Safety

Wet and contaminated runways: braking action, RCAM and the GRF

October 2, 202613 min readFor pilots

Leer en español · Lire en français · Ler em português

In this article

Midway, 2005: the accident that changed the rules
From TALPA to the RCAM: how US reporting changed
The RCAM and runway condition codes
ICAO's Global Reporting Format
Landing distance at time of arrival and the 15% margin
Wet is not always "good": the wet-runway problem
Recent excursions: what investigators found
What it means for pilots

On the evening of December 8, 2005, a Southwest Airlines 737-700 landed on runway 31C at Chicago Midway in heavy snow, with a tailwind, and did not stop. It went through the airport perimeter, crossed Central Avenue and struck a car. A child in that car was killed. The NTSB's final report became the starting point for almost everything a US airline pilot now uses to judge a slippery runway: runway condition codes, standardized braking action terms and the landing distance check you run before top of descent.

The work is not finished. In May 2026 the NTSB adopted a safety study on wet-runway overruns that found the standard "wet" assumption can be optimistic in heavy rain. Overruns on landing still happen about ten times a year in the US, according to FAA and NTSB information cited in AC 91-79B.

Midway, 2005: the accident that changed the rules

The NTSB report (NTSB/AAR-07/06, adopted October 2, 2007) lays out a familiar chain:

Runway: 31C was 6,522 ft long, with 5,826 ft usable for landing. About 10 inches of snow had fallen, with visibility 1/2 SM in moderate snow and freezing fog.
Braking action: Controllers reported braking action "good for the first half, poor for the second half." The onboard performance computer (OPC) could not accept a mixed report, so the first officer ran it twice. With "fair," it showed 560 ft of stopping margin. With "poor," it showed 40 ft.
Tailwind: The FDR showed an 8–9 kt tailwind at touchdown. Southwest policy did not allow landing with poor braking action and more than 5 kt of tailwind.
Hidden assumption: The OPC gave credit for reverse thrust on the 737-700. Until a week before the accident, two of three places in the flight operations manual said it did not. Some pilots interviewed afterward did not know about the credit.
Deceleration: The airplane touched down about 1,250 ft past the threshold. It was the pilots' first landing with the airplane's newly introduced autobrakes, and the captain had trouble deploying the reversers. The first officer deployed them about 15–18 seconds after touchdown. Five airliners that landed in the 25 minutes before had full reverse within 4–6 seconds.

The airplane left the runway at about 53 kt. Its braking ability was about five times worse than on a bare, dry runway. The probable cause was the pilots' failure "to use available reverse thrust in a timely manner." Contributing factors included no safety margin in the arrival assessment, an OPC that hid its assumptions, and not diverting with poor braking and a tailwind over the limit. The Board found that timely reverse thrust could have stopped the airplane on the runway.

The investigation also exposed a reporting problem. Friction readings on 31C that evening were 0.67 Mu about 30 minutes before the landing and 0.40 just after it, yet pilots heard only fair or poor. No standard tied a braking report to what an airplane could actually do.

From TALPA to the RCAM: how US reporting changed

The FAA moved fast. SAFO 06012 (August 31, 2006) urged turbojet operators to assess landing performance using conditions at the time of arrival and to add at least 15% to the computed distance. The NTSB's final report also called for a 15% margin on arrival assessments, along with better standards for braking action reports.

In October 2007 the FAA chartered the Takeoff/Landing Performance Assessment Aviation Rulemaking Committee (TALPA ARC), about 40 members from manufacturers, operators, pilot unions and the FAA. Its scope covered landing performance assessment, contaminated-runway takeoff performance and airport certification issues. In 2009 the ARC recommended a field trial of a new tool: the Runway Condition Assessment Matrix (RCAM).

The RCAM went live across the National Airspace System on October 1, 2016. The main changes:

Airport operators assess each third of the runway (touchdown, midpoint, rollout) and publish runway condition codes (RwyCC) in a Field Condition (FICON) NOTAM, such as 3/3/2.
Mu friction reporting was removed and replaced with RwyCC.
Braking action terms became good, good to medium, medium, medium to poor, poor and nil. The old scale, with "fair," was retired.
A code of 0 marks nil braking and is never published in a FICON NOTAM. At federally obligated and Part 139 airports, a nil condition must be mitigated by closing the surface.

The RCAM and runway condition codes

The table below summarizes the FAA RCAM from AC 150/5200-30D, with code 1 and 0 rows from the NBAA TALPA quick reference card. Always use your operator's current version.

