Nall report released

Discussion in 'Aviation Mishaps' started by Cluemeister, Oct 12, 2019 at 3:47 PM.

  1. Cluemeister

    Cluemeister Line Up and Wait PoA Supporter

    Joined:
    Nov 11, 2015
    Messages:
    657
    Display Name:

    Display name:
    Cluemeister
  2. Doc Holliday

    Doc Holliday Pre-takeoff checklist

    Joined:
    Dec 9, 2016
    Messages:
    453
    Location:
    Tombstone
    Display Name:

    Display name:
    Doc Holliday
    The trends are improving because activity is down.

    It would have been useful to show a chart of GA activity over the past 10 years and the correlation.
     
    dtuuri and KA550 like this.
  3. OneCharlieTango

    OneCharlieTango Pre-Flight

    Joined:
    Dec 21, 2018
    Messages:
    79
    Display Name:

    Display name:
    OneCharlieTango
    According to the story on the Nall Report on AOPA Live This Week, activity is up.
     
  4. Cluemeister

    Cluemeister Line Up and Wait PoA Supporter

    Joined:
    Nov 11, 2015
    Messages:
    657
    Display Name:

    Display name:
    Cluemeister
    "While some areas are not improving as quickly as others, the overall fatal accident trend shows a large reduction and simultaneously an increase in GA activity (total flight hours flown). The FAA estimated 2016 flight time at 24.64 million flight hours, compared to 23.98 million flight hours in 2015."
     
    wilkersk likes this.
  5. Doc Holliday

    Doc Holliday Pre-takeoff checklist

    Joined:
    Dec 9, 2016
    Messages:
    453
    Location:
    Tombstone
    Display Name:

    Display name:
    Doc Holliday
    The last data sampling was 2016.

    Also interesting is the VMC into IMC rate.
     
  6. wanttaja

    wanttaja En-Route

    Joined:
    Jun 7, 2008
    Messages:
    2,878
    Location:
    Seattle
    Display Name:

    Display name:
    Ron Wanttaja
    The trouble is, this is based on the FAA General Aviation Survey, and the survey results are badly affected by the FAA re-registration process.

    The FAA estimate of annual flight hours is determined by A) The average hours reported by the respondents to the survey, B) The estimate of active aircraft based on the respondents to the survey, and C) The number of aircraft listed in the FAA registry.

    C) is varying wildly, from year to year. 40,000 aircraft were removed from the registry in 2013, vs. 4,254 in 2016. This produced, essentially, an "uptick" in aircraft registrations in 2016...which, of course, gets multiplied by the average hours to compute the total flight hours. This is purely artificial, driven by the re-registration process.

    Here's the number of registrations by year since 2000:
    [​IMG]
    The "total registered aircraft" for each year was based on the FAA registry as of 31 December of that year.

    So, yes...there was an increase in hours in 2016, compared to 2015, because there were more aircraft. But, again, that rise was mostly artificial. The results for the next two years don't bode well.

    Ron Wanttaja
     
    Velocity173 and Doc Holliday like this.
  7. Doc Holliday

    Doc Holliday Pre-takeoff checklist

    Joined:
    Dec 9, 2016
    Messages:
    453
    Location:
    Tombstone
    Display Name:

    Display name:
    Doc Holliday
    Thanks Ron for the explanation.
     
  8. Clip4

    Clip4 En-Route

    Joined:
    Jun 27, 2013
    Messages:
    4,630
    Location:
    A Rubber Room
    Display Name:

    Display name:
    Cli4ord
    Other than the Great Recession, what around 2009 that reduced accident rates 15% going forward?
     
  9. wanttaja

    wanttaja En-Route

    Joined:
    Jun 7, 2008
    Messages:
    2,878
    Location:
    Seattle
    Display Name:

    Display name:
    Ron Wanttaja
    The number of GA accidents has been dropping over the past twenty years....
    [​IMG]
    I don't track the specific statistics for overall GA. But I do have some specifics for homebuilts.

