Hand-Flying and Attention to Detail by Shem Malmquist

This talk proposes that the concepts developed for Just Culture may provide an avenue to broaden the scope of accident investigation and move away from the "blame" outcome of most reports through the use of a simple Just Culture algorithm to mitigate cognitive bias on the part of the investigator.   Absent a formal strategy, cognitive bias has a high probability of occurring, and becoming integrated into the investigators subconscious during the early stages of an accident investigation.  Just Culture is becoming widely accepted, and as such the transition to integrating an investigative model utilizing the concept should be easier to implement and may encounter less political push back than some of the more complex approaches proposed in recent years, yet still provide a robust path to causality and human factors aspects that is more comprehensive than that offered through the traditional models that are still in use by most organizations.

Slow-Onset Hypoxia: An insidious killer

Slow-Onset Hypoxia: An insidious killer

Captain Shem Malmquist and Dr. Paul Buza, Florida Institute of Technology

Have you ever ran a checklist, and then looked up a few minutes later and realized that something that you thought you had done was not done?  Is there a professional pilot with any experience who has not had this experience?  The point is that as pilots we often read reports and dismiss them with the thought that “it won’t happen to me”  If you could honestly say that you have never missed a checklist item, then you can skip this article as you are likely not a pilot!

Hypoxia.  That’s one of those things that you have to worry about flying little airplanes.  In our modern transport aircraft, that’s not a real risk, is it?  

We all learned about hypoxia early in our flying careers.  Part of that discussion likely involved a chart that showed the time of useful consciousness at various altitudes in the event you lost pressurization.  Like most pilots, I ran through the scenarios in my head.  I had climbed mountains over 14,000 feet, so it did not seem that bad, so how bad could it be if we lost the cabin?

Time of Useful Consciousness Chart from Pilot's Handbook of

Aeronautical Knowledge

The truth is that a sudden, rapid loss of pressurization is not likely to lead to a problem of hypoxia.  The reason being that the crew is going to have no question as to what is happening and the need to don oxygen masks.  While this is the type of event that seems like the “worst case” scenario, like many risks, the one with the most salience is not the one that is the most likely to kill you.

On August 14, 2005, the crew of a Helios airlines B-737 were preparing for departure[i].  Things moved quickly during the preflight and departure stage, as we have all experienced.  Flows were ran, checklists completed, and the flight departed.

A few minutes after takeoff, the Captain called back to their dispatcher to discuss a problem that appeared to be the air-ground shift.  The crew had an intermittent warning horn, a master caution and an equipment cooling alert.  The Captain inquired as to where the circuit breakers were for the equipment cooling, not surprising as the aircraft maintenance history showed that it had a history of equipment cooling problems.

Helios B-737

Expectation bias is the tendency to believe or see results that agree with those details that agree with our expectation, and to disbelieve, discard or downgrade the cor­re­spond­ing weight­ings for the data that appear in con­flict with those expec­ta­tions[ii].

The crew continued to trouble-shoot the issue while they climbed.  13 minutes after takeoff there were no further communications from the flight crew.   The actual problem turned out to be that the pressurization was not properly set;  it was part of the checklist, but it was not accomplished.  The Captain was known for being meticulous on procedures and quiet.  According to the report, the first officer had a history of rushing through checklists.  This is not uncommon, although what is even more common (and likely something every professional pilot has done at one time or another), is “looking” at an item without really “seeing” it. 

Unlike a rapid decompression, the cues for “slow onset” hypoxia are not as clear.  Think of it as the “boiled frog” effect.  While the cabin slowly ascends there are the usual cues.  Ears need to be cleared, but for anyone that flies often, that is pretty much an unconscious act.  Other cues are more subtle, such as gradual decrease in the ability to discern color and a slow degradation in the ability to think clearly.  The crew of the Helios B-737 never sorted out what was happening.  The aircraft continued to fly on autopilot, finally running out of fuel and crashing about three hours later, killing all 121 aboard.  The flight crew never regained consciousness.  It appears that a flight attendant either maintained or regained consciousness and attempted to regain control of the aircraft just prior to the aircraft running out of fuel.

