Terrain and IFR

Sundays are normally my non-flying days. I’ll often be found on my bicycle on the roads around Altadena, Pasadena, and La Canada. This Sunday found me and cycling buddies climbing the Angeles Crest Highway (ACH) from La Canada to Clear Creek Junction. It was a cool, cloudy, and foggy morning. As we climbed through the clouds, visibility was very poor (sometime as low as 200 ft).

Poor visibility at Clear Creek (~3600 MSL)

Visibility at Clear Creek

Part way into our climb, a stream of fire and sheriff vehicles began passing us going uphill; we also heard a helicopter overhead above the clouds. Experience led us to expect another motorcycle crash on the road. Only when got to Clear Creek (3600′ MSL) did we hear they were looking for a downed airplane.  Details were non existent, but my first guess was VFR into IMC (pilots who aren’t trained to fly in the clouds ending up in the clouds with deadly results).

WARNING: My assessment is based solely on publicly available information. It will likely take the NTSB 1-2 years before they will publish a final report on what caused the accident (to the extent they can determine that). Much of this is educated guesses with the purpose of illustrating what can go wrong when flying IFR and how to avoid problems. One must not forget that a pilot lost his life in the events described here…

N133BW Track FlightAware.com

Once I was back into cell coverage and off my bike, I started to look for information. The first news reports came in just about this time, indicating a Cessna 182 had gone down “near Mt. Wilson” on a flight from Montgomery Field (KMYF – San Diego) to Santa Monica (KSMO). The first question that begs “why is the plane there on that flight?” – it’s no where near SMO nor the path between the two airports. FlightAware allows you to search even for general aviation flights based on departure and destination airports. I  found a likely flight and it’s radar track showed it heading northeast from SMO on a nearly straight line to the crash area. The routing and altitudes made it clear this was an IFR (Instrument Flight Rules – how one flies in the clouds to avoid other airplanes and terrain) flight – my original guess of VFR into IMC was wrong. This was a pilot operating legally in the clouds

To non-pilots and even many VFR pilots, this track seems strange. To an IFR pilot familiar with SoCal and SMO, this makes sense almost right up to the point where the pilot flies into the mountains.

SMO VOR-A Approach

SMO VOR-A Approach

The pilot followed a standard IFR route from MYF to SMO. The turn to the northeast is required to get into a position to fly the only available instrument approach (a way of getting to the runway under the clouds or in low visibility) to Santa Monica. The heavy line on the approach chart shows that airplanes approach from the northeast (starting near Glendale) on a descending profile lined up to runway 21 at Santa Monica. The radar track of the airplane as it turns northeast is paralleling this path with the intention of being turned around back for this approach.

But the turn never happened. 4000′ is a reasonable altitude for that turn, but becomes increasingly dangerous as you continue too far northeast and into the San Gabriel mountains. The pilot had 8-12 miles from when the turn should have happened before reaching terrain above his altitude of 4000′. At 150 kts, that’s only 3-5 minutes.

A normal flight with expected turn

A normal flight with expected turn

With LiveATC, we can get some insight on the communications. All seems normal with the path flown and communication from the pilot until the point where the turn should have been made. The controller can be heard giving the instruction for the turn, but the pilot never acknowledges and never makes the turn. For several minutes after that, the controller continues to try to reach to pilot to have him climb and/or turn in an effort to avoid the rising terrain ahead – all with no effect. Because the radar continues to show a constant heading and altitude, we can probably rule out loss of control of the aircraft, or major mechanical or electrical failure.

It is possible that we may never know why the pilot failed to respond and turn. With what is currently known, we can’t rule out incapacitation of the pilot (with an autopilot maintaining heading and altitude). Another theory would be a loss of radio communication – it could have been failure of a radio, headset, or an accidental change of frequency or volume, or even other interference like nearby transmitters or intervening terrain.

IFR procedures anticipate the possibility of the loss of two way radio communications between controllers and pilots. It becomes the pilot’s responsibility to realize communication has been lost, be aware of current position, and implement the next step in the plan. In this case, the pilot would have been expected to squawk 7600, make the turn and execute the instrument approach while continuing to try to reestablish communication.

