A Runway Excursion Excursion
10/15/2024
Read time 10 minutes.
The NBAA published a guide in late 2023 titled "Reducing Runway Excursions in Business Aviation." The 25-page PDF contains interesting data points about takeoff and landing incidents and suggestions for operators to reduce risk.
An excursion is an overrun or veering off of the runway. Based on the data, this is the lowest-hanging fruit for accident prevention. The report points out that there is no silver bullet and that the issue and mitigation are multi-faceted.
Here is what stood out to me:
From 2017-2022: 41% of corporate aviation accidents were runway excursions.
Takeoff
Takeoff = 21% of runway excursions
Most actual RTOs are non-engine-related
76% of RTOs are less than 80 knots
18% of RTOs are 80-100 knots
4% of RTOs are 100-120 knots
2% of RTOs are 120+ knots
RTOs at 100+ knots = most overruns
2/3rds of takeoff excursions are overruns
Landing
Landing = 79% of excursions
Approach and Landing = 65% of accidents
83% of these accidents could have been prevented with a go-around (54% of all accidents!).
A go-around was executed on 3% of unstable approaches.
Unstable approaches are the #1 common factor of runway excursions.
Unstable approaches + landing long + fast + no go-around = most significant factors.
65% of long/fast touchdowns result from unstable approaches
Vref + 10 = 20% increase in landing distance
A long flare = 30% increase in landing distance
Dipping below the glide slope to land closer to the end of the runway does not result in shorter landing distances due to the shallower approach angle and longer resulting float.
Source: NTSB
NBAA Suggestions
As an aviation professional, looking at this document is worth your time. It is dense with factors and suggestions, from human psychology to our operational environment. The NBAA report was compiled and concentrated from other resources (see references), so instead of rehashing what it offers, I suggest digging in yourself.
What is most interesting to me is what I can do in the heat of the moment during takeoff and approach/ landing without having a runway excursion. After arming myself with this information, I came to some conclusions, points to ponder, and practical applications.
Operational Baseline
First of all, there are assumptions about the environment in which you are operating. This is for Part 91 business operators, and we have some advantages over our 135 and 121 brothers and sisters. I.E., no reduced thrust takeoffs, lower V1 speeds, less mass, typically well under gross takeoff weights, shorter wingspans, excess brake energy, excess power, great automation, operational flexibility, and less regulation. Adversely, we can operate into shorter and more complex airstrips with fewer services on less notice.
Additionally, the way I operate is assumed. See SMS article. Things like always checking your fluids and flight controls, being well rested and not hurried, being well trained and practiced, and doing your pre-flight performance calculations are assumed - a big one: always have room to abort and continue. In other words, operate as a professional. Lastly, aircraft specifics come into play. Below, I will reference a Gulfstream G280.
With that in place, buckle up because this food for thought is an alternative to the prevailing industry standard.
Takeoff
Terms and background:
Vmcsg: Minimum controllable speed on the ground (G280 95 knots assuming no crosswind)
Vef: Engine failure speed. This is for aircraft certification and is the speed at which the engine fails to establish V1 values. It cannot be less than Vmcsg.
V1: Initiate RTO no later than speed (my paraphrase). The decision to abort should have already been made, and the first action should have been taken by this point.
Acceleration stop distance for certification = accelerate, the engine fails at Vef, continue to accelerate to V1, hold V1 speed for 2 seconds, and decelerate to stop.
Gulfstream: Vef + 1 sec = V1
G280 SOP callouts: '80 knots", "V1" (hand from power levels to yoke), "Rotate"
G280 automation: RTO auto brakes & auto-inhibit most CAS messages (to avoid nuisance messages) above 80 knots.
G280 has independent brake systems and dual wheels for redundancy.
Notice that with Gulfstream, there is a one-second lag from Vef to V1, and the FAA has a 2-second lag at V1 to build in the margin for accelerate stop distances. This coincides well with a Qantas study of pilot reactions during engine cuts (2-4 seconds to recognize and initiate action). However, this conservative approach to account for reaction time does not factor in continued acceleration faster than V1.
