The most repeated accident
Recently, a Fireboss landed on the runway with its gear up while fighting a fire in Coimbra, Portugal.
Over the course of 18 years, the same organization has experienced the same accident seven times.
Three of these accidents on the water resulted in a risk to the pilot’s life.
Four were minor accidents on the runway, where although it represents a mess for the airport and ATC, the damage is minor, only to the bottom of the floats, similar to what we see in the following video, where a Canadair 215 goes through the same event, but going around instead of stopping.
Here are the official reports of the four reported accidents.
- Gear down for water landing 3: (English)
- Gear down for wáter landing 2: (English)
- Gear down for wáter landing 1: (English)
- Gear up for runway landing 1: (Spanish)
As for the remaining three, they were either not reported or not investigated by the accidents investigation boards
The first question that comes to mind is:
Are 7 accidents in 18 years too many for the same organization?
Would you say yes or no?
At first glance, it appears that yes, but we need to measure accidents against flying hours in order to draw reasonable conclusions.
Analysis of aviation safety can be made more meaningful by matching accident statistics with flight hours. Fleet utilization is taken into account when calculating accident rates, which are expressed as accidents per 100,000 flight hours. Using this method, it is possible to compare one year to another accurately.
The following is an example from Safety Study NTSB/SS-01/01
Considering that there are few studies for General Aviation and Aerial Work, let’s attempt to determine the approximate number of hours flown during that period for the Fireboss fleet, to compare them with the industry average.
In firefighting, the average contract for a specific aircraft ranges from 120 to 180 hours for the season, although there are busy seasons where aircraft use extra hours – one example was 2013 in Portugal, where between 6 aircraft we counted 1195 hours, 360 sorties, and 5141 drops.
During other seasons and locations, there could be less than 100hrs / aircraft, sometimes as little as 50hrs.
Therefore, let’s shoot for 120 hrs/year on average. This also matches my personal pilot logbook, during 16 seasons as a firefighting pilot for different operators in different countries of the world.
As regards the number of planes for that other specific organization that had all the landing events, they went from a couple of planes in the early days of the model, to 10-12 units operating during the busier seasons, spread between Spain, Italy, and Portugal. Therefore, an average of 8 units operating every year sounds reasonable, which at 120 hrs per year, during 18 years, brings an estimate of 17.000 hrs.
According to simple math, 7 accidents occur every 17.000 hours, resulting in 41 accidents every 100,000 hours.
Considering that those 41 would only be for one category of accidents, and there were many other events in different categories such as LOC-W (loss of control in water) during scooping, or recurrent wire strikes, as I described recently in this entry. Therefore, that 41 could easily be doubled or tripled to reach 100 accidents or more per 10,000 hours.
When we compare 41, or 100, to the 7.2 rate of general aviation more than 20 years ago, as shown in the chart above, or to the 0.27 rate of airlines (Part 121 – scheduled air carriers), we begin to understand where we stand.
The following chart from AOPA confirms that the general aviation (GA) fixed-wing safety record tends to stay between 6 and 7 per 100.000 hrs flown.
In light of this, yes, it’s quite a lot for an organization, and it’s worth looking into.
Apples – Barrels and Hanlon’s Razor: Tracking down the root cause
The bad apple
Do you think all those pilots involved in the accidents were untrustworthy or stupid?
Having known them all, I can confirm they are far from stupid, they were among the best cohorts the organization had.
If the organizational safety culture is still to be developed, our tendency is to assume that organizations’ mega-complex systems would be perfect if not for some untrustworthy or just plain stupid people who cause procedural drifts. (Bad apple theory).
I have seen this happen.
The bad barrel
When the accident repeats over and over without an effective barrier being in place, we need to examine systemic and organizational issues.
The barrel could be rotten or in a bad state, contaminating the good -by nature- apples.
Employees of a company are not inherently ethical or unethical, but rather are influenced by the corporate culture and procedures.
People don’t go to work with the firm intention to perform poorly or cause accidents, except for violators who should be filtered and detected during selection processes.
In the accident “Gear down for water landing 2,” the pilot was and remains a good friend. Great pilot. In the wake of the accident, he was silenced, “this shouldn’t happen”, so we were unable to learn from it.