All humans are subject to cognitive biases of various sorts.  These biases are heuristic “short cuts” that we use to make sense of the world around us.  They are the basis of “intuition[iii]”, which allow us to make more rapid decisions in familiar situations.  The more careful thinking required to logically analyze a situation or concept requires considerably more energy.  Humans do not like to do this under the best of circumstances (preferring the much faster and easier intuitive method).  If our mental ability is compromised in any way, such as being fatigued, ill, impaired by medicines or alcohol, or even with low blood sugar, we will be even more inclined to think the “easy” way.  Hypoxia will have similar effects.

The effects of hypoxia are subtle, while the effects of biases are very strong.  Pilots have often continued flying without seeing or hearing warnings;  accident, ASRS and ASAP databases are full of these types of events.  Given a combination of circumstances, it is not hard to see how an event such as this happened to the Helios crew could happen to any of us.   Missing a checklist item, a warning or alert is common.  This has been a factor in so many different types of accidents, that it cannot be overemphasized. Regarding this issue, the Helios accident report, in discussing missed procedures or checklist items, states:

The Board was also sensitive to the fact that automatic execution of actions was very much affected by assumptions – in the case of performing a large number of verification steps, the assumption that all switches and indications were in the usual, normal for this phase of flight position. A superfluous green indication on the pressurization panel could be easily (inadvertently) overlooked when perception was biased by the expectation that it should not be present.

Exacerbating this tendency (expectation bias) is the rarity with which switches (especially, and directly relevant to this case, the pressurization mode selector) are in other-than-their-normal position. A pilot automatically performing lengthy verification steps, such as those during preflight, is vulnerable to inadvertently falsely verifying the position of a switch to its expected, usual position (i.e. the pressurization mode selector to the expected AUTO position) – especially when the mode selector is rarely positioned to settings other than AUTO…

…because they are performed repeatedly on the line, they are also performed by memory, typically in time-pressured circumstances (i.e. indirect pressure to maintain on-time  departures). For these two reasons, checklists are often performed in a hurried, automatic

fashion. From a human factors standpoint, rushing is known to lead to the inadequate allocation of attention to the task at hand – and thus to errors. Furthermore, like procedures, checklists are also vulnerable to “looking without seeing” because they are biased by the assumption that since each item verified an action performed only moments

ago, then it must be already in the desired position/set  (P.118-120)

These lessons alone can save your life.  The only way to break this tendency is to make a conscious, direct effort to focus on each item and ensure it is accomplished.  The failure to pressurize while on ascent, while the crew is attending to other issues, represents one of the most dangerous hypoxia situations.

Having now established how easily a crew might find themselves in a scenario where the aircraft is not properly pressurized, let us turn our attention back to the effects of hypoxia.  Unlike the rapid decompression situation, the “slow boil” hypoxia is not obvious, and there is nothing that “jumps out at you;” this is the situation that killed Paine Stewart.  Any sort of depressurization is rare, and one that is very subtle is even more dangerous.  The following chart outlines the effects:

Sequence of Events - Slow Onset Hypoxia

Notice how subtle the effects are.  The items up through 18,000 feet could be easily ignored, and the loss of cognitive functions from 18,000 through 24,000 feet are also things that we might not notice or might attribute to other things.  Tunnel vision, like loss of color vision, is much less obvious while a person is experiencing it, particularly when focused on trouble-shooting a task.  Once past that phase, things happen rapidly.  Unfortunately, the symptoms are impairing our ability to see the symptoms and recognize them!

Hypoxia can be a factor in other scenarios.  Cargo aircraft that have been depressurized as part of a smoke procedure could put crews at risk if they elect to go back and investigate without an activated walk-around oxygen mask.  The canular-type supplemental oxygen used in general aviation can put a person at risk at higher altitudes as well due to the insidiousness of the slow degradation.  A similar effect could occur due to a leaking mask or a diluted oxygen flow.

Hypoxia is one of those threats which most professional pilots seldom consider.  With modern aircraft, the chances of a rapid decompression are very low.  Many pilots do not worry about it and approach the entire issue casually.   Slow-onset hypoxia has received little attention over the years.  The thought of the pressurization not working on climb out is not one that strikes fear in our hearts or grabs our imagination like the thought of a door blowing off an aircraft leading to explosive decompression.  Despite that, it is the slow one which is more likely to kill you. 