Source: Flight Safety Foundation, Reducing the Risk of Runway Excursions
The above chart for takeoff has a whole bunch of causal factors. Generally, they can be divided into two categories: loss of control and RTO decisions. Runway excursions are the end result of a bad situation. If we want to get to the cause of RTO decisions, the situations should be divided into:
should have aborted but didn't
or
did abort but shouldn't have
or
correct decision, but still a bad outcome
Sometimes, the pilot cannot prevent an excursion, but in many accidents, the correct decision to abort or not can be improved.
The industry standard takeoff briefing is to abort everything below a certain speed (G280 is 80 knots), above 80 knots up to V1 abort for anything catastrophic (engine fire, engine failure, flight control malfunction, loss of control), and above V1 continue the takeoff.
Layer combination as of these possible factors:
nuisance, minor, and critical CAS messages
below 80 knots, between 80-100 knots, below and above V1
unannounced controllability issues, loss of control, and catastrophic failures
runway length, contamination, and crosswinds
crew alertness, response time, and fatigue
In the heat of the moment, we encounter the human brain's computational limits. There is insufficient time to remember, recognize, identify, assess, and act while rapidly accelerating down the runway. We need to simplify the decision-making process.
Muzzle Thump
It is a rare profession where life-or-death decisions are compressed into a short window of re-action and action. One tactic within military forces is when you encounter a person in close quarters who you do not know if they are a threat or not, you buy more time with a muzzle thump.
Instead of entering a room gun blazing, you prioritize your own and your team’s safety by incapacitating a potential threat by thumping a suspect in the chest with your weapon. It provides sufficient force for safety while not causing undue harm or loss of life if you’re wrong.
This aligns with the Patten adage, “A good plan, violently executed now, is better than a perfect plan next week.” This wisdom can be applied to aviation.
I propose a simplification of the takeoff brief and decision-making process:
The data shows that the most dangerous aborts are at 100+ knots. This theory removes the analysis of the situation at critical speeds to increase reaction time.
The G280 automatically excludes non-critical CAS messages above 80 knots. If you get a CAS above 80 knots, it's serious.
The G280 minimum controllable speed on the ground is 95 knots.
Most G280 V1 speeds are between 100 -125 knots.
Therefore
Any CAS below V1 = abort
V1 = continue
Loss of control at any time = abort
Using this method removes the troubleshooting portion of the decision-making process while you're barreling down the runway. The focus shifts to maintaining control of the aircraft, and the aircraft's automation is leveraged to assist with decision-making. The crew focuses on continually assessing if the aircraft will fly, and the go/no-go decision is a reaction based on CAS messages and airspeed.
You are muzzle-thumping the situation and avoiding catastrophic loss by executing sufficient force to keep you and your team safe. Since there is insufficient time to analyze the threat, you are making a timely, good-enough decision instead of the perfect decision too late.
This philosophy could cause unnecessary aborts at low speeds for nuisance CAS messages. Still, I would gladly trade a minor inconvenience to avoid reinforcing a human behavior that the data shows has catastrophic consequences at high speed (100+ knots).
Above V1, continue the takeoff if possible, meaning you can keep it going down the runway and accelerating. If you lose control or it cannot accelerate, it will not fly. This leaves no choice but to abort.
What if the brakes fail and you can't stop? With a G280, both brake systems, tires, and the parking brake would have to all go out simultaneously, which is unlikely. Also, when things break, you usually do not have enough information to make the perfect decision; you are a test pilot. Making a good enough timely decision is more important than making the best decision too late.
If it looks like your initial multi-engine training guidance, you're right. I think we've complicated it, which has led to many incorrect abort/go decisions or running out of time to make good decisions more than it's helped. Things should be simplified whenever possible.
