TBFM, TMI, GDP, GS, AFP, BOSOX, MIT, CTOP, ZNY? Unpacking the FAA's traffic management toolbox.
This is the third explainer in a trio that makes up an air traffic 101 of sorts. If you haven't read the first two blogs, we'd encourage you to do so: we covered the concept of airport capacity in our first explainer and demonstrated how delays are created in our second post.
When we previously created some delays at our hypothetical airport, we most closely simulated a ground delay program (GDP). As the title suggests, however, there's a veritable alphabet soup of traffic management initiatives (TMI) available to the Federal Aviation Administration (FAA) to manage an imbalance between demand and capacity. While we've largely focused on disequilibrium at airports so far, a volume of airspace can similarly get imbalanced. In such a case, an airspace flow program (AFP) will most visibly create delays. Like a GDP, an AFP assigns flights an expected departure clearance time (EDCT, though often referred to as a wheels up time) to align it with an arrival slot; whereas a GDP's arrival slot correspond to the time at which the flight can land on the airport's runway, an AFP's arrival slot correspond to the time at which the flight can transit that volume of airspace. Because both rely on an EDCT to dispense the delay, flights delayed by an AFP are also held on the ground at their origin station (though unlike a GDP, an airline can re-route the flight around the impacted airspace). AFP's typically capture flights during their en-route phase (i.e. generally at or above 12,000') and most commonly owe to thunderstorms (though seasonal volume to ski or beach destinations can also create a demand overage).
Let's provide a quick AFP example before we move onto the next traffic management tool. Consider a hypothetical summer afternoon flight from Chicago O'Hare (ORD) to New York LaGuardia (LGA). Weather is favorable on both ends, however clusters of thunderstorms are moving east out of Ohio into Pennsylvania. Commercial flights are not able to climb above our hypothetical thunderstorms (whose tops reach 50,000') and instead must pick through gaps in the thunderstorms. The FAA will respond by increasing the separation between aircraft operating in this sector of airspace: this ensures that they have adequate space to deviate in search of gaps. Increased spacing in a volume of airspace reduces its capacity (just like at an airport) and soon a mismatch develops between the time a flight was planned to transit that sector and when the sector will be able to accept it. Whatever delay that is assigned to our hypothetical ORD-LGA flight will be absorbed on the ground in ORD; alternatively the airline could re-route the flight so that it avoids that volume of airspace (perhaps out over northern Michigan, across Toronto and dropping in through western Massachusetts). The re-route doesn't eliminate the delay as much as it shifts it to the air (such a route would increase flying time by about 35%), though delays associated with EDCTs can occasionally be extensive enough that its sensible to elect the reroute.
GDP's and AFP's are both TFMS programs (traffic flow management system). Adjacent to TFMS in the FAA's toolbox is a set of options that is of the type TBFM (time-based flow management). TBFM is generally beyond the scope of this explainer, but we do want to mention metering. Metering functions almost exactly as a GDP, however it's geographical scope is limited to flights originating from airports that "neighbor" the impacted airport and the wheels up times are managed less formally. Unfortunately, the informality means metering is not included on the FAA's very handy (and recently redesigned) National Airspace System Status dashboard. While quite technical, metering would be included in the Current Restrictions page, which warrants its own explainer (for now, we might suggest use of Ctrl/Command + F to search for 'METER' and/or your arrival airport).
Let's shift gears to ground stops (GS), which we've mentioned in our disruption outlooks and also bear some resemblance to GDPs. Like a GDP, a GS results from a reduction in capacity at an airport and imposes their associated delays at flights' origin airports. Unlike a GDP, flights captured by a GS are not released on individual EDCT's but rather when the GS expires. GS's are often used as a last resort and may occur with little (or no) warning - their nature is more indefinite than a GDP as a result. When a GS is issued, an update time is published (as is a probability of extension), though its end time is undefined. GS's are liable to be extended multiple times before being allowed to expire or may cancel prior to an update time. GS's do, however, provide us an opportunity to cover the concept of tiers.
US airspace is divided into 21 air route traffic control centers (ARTCC, or just "centers") that marshal en-route traffic (plus 5 contiguous Canadian facilities). This link is a great resource for visualizing the boundaries for these centers, though [to the best of our knowledge] it is only accessible through the legacy National Airspace System Status page (which is being sunset this November). We've provided feedback to the FAA and hope it will be accessible going forward. It's also a great resource for understanding the concept of tiering, because you can select a center and it will highlight its first and second tier centers. For a given center, its first tier includes any contiguous center (and its second tier includes any center that touches its first tier). Tiering is important because GS's, more so than GDP's, rely on them to limit their geographical scope. For example, a first tier GS may be issued for LGA, which is within New York center (ZNY): this GS will only captures flights departing from airports in Boston center (ZBW, e.g. BOS or PWM), Cleveland center (ZOB, e.g. CLE and DTW) and Washington center (ZDC, e.g. DCA or RDU). As a result, GS's tend to disproportionately distribute delays to flights that cover shorter distances. GS's also offer less flexibility to airlines, as their is no ability to swap slots or "route out" of the constraint.
