• compacted snow (solid contaminant, its depth is irrelevant),
  
• dry or wet snow, depth at or more than 3 mm - 1/8 inch (*)
• water, slush, depth at or more than 3 mm - 1/8 inch (*)
• ice (solid contaminant, its depth is irrelevant).

In some situations though, the contaminant reported to be present on the runway may not make it possible to identify the corresponding performance level just by considering the contaminant type and depth. It is the case particularly when the contaminant is:

• too variable as to its impact on aircraft performance: e.g. volcanic ash, hydraulic fluid spillage. Operations cannot, in general, be supported with specific performance information;
  
• a common natural one, but outside of the temperature conditions where its characteristics are well known: e.g. compacted snow if the outside air temperature subsequently raises above -15°C. Indeed, compacted snow is a specially prepared winter runway when temperature is very low, at or below -15°C. Above that, there is a risk that some of the contaminant be no longer true compacted snow. A downgrade of performance should then be considered as risk mitigation to support safe operations.
• a piling up of layers of different contaminants: the few cases documented water on top of compacted snow, water on top of ice (or wet ice), or dry/wet snow over ice, have shown unacceptable impact on aircraft performance and operations cannot be supported, even adoption of the most conservative contaminant, i.e. ice, for snow over ice condition might be unsafe.

Eventually, the most common contaminants for which aircraft performance level can be defined have been synthetized into the Runway Condition Assessment Matrix (fig.1) that permits deterministic classification of the expected Landing Performance.

(fig.1) Example of a Runway Condition Assessment Matrix

Weather conditions evolve quickly and elude a forecast accurate enough to be compatible with the sensitivity of landing performance. As an example, Landing performance is defined as GOOD when the runway is normally wet (runways quickly drain water during showers with normal precipitation rates). It might drop to MEDIUM TO POOR with standing water accumulation (the 3 mm water depth criterion is a necessary simplification to represent this phenomenon).

Likewise, the estimated runway condition and resulting landing performance may be sensitive to temperature. It is the case especially when the temperature leads to a change of state of the contaminant: landing performance is poor on dry ice, but can become non-existent if ice surface is melting (here again, the -3°C temperature criterion is a necessary simplification).

Determining precisely when the precipitation accumulation will become critical or when the ice will start melting in significant proportion is already a challenge when nothing interferes with it. Yet in reality, a number of other factors do interfere with this weather dimension and make it even more difficult to determine the actual runway condition, not to mention an anticipation of it.

Although runways vary in size, 3 km long and 45 m wide give a representative indication of the surface area of a runway. On such a surface, the exact contamination may vary from one place to another. As an illustration, “patchy snow and ice” may be reported in some airports as representing less than 25% of runway coverage.

Whatever the actual state of the runway and its variability, it needs to be simplified to make a landing performance computation. Indeed, landing performance models can only consider a single contaminant evenly distributed on the runway.

Beyond these intrinsic difficulties of having an accurate representation of the runway condition, operations taking place on the runway modify the runway condition at least in some places of the runway. An aircraft landing on a runway may change the depth of a contaminant if not its nature. Indeed, it can for example induce a change of state at the touchdown point or along its deceleration path. The contamination will remain unchanged though on the un-trafficked last part of the runway or further away laterally from the landing gear.

An aircraft taking off might also induce changes in the runway contamination along its take-off roll, thereby increasing as well the heterogeneity of the contamination throughout the runway surface.

A more obvious case of impact of airport operations on runway contamination is any runway management action such as cleaning or de-icing. In many cases, de-icing fluids are applied only to a limited width along the runway axis.

Although this sounds obvious, it means that what pilots need to know is not the very physical details of the runway conditions but rather how the performance of the aircraft might be affected, thus what they will need to do to still perform a safe landing. In other words, what pilots really need is a translation of the runway condition into its practical effects on the aircraft.

Yet today, the information provided to pilots on runway condition is not directly a level of performance. One of the main challenges for pilots is to translate from their vantage point in the cockpit of an approaching aircraft the sometimes complex information provided to them on runway surface condition into a single classification of the runway condition landing performance level.

This translation is done by means of the Runway Condition Assessment Matrix (RCAM) introduced earlier. The RCAM includes, beyond DRY, WET and thin contaminants that are equivalent to WET, 4 discrete levels of contamination, each of which is associated with a landing performance level.

The information provided to pilots of runway condition may vary from one country to another and from one airport to another. Let’s review the three categories of possible information pilots may get on runway condition before discussing how they can be integrated to come up with a single, representative, performance level.

In accordance with ICAO standards, all airports around the world should provide this information to pilots prior to landing. It is the primary information about runway contamination (this reporting is even more essential for take-off).

Currently, the description of contaminants in SNOWTAMs is done through a combination of codes and free text/ plain-language remarks. There is no clear distinction between performance relevant contaminants and other runway surface conditions provided for situational awareness. The ICAO SNOWTAM codes correspond to a set of generic contaminants, thus are different from the RCAM landing performance codes agreed by the Takeoff and Landing Performance Assessment Aviation Rulemaking Committee (TALPA ARC, see article Safety First 10).

Providing the contaminant type & depth to pilots relies on measurements, especially that of contaminant depth. Performing these measures in a way that provides a representative view of the real depth is a challenge to airports. More generally, measuring runway contamination, whether it is to determine contaminant depth or to estimate the surface friction coefficient (see next section), can become challenging for a variety of reasons (see insert The challenge of providing measures on runway contamination).

