’s Director of the Air Navigation Bureau, Nancy Graham: to unveil the Aviation System Block Upgrade (ASBU) framework to be used by all states to upgrade their air transportation systems. The symposium allowed the community to engage in the process of defining what the ASBUs ultimately should be when the ICAO 12^th Air Navigation Conference (ANC/12) convenes in November of 2012. This will be the point at which the member states approve the definition and concept of block upgrades.

There has been some measure of angst around the manner in which the block upgrades were developed so far. They were developed by a small, closed committee of technical experts. The traditional ICAO processes weren’t used. Things were happening too fast. Perhaps they are, but this is signature Nancy Graham. She personifies the motto “Make dust or eat dust.” She hasn’t eaten dust in a long time. Over the years, people have complained about ICAO’s glacial pace. People have said it’s too much like the UN. Well, it is part of the UN. So, why the surprise? Nancy Graham is on a mission to change that. She’s confronted the need to address what happens in the rest of the world. Some have said that the rest of the world could simply follow the lead of the US, Europe or Japan. But, those same people haven’t spent enough time outside the US, Europe and Japan to realize that things are different in other parts of the world. Traffic levels. Pre-existing infrastructure (a blessing and a curse). Power and broadband. Political, institutional and cultural differences. One size doesn’t fit all.

The ICAO plan acknowledges these differences and provides flexibility in implementing the block upgrades depending on the individual state’s “need” and “readiness.” Regardless of the form of block upgrade implementation, however, each state must 1) clearly define measurable operational improvements, 2) install the required ground and airborne equipment with required operational and regulatory approvals, 3) have the standards and procedures for ground and airborne systems, and 4) have a positive business case. ICAO says that if states do this, “a level of investment certainty for operators, infrastructure providers and equipment manufacturers” will result. Sounds simple enough. Except, we haven’t been able to do this very well in the US and Europe.

Let’s take item 1, measurable operational improvements. To define measurable improvements, we have to clearly define the problem we are solving. It’s amazing to me how we seem to launch off on new programs without having a truly clear definition and understanding of the problem that must be solved. We wind up with poor requirements, functionality that isn’t laser-focused on solving the problem, bigger than necessary price tags, and a lack of adequate metrics to measure improvements. Improvement against what?

Let’s take item 4, a positive business case (items 2 and 3 are done as a necessary part of completing item 4). A positive business case requires the present value of future benefits to exceed the present value of all costs. Benefits are essentially dependent on the size of the operational improvements. So, to do item 4, you must go back to item 1. All roads lead from item 1 and circle back to item 1. One of the main reasons we have so much trouble developing a positive business case is that we fail to do an adequate job on defining the operational problem and the associated measurable operational improvements that fix them.

While there is flexibility in the way that ICAO will allow block upgrades to be made in each state, one critical thing to keep in mind is that it’s the same aircraft and crew flying from one state to the next. So, the block upgrades can depend on a state’s specific need and a state’s readiness, but it shouldn’t require aircraft to carry different sets of equipment or crews to use different sets of procedures when flying from state to state. To the extent that the aircraft and crew can be relied upon to do more, the need for ground-based infrastructure and workforce will be reduced. It’s also naturally faster to do technology insertion and workforce implementation on the airborne end than it is on the ground, particularly when thinking about this on a global scale. There’s a lot of money and time to be saved.

As we embark on a long overdue plan to upgrade the global air transportation system beyond just the US, Europe and Japan, let’s use a balanced approach that doesn’t over-emphasize the ground-based portion of block upgrades. If anything, we ought to place more emphasis on the airborne end. Let’s not over-emphasize the block upgrades themselves and lose sight of the fundamental problems that need to be solved. Even before block upgrades were uttered, there had been insufficient attention given to defining and understanding operational problems. When we seek to more clearly define and truly understand the operational problems we face, we likely will find that there are creative ways of solving good portions of those problems quickly and inexpensively.

A year ago, there was much angst about the massive disruptions caused by Eyjafjallajökull. The debate continues about who should compensate who for damages and suffering incurred. Some valuable time was lost as organizations seemed to avoid taking any corrective actions that might be viewed as accepting some responsibility for disruptions and the resulting damages. Nonetheless, some progress was made and an exercise was conducted in April 2011, with resounding success declared at its conclusion. So, how did we do with Grímsvötn?

