Wednesday, November 24, 2021

Glasair III, N291KT: Fatal accident occurred November 20, 2021 in Big Bear City, San Bernardino County, California

Federal Aviation Administration / Flight Standards District Office; Riverside, California

Aircraft crashed under unknown circumstances in mountainous terrain. 

FBI Glasair LLC


Date: 21-NOV-21
Time: 00:00:00Z
Regis#: N291KT
Aircraft Make: GLASAIR
Aircraft Model: III
Event Type: ACCIDENT
Highest Injury: FATAL
Total Fatal: 1
Flight Crew 1
Pax 0
Activity: PERSONAL
Flight Phase: EN ROUTE (ENR)
Operation: 91
Aircraft Missing: No
Damage: UNKNOWN
City: BIG BEAR CITY
State: CALIFORNIA

Those who may have information that might be relevant to the National Transportation Safety Board (NTSB) investigation may contact them by email witness@ntsb.gov, and any friends and family who want to contact investigators about the accident should email assistance@ntsb.gov. You can also call the NTSB Response Operations Center at 844-373-9922 or 202-314-6290.

A plane that went missing on November 20, was found to have crashed in the mountains east of Big Bear City with the pilot dying in the impact, authorities said.

The remains of the Glasair III were discovered about 40 miles east of Big Bear City in mountainous terrain on November 21, according to the National Transportation Safety Board and Ian Gregor, spokesman for the Federal Aviation Administration. The pilot, the plane’s sole occupant, was the lone fatality in the crash, Gregor said.

The NTSB was investigating.

The San Bernardino County Coroner’s Office received the pilot’s remains on November 23, and was working on identifying the body, said Gloria Huerta, spokeswoman for the county Sheriff’s Department.

The single-engine plane crashed under unknown circumstances at an unknown time Saturday, the FAA said. The incident was characterized as “accidental” by the FAA.

It was neither immediately clear where the plane took off from nor where it was headed.

Both the San Bernardino County Fire Department and Sheriff’s Department assisted with the search and rescue effort on November 21.

The Sheriff’s Department’s search and rescue team hiked six hours to the crash site and the sheriff’s aviation unit hoisted down personnel with the NTSB, Fire Department and law enforcement, Huerta said. The Fire Department assisted by establishing and securing landing zones for all aircraft involved in the search, said Eric Sherwin, spokesman for the Fire Department.

36 comments:

  1. https://photos-e1.flightcdn.com/photos/retriever/d05fb676caf6710be952a46e4c34f7250de6d269

    ReplyDelete
    Replies
    1. Time of crash was at zero hours of UTC day 21 November. Easy to misunderstand 20/21 November track log calendar day if not paying attention.

      Flightaware track:
      https://flightaware.com/live/flight/N291KT

      Adsbexchange, UAT data as received or as ADS-R re-transmit:
      https://globe.adsbexchange.com/?icao=a2fa24&lat=34.189&lon=-116.987&zoom=10.4&showTrace=2021-11-21&trackLabels&timestamp=1637456040

      Big Bear City Airport nearby is elevation 6752' MSL. Altitude changed from cruising at 9300 to 8000 near Big Bear City. Flew 8600 as cruise altitude in that area on an October 18 flight AZ-->CA:

      https://globe.adsbexchange.com/?icao=a2fa24&lat=33.776&lon=-116.924&zoom=8.3&showTrace=2021-10-18&trackLabels&timestamp=1634520149

      Delete
  2. Replies
    1. The flight aware data is correct.

      Delete
    2. This comment has been removed by the author.

      Delete
    3. Keeping tourists away seems likely.

      Mapped the last recorded GPS data points to visualize. Sure enough, if N291KT ADS-R/UAT position and altitude data is accurate, CFIT was assured when compared to topo map contours adjacent to the hiking trail at San Bernardino Peak.

      KSBD altimeter was 30.03 at the time, making the 29.92 referenced ADS altitudes reported by N291KT about 100 feet below true.
      METAR: KSBD 202352Z 23004KT 4SM HZ BKN250 22/03 A3003,M

      Checking cloud cover from the satellite image at the time, the accident area shows clear sky (working from the labeled Indio position to correlate).