ICAO and national authorities have progressively shied away from reporting measured friction to pilots. In fact, there is no established meaningful correlation on most contaminants between estimated surface friction established by ground measurement devices and aircraft performance. Therefore, reporting ESF is strongly discouraged by ICAO on contaminants for which it is now known that it may be dangerously biased (fluid winter contaminants as snow or slush, i.e. dry or wet snow or slush). (see insert The challenge of providing measures on runway contamination). Yet, it is a secondary information pilots may get in some areas of the world.

ESF can be reported in different formats. Either under the terminology: GOOD / GOOD TO MEDIUM / MEDIUM / MEDIUM TO POOR / POOR by third of runway length, or through a figure e.g. 26μ. When the surface friction is expressed through a figure, it may give the illusion that it is an accurate measurement although it still remains of limited practical use in characterizing winter runway conditions for aircraft operations. Indeed, no related landing performance level can reasonably be derived from the sole figure.

The last secondary information pilots may get on runway conditions, although its use largely varies regionally, is through the air traffic controller in the form of a Pilot Report or PiRep of Braking Action. PiRep of BA are encouraged in some countries. These reports are individual perceptions that may be influenced by a number of factors: whether the pilot is familiar with contaminated runways and this particular type of conditions or with the type of aircraft or the use of deceleration devices. It is also easy for a pilot to mistake aerodynamic and reverse thrust deceleration forces for braking forces. However, the usefulness of such subjective reports should not be underestimated, as they often (but not always) provide the most recent information available under dynamic weather, and resulting runway surface conditions. PiReps should always be communicated to the approaching pilots with a time and emitter of the report including the airline and the aircraft type.

PiReps of Braking Actions are also reported using the terminology: GOOD / GOOD TO MEDIUM / MEDIUM / MEDIUM TO POOR / POOR, and can also be reported by third of runway length.

In countries where PiReps of Braking Action are transmitted to following traffic, it is the sole responsibility of the pilot performing the In-Flight landing performance assessment to determine whether the transmitted information can be considered reliable or not.

Eventually, pilots need to integrate all the pieces of information they receive in relation to runway contamination to come up with a single level of landing performance. They can receive up to three different information types, coming from different sources:

• Runway contaminant type and depth: mandatory as primary information;
• Estimated Surface Friction (ESF): not systematic as secondary information;
• Pilot Report of Braking Action (PiRep of BA): not systematic as secondary information.

Some rules do exist for pilots to integrate these various types of information.

As a general rule, the Related Landing Performance level derived from the primary information (contaminant type & depth) prevails if considering other sources of information would lead to being less conservative than EASA regulation.

When ESF is lower than the performance associated to contaminant type and depth in the RCAM, it should be used to determine the Related Landing Performance Level for in-flight landing performance assessment (downgrade). When ESF is higher than the performance associated to contaminated type and depth in the RCAM, its use to determine the Related Landing Performance Level is not supported (no upgrade).

When PiRep of BA is lower than the performance associated to contaminant type and depth in the RCAM, it should be used to determine the Related Landing Performance Level for in-flight landing performance assessment (downgrade). When PiRep of BA is higher than the performance associated to contaminated type and depth in the RCAM, its use to determine the Related Landing Performance Level is not supported (no upgrade) by EASA, but under pilot responsibility in USA.

Even if under EASA regulation, landing performance is calculated based on the probable contamination before dispatch, it is necessary to re-evaluate the landing performance prior to landing. Dispatch considerations will most probably no longer apply to the actual conditions at the time of landing. In addition, should the conditions be exactly the ones anticipated, the most recent in-flight landing performance models can lead to longer distances. Indeed, the in-flight landing performance models used today rely on more realistic assumptions thus allow for deriving more realistic, though often more conservative, landing distances.

The model used for all Airbus aircraft for In-Flight Landing Distance assessment is based on the comprehensive work of the TALPA-ARC group. This work relies itself on the contaminants characteristics described in EASA CS25.1591 (see SAFETY FIRST n° 10 August 2010 P8-11). Airbus concurs with the FAA in recommending a minimum margin of 15% on these distances, achievable in line operations when no unexpected variations occur from reported outside conditions and assumed pilot technique.

The improvements brought by the RCAM are so widely recognized that they allowed EASA, in combination with a minimum margin of 15%, to accept a new still safe but more realistic (better) performance level for POOR. This level is consistent with ICE (COLD & DRY) rather than with WET ICE (as previously), for which the RCAM prohibits operations. These new computation options have started to appear at the end of 2014 on the Airbus fleet and will continue progressively.

While performing the in-flight check on landing performance, anticipating all the realistic degradation or aggravating factors and determining the thresholds below which a safe landing can still be performed is a way to cope with the uncertainty of the information available in approach, hence remove a potential element of surprise should one or more parameters evolve by the time you actually land.

For example, if it is snowing and the latest airport report states less than 3 mm (1/8 inch) of snow, asking yourself: “is it going to exceed the critical depth of 3 mm (1/8 inch)? If it does, am I still safe?” is a way to proactively get prepared to a safe landing. Likewise if it is raining, “what is the maximum cross-wind under which I can still perform a safe landing” is the kind of question that contributes to a good preparation to a safe landing.

As mentioned earlier and illustrated in SAFETY FIRST n°10 fig.5, a 15% margin is to be integrated in the calculations of In-Flight Landing performance, on DRY, WET and on contaminated runways (Factored In-Flight Landing performance), except in case of failure. This margin is meant to cover some uncertainty related to a variety of aspects:

• Pilot achievement of the assumed touch-down location and touch-down ground speed
• Pilot timely activation of deceleration devices assumed (brakes if no Auto-Brake, reversers)
• Lower performance than expected (even if friction models of CS25.1591 are generally conservative)

If the 15% margin is fully “eaten” by the sole effect of runway conditions worse than expected, there is no margin left for any other deviation as a slightly long flare or slight pilot lag in applying deceleration means.