First thing to note is that the UK Met Office is still under attack. This time, as last time, the most vociferous and vituperous is Ryanair. The UK Met Office issued forecasts, providing cloud position and densities in 6 hour snapshots for three altitude bands (surface to FL200, FL200-FL350, and FL350-FL550). Ryanair said it flew an aircraft at FL410 and determined that there was no appreciable ash to be found. It said that the “mythical red zones” (denoting ash concentrations above the safe limit of 4000 micrograms per cubic meter) were a “misguided invention” of the UK Met Office. While it’s hard to know exactly what happened, it appears that the large red areas denoted on UK Met Office forecasts were in the lowest altitude band. The upper band in the UK Met Office forecast that includes FL410 showed no red areas close to the UK or mainland Europe. So, it would seem that Ryanair merely confirmed the UK Met Office’s forecast in the upper altitude band. It’s not uncommon for people to talk past each other in the midst of all the confusion during a crisis. Perhaps this is what happened.

Figure 1. UK Met Office volcanic ash forecast for surface-FL200, FL200-FL350, and FL350-FL550.

The most significant point is that, unlike the US, European regulators are still closing the airspace. In the US, based on forecast meteorological conditions (including volcanic ash), the aircraft operators are responsible for determining where they want to fly and if they should fly. Airspace normally is not closed per se. If Europe were to adopt the same approach used in US airspace, Ryanair would have to decide where and if it wanted to fly. Following the Eyjafjallajökull eruption in 2010, it appeared that Europe would adopt an approach similar to the US. But, this is a pretty fundamental shift in operational, regulatory and legal terms, so it will take some time to do — if it happens.

The second thing to note is that the no-fly ash density threshold has changed from 2 mg/m³ (2000 micrograms per cubic meter) in April 2010 to 4000 micrograms per cubic meter. More specifically, in May 2010, European authorities decided that 4000 micrograms per cubic meter should be the no-fly threshold and that some time-limited flying could be done at ash densities between 2000 and 4000 micrograms per cubic meter, provided aircraft operators established a more rigorous inspection program in coordination with their engine and airframe manufacturers. I’m happy to see that duration of exposure is being considered in addition to ash concentration. But, I suspect a lot more work will be needed in this area, especially as new, higher efficiency engines hit the marketplace. These new engines will run hotter and are expected to be more susceptible to volcanic ash. Also, we need to keep in mind that these thresholds aren’t really “safety” limits. In reality, these should be thought of as “maintenance” or “economic” limits. Even modest ash damage to jet engines, which might present little or no safety risk, can result in millions of dollars of overhaul expense. These expenses would be borne by aircraft operators. So, this is another good reason to allow the aircraft operator to decide where, when or if the aircraft should fly.

Figure 2. Damage to NASA CFM-56 engines from 7 minute exposure to unspecified volcanic ash density from 2000 Hekla eruption in Iceland. Resulted in over $3M repair costs. Source: NASA

Lastly, despite the relatively limited nature of the Grímsvötn event, early estimates suggest that over 1000 flights will have been canceled. Had it been more like Eyjafjallajökull in 2010, where over 100,000 flights were canceled, what would have happened? Much more needs to be done to allow all of the stakeholders to exchange and collectively evaluate critical information, interact with each other, share intent information, and develop and execute coordinated plans. As we gain a better understanding of the effects of volcanic ash density and duration of exposure on jet engines, aircraft operators will require far more accurate and dynamic information to plan and operate their flights in a way that minimizes disruptions while avoiding huge, unexpected maintenance losses. Volcanic and other disruptions usually can’t be controlled, but there is so much more we could and should be doing to better prepare and cope with these difficult situations when they do happen.

In the mid-1990s, FAA’s Western-Pacific Regional Administrator Carl Schellenberg came to FAA Headquarters. Carl assumed responsibility for the Airport Capacity Office, whose mission in life was to increase airport capacity – figure out where it was needed or going to be needed, and then work with the right people and do whatever had to be done to build it. Carl was a lawyer by training and had acquired a wide range of critical skills that made him the right person for that job. When you study the system from this perspective, you come to recognize that airports are just one component of a complex system. It’s not just concrete and terminal buildings, but airspace, air traffic management services, airlines, passengers, environmentalists, and the local economy that relies on and supports aviation. Shortly after getting settled in his job, he began to question the adequacy of existing performance measures. He said that safety and efficiency/capacity/delay were important, but not enough. He began promoting predictability, flexibility and access as additional measures and goals to strive for. While these ideas resonated with many, very little progress was made to quantify them. So, after a while they faded from discussion, with safety and efficiency/capacity/delays continuing to serve as the main institutional measures of system performance.

It had been over ten years since I had focused on predictability, flexibility and access. A lot has happened since Carl first introduced them. I’ve come to realize that I have seen examples over the years that may allow us to begin quantifying these elusive measures. It could very well turn out that these measures are associated with large pools of economic benefits that ought to influence the design of NextGen policies, concepts and technologies.

