Fog vs Thunderstorm Flight Delays: Predictability Profiles
Fog and thunderstorms produce opposite flight delay profiles — one short-horizon and high-confidence, one long-horizon and dispersed. How to read both.
Fog and thunderstorms sit in the same “weather” column of every delay report, and that is roughly where their similarity ends. For a trader pricing flight delay event contracts, the two phenomena pose opposite information problems: fog is a short-horizon event that forecasts resolve with unusual confidence, while convective weather announces itself days ahead and then refuses to commit to an hour, an airport, or even a route. This guide builds the two predictability profiles from primary aviation-weather sources and maps what each one means for reading delay markets.
Why “Weather Delay” Is Not One Thing
Official statistics treat weather as a single cause. The U.S. Bureau of Transportation Statistics counts a flight as delayed when it arrives 15 or more minutes behind schedule, and its delay-cause reporting sorts delays into five categories — air carrier, weather, national aviation system, security, and late-arriving aircraft — prorating each cause by the minutes it is responsible for. That framework was built for accounting after the fact, not for anticipation before it.
Anticipation is what a market participant needs, and it depends entirely on the physics of the specific phenomenon. Two questions define a predictability profile: how far in advance reliable information exists, and how tightly that information constrains the outcome. Fog scores short on the first question and tight on the second. Thunderstorms score long and loose. Everything else in this comparison — morning versus afternoon clustering, airfield versus route exposure, which instruments deserve screen space — follows from those two answers.
Fog: The Short-Horizon, High-Confidence Profile
Radiation fog, the variety that most often blankets inland airports, is produced over land “when radiational cooling reduces the air temperature to or below its dewpoint,” in the FAA’s definition, which makes it “generally a nighttime occurrence” that often persists past sunrise (FAA AC 00-6B, Aviation Weather, ch. 16). The conditions that favor it are all observable the evening before: a shallow layer of moist air near the surface, clear skies, and light winds. The FAA adds a usefully hard number — fog seldom forms when the temperature-dewpoint spread exceeds 2 °C (4 °F).
That is what a high-confidence setup looks like. By late evening, the ingredients are either present or they are not. Better still, the event carries its own expiry: per the same circular, ground fog “usually burns off rather rapidly after sunrise,” and other radiation fog “generally clears before noon unless clouds move in over the fog.” A radiation fog event is bounded on both ends — a formation window overnight, a clearing window before midday — which is exactly the kind of structure a short-dated event contract can be priced against.
Advection fog, the coastal variant, forms when moist air moves over a colder surface. It is less punctual: it can arrive at any hour, survives wind that would disperse radiation fog, and actually deepens as wind speed increases up to about 15 knots. But it is geographically honest — the FAA notes it is most common along coastal areas — so exposure is a property of a handful of marine-influenced hubs rather than of the whole network.
Visibility, RVR, and the Approach-Category Ladder
Fog rarely closes a major airport outright. What it does is push operations down the instrument-approach ladder, one measurable step at a time. U.S. ILS minimums run from Category I — decision height 200 feet and runway visual range 2,400 feet, or 1,800 with specified lighting or flight-deck equipment — through Category II at 100 feet and RVR 1,200, down to Category III, which bottoms out at IIIc with no decision height and no RVR limitation (FAA AIM, Ch. 1 §1-1-9).
Two features make this ladder valuable to a delay trader. First, it is discrete: each rung is defined by published numbers, not adjectives, and not every aircraft, crew, and runway is certified for the lower rungs. Second, it is observable in near real time: visibility and RVR are measured continuously and published in surface observations, so the current rung is public information. In a fog event the open question is rarely “how bad is it right now” — the sensors answer that — but “how fast does it lift,” and radiation fog physics keep that answer inside a narrow morning window far more often than convection ever allows.
Where and When Fog Flight Delays Concentrate
Everything about fog is local. Radiation fog needs calm, clear nights over land; advection fog needs a marine moisture source and a cold surface to condense on. Neither travels across a continent the way a frontal system does. The diurnal signature follows directly from the physics: formation overnight and clearing before noon means fog flight delays concentrate in the early departure bank and the first arrival waves, with afternoons typically back to normal operations.
For contract selection, fog risk therefore lives at specific hubs during specific hours. Ranking which airports are structurally delay-prone overall is a different exercise — that is the subject of the airport delay hotspots guide — and how weather regimes rotate through the calendar belongs to the seasonal flight delay strategy. This profile is deliberately about neither ranking nor season: the point is that when a fog setup exists, it is confined to a known place and a known window, and that confinement is what makes it tradable on a short clock.
Thunderstorms: The Long-Horizon, Low-Confidence Profile
A thunderstorm cell lives fast. The FAA describes a three-stage life cycle — towering cumulus, mature, dissipating — with a total duration of “typically about 30 minutes” (AC 00-6B, ch. 19). Single-cell storms are rare; almost all thunderstorms are multicell clusters or lines in which new cells continuously replace decaying ones. And because convective currents are most active on warm afternoons, airport thunderstorm delay patterns cluster in the afternoon and evening, the mirror image of fog’s morning peak.
Forecast products are shaped by that churn. The strategic planning product used by U.S. traffic-flow managers — the Collaborative Convective Forecast Planning guidance, or CCFP — draws forecast polygons at 2, 4, 6, and 8 hours after issuance, and a polygon qualifies with areal coverage of at least 25 percent and forecaster confidence of at least 25 percent over an area of at least 5,000 square miles (FAA AC 00-45H, Aviation Weather Services, §5.6.3). Read that specification again: the official strategic convection forecast deals in five-thousand-square-mile shapes and one-in-four confidence. That is not a defect of the product; it is an honest statement of what convection permits.