Caveats
Is the above an oversimplification? Yes. First, if your aircraft does not automatically inhibit CAS messages, you must react and respond in the heat of the moment. Secondly, while the above does simplify the most demanding executive functions of aircraft issues and notifications, it does not address aborts called by ATC or instances when there are no CAS messages, such as smoke in the cabin or another aircraft on the runway. You still must aviate.
Lastly, the above is highly aircraft-specific. For example, a fully loaded G600 will have a much longer dwell time from 80 knots through 100 knots and on to V1. You have to decide for or against simplicity. Do you keep the simple V1 go/no-go decision, or do you further subdivide between 100 knots—V1 of aborting for control issues only? While I would argue for simplicity, this is going to have to be your own thought exercise.
Landing
Landing excursion data is straightforward: energy mismanagement. Aircraft are not arriving on speed and on point.
Over and over, "the pilot failed to fly a stabilized approach and failed to execute a go-around." The industry response is to "fly stabilized approaches!" Defined as on speed, configured, aligned, and steady descent by 1000 ft (500 ft VFR) or go-around. Be a perfect pilot every time.
Source: Flight Safety Foundation, Reducing the Risk of Runway Excursions
Stick and rudder skills
Instead of blaming the pilot's skill, my question is, how has the system contributed to this energy management issue?
A stabilized approach is an ideal scenario under ideal conditions. It works excellent for airliners vectored to 10-mile ILS approaches at radar-controlled airports. However, in business aviation, disturbances regularly occur to the ideal. For example, you're held high for lower crossing traffic heading into an uncontrolled field, tower requests a tighter base than the 5-mile final, traffic is in the pattern, or the terrain or airspace constraints around common destinations (PWK, SDL, TEX, ASE, etc.).
When the environment forces us into a less-than-ideal position, the pilot is now in a unique and UNPRACTICED situation. The pilot may not have the experience to know whether or not a recovery to the ideal is possible.
The alternative solution is not to train to avoid abnormal situations but to practice them intentionally. Use every deadhead opportunity when there is a performance margin to develop energy management skills. Start high, low, or fast; fly a curved approach to a 500 ft final, and practice approaches without backing them up. Practice how long and how much altitude it takes for your aircraft to decelerate on a 3-degree glideslope.
The idea is that instead of restricting what an acceptable approach is, you become more experienced, so you can expand the criteria for a satisfactory approach and become better equipped to identify that you are in a safe energy management window. Opposed to the avoidance strategy, this is an engagement strategy.
Go-Around
You always have the option to go-around (mechanical defects and bush flying aside). According to the data, being long and fast are the killers. These two factors should be the primary energy management data for making a go-around decision.
Currently, 3% of unstable approaches result in a go-around. This is due to the definition of a stable approach and pilots ignoring stabilized approach gates. What isn't captured in the data is that even though an approach may be flagged as unstable at 1000 ft, most of the time, there was enough margin for the pilot to correct before touchdown and make a safe landing.
The currently defined stabilized approach is practically ignored and, therefore, useless. The industry push to conform operations is not working. Assuming pilots take the opportunities available to expand their skills described above, I propose a new set of acceptable approach criteria:
At 200 ft above minimums on an instrument approach
or
200 ft above touchdown on a visual
You are on speed, on point, and aligned.
If not, go-around.
The goal is still to be stabilized early (500 or 1000 ft). I am not advocating aiming to be stabilized at 200 feet. The difference is defining a red line that pilots agree not to cross. It's a hard cap: if the approach is unstable, fast, or high, we have already decided to go around at this point rather than attempt to save it.
Aviate
Discuss takeoff briefings, stabilized approaches, and go-arounds within your flight department.
What takeoff decisions can you simplify? Are you ignoring stabilized approach criteria? What is your red line for go-arounds?
Message me your feedback on this approach to takeoff and landings; I want to hear criticisms.
Navigate
Resources
Reducing Runway Excursions in Business Aviation - NBAA
Reducing the Risks of Runway Excursions - Flight Safety Foundation
Rejecting a Takeoff After V1...Why Does it (Still) Happen? - NLR Air Transport Safety Institute