Last year we were doing human-in-the-loop simulations with airline dispatchers interacting with the FAA in rerouting major traffic flows around large, convective weather systems. We wanted to see how well the airlines were able to cope with large numbers of rerouted flights. We anticipated that reroutes would be designed to keep the flight path as short as possible to save fuel and time. Instead, we saw instances where the airlines were selecting reroutes with bigger deviations that resulted in longer and seemingly more costly flight paths. On further inquiry, we learned that these flights were rerouted on a longer path because it was more important to have a predictable arrival time, since the destination runway, gate, ground crews and equipment were in short supply. In other words, the airlines were flying further, burning more fuel and taking more time to be more assured of arriving when planned, so that all of the resources needed were available when needed. When we begin to think about the system more as a supply chain system with a myriad of interdependent pieces, taking a longer and seemingly less efficient reroute begins to actually make sense. The overall costs in the end are lower.    How much is predictability worth? We begin to get some sense of its worth when we add up all of the fuel, crew and equipment time spent to reduce the uncertainty in arrival times.

Flexibility is provided to some degree today. During schedule disruptions, working through the FAA’s collaborative decision-making processes, the airlines have the ability to swap their flight positions in a flight schedule. They can decide to take delays on the ground at the origin airport or leave earlier and risk taking delays in the air, where they’re more costly. These decisions enable them to better achieve their individual business objectives, which include passenger satisfaction, utilization and cost of the many different types of resources needed to support a flight, and impact to the overall network schedule. In the next couple of years, they will be able to request multiple flight trajectory options together with the valuation of each option and the conditions under which they apply. The FAA will use this information in automated processes to more dynamically respond to changing capacity conditions. In the end, the airlines will get more of what they want and the FAA will make the most of limited, dynamic airspace resources. In the longer term, flight planning will build more flexibility into flight plans. The increased flexibility in flight plans will allow them to better deal with uncertainties in weather, traffic congestion, and other phenomena. Each of these forms of flexibility enables the airlines to better meet their business objectives and is worth a lot.

On the issue of access, I’m going to look at access from the perspective of communities. Hub and spoke networks have inherent strengths. The greatest of these is their ability to serve so many origin-destination pairs. One schedule bank containing n aircraft in a hub and spoke network can theoretically serve up to n(n-1)city pairs. A typical bank might include 30 aircraft, so for n=30, n(n-1) is 870 city pairs. One downside of the hub and spoke is that to maximize connectivity and efficiency, you have to maximize the flights in a bank and minimize the time it takes to land the bank, unload, service, reload and depart. To work well from an airline perspective, you want the demand pattern to be as peaky as possible. From air navigation service provider and airport operator perspectives, peaky is very bad, since it drives the sizing and cost of your infrastructure and staffing. It’s like feast or famine, so the infrastructure and staff often suffer peaks and valleys and inefficiencies in utilization. Another disadvantage of hub and spoke is that it’s fragile. A few flights get out of sync and things unravel in a hurry. As a result of criticism and problems, the airlines have reduced over-scheduling and de-peaked. As these reductions have occurred, consider what happens when you go from 30 to 20 aircraft in a bank. For n=20, n(n-1)=380, which means 10 fewer aircraft in a bank and up to 490 fewer city pairs served. Many smaller communities fall off, either losing more frequent service or losing scheduled air service all together. What are the regional and aggregated national losses associated with that?

I think Carl was right. It’s time to take a harder look at the economic value associated with predictability, flexibility and access. NextGen needs to support performance in these terms in addition to safety and delays. Have we been ignoring the need for these potential high-value outcomes? We very well may be sitting on top of sizable, untapped pools of benefits. Focusing on predictability, flexibility and access could make NextGen all the more robust and valuable.

As I’ve said before, the benefits story is one that needs to be developed and boiled down into a form suitable for USA Today. This is one attempt to begin that process, trying to answer the question: Why should Joe Shmoe care about NextGen?

The nation’s air transportation system is subject to the fundamental forces of supply and demand. While I’ll mainly talk about supply and demand, and capacity and efficiency, we should note that there is more to NextGen than just capacity and efficiency; there’s safety, security, and the environment too. But, let’s save those for another time.

The demand for air transportation grows with the Gross Domestic Product. Generally speaking, as the economy grows, demand grows. While we’ve had an economic downturn of substantial proportions since 2008, we have seen increasing signs of a recovery and resumption of GDP growth. So, it’s only a matter of time before air transportation demand returns and resumes its steady, long term growth.