      Unless haze was a factor, the ADS & weather sat image suggest that the 4 PM PST crash was clear sky CFIT. Sunset was at 4:42 at San Bernardino that day.

      Links:
      Mapped final position from Flightaware:
      http://maps.google.com/maps?t=k&q=loc:34.1240+-116.9437
      Mapped final position from Adsbexchange:
      http://maps.google.com/maps?t=k&q=loc:34.123+-116.936

      Topo map (move north/zoom to find square 36 after opening):
      https://www.topozone.com/california/san-bernardino-ca/city/forest-falls/

      Weathersat image at 2350Z, 10 minutes before the accident:
      https://weather.us/satellite/california/satellite-superhd-15min/20211120-2350z.html

      Delete
    4. ^^ Correcting: "the 4:56 PM PST crash was clear sky CFIT."

      Delete
  3. The aircraft departed Camarillo Airport at 4:25 pm. The Flight Aware log ends at 4:56 pm.

    This area was burned in the El Dorado fire in September, 2020. The mountains probably appeared black.

    ReplyDelete
  4. The flight path looks pretty straight and level at 9500 ft and a heading of 095, which suggests he was on autopilot. (Per ADSBX, the barometric derived altitude did fluctuate between 8000 and 9100 feet, but the GPS derived altitude was fairly constant at 9500 feet.)
    I'm wondering if the pilot was medically incapacitated. His cruising altitude of 9500 would not have cleared any of the terrain north to 10 NM south of his route of flight. That would be extremely poor flight planning to not realize that, especially with the many EFB apps available today.

    ReplyDelete
    Replies
    1. This comment has been removed by the author.

      Delete
    2. This comment has been removed by the author.

      Delete
    3. Waiting for word on who was piloting on the accident day as the Lt. Col. and his son are both ATP pilots, have three aircraft and neither was aboard his Grumman AA-1B in 2017 when it had an accident.

      Can't agree with the "in the system" speculation. Another poster has suggested incapacitation and pointed out the discrepancy between geometric altitude (visible in the Adsbexchange side panel data) and the barometric altitude. There may not actually have been a descent to 8000' at all.

      ATP flying is only a part of the experience. The Lt. Col. competed in N291KT at Reno air races in September, got 1st place in one heat and finished ahead of Elliot Seguin in 2 out of three heats. Flew his RV-8 in the races for several years previous to N291KT, with a remarkable recovery from mid-air contact with another racer that sliced off an aileron and the tip of the RV-8 wing in 2017.

      There is a video on Facebook apparently taken from a closely following drone of the Lt. Col. flying the RV-8 off dirt for an Arizona Flying Circus event early this year that is among the most artistic pov perspective sequences ever seen. Would link it now if his name was not in the title.

      It is sad news, no matter who was piloting, but don't let your imagination run wild with so little information to work from.

      Delete
    4. It's pretty unlikely the aircraft actually descended to 8000 if you take a close look at the ADS-B data on ADSB Exchange here https://globe.adsbexchange.com/?icao=a2fa24&lat=34.201&lon=-117.517&zoom=10.4&showTrace=2021-11-21&trackLabels&timestamp=1637455611 you will see that:
      1. At 16:48:28L the barometric altitude returned was 9200. Two seconds later, the barometric altitude returned was 8000. There was no way this plane dropped 1200 feet in 2 seconds and then continued straight and level, especially since the speed did not deviate. The broadcasted barometric altitude continued to wildly fluctuate between 8000 and 9300 for the remainder of the track. This clearly indicates the barometric altitude that the ADS-B transmitter was sending was not accurate for that portion of flight. Many low cost ADS-B transmitters derive the barometric altitude from detecting and decoding the Mode C transponder replies, which is not always a robust process.
      2. Further evidence of this is that the GPS-derived geometric altitude returned remained pretty solidly at around 9500 MSL the entire time the barometric altitude was fluctuating, including the last return at 16:56:26.
      3. This data suggests that the flight remained at 9500 the entire time.
      4. The ADS-B data also shows that the flight was squawking 1200 the entire flight, which indicates they did not pick up an IFR clearance nor were they on flight following, as the first thing they would have been given when contacting ATC would be a discrete squawk code
      5. An IFR clearance would never be just "Proceed to (IFR way point) under VFR conditions and then cleared to PHX." without also specifying an altitude to climb to. Also, ATC would provide an altitude alert to traffic at a dangerously low altitude, even if they aren't on an IFR clearance yet.