    Over the period 1998 through 2017, the homebuilt fleet increased by about 15%. During that same period:

    - Total accidents decreased by about 19%
    - Accidents due to loss of engine power decreased by 41%
    - Accidents due to Pilot Miscontrol (stick and rudder errors vs. judgement errors) decreased by 16%
    - Accidents due to Builder error dropped by 60%
    - Accidents due to Maintenance error dropped by 36%
    - Accidents due to pilot judgement issues (overshoot, undershoot, bounced landings, etc.) dropped by 40%
    - Fuel exhaustion accidents dropped by 83%
    - Maneuvering at low altitude accidents dropped by 43%
    - Accidents due to continued VFR in IFR conditions or midairs weren't a significant proportion of homebuilt accidents.

    Note that these percentage decrease is in the raw NUMBER of accidents, not percentage of total.

    So, basically, what we're looking at in the homebuilt world is that the aircraft have better reliability, builders and maintainers are getting better, pilots are showing better judgement, but show no improvement in the physical control of the aircraft (which is the leading cause of homebuilt accidents).

    Ron Wanttaja
     
  10. wanttaja

    wanttaja En-Route

    Joined:
    Jun 7, 2008
    Messages:
    2,878
    Location:
    Seattle
    Display Name:

    Display name:
    Ron Wanttaja
    This is another view of the effects of the FAA re-registration program. Starting in 2010, aircraft owners were required to confirm their aircraft registrations every three years. The program was phased in over the following three years so that not everyone's aircraft came up for renewal at the same time.

    If the owner didn't confirm, the aircraft registration was administratively cancelled. This plot shows the number of cancelled registrations each year.
    [​IMG]
    What's interesting is that, with a three-year cycle, it appears the variation is actually appearing on a six-year cycle. Note the number of aircraft deregistered in 2017 and 2018 exceed those leading to the previous peak (2011 and 2012).

    Obviously, the ten-year low in deregistrations in 2016 contributed to the "spike" for that year in the fleet size plot I posted earlier. This produced an similar spike in estimated hours flown, which, of course, leads to a reduction in the fatalities-per-100,000-flight-hours rate. But as the above plots show, the results for 2017, 2018, and, probably, 2019 are going to come out worse. Just because the FAA paper chase removes more aircraft from the registry.

    The basic problem is that this issue is administrative. It's a paperwork game, and it has little do with how many active airplanes there REALLY are.

    The problem comes when you try to combine this registration paper chase with the blood-and-iron reality of aircraft accident statistics analyses. I greatly respect FAA Survey people and Nall Report folks, but you can't draw consistent conclusions when you try to mash these separate components together.

    Ron Wanttaja
     
    mcmanigle likes this.
  11. wanttaja

    wanttaja En-Route

    Joined:
    Jun 7, 2008
    Messages:
    2,878
    Location:
    Seattle
    Display Name:

    Display name:
    Ron Wanttaja
    The effect of this is hard to understand with a bunch of hand-waving. I've generated a set of graphics to show the effect.

    This first image shows a notional breakdown of the FAA registry. Let's assume there are 100,000 airplanes in it (actual is about 300,000). Prior to 2010, aircraft were only removed from the registry upon their owner's request. So the only entities actually cancelling registrations were corporations and insurance companies. So Great-Uncle Jim's Piper E-2 might still be listed, even though he scrapped it before WWII.

    The first figure shows the notional breakdown of our 100,000 airplane set. Half the registrations are existing aircraft, the other half are aircraft still on the registry but with no valid owner associated with them (again, these numbers are for illustration, not actual estimates).
    [​IMG]
    Of the 50,000 "real" airplanes, let's assume 75% of them are actively flying, and the other 25% have valid owners who aren't active.

    The FAA General Aviation Survey uses the FAA registry to send surveys to a sample of the registrants. The recipients are chosen at random. Let's assume the FAA sends out surveys to 10% of the owners of record on the registry. Let's assume that 10% of the owners of valid-registration aircraft return them.

    That means the Survey will receive 375 responses from owners of active aircraft, 125 from owners of inactive aircraft....and none from the old, out of date registrations.