Source: Bureau of Economic Analysis

When it comes to capacity, we can look at the interplay between demand levels and delays. What we see is that delays are quite closely correlated to demand. As demand rises or falls, so do delays. The problem here is that delays swing quite dramatically with small changes in demand. This is a sign of a system that is approaching the limits of its capacity. If you’re interested in reading a lot more about delays, the NEXTOR Total Delay Impact Study, published in October 2010, is the best available and most comprehensive study of air transportation delay impacts and costs. It concluded, for instance, that the total cost of delays was $32.9B in 2007 alone. Some of these delays were unavoidable. But, this estimate provides a sense of the scale of one of the big problems NextGen is striving to solve. 

The recent economic downturn has caused demand to fall off. So, delays are down for the moment. But, as the economy recovers and demand grows to entirely new heights, we can expect that delays will increase much faster than demand. That total cost of delay of $32.9B in one year will be a much bigger number when the economy recovers.

Source: Bureau of Transportation Statistics

Why are delays so important? The conventional thinking has been that delays are costly. Delays cause more fuel to be burned, crews have to work longer, and passengers incur more cost because of the cost of their lost time. But, especially as we approach the limits of a system’s capacity, there are even more dire consequences. Delays introduce a significant layer of uncertainty. Schedules become far less predictable. Airlines typically operate each aircraft on a series of sequential flights each day. They are striving to operate each flight according to a published schedule. But, delays cause everything downstream to be disrupted, not just the flights and passengers, but gates, ground crews, and other scheduled and constrained resources, like airspace and runways. Striving to minimize the disruptions potentially adds even more uncertainties. Safety rules limit the total work hours and require prescribed amounts of rest for flight crews, so matching crews to all of these planned flights is a real challenge. Aircraft require prescribed inspections and maintenance actions, which are not readily available at each location. So, ensuring that each aircraft is at an appropriate location for inspections and maintenance is another major challenge. When disruptions and uncertainties are introduced by delays, you reach a point where you can no longer plan and execute effectively. Things begin to unravel. Airlines are not able to make a profit unless they can plan and execute. They have no choice, but to throttle back. They add more time to the published trip times to reduce the effects of delay uncertainties (ever wonder why you often arrive much earlier than planned?). They de-peak hub and spoke operations to reduce the total hourly load on an airport. They increase the connection times between flights, so fewer people miss planned connections. The net effect of all of these kinds of adjustments is lower overall capacity and efficiency of the system. The system is less capable of moving as many people from point A to point B each day.

Reaching the limits of system capacity effectively means that we run out of supply while demand continues to grow. When there is huge demand in the face of limited supply, prices go up – often a lot. They go up until supply and demand again reach equilibrium. How about a $10,000 economy class ticket to get from NY to LAX? In the past, when prices went up, new airlines would form to increase supply and compete on price. But, in the future, with capacity reaching its limits, there will be far less room for new entrants. A few will emerge and fill in wherever available capacity can be found, but soon all of the capacity will be consumed. We can expect other kinds of changes as well. Larger aircraft. Lower frequency of service between cities. Less service or no service to smaller cities and communities. Less access by air will result in less regional economic growth, leading to fewer jobs and economic decline. People will use other viable modes of transportation when feasible. They will drive longer distances to catch a flight, and they’ll pay a whole lot more.

NextGen is about creating a lot more capacity and getting much more use out of existing and new capacity. It’s about battling the shortage on the supply side of the equation. Increasing supply will allow a greater part of the demand to be served. With fewer shortages, the system will become more predictable. The airlines should have an easier time planning and executing their plans, which in turn should allow them to be more profitable. Travelers should see NextGen benefits in terms of service quality and prices. Service quality will come from more predictable and on-time performance, less delay, higher frequency of flights between cities, more destinations to choose from and shorter connection waiting times. It will also come from competition, because there will be more room for competition (maybe one day we’ll actually look forward to flying again). Greater overall supply as well as competition will keep the lid on prices, so travelers will continue to be able to afford to fly.

Clearly the air transportation system is far more complex than this. But, perhaps we’ve allowed this complexity to stand in the way of communicating in an effective way with the broader public. Straight talk is vital to preserving the fragile support for NextGen and transforming it into more robust support. That support can only come when people make a clear connection to what NextGen will do for each of us.