      Delete
    5. Re "It's pretty unlikely the aircraft actually descended to 8000":

      Thought that also until the stale geometric update pattern was revealed. Examination of geometric data in comments down-thread makes it impossible to assert anything by relying on geometric altitude in this case.

      Still might be true that no descent to 8000' occurred and that would make sense from an on-autopilot CFIT expectation perspective, just not credible to assert without FAA radar data.

      Delete
  5. I knew the mishap pilot very well. We flew together in F-16s and then Southwest Airlines. Mishap pilot commuted to work in N291KT once a week between Phoenix Deer Valley (DVT) and LAX. I do not agree with most of the creative speculation here, except for the comment above about the mountain appearing black. Time of day and terrain color could have made it difficult to see threatening terrain, thus increasing the risk of CFIT.

    Another factor to consider is that the mishap pilot was very accustomed to flying twice per week through the Banning Pass. It is plausible that route planning complacency was a factor here. The direct line between Camarillo and DVT takes an aircraft over high terrain to the north of Banning Pass. The minimum safe altitude on this direct route would be higher than the aircraft's altitude on the night of the mishap.

    Summary: 1) Sun angle due to time of day, and 2) lower than minimum safe altitude, and 3) *possibly* complacency due to being accustomed to traveling a straight line through Banning Pass = CFIT.

    ReplyDelete
    Replies
    1. Re the suggestion that he took a direct line between Camarillo and DVT and simply did not choose high enough altitude, seems plausible comparing to a 2020 example track through Banning pass.

      Here are the 2020 Banning pass example track and accident track presented such that you can open in two windows and switch back and forth as overlays:

      2020 example through Banning pass:
      https://globe.adsbexchange.com/?icao=a2fa24&lat=33.940&lon=-117.624&zoom=8.9&showTrace=2020-10-30&leg=2&trackLabels&timestamp=1604089847
      Accident flight:
      https://globe.adsbexchange.com/?icao=a2fa24&lat=33.940&lon=-117.624&zoom=8.9&showTrace=2021-11-21&trackLabels&timestamp=1604089847

      Zoomed out further to visualize straight line to DVT:
      2020 example through Banning pass:
      https://globe.adsbexchange.com/?icao=a2fa24&lat=33.940&lon=-117.624&zoom=7.3&showTrace=2020-10-30&leg=2&trackLabels&timestamp=1604089847
      Accident flight:
      https://globe.adsbexchange.com/?icao=a2fa24&lat=33.940&lon=-117.624&zoom=7.3&showTrace=2021-11-21&trackLabels&timestamp=1604089847

      FAA'S radar history will sort out the discrepancy between ADS reported geometric altitude (visible in the Adsbexchange side panel data) and barometric altitude near the end of the accident flight.

      Delete
  6. There are some of us who are very close to the situation. We are doing our best to keep this out of the news until all the legal requirements are set for the family. We appreciate the restraint we've seen so far.

    I'm specifically looking to see if anyone can reconcile the altitude differences. It will be another 1-3 weeks until the NTSB releases anything. I understand barometric changes, but not causing 1000 feet in a that short amount of time.

    ReplyDelete
    Replies
    1. Regarding reconciling the altitude profile:

      The noted discrepancy between ADS reported geometric altitude (visible in the Adsbexchange side panel data) and barometric altitude shows up in a repeating pattern throughout the accident flight.