A few years later, on “Gear down for water landing 3”, it turned out that he had also been a good friend and an excellent pilot. We were once again not in a position to learn much from it, and no significant changes were made in procedures, which meant we were still far from the root cause of the problem.
Why is the barrel rotten? Is there a reason why they don’t fix the issues? Is this a group of evil people?
The way accidents are normalized every season may have led many of us to think the showrunners are to blame. And they are to some extent, but not probably the way I envisioned.
Nowadays I still like to keep them as far as possible, but instead of believing they always act deliberately, or bringing a moral debate impossible to win, I would like to think, in order to preserve some degree of faith and drive, that we should not attribute malice to things that could be explained more adequately by ignorance (Hanlon’s razor concept).
No conspiracy, no evil.
In other words, those running the show high up in the hierarchy, have no idea what they’re doing, as we’ll see in the next point; mitigation
As mitigation after the last gear-up runway landing, they are adding extra flybys prior to a runway landing.
Imagine an Iberia, Air France, or American Airlines performing a routine low approach and go around to cross-check with the tower that they did not forget the gear up.
Probably not the most effective way to tackle the issue.
Before that, it was the sticker on the flaps. That was another point of the master plan.
As there is no flap indicator in the cockpit, and the flap setup is done visually, they thought that reading “Check Gear” on the flap would solve the problem.
Surprise, surprise, the idea did not work.
The procedure to check the landing gear is a memory scan flow, repeated in the form of a crosscheck between the aircraft flying in the same group. There has been no success with this system. When pilots speak, they don’t always look, and when they do, they aren’t always seeing. There is a difference between looking and seeing.
We know pilots had trouble seeing and identifying the landing gear indicator and there were recommendations from the authorities, to train them more effectively. Why would they see the sticker if the problem is not what they check but the way they perform the check?
Their approach is to use the same tools that did not work in the past to fix the symptoms.
It is called a checklist, and it has been around forever. If you are used to another type of aviation, it will seem strange that there is no checklist and everything is done by memory.
There is an inherited belief that checklists are not necessary for single-engine low-level aircraft because we don’t have time to read them.
This is true for emergency procedures, and we can even find it in the official Air Tractor literature.
However, for normal procedures, we can develop an abbreviated checklist in the form of a flip card that is easy to see and follow.
In our operational checklist (flip card), we should determine what are the killer items, which are less relevant, what flow makes more sense, which items should be “Read & Do”, and which ones can remain in memory. The items that must be crosschecked within a formation or group of aircraft should also be clarified.
Checklists should be brief. Overpacking will result in not being used or being vomited without proper checks.
Here is an example of a self-explanatory flip card that was implemented on a new operator, where we were able to start the operation from scratch, leaving inherited paradigms behind.
The blue is the flip card for water operations and fire. The green is the flip card for land operations when interacting with a runway.
This part, extracted from the flip card is of utmost importance:
It must be crystal clear how we wish to operate through the checklist, what to read and do, what to check in memory, and what to crosscheck from one aircraft to another.
The old school way of thinking is rooted in the agricultural view when aircraft were simpler and scenarios different.
Flying a single-engine air tanker locally is not the same as operating a group of amphibious scoopers overseas.
The field of human factors psychology has conducted a great deal of research on checklist design and effectiveness. A key component of human factors research is looking beyond specific technologies to the guiding principles that govern human behavior. CRM and non-technical skills play a much more substantial role, as do checklist procedures and their training.
Therefore, pilots should be trained on checklist procedures, make sure they are aware of their systems (looking but also seeing), or intentionally interrupt them so they will start over.
No checklist, it will bite you sooner than later. You are not different from all of those who bite the dust (or swallow water in this case).
Soft skills – CRM
Here is an example, extracted from a competence-based program, on how to train those specific CRM items:
If during the training or the supervised real operation, the pilot gets caught speaking but not looking, or looking without awareness, not seeing things, I am sorry but you are not competent yet.
The same thing goes for not managing to recover effectively from interruptions, distractions, variations, and failures.
There is a lot more going on than individual mistakes, as you can see.
In an organization, safety culture determines performance and results.
Once the modus operandi is well established, it’s easier for an individual to be influenced by the culture than for the wrong culture to be changed.
Let´s keep dry out there, and our keels unscratched!