• Forecasts of future volcanic cloud positions and densities
      -More accurate, timely and frequently updated
      -Finer temporal and spatial resolution
      -Forecast models augmented with satellite and other sensor and
       observation data
• Optimal rerouting of aircraft to:
      -Avoid volcanic clouds
      -Provide predictable departure and arrival times
      -Determine the trajectory to be flown, load to be carried and fuel required
      -Verify that the aircraft and equipment can fly the planned route, including 
       extended twin engine operations as well as prolonged exposure to low
       levels of volcanic contaminants
      -Crews, equipment, and resources are available to support the flight
• Sufficient airspace capacity to accommodate demand based on:
     -Airspace sector configurations
     -Communications, navigation, surveillance and automation support systems
     -Controller staffing and scheduling
     -Capabilities to manage the supply-demand problem
• Airport services:
      -Ability to handle arrivals, departures, and other critical services
       (fuel, de-icing, security, etc.)
      -Surface capacity and scheduling
      -Ramp and gate capacity and scheduling
      -Ground equipment and personnel
      -Terminal capacity and scheduling
      -Intermodal connectivity

This is not an exhaustive list, but you get the idea. All of the parts need to work together in a reasonably coordinated way. This means we need to develop coordinated plans and execute those plans in close coordination. Each stakeholder in the system – the aircraft operators, the air navigation service provider, and the airports – has its own set of interests and objectives. To the extent possible, our approach should enable each stakeholder to achieve its goals. In the end, if one of the pieces is not an integral part of the plan, it would be like pulling out one of the gears in a clock – it will stop working.

Management guru Stephen Covey says “Begin with the end in mind.” If we are to develop a coordinated approach that is capable of coordinated planning and execution, we need to begin with the end in mind. We need to deal with all of the stakeholders and all of the moving parts in an integrated way. Instead, we seem to be too focused at any given time on one facet of the problem. We seem to ignore the dynamic nature of the problem. We seem to unnecessarily constrain the problem by imposing the limitations of today’s or yesterday’s system; for example, discussions have focused on the need to have the text and graphics of a  SIGMET match exactly and the definition of restricted areas limited to no more than seven points (see recent SIGMET from the Darwin VAAC concerning the Merapi eruption). We seem to be swept away by abstractions and abstract processes that sometimes distract us from what is going on in the physical world and the physical problem we should be solving. Perhaps because we come from a community of specialists and experts and seem to thrive in stovepipes, we have a natural tendency to want to limit our work to our own backyard and drill down ‘til the cows come home. Unfortunately, these kinds of problems don’t lend themselves to being solved a piece at a time. Even a very rough, integrated approach is likely to have much more success.

There have been years of substantial investments in satellite sensors and imagery, transport and dispersion modeling, air traffic management, and stakeholder collaboration in operational planning and execution. I believe we have all of the foundational pieces needed to develop a far better operational capability to coordinate planning and execution. I’m not saying there isn’t room for more research and improvement across the board. There absolutely is a need for that. But, what we need more urgently is an integrated approach. With a modest amount of additional investment focused on integration, we could have an initial operational capability within 1-2 years. 

The lesson for NextGen is the same. We need to be more top-down. We need to stay focused on the physical world and problem, using abstractions to help rather than distract us. We need a more integrated view and use more integrated approaches. We need involved stakeholders. We need to leverage existing capabilities. We need to deliver valuable operational improvements in 1-2 years, and we can.

RNP procedures have been adopted all over the world. It all started in Alaska in the mid-1990s. There were numerous airports surrounded by difficult terrain. Traditional approach procedures were unsuitable for getting aircraft in and out of these areas with tight terrain under instrument conditions. The use of GPS-based RNAV with special crew training enabled aircraft to thread their way through terrain-challenged areas to land during instrument conditions at airports that would be otherwise inaccessible. The use of RNP procedures grew as many other terrain-challenged locations in the U.S. and internationally began using them.

As time went on, since RNP procedures take advantage of the greater accuracy available with GPS and therefore take up less airspace, people realized that RNP procedures could be used to squeeze new approach paths into tighter spaces. They could fit into spaces that previously were too small for conventional approach paths, and they typically were shorter, so they reduced fuel, carbon and noise. Although little has yet been done, PBN offers the potential for similarly impressive gains during the departure and cruise phases of flight.

Air Traffic Flow Management (ATFM) was introduced in the U.S. in the 1970s in response to the fuel crisis of that era. The airlines complained in particular about the excess fuel they carried and burned due to the excessive holding over New York. A group of innovative individuals out of the FAA’s Jacksonville Center developed some of the initial technologies and procedures to begin balancing demand against available capacity at congested airports. Soon, holding was nearly eliminated. In 1981, when the air traffic controllers went on strike, ATFM took on crucial importance. The system was being propped up by a fraction of the controller workforce, mostly supervisors who worked long hours to keep flights moving. All flights were held at the gate and could not depart until explicitly approved, so that demand could be safely handled by the capacity-limited system.