      Once you see the pattern, it becomes apparent that the reported geometric altitude is not updated at every recorded data point. Here is an example stretch from the climb earlier in the accident track, at sequential data points:

      MM:SS /BaroAlt/GeoAlt
      30:08 / 5900 / 6075
      30:25 / 6200 / 6075
      30:36 / 6400 / 6075
      30:52 / 6700 / 6075
      31:02 / 6800 / 7000
      31:13 / 6900 / 7000
      31:30 / 7100 / 7000
      31:47 / 7300 / 7000
      32:00 / 7500 / 7575
      32:08 / 7600 / 7575
      32:20 / 7700 / 7575
      32:23 / 7700 / 7975
      32:33 / 7900 / 7975
      32:44 / 8000 / 7975
      32:53 / 8100 / 7975
      33:00 / 8200 / 8450

      Not having a fresh GPS altitude delivered into Adsbexchange's recording at each captured data point of this accident flight means that geometric altitude values are not an authority over the observed performance of the barometric altitude.

      It is also important to realize that the designation "Source:ADS-R or UAT" indicates N291KT is transmitting ADS-B on 978 MHZ, which gets received by a FAA ground station and rebroadcast on 1090 MHz. The FAA records what N291KT transmitted directly. Non-FAA receivers feeding third party ADS-B sites are most likely recording the 1090 MHz rebroadcast signal.

      The example 2020 track through Banning pass had loss of signal a short distance beyond the last data of the accident day. That example point of lost signal is a reminder that the last few data points recorded in the accident track are subject to weak signal error processing in whatever manner the decoding algorithms use.

      Reconciling the altitude profile will require a look at the FAA radar data. All that can be said for certain at the moment is that recorded geometric altitudes do not have any authority over what was seen in the recorded barometric altitude.

      Delete
    2. This comment has been removed by the author.

      Delete
    3. @D4T - Have to examine during a climb sequence because being in steady state cruise flight masks seeing how the "update on change" processing of data is being handled.

      Some people will be unaware that the geometric altitude is not being updated at every recorded data point. Being unaware leads to mistaken assertions that potentially stale geometric altitude values have more relevance than those recordings should be given.

      That is how the observed "stale" data pattern relates to this accident. As an example, stale geometric altitude data cannot be relied upon to challenge your evaluation of the brief drop to 8000'.

      Delete
  7. Replies
    1. Another significant barometric altitude data anomaly example is seen in leg 2 of July 14, 2021. As N291KT passes over Catalina Air-Sea Heliport shortly after takeoff from KTOA/Zamperini Field (Elevation:103 ft, Local altimeter = 29.97 at the time), the 29.92-referenced ADS-B baro altitude changes from 150' MSL to 2400' MSL over a period of five seconds between data points, which would be a vertical rate of +27,000 ft/minute.

      Occasional "glitching" anomalies seen in recordings that third party ADS-B folks so generously make available may be an unavoidable result from data processing ADS-R rebroadcasts for UAT-equipped aircraft.

      UAT data path after transmission from the aircraft is FAAreceives978-->decodes, FAAencodes-->transmits1060, then third party ADS-B feeder receives1060-->decodes and records except for a relatively small number of feeder receivers that directly record UAT 978 MHZ transmissions.

      Here is the data sequence from leg 2 on July 14, 2021:

      MM:SS /BaroAlt/GeoAlt/GNDspeed
      51:41 / 200 / 50 / 136
      51:43 / 400 / 50 / 136
      51:46 / 300 / 50 / 133
      51:53 / 400 / 50 / 132
      51:58 / 400 / 75 / 133
      52:05 / 400 / 75 / 140
      52:12 / 400 / 75 / 141
      52:18 / 400 / 75 / 145
      52:20 / 400 / 100 / 154
      52:23 / 400 / 100 / 156
      52:27 / 100 / 100 / 170
      52:39 / 100 / 125 / 175 *
      52:39 / 100 / 125 / 185 * time tag example 1
      52:48 / 100 / 125 / 181
      52:51 / 100 / 125 / 174
      52:56 / 125 / 125 / 169
      52:59 / 125 / 125 / 165
      53:05 / 150 / 125 / 163 *
      53:05 / 150 / 125 / 153 * time tag example 2
      53:08 / 150 / 150 / 151
      53:13 / 2400 / 150 / 151 <-- +27,000 ft/minute
      53:15 / 2500 / 150 / 148 *
      53:15 / 2300 / 150 / 148 * time tag example 3
      53:20 / 2400 / 2350 / 149
      53:24 / 2500 / 2350 / 148
      53:30 / 2600 / 2350 / 147
      53:36 / 2700 / 2350 / 149
      53:41 / 2700 / 2775 / 150

      Note the time tag examples. Position along track does change between the data points, but time of tag repeats, may represent latency between two receivers where the composite track is spliced together.