The degradation of our cognitive functioning which occurs with hypoxia can lead to a fixation that is so strong that the last thing you will ever know is “what a great pilot you are”.

Mitigations

Awareness is the key.  Each persons’ hypoxia experience will be unique to them. Crew communication (i.e., good CRM) becomes a vital component in identifying the problem prior to a loss of situational awareness and subsequent incapacitation.  Furthermore, flight attendants should be trained to have a fundamental understanding of the sequence of events once the “rubber jungle” drops and should proactively participate in the CRM with the flight crew.

---

[i] HELLENIC REPUBLIC MINISTRY OF TRANSPORT & COMMUNICATIONS. (2006). AIRCRAFT ACCIDENT REPORT HELIOS AIRWAYS FLIGHT HCY522.

[ii] Dismukes, R., Berman, B, and Loukopoulos, L. (2007). The Limits of Expertise. Ashagate: Farnham, England.

[iii] Kahneman, D. (2012). Thinking Fast and Slow. Ferrar, Strauss and Giroux: New York.

Paris info

2nd Page of Paris info

Not safety related, dated info, but still good, courtesy of one of our now retired pilots.

Airborne Weather Avoidance

Airborne Weather Avoidance

By Captain Shem Malmquist

You are on approach with convective weather in the area.  Lightning is present, and you can see the storm in front of you has some lightning in it.  You are in the “conga line”, and aircraft are landing in front of you with no problems.  Your radar is showing just light to moderate rain between you and the airport.  Will you continue the approach?

On August 2, 1985, Delta 191 approached the DFW area[i].  The ATIS described the weather as benign, scattered clouds at 6,000’, 10 miles visibility and calm wind.  There were scattered thunderstorms in the area, and the flight made a few deviations accordingly.  At 1756 CDT, ATC transmitted that “…there’s a little rainshower just north of the airport…”.   The Delta 191 crew told ATC they were at 5,000 feet at 1800. At 1802 they were 6 miles from the outer marker.  At 1804:18 the first officer stated that there was lightning coming out of the cloud in front of them.  They reached 1,000 AGL at 1805:05, and the aircraft crashed at 1805:58.   During this same time period, NWS radar showed a level 3 cell off the end of the runway at 1756, which had intensified to a level 4 cell by 1804.  But what did the crew see on their radar?

Where do you set your radar tilt control on approach?  If you’re like most pilots, it is likely that your radar is set to less than 5 degrees nose up. A tilt setting of 5 degrees nose up will place the radar beam at 5,000 feet above your altitude at a distance of 10 miles (the beam moves 1,000 ft/degree at a distance of 10 miles).  On final, that means that most aircraft will be scanning the weather between themselves and the airport that is below 10,000 feet. If you are flying an aircraft with the “auto” setting, odds are that you are leaving it in “auto”.  What altitudes is it scanning?  If you do not know the answer, a review of the system manual might be worthwhile.

Why does this matter?

Doctor Fujita, who created the tornado scales, also did a significant amount of research on microbursts.  What he found was that during the building stages of a thunderstorm extreme updrafts can literally hold the rain at relatively high altitudes (15-20,000 feet).  Eventually the water gets so heavy that it overcomes the updraft and it essentially dumps.  It is not unlike one of those big buckets at a water park that eventually reaches the point where it is full enough to tip over and dump its contents.

The column of water is now falling downwards rapidly, and as it encounters warmer air, starts to evaporate.  The evaporation causes cooling, which further accelerates the vertical column of water and air.  It eventually hits the ground and fans out – a classic microburst.

Modern Terminal Doppler Weather Radar (TDWR) and Predictive Windshear systems will detect the event when it starts to fan out.  Both of those systems are designed to measure horizontal (x-axis) wind shifts.  However, they will not “see” the column of water when it is still suspended in the updrafts or when it first starts downwards and has not started to fan out.