In the mid-1990s, the FAA and airline industry began working together to create the Collaborative DecisionMaking (CDM) environment. Prior to CDM, the FAA had limited visibility into the airlines’ operation. The only flight schedule information they had was the Official Airline Guide, which came out every two months and contained outdated information when it was printed, not to mention two months later at the end of the cycle. Little or no information was shared, since there was no motivation to do so. In fact, there were outright disincentives; more on that some other time. CDM took off because some of the fundamental incentives were changed to encourage the FAA and airlines to share more timely and accurate information. CDM now includes business aviation, general aviation and the military. The FAA and users each were better able to achieve their respective objectives and better support each other in the process.

During the last 10 years, ATFM has further evolved. The initial focus on balancing demand and capacity across the system has expanded to include an integration of all of the phases of flight; gate-to-gate, as well as actions well before and after gate-to-gate. ATFM today includes synchronization across the system and across all phases of flight – gate, ramp, surface, departures, cruise, arrivals and surface, ramp and gate at the destination. It optimizes the use of scarce resources and addresses constraints during normal and off-normal operations, all in an FAA-user collaborative environment.

As we begin to implement NextGen, the marriage of PBN and ATFM will produce substantial synergies. PBN provides greater efficiency and effective capacity by consuming less airspace and enabling higher traffic throughput in constrained airspace. ATFM balances overall capacity and demand, integrates and synchronizes all phases of flight and supports FAA-user collaboration to enable each stakeholder to better attain its own objectives. Together, PBN and ATFM can enhance capacity and make far better use of that new-found capacity. PBN use should be expanded to the departure and cruise phases of flight. ATFM can use that additional capacity to satisfy more demand and extract more value out of that capacity through synchronization and collaboration.

The marriage of PBN and ATFM provides an early opportunity to implement the best-equipped, best-served (BEBS) concept. ATFM concepts, technologies and procedures can provide a higher level of service to those flights that have higher levels of PBN capabilities. These incentives could encourage users to equip sooner. A faster rate of equipage means more benefits at an earlier time, which always makes for a stronger business case, not to mention a happier and more supportive industry.

Early in the meeting, the airlines criticized the UK Met Office for faulty models that forecast ash clouds that led to shutting down airspace across Europe, while the skies were blue and clear of ash. These critics called for less reliance on models and more reliance on actual observations. Also, they called for routine exercises to test out procedures in the event of another eruption. At the conclusion of that first session, as an ad hoc addendum to the agenda, the UK Met Office took the podium and fired back. They said that their transport and dispersion model had been developed transparently over the last 20 years with the participation of some of the best minds in the world. They said that the model was not in error; the ash was in fact present as predicted, but was difficult to see looking up vertically, but ash clearly was present when at altitude looking at it horizontally. And, oh by the way, there had been regular exercises held. Stakeholders had been invited to participate, but most had chosen to ignore invitations. They called for statements based on fact and data, not fiction. Thus the stage was set for an interesting meeting.

I’m not siding with the UK Met Office, but let me address the airlines’ statement about greater reliance on observations versus models. Dealing with volcanic disruptions requires rerouting traffic to avoid volcanic clouds. You have to know when to depart, what route to fly, how much fuel and load to put on the aircraft, and have a reasonably predictable arrival time. You must know where the flights will travel and the capacities of the airspace that will be traversed; capacities are dictated by things like controller staffing, the geometry of fixes and routes, navigational aids and communications, etc. You must know what the demand and capacities look like. You must know what the weather is going to be. All of this merely is meant to say that you need to be able to develop a plan that works for a lot of different organizations and execute that plan in a coordinated way. It’s a complex problem of harmonization, collaboration and synchronization – a ballet. To plan, you must know what is going to happen. Observations don’t tell you what is going to happen; they only tell you what the conditions are right now. So, to plan you need a forecast. To have a forecast, you must have a model. That’s the reason we are advocating the use of each and every observation to continuously and rapidly update forecast models.