      Link to leg 2 on July 14, 2021:
      https://globe.adsbexchange.com/?icao=a2fa24&lat=33.803&lon=-118.338&zoom=10.2&showTrace=2021-07-14&leg=2&trackLabels

      Delete
    2. This comment has been removed by the author.

      Delete
    3. Yes, a further example, similar to the glitch you evaluated.

      Delete
    4. This comment has been removed by the author.

      Delete
    5. @D4T - Agree that the subsequent shallow descent to 8,000 just before the crash looks like reliable data, but the five minute time period there is much longer than the typical four stale data points in a row pattern that is seen before the geometric altitude changes. Not sure what to make of that.

      The baro AND geo altitude data leaving KTOA/Zamperini in July doesn't make sense during that 100 foot MSL portion passing over the port's container yard before the +27,000 ft/minute climb.

      Local altimeter of 29.97 vs 29.92 ADS-B reference means that N291KT's ADS baro transmissions should have been pretty much true within the tolerance of it's ADS baro instrument during the KTOA departure.

      Difficult to interpret the last five minutes of the accident flight after observing all of that.

      Delete
    6. This comment has been removed by the author.

      Delete
    7. @D4T - Keep in mind that an effort was being made to simply reconcile the altitude profile, not construct an elaborate scenario.

      The examples found of improper/mismatched baro and geo altitude data in the tracks are what make it difficult to be certain about true altitude profile in the final minutes of the accident flight.

      You can go beyond Flightaware's log and learn about how adsbexchange shows "Squawk: 1200" as you move through data points in the N291KT accident flight track.

      Compare to the track from the Centurion N6209U track linked below that displays controller assigned squawk codes in that flight. (The squawk code shows in the upper left side panel in accordance with each data point that you place the little airplane symbol over).

      The Centurion squawked 7405 on leg 1, 1141 on leg 2 and 5232 on leg 3 during that flight day. Give that track a try to see how that functions, will help you work from a better informed basis next time.

      Then look anew at how 1200 being squawked by N291KT may alter your speculation, if you insist on constructing a scenario.

      Centurion N6209U track:
      https://globe.adsbexchange.com/?icao=a819e3&lat=41.225&lon=-79.447&zoom=13.6&showTrace=2021-11-24&trackLabels

      Delete
    8. This comment has been removed by the author.

      Delete
    9. @D4T - For reference, the ADS-B equipment latitude/longitude position accuracy requirement is within .05 nautical miles (304 feet) or better for operation where required. (See FAR 91.277 & AC_20-165B linked below)

      The final position reported by Adsbexchange of 34.123, -116.936 is further along the flight path than the Flightaware final point and corresponds to the 9600 foot elevation line on the topo map.

      Did you miss seeing the map-pinned locations in the earlier post?
      (Monday, November 29, 2021 at 4:09:00 PM EST)

      ADS-B Regulation links:
      FAR 91.277
      https://www.ecfr.gov/current/title-14/chapter-I/subchapter-F/part-91/subpart-C/section-91.227
      AC_20-165B
      https://www.faa.gov/documentLibrary/media/Advisory_Circular/AC_20-165B.pdf

      Delete
    10. This comment has been removed by the author.

      Delete
  8. I fly a Glasair out if KDVT and had the good fortune of talking to this pilot a few times on the phone when I first purchased my Glasair. I won't mention his name or any more personal data but a very nice guy. Heartbreaking to see this happen. Could've been anything really, so I gather we just need to wait it out. It's so sad to see this sort of thing happen to such a good guy and great American!

    ReplyDelete