RwyCC	Typical runway condition	Pilot braking action
6	Dry	—
5	Frost; wet (including damp, water 1/8 in or less); 1/8 in (3 mm) or less of slush, dry snow or wet snow	Good
4	Compacted snow at OAT −15 °C (5 °F) or colder	Good to medium
3	Slippery when wet; dry or wet snow over compacted snow; more than 1/8 in of dry or wet snow; compacted snow warmer than −15 °C	Medium
2	More than 1/8 in (3 mm) of water or slush	Medium to poor
1	Ice	Poor
0	Wet ice; slush over ice; water over compacted snow; dry or wet snow over ice	Nil

Note how thin the line is: only 1/8 inch of water separates code 5 (good) from code 2 (medium to poor). AC 91-79B warns that standing water can cut braking force by up to 77% and make directional control unpredictable.

The AIM asks pilots to use the six standard terms, name the part of the runway and give the aircraft type. When reports of medium, poor or nil come in, ATIS broadcasts that braking action advisories are in effect. AC 91-79B adds that a report is treated as valid for the whole runway unless you name a section, and that turbojet crews should judge wheel braking only, not the help from reversers.

ICAO's Global Reporting Format

ICAO built its own version on the same idea. The Global Reporting Format (GRF) became applicable worldwide on November 4, 2021. It was first set for November 5, 2020 and was postponed a year because of COVID-19. The changes touched Annexes 3, 6, 8, 14 and 15, plus PANS-ATM, PANS-AIM and PANS-Aerodromes (Doc 9981). Guidance sits in Circular 355 and the Aeroplane Performance Manual (Doc 10064).

Under GRF, a trained observer assigns a runway condition code and contaminant description to each runway third. The result goes out as a Runway Condition Report (RCR) in a SNOWTAM, on ATIS and from ATC. Circular 355 is explicit that measured friction is no longer passed to flight crews.

What differs from the US version:

Lowest term. ICAO's braking action scale ends in LESS THAN POOR. The FAA uses NIL.
Depth threshold. GRF uses 3 mm for contaminants. The FAA uses 1/8 inch, about the same depth.
Europe started earlier. EASA states applied GRF from August 12, 2021. EASA's webinar Q&A notes that GRF-format SNOWTAMs from those states are valid for 8 hours, not 24.
Pilot reporting. EU rules require a report when braking on the landing roll is worse than reported. The guidance sits in AMC1 CAT.OP.MPA.311, added by ED Decision 2021/005/R, which also set out what data may be used for landing distance at time of arrival.

Landing distance at time of arrival and the 15% margin

The data side changed too. AC 25-32 (December 22, 2015) gives manufacturers a standard way to publish time-of-arrival landing data by runway condition code. That data is meant to reflect procedures that trained line crews of average skill actually fly. By default it assumes an air distance of 7 seconds from the threshold.

The FAA's current guidance is AC 91-79B, dated August 28, 2023. It canceled AC 91-79A and two 2019 SAFOs, including SAFO 19001, which had itself replaced SAFO 06012. The core advice:

Compute operational landing distance with the conditions at arrival, then add at least 15%. Outside an emergency, do not land on a runway that fails that test.
If you expect more than 1/8 inch of water in heavy rain, use the data for medium to poor (RwyCC 2).
If manual braking gives the 15% margin, crews may combine autobrakes and manual braking even when autobrake-only data falls short, given operator procedures and training.
With standing water, tire cornering limits reduce braking, and the crosswind limit can become the deciding factor.

The same AC puts numbers on the usual culprits:

Factor	Effect on landing distance (AC 91-79B)
Approach speed 10% fast	About 20% longer
Each 10 ft above 50 ft at the threshold	About 200 ft longer
First 10 kt of tailwind	About 21% longer
2-second delay in deceleration devices at 118 kt	Almost 400 ft longer
1% downhill slope	About 10% longer

Europe is moving the same way. EASA's CS-25 Amendment 27 added CS 25.1592, which requires landing distance data for assessment at time of arrival on dry, wet, slippery wet and contaminated runways. EASA frames this data as what to expect in operations, not certification test minimums. In July 2022 an FAA advisory working group recommended codifying the same thing in Part 25, with a 1.15 factor consistent with EASA. In AC 91-79B, the 15% margin is written as guidance, not as a rule.

Wet is not always "good": the wet-runway problem

The RCAM puts an ordinary wet runway at code 5, good. The data behind that code is the wet braking model in 14 CFR 25.109(c). Several bodies have now said that model can be too generous.

Certification. In March 2018 an FAA advisory working group delivered a final report on wet runway stopping performance. It found cases where wet landing distance came close to, or exceeded, the standards, and proposed a physics-based wet-runway rule in place of factoring dry distance by 1.15. Extra credit for grooved and porous friction course (PFC) runways would depend on airport maintenance.
Grooving does not make a runway dry. AC 91-79B corrects a common misconception: a grooved or PFC runway is not treated as dry for preflight planning. The 15% wet increment still applies. SAFO 19001 flagged non-grooved wet runways as needing closer assessment.
Heavy rain. The NTSB's AIR-26-04, adopted May 12, 2026, studied 11 wet-runway overruns from 2008 to 2022. In all but one, the braking friction achieved was substantially less than the RCAM wet value, which comes from the 25.109(c) method. Rainfall in six of the events was 1.3 to 20 times the 0.30 in/hr heavy rain threshold, and the crews had nothing telling them to use a code worse than 5. The NTSB noted that AC 91-79B itself acknowledges the 25.109(c) method can produce wet landing distances that are too short. One case is the 2019 737-800 overrun at NAS Jacksonville, where the probable cause cited viscous hydroplaning on an ungrooved runway.