NASA contracted with MIT’s Lincoln Laboratory to conduct a study of aircraft thunderstorm penetrations in the terminal area. [ii] The study found that flight crews were regularly penetrating severe storms, and the only reason we do not have more accidents Is chance.  The probability of a microburst occurring over the close in final approach course of an airport while an aircraft is there is very low.   Pilot training as well as TDWR and PWS have led to crews avoiding those microbursts that were in the mature stage.  However, there is little or no protection against crews flying under the rapidly falling column of water that has yet to start fanning out.  Our radar training and the wind shear systems do not address this type of scenario.

To combat this, consider varying the tilt control.  If you tilt the beam upwards to the maximum limit and get a red return, but find that the return is showing much less precipitation at lower altitudes, that might be an indication that something is, literally, “up”.  There are only a handful of possibilities that would lead to this.  One, as has been described, is that the water is (or was) suspended up high due to updrafts – meaning that at some point gravity will win that tug of war. The second might be VIRGA.  That will show rain above, but below it is going to dry out.  The last might be hail, where wet rain up higher is freezing, so not showing on your radar when you are scanning the lower altitudes.   Any of these can be a bad scenario to fly into, so if you see that, be very aware and consider all the indications you are seeing.  Proceed with caution.  Consider asking ATC what they are depicting and pay close attention to any secondary indications of hazards.

Delta 191 encountered storm that was developing significant weather at higher altitudes, which, once reaching critical mass, would essentially dump the rain downward, along with the significant wind, which would doom the flight.  At 1800 CDT, and 5,000’, the flight was about 20 miles from the airport.  If, like many crews, they had their radar tilt less than 5˚ nose up, they would be viewing weather at around the 15,000 foot range.  It is probable that this was not high enough to detect the severity of the storm at that time.  Tilting the radar higher would have yielded the storm’s deadly secret of having a lot of moisture up high, and little down below.  The recipe for a microburst.

What if you’re already in the rain and your radar indications are limited?  Unlike airborne units, ATC radar can power through the strongest precipitation without attenuation can be of value if you’re not sure what might be on the other side of the radar echo you’re seeing.  The aircraft radar runs about ten times the frequency of ATC radar, and it ismuch less powerful (peak outputs 150 watts vs. the 25KW range).  This means that our airborne radar attenuates easily.  On the plus side, though, our higher frequency allows the airborne system to “see” smaller water droplets and provide more definition than what ATC can get.  We also can focus it to get an idea of the vertical development of the weather.

 Enroute

We want to look for weather that is extending vertically into the flight levels.  Use the tilt formula, and find the altitude where the bottom of the beam is first hitting the weather.  If you are at FL200 and the bottom of the beam is hitting weather aligned with your altitude, you are probably looking at convective weather.  If you are at 10,000’, you will want to rotate it upwards.  Let’s assume the weather is 50 miles in front of you.  If you are hitting weather with the bottom of the beam 2˚ above level, you are seeing something that is extending up to 20,000’.  If you are 30,000’, you will want to tilt it down to ensure that you are capturing the weather.  The reason is that the precipitation at FL300 will likely be frozen, and may not be picked up by your radar.  Momentarily turning your gain off theCAL position might help with this, but frozen precip just does not reflect very well. It is, therefore, very important to remember that just because you are not showing precipitation at your altitude, that does not mean that you do not have severe weather at your altitude.  The rain shaft might stop at FL 250, but the severe weather may extend thousands of feet above that, and just not be reflecting due to it being frozen.

Center weather depiction on the individual controllers scope utilizes a Weather and Radar Processor (WARP), which integrates weather data from one or more NEXRAD sites.  It does not display anything less than moderate precipitation.  Controllers can set their scopes to display weather in three altitude blocks, starting at the surface, starting at FL240 and starting FL 330.  In each case, the display will include all the weather above the floor selected.   You can ask them to toggle through their altitude blocks to get a better idea about the extent of a storm.

Our airborne radar has an advantage where we can get a lot better definition and see gaps that are appearing at our altitude – just be cautious to ensure that the gap is not just a dry pocket within a storm.   We lose that advantage, however, if we are flying along in the teens in an area of general heavy rain.  In this scenario our radar is likely to just depict precipitation in a solid arc of weather across the display.  It can be just a little better than useless.