Two engine manufacturers, Rolls Royce and Snecma, gave talks about how much volcanic ash turbofan engines can withstand. The work they described covered April 15-18, that first weekend after the eruptions began shutting down airspace across Europe. Previously, the manufacturers’ standing instructions were to simply avoid volcanic ash, period. During that first weekend, it was clear that simply avoiding all ash meant that no flights would be able to operate, and it was unclear how long the eruption might last. Under immense pressure, they scraped together all the data they could find, which was not much. They developed a graph showing ash density vs. degree of impact (similar to a risk-impact matrix in safety risk assessment). Then, they began placing any available data on the chart. There were the two 747 incidents from the 1980s, both at 2 g per m³ and both causing flameouts – high ash density and high safety impact. There was a NASA DC-8 incident in the year 2000, where volcanic ash had no discernable operating impact, but upon later inspection showed serious damage to the engines that required very costly overhaul. In this incident, they had information on volcanic ash particle sizes, but were not able to determine density. No help from this incident. There had been many other volcanic ash incidents, but no useable data on ash densities. So, they turned to some engine tests involving sand ingestion experienced in Middle East operations. The density of sand happened to be in the range of 1-2 mg per m³. Sand is less abrasive and melts at a different temperature than volcanic ash, but at 2 mg per m³, engines were not experiencing significant damage from sand. So, this became the basis for setting a 2 mg per m³ threshold, which then enabled aircraft to begin flying again around April 19. While some operators reported ash deposits in their engines, there were no reported cases of permanent engine damage.  ICAO suggests delaying the adoption of this density threshold until further study can be completed on engine performance and variations in volcanic conditions on a global basis, but the European standard appears to be set now at 2 mg per m³. If we have another event like this in Europe, it will be interesting to see what happens. Maybe everything will be fine, but the paucity of data leaves open the possibility for massive numbers of engines being damaged. Stay tuned.

As we watched the Eyjafjallajökull eruption and ensuing disruptions, we felt that some of our Air Traffic Flow Management (ATFM) technologies could be adapted to improve the safety and efficiency of operations in the presence of large, moving volcanic clouds.

We believe that ATFM facilities should be leveraged as part of the solution to volcanic ash disruption problems. On one level, rerouting traffic around volcanic clouds is analogous to strategically rerouting traffic around severe, convective weather systems. This is a job that ATFM has been doing well for many years. On another level, ATFM has nurtured a Collaborative Decision-making (CDM) environment, which enables all of the stakeholders to jointly respond to operational problems. CDM enables air navigation service providers to engage the airlines, general aviation, the military, airport operators, and regulators to collaboratively develop a plan that works for all stakeholders and support the coordinated execution of those plans. One of the advantages of relying on ATFM facilities and personnel is that ATFM is a 24×7 operation. Only modest expansion of functionality and responsibility would be required to support volcanic disruptions. Processes, procedures, people, relationships, trust and the infrastructure are in place and in daily use. These facilities exist or are being built in the US, Canada, Europe, South Africa, Japan, Australia, Mexico, Brazil, Colombia, and several other nations. In contrast, a crisis cell that isn’t used continuously, but only when emergencies occur, may not be able to function nearly as well due to learning curve problems and lack of practice.

To effectively respond to volcanic disruptions, an optimized traffic flow strategy is needed. We invested in some internal R&D to adapt one of our technologies to use in solving the volcanic ash disruption problem. We prototyped a tool that begins to synthesize optimal traffic flows to maximize the use of open airspace and avoid volcanic clouds while minimizing the cost of flights. This is a sample of an optimized flow strategy synthesized to respond to a forecast of Icelandic volcanic cloud positions and densities. Much more could be done to enhance the performance of this prototype, but it’s clear that it generates a decent solution to a pretty complex problem.

[youtube]http://www.youtube.com/watch?v=5Ojr4_3xwtk[/youtube]

Our prototype that synthesizes optimal flow strategies is off to a great start. But, more work is needed to develop accurate volcanic cloud forecasts. As I mentioned at the outset, forecasts are generated by transport and dispersion models. Models need to know wind forecasts from the surface through cruising altitudes. Temporal resolution of forecasts needs to be down in the range of a few minutes, certainly more frequently than hourly, and spatially at each Flight Level. Models also need input data about the volcanic eruption: the mass of materials injected into the atmosphere, the altitude, the distribution of particle sizes, etc. These are extremely difficult to obtain, so wild guesses or default values are often used in the absence of anything better. For this reason, the first forecast coming out a model will be suspect. But, a wealth of satellite sensors are available, the data from which observations may be obtained. In addition, there are ground-based sensors, balloons, pilot reports, and eventually there may be many more aircraft-based sensors. Observations could be compared with modeled forecasts to enable adjustment of the original model inputs, so that the modeled forecast matches the latest observations. In this way, future forecasts would be much more accurate. This is done today, but the process is largely manual, an art form of sorts. The process needs to be automated and fast, so that each and every observation can be used immediately to improve the accuracy and reliability of forecasts. There has been much good work done in the past that provides many of the required technological capabilities: satellite sensors and processing, ground lidar networks, transport and dispersion models, etc. But, little has been done to integrate these technological capabilities to produce a greatly needed operational capability to support aviation. After studying this problem for the last few months, our conclusions have been further reinforced through the Eyjafjallajökull Conference. We must find a way to leverage huge prior investments and integrate the technologies – connect the dots — for a valuable operational capability.

