[Thanksgiving 2021] Sunday Nov 28 update
We're back with refined delay predictions for air travel tomorrow! As we discussed yesterday, airlines and airports are set to face their first quiz of the winter season (we wouldn't quite call it a test). At the same time, we're eager to see if TSA throughput crosses the 2.5 million passenger mark for the first time since pre-COVID. The weather backdrop we outlined in our first look remains largely intact and some features are already impacting operations tonight - let's dive in.
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. We'll frequently refer to airport arrival rates, queuing delays in the airspace and different tools to distribute those delays. If these concepts are unfamiliar (or intimidating), we'd encourage you to check out our explainers.
Before we refresh our winter weather outlook, we'll start with Seattle (SEA) again, where the plume of moisture associated with another atmospheric river has already moved inland. Low end visibilities for tomorrow have been raised and their onset pushed later in the day since we published our first look. This allows us to knock down our probabilities for a depressed arrival rate from yesterday's estimate of 1 in 5 to an updated 1 in 10 chance. It also allows us to narrow our focus to evening operations, including 44 and 45 scheduled arrivals in the 6 and 8 p.m. hours, respectively. Weather forecasts don't quite offer enough granularity to distinguish between those two hours, but clouds are forecast to lower during this time. It's possible ceilings will hold out for the 6 p.m. hour but conditions will have deteriorated by the 8 p.m. hour. That said, to provide a reasonable worst case scenario, we'll assume SEA's arrival rate drops to 38 before 6 p.m.
In such a scenario, delays would average about 10 minutes for arrivals between 6:00 p.m. and 10:59 p.m. There's some question as to how these delays would be applied - either metering or a ground stop would disproportionately apply delays to flights originating within approximately 600 miles of SEA (i.e. first tier). Average delays could approach 25 minutes for first tier flights in either case (though effects of a ground stop would be more readily visible). Airborne holding would be the other option to administer delays, which would dispense delays on a last-come, last-serve type basis (and maximum delay would be around 25 minutes). 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). And of course, delays resulting from aircraft servicing, airline staffing, network effects, etc. are also always lurking (and not included in our modeling).
We'll start the winter weather coverage by briefly mentioning Detroit (DTW), where any moderate snow should taper off shortly after midnight tonight as the clipper exits into Ohio. A round of lake effect snow is possible Sunday afternoon and evening, however intensity will be more showery and additional accumulations look to be less than half inch. We maintain our stance from yesterday's first look: air traffic delays at DTW should be virtually zero, though de-icing delays should be expected in any snow showers. We'll similarly reiterate our prediction of virtually zero air traffic delays for New York-Kennedy (JFK).
Newark (EWR) and LaGuardia (LGA) warrant more discussion, however. It appears precipitation associated with this system will come in a couple rounds. Though forecasting precipitation type continues to be a challenge, precipitation amounts should be light. Some showers - a mix of rain and snow - are expected Sunday morning into the midday on the front end of the system. By Sunday night, a secondary low develops offshore and brings another chance for precipitation. Departure delays to allow for de-icing can be expected anytime frozen precipitation is actively falling. Otherwise, precipitation type doesn't figure prominently in our estimates, as any snow accumulations fall short of requiring runway closures for treatment. Nevertheless, we still like yesterday's 1 in 10 chance for a depressed rate at EWR on account of ceilings/visibility, though think a 36 rate is a better floor than 38. If the FAA elects a wait-and-see approach, they'll be required to lean on ground stops if lower rates materialize; in this scenario, delays for first tier arrivals could approach 60 minutes for much of the late afternoon and evening. Alternatively, the FAA could elect to implement a GDP earlier in the day, which likely casts a wider geographic net; with the same amount of delay spread delays across more flights, delays for all flights arriving after 13:00 average about 15 minutes (peaking at 22 minutes in the 8 p.m. hour).
LGA's forecast is a bit more squirrelly, as earlier guidance hinted at a northerly component to winds. That northerly component has since been removed and we'd currently put chances around 1 in 20 that LGA delivers a 36 rate or worse. There is less demand on LGA's capacity, however, so a 36 rate would be less destructive relative to EWR: there's really just one notable overage in the 1 p.m. hour that we'd bet is managed with a ground stop. We've modeled a first tier ground stop in the 1 p.m. hour that produces average delays of about 20 minutes for captured flights (with delays lingering into arrivals scheduled for the 2 p.m. hour). If the northerly component returns in future forecasts, there's additional downside risk.
More uncertainty still exists in Boston's (BOS) forecast than we would prefer at 24 hours out, though at least there's no longer disagreement on large scale features. The remaining uncertainty is related to the track of the secondary low as it either crosses over or east of Boston - a track that is farther offshore favors lower snowfall amounts. Before the low lifts over or east of the area, much of Sunday should be quiet and dry. Precipitation chances increase during the evening and high-end snow amounts are still only around half inch. We're confident the arrival rate lowers to 36 once snow starts, though don't foresee further reductions for snow removal. A 36 rate won't create any arrival demand overages, however de-icing delays should be anticipated for departures after 8:00 p.m or so.
As we mentioned yesterday, there's likely to be a pair of themes that compete for influence as tomorrow plays out. Airlines and airports typically over-prepare for the first snow event of the season, which should provide a degree of insurance against high end snow amounts; unfortunately, we'd bet airlines will see an uptick in sick calls for baggage handlers on Sunday (watching football on a couch will prove too attractive relative to a cold, wet shift for some), which will exacerbate existing staffing shortages. Though the wheels shouldn't come off as a result, we think the influence of staffing outages narrowly wins out. We'll also introduce a third theme on account of difficulties forecasting precipitation type: mixed precipitation types (e.g. sleet, a rain/snow mix, ice pellets, freezing rain) can complicate aircraft de-icing. An explainer on de-icing [and holdover times] is definitely beyond the scope of this outlook, so we'll point you to Flightradar24's well-done blog on the topic. For now, you should know that mixed precipitation types can require that aircraft be de-iced multiple times. Though forecasted probabilities are slim to none for freezing rain or ice pellets, in our experience, they find their way into METARs when the rain/snow line is in the vicinity of the airport.
In terms of recommended actions for travelers, we're most concerned about any trips that include a planned layover in EWR under 45 minutes (especially if originating from the Mid Atlantic, eastern Great Lakes or New England). We think it'd be prudent to at least explore rebooking options if this describes your itinerary. The good news is airlines are increasingly offering rebooking flexibility (including no change fees on most tickets; we've linked to United's same-day change policy). Similarly, readers starting their trip in the Mountain West or NorCal and connecting in SEA during the evening should ensure their layover can absorb an arrival delay of approximately 25 minutes (we'd expect fewer itineraries fit this definition). Otherwise, de-icing delays should be expected for much of the Great Lakes, New England and Tri-State area, though we think adjusting expectations should suffice in most cases.
For the moment, this concludes our Thanksgiving outlooks, though we'll be active on Twitter tomorrow - we appreciate all the readers who followed along! (We're also keeping an eye on some wraparound precipitation for the Northeast on Monday and will publish an outlook if delays are likely).