If you find yourself in this situation, flight planned to fly at 10-15,000 feet, stuck in the rain, another strategy might be to continue up to the lower flight levels, to a point that you are out of the rain and able to utilize the airborne weather radar.   Another choice, and an equally valid one, would be to request ATC’s assistance in this scenario.  ATC will not be faced with the attenuation issues, and should be able to identify areas of more intense weather.  As previously stated, you might consider asking Center to toggle their display to only depict the weather above FL 240, as anything that was depicting at that altitude is very probably significant convective weather.

ATC weather depiction ability in the U.S. has improved greatly since Delta 191, but one thing that has not changed is the ATC primary function of separating known traffic.   Telling us about the weather is not ATC’s primary responsibility, but controllers do want to help where they can.  Many controllers are not aware of the limitations of our radar displays, and they are also limited by their equipment and FAA legal has placed additional limitations on what they are, and are not, allowed to tell us.  Through understanding of what they can depict, we can directly request the information that we need.

[i] http://www.airdisaster.com/reports/ntsb/AAR86-05.pdf

[ii] Rhoda and Pawlak, An Assessment of Thunderstorm Penetrations and Deviations by Commercial Aircraft in the Terminal Area.  1999

Airborne Weather Avoidance - What Air Traffic Controllers should know

Weather Depiction

By Captain Shem Malmquist

“Flight 1210, I am currently depicting an area of extreme precipitation from your 10 o’clock to 2 o’clock approximately 40 miles in front of you”.  We were enroute from TPA to MEM at FL380.  Our radar was set to about 2 degrees down, and we were showing absolutely nothing on the radar.  Suspecting snow, I turned the radar’s gain all the way to maximum, and at that setting, a few scattered green (lowest level of intensity) very small dots (more like pin-points) appeared in the general area where the controller said he was depicting weather.  Why such a difference?  When pilots ask for more information, we are often met with a response from ATC with words similar to “Your weather depiction on your radar is better than mine”.   The statement highlights the following fact: Many pilots and many controllers do not fully understand each other’s capabilities. 

So, who does have the better display? The answer is “it depends”.  First, a caveat. Aircraft have a wide variety of equipment. Many non-airline aircraft have various uplinked weather products.  Few (if any) U.S. airline aircraft have more than airborne weather radar.  As this article concerns U.S. air carriers, we will not discuss those tools that may be available to corporate and general aviation aircraft.   

Airline aircraft use x-band radar, which is very good for getting sharp definition of wet rain, but attenuates very easily and is virtually unable to get a return on a dry particle (snow or hail).  It is also low power.  Unlike the radar that ATC uses, which has a peak power output in the 25kw range, the modern airborne weather radar units only have a maximum output of 150 watts. Yes, you read that right.  They put out less energy than some of the light bulbs likely installed in your house!  They make up for that by having very sensitive receiving antennae and good processing computers.  Still, that small amount of power only can go so far – is it any surprise that it cannot pass very far through a band of weather?

This capability is even worse when we are flying in the precipitation itself.  The rain coats the radome and greatly reduces the ability for the airborne weather radar to “see” what is happening.  Couple that with the attenuation once the beam leaves the aircraft and it is attempting to pass through a wall of water.  Flying in the precipitation, sometimes all we see is an arc of red extending only a few miles in front of the aircraft.  During these times, what is interesting, is that the most intense storms will actually attenuate the radar even more, and depict as a thinner spot, with a “bow” inward in the depicted arc.  If the pilot is not sharp, what it looks like is that the fastest way out of the weather is to turn towards that bowed in area.  On April 4, 1977, Southern Airways 242, a DC-9, was sucked into such a trap.  Attempting to find a way through severe weather, they turned towards that bow like a moth to a light.   They flew into the jaws of a Level 5 thunderstorm, crashing a short time later after both engines flamed out in the heavy rain and hail.