The convective profile therefore inverts fog. Awareness arrives early — model guidance can flag a convectively active day well ahead — but precision never fully arrives at all. Which hour, which terminal, which arrival fix: those remain probabilistic until radar answers them in real time.
Convective Weather Delays Hit the Route, Not Just the Airport
A fog bank sits on the airfield. A squall line crosses departure corridors, en-route sectors, and arrival fixes that dozens of airports share. The advisory system is built around that reach: Convective SIGMETs, the advisories issued for thunderstorms over the contiguous U.S., are published hourly at 55 minutes past the hour, valid for two hours or until superseded, and trigger on features such as a line of thunderstorms at least 60 miles long with storms affecting at least 40 percent of its length (AC 00-45H, §5.1.4). These are en-route hazard products, not airport bulletins.
The operational consequences propagate the same way. The FAA describes the CCFP as a strategic tool for air traffic flow management that “aids in the reduction of air traffic delays, reroutes, and cancellations due to significant convection” — delays, reroutes, and cancellations in a single breath. In Europe, the equivalent function is centralized in EUROCONTROL’s Network Manager, whose flow measures surface as ATFM delay — a signal traders can read directly through the ATFM delay guide. And because rerouted and held aircraft arrive late somewhere else, convective weather delays export themselves downline; the cascade mechanics are a study of their own.
For predictability, the key implication is that a convective event has an open-ended exposure set. A storm that never touches your departure airport can still delay your flight by squeezing the routes into it.
What TAF and METAR Resolve for Fog — and Not for Convection
How to read TAF and METAR from a standing start is covered in the entry-window guide; the narrower question here is what the routine forecast cycle actually pins down for each phenomenon. Scheduled TAFs are issued four times a day, every six hours, with 24- or 30-hour validity (AC 00-45H, §5.11.3). For fog, that cadence is close to sufficient: the setup is diagnosable the evening before, and the event — formation through dissipation — typically completes within one or two TAF cycles.
The observation side shows the same asymmetry. Routine METAR reports arrive on an hourly schedule, and a SPECI — in the FAA’s wording, “an unscheduled report” — is taken whenever conditions change past defined criteria. For a fog watcher, that stream is nearly ideal: successive observations either step the airfield down the approach-category ladder or confirm the clearing trend, and the hourly rhythm matches the pace at which fog actually evolves.
Convection outruns the cycle by an order of magnitude. A cell’s roughly 30-minute life fits a dozen times into a single TAF interval, and can begin and end entirely between two routine observations. A TAF can and does carry thunderstorm groups, but it cannot commit to timing at delay-relevant resolution — which is precisely why the advisory layer refreshes hourly and the strategic layer reissues area polygons rather than point forecasts. The structure of the instrument stack is itself the tell: fog is a forecast problem that forecasts can win; convection is a nowcast problem that scheduled forecasts can only frame.
The Two Predictability Profiles Side by Side
| Dimension | Fog | Thunderstorms |
|---|---|---|
| Reliable lead time | Hours — evening before, for radiation fog | Days for “a convective day,” never for the specific hour |
| Forecast confidence | High once the setup is visible | Probabilistic at every horizon (CCFP threshold: 25%) |
| Diurnal cluster | Overnight formation, clearing before noon | Afternoon and evening |
| Spatial footprint | Airfield-local, narrow hub geography | Airport plus shared routes and sectors |
| Live observability | Discrete: visibility and RVR against CAT I/II/III steps | Continuous churn: radar plus hourly advisories |
| Event resolution | Bounded — burns off on a physical schedule | Open — new cells form while ingredients persist |
The observability row deserves emphasis. Fog degradation is measured against fixed numeric minimums, so the market’s central input is a published measurement that updates in discrete, legible steps. Convection’s central input is a moving radar picture that reprices continuously and can invalidate an hour-old read. The first rewards preparation; the second punishes complacency at any hour of a convective afternoon.
What Each Profile Means for a Delay-Contract Trader
The profiles do not argue for trading one phenomenon and avoiding the other; they say the available information has a different shape in each. In fog, information is concentrated: a narrow set of hubs, a morning window, and a central question — does it lift on schedule — that resolves against an observable physical process on a known clock. In convection, dispersion is the point: outcomes across On time, Delayed, and Cancelled stay genuinely open deep into the day, exposure extends to flights whose airports never see a storm, and any position is a statement about a distribution rather than about an event you can watch form.
The horizon difference matters as much as the confidence difference. A fog thesis is testable within hours — the atmosphere grades it by noon. A convective thesis can look right at breakfast and still be unresolved at dinner, because the phenomenon that settles it does not exist yet. Neither profile is better for trading; they simply demand different levels of comfort with waiting, and different respect for how stale a forecast can get.
Winter storms form a third, logistics-driven profile with its own dynamics, covered separately in the winter storm strategy guide.
Gaduin lists event contracts on flight, train, and vessel delays with settlement in USDT, and each flight market resolves to a published outcome — On time, Delayed, or Cancelled — under the process described in how outcomes are verified. Whichever weather profile a market sits in, the settlement question is the same; what differs is how much of the answer the atmosphere is willing to reveal in advance.
This content is provided for informational and educational purposes only and does not constitute financial, investment, or trading advice. Event contracts on Gaduin involve risk of loss and are not suitable for all participants. Past performance does not indicate future results. US persons are not eligible to participate. Review the full Terms of Service and User Agreement before opening any position.