Update: President's Day go travel
Update: President's Day go travel
Wobble in placement of heaviest snow for ORD; deteriorating picture for Northeast
Welcome to any new readers—we're grateful that you're here! We're building deep learning algorithms to democratize flight delay predictions; until we launch, we're eager to synthesize things manually in our outlooks. These feature several recurring themes that we recognize may be unfamiliar or intimidating, so we’ve written explainers that tackle airport arrival rates, queuing delays in the airspace and different tools to distribute those delays. If there’s a topic or mechanism you’d like to see unpacked, please let us know (same goes for special travel occasions).
In yesterday’s first look, we estimated that the TSA would screen 2.20 million travelers on Friday. With the benefit of one more data point, we’re nudging our forecast for tomorrow’s TSA throughput up to 2.26 million travelers. Even having bumped up our prediction, we still lean towards checkpoint numbers outperforming our estimate owing to influences from NYC Public Schools’ midwinter recess and the Daytona 500. While the large-scale weather pattern outlined yesterday is still intact—low pressure over the Souther Plains is already racing northeast along an attached cold front—some finer details have importantly come into focus.
Chicago O’Hare (ORD)
We’ll start again in ORD, where they’re currently dealing with a [hopefully] quick burst of heavy snow. Yesterday we had expressed some concern about how they might emerge Friday morning from today’s snow event—though damage has been done to the integrity of tonight’s operation, the outlook for tomorrow has improved. ORD was already on the northern periphery of the highest forecasted snow totals when we published yesterday; overnight, the placement of the heaviest band was shifted southward by a not insignificant amount (about 50 miles). While there’s since been some northward wobble to heavier snow, forecasted snow accumulations for ORD are now more like 2-4”, with 4-6” occupying the high end of the range. Snow is still forecast to end late this evening. Moreover, winds are forecast to lie down after midnight for a period, which assuages our anxiety that blowing snow would hamper efforts to clean-up the airfield overnight. Taken together, we feel reasonably confident in revising our estimate of ORD’s opening arrival rate tomorrow up to 1021. This should be sufficient capacity to accommodate the 88 scheduled arrivals in the 7 a.m. hour without delay2.
Philadelphia (PHL)
Before we get to PHL, we do want to close the loop on Atlanta (ATL) and Charlotte (CLT). The cold front and any associated convection still looks clear the region by 6 a.m. tomorrow, thus we maintain our position that air traffic delays should be zero.
Unfortunately, the Northeast seems set to pay for the sins of an improved ORD outlook. As the front comes through early morning, blustery winds will shift to the west-northwest. While the strongest winds are forecast in the immediate wake of the frontal passage—with gusts to 38 knots at PHL—winds diminish only somewhat into the afternoon. In what will be a recurring theme, we’re prompted to revise downward our estimate of PHL’s arrival rate floor to 32 (less than 1 in 10 chance for lower rates). In what will also be a theme, the floor itself is not a fringe scenario, with a nearly 1 in 5 chance for rates equal to 32 at PHL. There’s 3 hotspots where a 32 rate would be problematic: the 8 a.m. hour (45 scheduled arrivals), noon hour (37) and 6 p.m. hour (46). In the case of the morning and early evening overages, we’d bet the FAA reaches for a first-tier ground stop, which distributes delays disproportionately to those flights originating from within 700 miles or so; we’ve modeled average delays3 of approximately 28-33 minutes for captured flights in these scenarios. In the case of the midday imbalance, we’d bet miles-in-trail could handle the overage; we’ve modeled average delays of 10 minutes for all flights in this scenario.
New York City
LaGuardia (LGA) is forecast to see the strongest winds of the NYC airports, with gusts to 45 knots. We think there’s a 1 in 8 chance for arrival rates in the range of 26-28 and at least 1 in 4 chance for arrival rates equal to 30. We’ve modeled a first-tier ground stop in the 7 a.m. hour that rolls into an all-day ground delay program (28-rate for the 8 a.m. hour, followed by a 30-rate until 5 p.m., then 36-rate thereafter). These assumptions produce average delays of 39 minutes across all the day's arrivals, with average hourly delays peaking at 60 minutes in the 3 p.m. hour. While we wouldn’t bet on the day unfolding exactly as prescribed, it does convey a sense for potential disruption.
Across the river at Newark (EWR), winds are forecast to gust to a still-impressive 41 knots. Here, we’ll set our arrival rate floor at 35 (less than 1 in 10 chance for lower rates). Though tail risks, we should mention two particularly consequential possibilities:
There’s a relatively narrow window (approximately 7-9 a.m.) after winds have shifted to the northwest but before ceilings scatter out that some arrivals may not be able to land4, resulting in airborne holding and potential diversions.