Over the years, I occasionally have heard of volcanic events occurring in various parts of the world. Typically I heard only about the severe ones. You may recall the two major events from the 1980s involving Boeing 747 aircraft that encountered volcanic clouds. The clouds were of sufficient density to cause engine flame outs and a complete loss of engines. The first one occurred in 1982 when a British Airways 747 flying to Perth, Australia, encountered a volcanic cloud from an eruption in Indonesia, lost 20,000 feet and fortunately was able to restart three of four engines for an emergency landing in Indonesia. The second occurred in 1989 when a KLM 747 encountered a volcanic cloud from Mt. Redoubt south of Anchorage, lost 26,000 feet before managing to restart its engines. Although no lives were lost, the severe 1989 encounter resulted in $80M in repair costs to a $125M airplane – replacement of engines, aircraft skin, windows, and more.

Taking a closer look at aircraft damage, it’s interesting to note that in February 2000 an instrumented NASA aircraft encountered a volcanic cloud north of Iceland. It was flying from Sweden back to California at night following an ozone data collection mission. The DC-8 had been re-engined with modern, high bypass ratio, CFM-56 engines. It was packed with air quality measuring equipment, which was operating during the flight. The flight flew 200 miles north of the reported extremity of volcanic clouds emanating from Hekla, another Icelandic volcano. The sensitive air quality instrumentation detected an encounter with the volcanic cloud, which lasted for seven minutes. Since all engine functions and instruments appeared normal, the flight continued back to California without diverting. However, upon post-flight inspection, it was discovered that the engines had incurred some damage and required overhaul and refurbishment at a cost of over $3M.

While the public hears about the severest of events, few hear about the more subtle encounters. Yet, the magnitude of maintenance costs, even from brief encounters, can be substantial. How many revenue flights are needed to generate $3M in profit, the cost of overhauling engines from one incursion?

One of the problems encountered by the Europeans was the lack of useful aircraft and engine specifications for ash exposure. The current guidance from the manufacturers had been to simply avoid encounters of any kind. Simple enough to say, but seemingly too cavalier when an entire continent is shut down for days at a time. After several days of severely curtailed operations and mounting public pressure to deal with millions of stranded travelers, the manufacturers and regulators arrived at an interim ash density threshold of 2mg per cubic meter (the estimated ash concentration from the Mt. Redoubt encounter in 1989 was 2g per cubic meter, or 1000 times greater than the interim adopted level). This interim step at least allowed traffic to start moving again. No safety problems appeared to occur as a result.

However, considering the results of the NASA DC-8 encounter in 2000, a scientifically based set of standards appears to be vitally needed. Flying around volcanic clouds will likely be driven more by maintenance costs than safety, since maintenance costs will dictate ash density limits that are far lower than any safety limit. In addition, instead of a single density threshold, standards likely need to be a combination of density and duration of exposure. Finally, while most of the media attention has been on volcanic ash, it should be noted that volcanic clouds often contain significant amounts of sulfur dioxide or SO₂. Sometimes the ash and SO₂ are not in the same place. When combined with water in the atmosphere, SO₂ forms sulfuric acid, which corrodes aircraft skin and damages windows and other components, and thus results in substantial costs due to reduced aircraft service life. So, we have to pay attention to both ash and SO₂.

In response to the severe encounters in the 1980s, in 1995, ICAO established Volcanic Ash Advisory Centers (VAAC) as part of the International Airways Volcano Watch program. Nine VAACs were established under their respective host nation’s weather forecasting agency. The US has 2 VAACs (Anchorage and Washington) operating under NOAA. While VAACs have provided some degree of warning, they have a way to go. For instance, when Mount Pinatubo erupted in 1991, 20 aircraft were damaged due to inadvertent encounters, most over 600 miles from the eruption. There are numerous aircraft encounters in the absence of any warnings or advisories, sometimes requiring a lot of detective work to determine where the volcanic cloud might have come from.

In the case of Eyjafjallajokull, there was no confusion about the location of the eruption. While the images broadcast by the media might have suggested that there was complete understanding of the position and density of the cloud, in fact, there was great uncertainty about the current and forecast position and density. Many of the satellite sensing data required significant time to process before the resulting images could be produced. There was a lack of timely, accurate information and a lack of common situational awareness to support planning by air navigation service providers, airlines and other users, airport operators and regulators.

Many experts in European aviation, including the people who were directly involved, said that they were unprepared for this kind of event. Was this a rare event? Hardly. The diagrams below provide an idea of where active volcanoes are located, how often eruptions occur and how big they potentially could be. The USGS tracks 1500 active volcanoes worldwide.

Should we be doing more in the US and globally to be better prepared for these kinds of events? Do the financial losses incurred in the wake of Eyjafjallajokull suggest more investment to be better prepared? You be the judge.