Its three recommendations to the FAA (A-26-61 to -63) ask for lower runway condition codes as rainfall intensity rises, for new rain intensity descriptors above heavy in aviation weather reports, and for those descriptors to be built into the RCAM.

Recent excursions: what investigators found

Recent reports repeat the same themes: conditions worse than reported, long or fast touchdowns, and deceleration devices used late or not at all.

Kingston, Ontario, November 30, 2021 (report February 2023). An Embraer EMB-505 overran runway 19 and stopped in the grass about 440 ft beyond the end, according to the TSB. As summarized by the Kingstonist, the TSB found the last runway report, issued about an hour before landing, showed 5/5/5. With snow falling and the temperature dropping, the wet surface had frozen, and the crew had no current information.
Pittsburgh, March 9, 2022 (final March 2024). A HondaJet landed long and fast on wet concrete with patchy slush, after a "good" report from a previous airplane. The NTSB cited the crew's continuation of an unstable approach. A contributing factor was the captain's full application of the emergency brake, which caused hydroplaning.
Houston, March 8, 2024 (investigation docket public). United Airlines flight 2477, a 737 MAX 8, departed the paved surface near the end of runway 27, and the left main gear separated. The NTSB's operations factual report notes the field reported 3/3/3, slippery when wet, and that the operator's procedures called for autobrake 3 or higher, or aggressive manual braking, on that code. The captain selected autobrake 2, then changed to autobrake 1 after the tower approved rolling to the end and asked the crew to keep their speed up, and stowed the speed brake lever shortly after touchdown, which disarmed the autobrakes. The performance study notes the captain thought the runway looked dry; the braking data matched wet pavement. Reversers stayed at idle. In 24 of 29 alternate cases modeled, the airplane stopped on the runway. We found no probable cause published yet.
Miami, September 6, 2026 (under investigation). A 21 Air Boeing 767 freighter left runway 30. The NTSB's September 9 update says it touched down at 158 kt groundspeed with no speed brakes or reversers deployed. The update does not describe the runway surface.

Reverse thrust needs care too. In the Flight 1086 accident on a snow-covered LaGuardia runway in 2015, the captain reached about 2.0 EPR in reverse against a 1.3 EPR target. That degraded the MD-88's rudder effectiveness and cost the crew directional control.

What it means for pilots

Run the numbers at time of arrival, every time the runway isn't dry. - Use the latest FICON or RCR, not the one from dispatch. At Kingston, a 5/5/5 report was an hour old and wrong. - Use time-of-arrival (AC 25-32 type) data and add at least 15%. If braking reports are mixed, use the more restrictive one, as Southwest's policy already required at Midway. - Know what your tool assumes about reverse thrust and autobrake setting. - In heavy rain, don't trust code 5 by default. Use RwyCC 2 data if you expect more than 1/8 inch of water.

Choose autobrake or manual braking deliberately. - Pick the autobrake setting your data and procedures call for on that code. Don't step down to keep your speed up for a distant exit. - Know what cancels your autobrakes. At Houston, stowing the speed brakes did. - If you transition to manual braking, apply firm, steady, symmetric pressure and let the anti-skid work. At Pittsburgh, full use of the emergency brake disabled anti-skid and led to hydroplaning.

Get reverse thrust out early and keep it as needed. - Reversers work best at high speed. Midway's 15–18-second delay was the probable cause, and AC 91-79B puts a 2-second delay at almost 400 ft. - Use the reverse setting your procedures call for. Flight 1086 shows that too much reverse can cost rudder authority.

Respect crosswind limits on slippery runways. - Use the crosswind limit for the reported condition, not the dry-runway number. - Airbus found 22 of 25 lateral excursions it studied involved a wet or contaminated runway. Airbus advises considering reduced reverse thrust if directional control becomes a problem.

Fly the energy, not just the approach. - A long, fast touchdown wipes out a 15% margin quickly. - If the margin isn't there, divert.

Report braking action accurately. - Use the six standard terms, name the runway third and give your aircraft type. - Report wheel braking only. Leave reverse thrust out of it. - Say so when it is worse than reported. Under EU rules, you must. In the US, a nil report closes the runway at Part 139 airports. - Be precise. Before Flight 1086, two preceding airplanes had reported good braking; the crew found a snowier runway than expected.

Sources

This is general information, not operating guidance. Your company's procedures, your aircraft's manuals and the regulations come first.