Flying Weather Index Methodology

The default study period is January 1, 2006 through December 31, 2025 — 20 full years. The period is configurable; the version of every ranking is recorded so results stay reproducible when it changes.

We use archived hourly and special surface observations (METAR/SPECI-format) from the Iowa Environmental Mesonet ASOS/AWOS archive, which redistributes the underlying NOAA/NWS observation record. See the data sources page for attribution. We use observed conditions only — never forecasts.

Each observation is classified by the worse of its ceiling and visibility:

• VFR — ceiling above 3,000 ft and visibility greater than 5 statute miles.
• MVFR — ceiling 1,000–3,000 ft and/or visibility 3–5 statute miles.
• IFR — ceiling 500–999 ft and/or visibility 1–2.99 statute miles.
• LIFR — ceiling below 500 ft and/or visibility under 1 statute mile.

A sky with no ceiling layer reported (clear, few, or scattered only) is treated as unlimited ceiling, so its category is set by visibility alone. A vertical-visibility (VV) group counts as a ceiling. Published category percentages use one representative observation per clock hour, so that bursts of special reports during deteriorating weather don't over-weight the worst conditions.

An estimated VFR flying day is a local-calendar day with at least six consecutive daylight hours of observed VFR conditions. Daylight is computed from each airport's latitude, longitude, and date; local time uses the airport's IANA time zone, with daylight-saving handled automatically. We also compute alternative definitions (at least one VFR hour; a majority of daylight hours VFR; all daylight hours VFR; four- and eight-consecutive-hour variants); the public default is the six-hour definition.

Wind statistics use sustained surface wind. Calm and variable-direction observations are excluded from directional and crosswind measures rather than counted as zero. For crosswinds, each windy observation is tested against every open runway end at the airport using true headings; we keep the smallest crosswind component — the most favorable runway. Historical runway-closure data is not available, so runway availability is assumed constant across the study period. Crosswind metrics are only computed where runway data exists.

We do not publish the single largest gust value on record. A reported gust is treated as credible only when it has a plausible gust factor relative to the sustained wind and is corroborated by a neighboring observation within a few hours. The published figure is labeled the “highest credible observed gust,” and suspicious outliers are flagged and excluded — this is why some raw record gusts do not appear here.

Fog, thunderstorm, snow, and freezing-precipitation frequencies come from the present-weather codes in each observation, counted as estimated days per year with at least one qualifying report. Density altitude is computed per observation from field elevation, temperature, and altimeter setting using the standard E6B formula; the summer-afternoon figures reflect June–August local-afternoon observations. Historical density-altitude figures are educational and are not a substitute for aircraft performance calculations.

The four ATIS Alerts scores (VFR, Wind, Weather Reliability, Flight Training Weather) are each 0–100. Every input metric is normalized across the eligible airport set — winsorized at the 2nd and 98th percentiles so a single extreme airport can't compress the scale — then combined with fixed published weights. Scores are therefore relative to the field of ranked U.S. airports. A high score means favorable historical weather; it does not mean an airport is safe. Score formulas are versioned so they can be improved without silently changing past results.

Ranking ties are broken by airport identifier for a stable order. The dataset is recomputed when a new year of observations completes; the methodology and score versions are shown on every page and tracked on the changelog. Airport eligibility rules are on the eligibility page.