    That's not a problem with the Survey. They base their results on the data on returned surveys, and don't make any assumption about the ones that didn't get answered (remember, many of those went to valid registrations and the owner didn't bother to answer).

    Let's assume every active owner says he or she flies 100 hours per year. The Survey folks take the responses, and compute that 375 out of 500 GA aircraft is active. So 75% of the GA fleet is active, and since they all fly an average of 100 hours per year, the total GA fleet flying time for that year was 750,000 hours.
    [​IMG]
    OK, let's turn things over to the Nall Report folks. They look at the NTSB records and determine that there were 100 accidents that year. Couple that with the FAA Survey's estimate of 750,000 GA flight hours, it's obvious that there is a GA accident every 7,500 hours. Again, perfectly acceptable methodology.
    [​IMG]
    With me so far? Let's start looking at how the FAA re-registration process gums up the works.

    After 9/11, the FAA had to admit that they'd lost track of tens of thousands of aircraft in the US. Aircraft that were still on the registry, but they didn't have a valid owner listed. Congress was not happy.

    Starting in 2010, the FAA requires that owners of aircraft confirm their registration and address every three years. They get a notice from the FAA, and have to return a confirmation and $5 to retain their aircraft's license. If they DON'T do this, their registration is cancelled and the plane removed from the FAA registry.

    The FAA phased in the implementation of this, so that a third of the US fleet comes up for confirmation every year.

    This graphic shows how this works. The Valid registrants, both active and inactive, respond to the re-registration demand and their airplanes remain on the registry. Of course, no response is received from the invalid registrations. So on this first year, the FAA cuts 16,600 aircraft from the registry.
    [​IMG]
    Note that this ONLY affects aircraft with bad registrations. THERE IS THE SAME NUMBER OF VALID REGISTRATIONS, because the owners of these aircraft receive the FAA notice and respond.

    (in reality, there is always some subset of of owners that DON'T respond, and end up getting their registrations cancelled. They can re-activate them with little effort for the first five years; beyond that, they risk having their N-numbers reassigned.)

    The main point here is that the number of flying, active aircraft HAS NOT CHANGED. All the aircraft removed are obsolete registrations; many of the planes haven't existed for years.

    So, what happens the next time the FAA Survey rolls around? One good thing is since there's less "deadwood" in the registry, more real owners receive surveys. But the percentage of active vs. inactive aircraft hasn't really changed. It *still* comes out at 75%.

    The problem is, the overall fleet size (which includes the invalid aircraft) is lower... so the computation of the total hours flown by the GA fleet is lower as well.

    [​IMG]
    This, of course, is NOT the actual case. We've got the same number of active aircraft, flying the same number of hours per year. But because the "official" fleet size decreased, the result is a lowered estimate of annual utilization.

    You can see where this is going, when we hit the next Nall report. It's the same number of active aircraft, and the SAME number of accidents...but because the estimated annual utilization has gone down, the GA accident rate goes up.
    [​IMG]
    So that's how the re-registration process is affecting the estimate of the GA accident rate.

    The FAA Survey folks and the Nall Report people are eminently qualified. Me, I'm just an 'ol retired engineer who likes working with numbers. It's just that their processes were developed in an era before the FAA re-registration kick.

    Over the long term, it doesn't matter...but in the short term, we're going to see some spikes and valleys in the accident rates that have little do do with actual safety. Let's not get ourselves too wound up about it.

    Ron Wanttaja
     
  12. James331

    James331 Touchdown! Greaser!

    Joined:
    Apr 18, 2014
    Messages:
    19,368
    Display Name:

    Display name:
    James331
    Did I not see it, but where is the breakdown per 1000 flight hours or whatever.

    I mean 14 crashed by a ATP, 8 by a CPL, but how many hours did the ATP group vs CPL group fly over that span of time, or is it I just didn’t see it on my phones screen?
     
  13. GeorgeC

    GeorgeC En-Route PoA Supporter

    Joined:
    Dec 5, 2010
    Messages:
    2,780
    Display Name:

    Display name:
    GeorgeC
    2020 will provide a new, if arbitrary, chalk line; should be able to measure hours independently using data from e.g. adsbexchange.