After ceilings scatter out, but before winds meaningfully diminish—let’s say mid-morning into the early afternoon—arrivals and departures may be required to share Runway 29 at an arrival rate around 24. Thankfully, delays would be somewhat offset by the fact that this midday period corresponds to a relative lull in EWR demand (and departures figure to bear the brunt of delay in this scenario).
In terms of more likely outcomes, we estimate there’s at least a 1 in 4 chance for arrival rates in the range of 35-38. In this case, a ground delay program looks to be likely, but we wouldn’t be surprised by a 1400-mile scope (rather than the entirety of the national airspace system). Assuming a 35-rate until 5 p.m. then a 38 rate thereafter, hourly average delays would peak at around 30 minutes for flights originating with 1400 miles.
Boston (BOS)
Up at BOS, gust speeds are forecast to match LGA’s 45 knots (and sustained speeds are actually forecast to be even stronger). While we think rates less than 32 are improbable (1 in 10 chance), we wouldn't bet against arrival rates in the range of 32-34 (1 in 4 chance). BOS should be able to navigate a 32-rate without much delay into the afternoon, however by late afternoon, some overages would be present. With 36 arrivals scheduled in the 4 p.m. hour then 32 in each of the next two hours, the difference between a 32-rate and 34-rate will be appreciable. We think a first-tier ground stop would be required if stuck at 32, with delays averaging 29 minutes for captured flights; if they can squeeze out a 34-rate, we bet they manage with miles-in-trail, with resulting delays being largely imperceptible.
Recommendations
Though additional ORD cancellations are going in as a result of this burst of heavy snow, we still think whatever’s left on the books tomorrow morning should have a clean airfield to navigate. Unfortunately, predicability is lower than normal with respect to what might yet cancel, given that this is a fairly reactive schedule reduction. If readers holding an itinerary that involves ORD tomorrow don’t wish to wait-and-see if their flight gets cancelled, we don’t blame them—in these cases, United and American have both issued travel alerts that waive any fare difference.
Predictability is better for travelers transiting PHL, LGA, EWR and BOS. The good news is, unless one of the EWR tail risks verify, we don’t expect disruption to necessitate cancellations. Except for cases where there’s little buffer on the arrival end (perhaps a belated Valentine’s reservation), non-stop itineraries should remain useable, albeit with some adjustment to delay expectations being appropriate. There are a couple connecting itineraries that would prompt us to consider alternatives:
a layover in PHL around 9 a.m. that originates from one of the first-tier airports highlighted above;
though there’s relatively less connecting activity at LGA, any layover that cannot absorb a delay of at least 40 minutes;
evening layovers in EWR that cannot absorb a delay of 30 minutes;
a layover in BOS around 5 p.m. that cannot absorb a delay of 30 minutes.
While a quick check of weather waivers revealed no applicability for Northeast airports, airlines have meaningfully improved general rebooking flexibility by eliminating change fees for most tickets (though a fare difference may still apply). We’ve linked to the same-day chance policies for American, Delta and United.
We’re now using January 28, 2022 as a proxy, when ORD received 3.3 inches of snow leading into their start-up: on this day, a 102 arrival rate was restored 4 hours after the end of snow.
Cargo airlines as well as private jets are not included in scheduled demand and only become apparent when they file a flight plan (generally day-of). This unforeseen demand introduces the risk that delay probabilities/intensities are under-forecast. Unscheduled demand added 9.8% to scheduled demand for Core 30 airports between 1/3/2022-2/16/2022. For the airports we’ll consider today, only PHL (17.5%) is above average in this regard.
Additionally, our efforts are aimed at diagnosing air traffic delays (i.e. those that result from an imbalance between capacity and demand). Though not the focus of our efforts (yet), delays owing to aircraft servicing, airline staffing, network effects, etc. are always lurking.
While our modeling is aimed at tackling arrival delays, there's a strong correlation to departure delays (albeit with some lag and/or possible alleviation). Consider a scheduled "turn" at an airport: the inbound flight is scheduled to arrive at 2:19 p.m. and departs at 3:30 p.m. (71 minutes of turnaround time). Let's say the inbound is delayed by 40 minutes and instead arrives at 2:59 p.m. We'll further assume that the airline doesn't need the full 71 scheduled minutes to turn the aircraft and can accomplish the turn in 45 minutes if they hustle - the departure will push back from the gate at 3:44 p.m. (delayed by 14 minutes). In this example, a 40 minute arrival delay in the 2 p.m. hour is partially passed through to a departure in the 3 p.m. hour. Had the turnaround been scheduled at 45 minutes instead (i.e. no turnaround buffer), the lag between arrival and departure delay would still exist, however the delay would be fully passed through.
Wherein the crosswind for 4R-22L would require a 29 arrival, however ceilings require an instrument approach—these instrument procedures may be unapproved on a carrier-by-carrier basis.