How to Read a TAF: Times and Change Groups

This guide covers US NWS TAF conventions for private-pilot study. The supplied four-scene hero illustrates change through time; neither it nor the worked forecast below is current operational weather.

What are the key TAF facts?

Read the times first, because familiar weather codes can belong to different parts of the forecast.

Wind directions are true, speeds are knots and cloud heights are hundreds of feet AGL. Those familiar METAR conventions carry into a TAF; the time logic is the new part.

How is a TAF different from a METAR?

A METAR reports an observation; a TAF forecasts a terminal's weather over its stated valid period.

Use the newest applicable observation to understand what has happened, and the newest applicable forecast to examine what may happen. If they disagree, investigate the trend and updated products rather than treating either report as a guarantee. A destination TAF does not describe terrain obscuration or all en route hazards.

How do you read the issue and valid times?

Separate the time the forecast was issued from the interval it forecasts.

Here is an original US-format practice TAF for hypothetical station KAAA. It is not an actual forecast:

TAF KAAA 181130Z 1812/1912 18008KT P6SM SCT030

FM181800 22012G20KT P6SM BKN040

TEMPO 1820/1823 3SM SHRA BKN020

FM190000 28008KT P6SM SCT050

PROB30 1903/1906 2SM TSRA BKN015CB

181130Z means issued on day 18 at 1130 UTC. 1812/1912 means valid from 1200 UTC on day 18 to 1200 UTC on day 19. These codes do not include month and year; the report context supplies them. Do not read 1912 as 19:12 on the issue date.

For an October Chicago study conversion, 1200 UTC is 7:00 AM CDT. Keep the UTC timeline while decoding, then convert the relevant date and interval if useful. Local dates can differ from UTC dates near midnight.

Original hypothetical timeline. FM boundaries change prevailing conditions; temporary and probability groups are separate overlays.

What do the initial conditions and FM groups mean?

The initial line applies from the beginning of validity; each FM line replaces the previous prevailing forecast at its specified time.

From 1200 to 1800 UTC on day 18, the example forecasts wind from 180° true at 8 knots, visibility greater than 6 SM and scattered clouds at 3,000 feet AGL. From FM181800, use wind from 220° at 12 gusting 20, visibility greater than 6 SM and a 4,000-foot broken ceiling.

At FM190000, the new prevailing forecast starts at midnight UTC on day 19: wind from 280° at 8 knots, visibility greater than 6 SM and scattered clouds at 5,000 feet. The earlier BKN040 does not silently carry forward into this complete FM group. FM is followed by day, hour and minute; FM190000 is not 1900 UTC.

For this example, the working table is:

Use the other groups as overlays on that table. This avoids reading every line as another complete forecast that runs independently to the end.

How does TEMPO change the forecast?

TEMPO forecasts temporary fluctuations within its stated interval, with unmentioned elements continuing from the prevailing forecast.

TEMPO 1820/1823 3SM SHRA BKN020 applies from 2000 to 2300 UTC on day 18. During temporary fluctuations, expect 3 SM visibility, rain showers and a 2,000-foot broken ceiling. The wind remains the prevailing 22012G20KT because it is not changed in that group.

The NWS instruction defines these events as lasting one hour or less each and, together, less than half of the specified period. It does not say the lower conditions last for the whole three-hour window, nor does it provide their exact start minute. They can still coincide with an arrival you intended to make within that window.

Interpretation of change groups, not a forecast of when an actual event will occur. Use the example's complete time windows.

An improvement in the prevailing line is not a reason to discard a worse TEMPO group affecting your planned time. Compare both with the intended operation and realistic alternatives.

What does PROB30 mean, and why not ignore it?

PROB30 gives a 30 percent probability of the specified weather event during the stated interval; it does not mean the weather occurs for 30 percent of the time.

In PROB30 1903/1906 2SM TSRA BKN015CB, the example forecasts a 30 percent chance of thunderstorms with rain, 2 SM visibility and broken cumulonimbus at 1,500 feet AGL between 0300 and 0600 UTC on day 19. A lower probability is still relevant when the event's consequences are serious.

NWS domestic TAFs use PROB30, rather than PROB40. The October 2024 NWS instruction removed the former restriction on PROB30 in the first nine forecast hours; do not reject an early probability group using that older rule. Do not generalize US NWS conventions to every international or military TAF: BECMG and other conventions can appear elsewhere (NWS instruction, FAA AIM).

How do ceilings, wind shear and amendments fit in?

Decode the weather elements for the applicable interval, and check whether the forecast has been replaced or amended.

BKN020 forecasts a ceiling at 2,000 feet AGL; SCT020 alone does not establish a ceiling. VV008 forecasts vertical visibility of 800 feet into total obscuration. Read the METAR hub for the shared cloud and weather vocabulary, and keep cloud heights distinct from MSL cruise altitudes.

A separate practice code, WS020/27040KT, indicates nonconvective low-level wind shear with wind at 2,000 feet AGL from 270° true at 40 knots. It does not replace the surface-wind group or mean a surface gust of 40 knots. Convective hazards also require attention even when no separate WS code appears.

TAF AMD marks an amended forecast. Use the amendment when issued; do not wait for its validity-start label to decide the old forecast has been superseded. COR indicates a correction. Check the latest applicable product before departure and during planning changes (AWC).

Can a favorable TAF establish that a VFR flight is safe?

No: it is one forecast input, and its terminal coverage does not replace a route-wide briefing.

For an original example arrival at 2130 UTC on day 18, the prevailing row is P6SM BKN040, but TEMPO 3SM SHRA BKN020 also covers that arrival window. A student should identify both, examine the broader weather and discuss margins and alternatives with the instructor. This is a decoding exercise, not a go/no-go decision for a real flight.

Compare the VFR minimums guide and Class E/G guide for separate legal conditions. 91.103 requires familiarity with available preflight information; a favorable forecast does not cancel aircraft, pilot, instructor or weather restrictions.

Check yourself

These are original study scenarios, not released FAA questions.

Frequently asked questions

Is P6SM exactly six miles?

No. It means greater than six statute miles. Keep the P attached to the value.

Does a TAF give the exact minute a temporary shower starts?

No. A TEMPO group supplies an interval and temporary conditions, not their exact onset minute at your arrival.

Does a new FM group keep the old ceiling automatically?

No. An FM group supplies a new complete prevailing forecast. Read its cloud groups rather than carrying the old layer forward.

Can I use a nearby airport's TAF as the destination forecast?

It can be relevant context, but a TAF's terminal coverage does not expand automatically to a different airport or the whole route. Use the relevant additional forecast products and briefing information.

What should you study next?

Return to the METAR weather-reports hub to compare a forecast with an observation. Review VFR weather minimums and Class E/G floors and conditions before applying decoded values to an airspace scenario.

Practise aviation weather with FirstSolo.ai. It is solely a study tool and does not issue endorsements now or in the future.

Sources and review notes

• FAA AIM 7-1 and AWC product help: TAF structure, terminal scope, shared weather codes and amendments.
• NWS Instruction 10-813, October 30, 2024, Appendix B: prevailing forecasts, FM, TEMPO, PROB30 and US conventions.
• NWS METAR/TAF key: decoding reference; 14 CFR 91.103: preflight action.

Facts checked October 18, 2026. The station, forecast and timeline are hypothetical. Human technical review remains pending before publication.

FirstSolo.ai is independent and is not affiliated with or endorsed by the FAA.