Contrast this against the WARP (Weather Radar and Processing Unit) on the ARTCC displays.  The higher power, multiple sites and integrated NEXRAD entirely eliminate the attenuation issue for all practical purposes.  Although the display can be a few minutes old due to the processing time, it is very valid information.  Further, the display can be selected to change the base altitude from the surface, FL240 or FL 330.  This is important.  Storms with vertical development can be analyzed this way to get some idea of what is being looked at.  Further, remember my story earlier about not depicting anything while ATC was showing an area of intense precipitation?  I asked the controller what altitudes he had selected.  He was showing weather FL 330 and above.  Most precipitation at that altitude will be frozen.  So, why could I not see it on the airborne unit?  The reason is that it was snow.  The ARTCC display shows dry precipitation.   Airborne radar does not.  This means that a controller can potentially know where the thunderstorms really are.  If there is a wide area of rain, with some thunderstorms popping out of it, by displaying the higher altitudes, the actual individual storms can be depicted.  As most of the precipitation at that altitude is likely snow, the airborne system is very limited in what it will display.

So, when is the airborne system better?  Well, if we are out in the clear, we can tilt to remove the lower altitude stuff and differentiate that way, looking at the FL250 range to try to find the part of the storm that is not just frozen precipitation. We can also adjust the beam tilt to show the weather that is at our altitude, and not below it.  The other advantage is much higher definition, which, if we are clear of the storm, allows us to come much closer to the edge of the weather or thread through and area of storms.

In the end, both ATC and airborne systems have their advantages.  By letting airline crews know the altitudes and intensities you are depicting the weather, the crew will be in a lot better position when making safety critical decisions in the enroute environment. 

A similar difference exists in the terminal environment, but for different reasons.  Not only is the airborne radar system subject to attenuation, but it is also limited by tilt.  Airborne radar on airline aircraft send out a beam that is approximately 3.5 degrees wide.  It sweeps to the left and right, usually about 45 degrees to each side.  The pilot can adjust the tilt of the beam from 15 degrees down to 15 degrees up.   Using trigonometry, for every 1 degree of tilt change, the beam is moving 1,000 feet at a range of 10 miles from the aircraft.  This means that at 20 miles the beam is moved 2,000 feet, and at 5 miles it is moving just 500 feet per degree.

  Using the same formula, we find that 15 degrees nose up tilt will put the beam at 15,000 feet at the 10 mile range, and 7,500 feet at the 5 mile range.  This means that the ability for an aircraft to depict whether the rain is a thunderstorm or just low altitude rain is very limited for close-in weather when the aircraft is on the ground, or close to the airport. 

Unlike the aircraft, ATC radar is not tilt limited, so ATC has a much better depiction of weather that is close in to the aircraft.  Just something to keep in mind.  On August 2, 1985, Delta 191 approached the DFW area[i].  The ATIS described the weather as benign, scattered clouds at 6,000’, 10 miles visibility and calm wind.  There were scattered thunderstorms in the area, and the flight made a few diversions inbound.  At 1756 CDT, ATC transmitted that “…there’s a little rainshower just north of the airport…”. The Delta 191 crew told ATC they were at 5,000 feet at 1800. At 1802 they were 6 miles from the outer marker.  At 1804:18 the first officer stated that there was lightning coming out of the cloud in front of them.  They reached 1,000 AGL at 1805:05, and the aircraft crashed at 1805:58.   During this same time period, NWS radar showed a level 3 cell off the end of the runway at 1756, which had intensified to a level 4 cell by 1804.  But what did the crew see on their radar?

This was a storm that was developing significant weather at higher altitudes, which, once reaching critical mass, would essentially dump the rain downward, along with the significant wind, which would doom this flight.  1800 CDT, and 5,000’, the flight was about 20 miles from the airport.  If, like many crews, they had their radar tilt at around 5˚ nose up, they would be viewing weather at around the 15,000 foot range.  It is probable that this was not high enough to detect the severity of the storm at that time.

There is no doubt that a good air traffic controller can bring a lot to the table in improving flight safety, and, as been highlighted in this article, ATC has capabilities that pilots do not have access to.  There is also an aspect to consider that is not related directly to the equipment.  Pilots, being human, can get caught in something called “plan continuation bias”, where the mindset of continuing is strong enough that they discard any information that goes against the plan.  ATC can sometimes be in a better “big picture” position, and by providing more accurate information regarding the physical nature of the storms, may be able to break through that bias.   Sometimes a query is all that is needed to prevent a deadly dynamic.

---

[i] http://www.airdisaster.com/reports/ntsb/AAR86-05.pdf