To this point, we've concerned ourselves with flights that are inbound to an impacted airport or airspace. Air traffic delays, however, are also possible owing to the departure phase of a flight. When a flight takes off, it's initially under the control of the local airport tower before it's quickly handed off to a TRACON facility; TRACON then handles the flight through the first 30 to 50 miles of its path. TRACON, in turn, hands off the flight to the previously mentioned ARTCC's for the flight's en-route phase. The handoff from TRACON to an en-route center takes place at a departure fix, which is a specific latitude and longitude along the boundary between TRACON and ARTCC. As an aside, fix naming is one of the quirkier parts of the national airspace system. The FAA requires fix names be 5 letters in length, but otherwise encourages that it be an easily understood word that is representative of the city. Boston TRACON, for example, leans heavily on the region's sports teams and you'll find departure fixes named BOSOX, BRUWN (BRUIN was already taken), CELTK and PATSS. If you're curious, you can check out the terminal procedures diagrams for your home airport (select one of the diagrams of type DP - for departure procedure - that does not include RNAV in the procedure name and fixes will be identified by a triangle).
If thunderstorms populate the space between an airport and a departure fix (or sit directly over the fix), then mechanisms similar to our AFP example play out to reduce that fix's capacity. Moreover, not only can a fix's capacity be partially reduced, but departures via a fix can be suspended entirely. Should capacity only be partially reduced, the FAA can space departures via the same fixed based on miles (called miles-in-trail, i.e. MIT) or minutes (MINIT). Let's return to that hypothetical summer day in our AFP example, however we'll fast forward a couple hours to when the thunderstorms have moved just inside the Pennsylvania-New Jersey border. We'll also consider a flight in the opposite direction this time - leaving LGA, bound for ORD. Fix capacities are far more fluid than airport or en-route airspace capacities and it's not uncommon for a fix to be shut down in the time between a flight pushing back from the gate and being ready to takeoff. Let's assume our LGA to ORD flight has exactly that happen - during its taxi to the runway, New York TRACON's westbound fixes are stopped entirely. The airline can elect to wait it out, though an estimated resumption time is not published, or re-route the aircraft to depart via an available fix (perhaps southbound, in this case). Re-routing a flight to depart via a different fix is less consequential than re-routing a flight to avoid an en-route airspace (as in our AFP example): in this scenario, it might mean heading south to Atlantic City before jogging west to Chicago (increasing flying time by approximately 10%).
Unfortunately, because these re-routes are liable to happen whilst taxing to the runway, it's also common that the flight turns around for a gate to receive additional fuel (... by which time their original fix might have re-opened). While observability can be low (like metering, constrained fixes can only be found on the current restrictions page), departure fix capacity is nonetheless one of the trickier elements of an operation to manage. Thankfully, improving adoption of the FAA's collaborative trajectory options program (CTOP) should yield increasingly imperceptible re-routes, as flights push back from gate with not just one route, but multiple [for which paperwork is already done and the plane is fueled for].
The last TMI we'll cover is airborne holding (mercifully, this one is not commonly turned into an acronym). Much like flights exit TRACON airspace via a fix at the end of the departure phase, flights similarly enter TRACON airspace via a fix at the start of their arrival phase. The stakes are higher, however, when it comes to arrival fix equilibrium - while delays associated with reduced capacity of a departure fix are applied on the ground (whether at the gate or during taxi), delays resulting from imbalanced arrival fixes are absorbed in the air. If a flight is made to queue in the oval holding pattern for too long, the pilots will eventually exhaust their "hold" fuel and be required to divert to an alternate airport. Because diversions are a particularly severe type of disruption, airborne holding is only used in a planned capacity when the imbalance is expected to be short-lived. Such instances typically owe to bunching of demand rather than a reduction in capacity. Imbalances that result from a reduction in arrival fix capacity (e.g. thunderstorms shutting a fix down) are much more likely to result in diversions, as any forecasted improvement in capacity - and therefore the duration of the airborne hold - is less certain. Airborne holding ranks with GS's in terms of sub-optimality, as it offers little room for an airline to choose how they distribute delays.
Whew! That was a long (and occasionally technical) one - we appreciate you sticking with us! Send us a suggestion for what topic you'd like to see us cover next.