The scenario in which we have the ability to affect the outcome is the one of loss of radio communication. If I were a pilot in a similar situation and had a loss of communication when being vectored (guided by specific compass headings) out parallel to the final approach, but towards rising terrain, my recommendations would be:

  1. Slow down. The radar shows the airplane continuing away from the airport at 150kts (consistent with cruise speed on the rest of the flight). That speed simply increases the distance you fly away from the airport and the speed with which you approach terrain
  2. Use a GPS for positional and terrain awareness. Built in GPS units like the Garmin 430W/530W/G1000 all give some insight to terrain. Most newer apps and portable GPS systems will as well. Get one and know how to use it.
  3. Understand when you should expect to be turned. Review of the approach should indicate to an IFR pilot that the turn should occur prior to the DARTS intersection. The chart indicates the proximity of rising terrain northeast of DARTS.
  4. Be aware of sudden changes in communication patterns. The SoCal Approach frequency in that area is pretty busy – were I to stop hearing fairly constant communications, I’d would diagnose, try a different local frequency, or try contacting an airplane you can hear.

There are always points in an IFR flight where loss of communication would be most critical – being vectored at low altitudes toward rising terrain is definitely a critical time. In such a case, a turn to safety, even if not instructed by controllers is preferable to CFIT (controlled flight into terrain – the FAA’s term for this sort of accident).

What the press gets wrong

Reading, listening, and watching various news reports has been really tough as they get critical details really wrong or espouse theories that have no basis in reality. Some examples

  • “He didn’t file a flight plan” – he was on an IFR flight plan and in constant communication with air traffic controller up until 5-10 minutes before the crash.
  • Describing the airplane’s turn away from Santa Monica as “being off his flight plan”. That turn was an instructed turn by air traffic controllers and completely normal.
  • Describing the conditions as “inclement weather” – while the clouds and fog would keep the pilot from seeing the mountains, the overall conditions were rather benign as instrument flying conditions go.
  • Trying to explain the airplane dropping off radar as possibly having been caused by catastrophic engine failure and “dropping out of the sky”. It’s pretty clear the aircraft flew right into the mountain at 4000′ at cruise speed. Even if the engine had failed, airplanes do not drop out of the sky – they glide.
  • Failing to make online corrections when facts are shown to be false. There was an early report (~9AM) that a Sheriff’s helicopter had spotted the airplane. That proved to be false, but was included in many later reports and it wasn’t until around 4:30PM that the wreckage was spotted.
  • Describing the last radar contact as coming at 8:37. Most reports seemed to be using FlightAware. While the last entry in FlightAware is 8:37, that appears to be the system just “assuming” that airplane must have landed and closing out the radar track [since changed by them to show “undetermined” as the outcome of the flight]. The last reliable radar information I find is more like 8:30.

Timeline of Events

My audio information comes from the publicly accessible archives at LiveATC.net These archives are only kept for limited time (often 30 days), so the links will cease to function at some point. Times are UTC (aka GMT or Zulu, denoted with a “Z”), Pacific Daylight Savings Time is 7 hours behind. I find the LAX WebTrak (a system put in place mostly for tracking of airplane noise and complaints) gives some of the best location information and shows  N133BW continuously from the area of LAX until radar contact is lost. I have not attempted to track the pilot for the whole flight, but rather as he is approaching the crucial part of the flight:

15:12:30Z (8:12 AM PDT) – the pilot checks in descending out of 4200 for 4000′ and is given the LAX altimeter setting, which he reads back correctly.

15:16:05Z (8:16 AM PDT) – the pilot is instructed to turn to 310 (northwest); he reads this back as 210; the controller corrects him, and the pilot reads back the corrected heading and complies.

15:19:30Z – the pilot is assigned heading 360, reads back and complies. At this point, he is almost directly over LAX and the controllers are starting to vector him for the SMO approach.

15:20:00Z – the pilot is assigned heading 030; he reads back and complies. The pilot is now over Westchester and has been turned roughly parallel to the Santa Monica VOR-A final approach course (but outbound).

15:20:54Z – the pilot is given a frequency change to 135.05, he reads the frequency back correctly.

This new frequency is the last frequency the pilot would use prior to being handed off to the Santa Monica tower controller: http://archive-server.liveatc.net/kbur/KBUR2-App-May-15-2016-1500Z.mp3

15:21:30Z – On the LiveATC.net feed, we can’t hear the pilot, but the controller acknowledges a “check in” and gives the pilot the local altimeter setting (standard procedure). This is the last point at which I have some confidence that pilot and controller are successfully communicating. It is possible that controller thinks he hears this pilot but is hearing another, but the FAA and NTSB will have access to the actual controller tapes to confirm. For the next 2.5 minutes, the pilot is flying parallel and outbound in order to get far enough away to descend safely on the instrument approach.

15:24:00Z – the controller assigns a turn to 290 (WNW) and a descent to 3000′; there is no response. About 15 seconds later, the controller repeats the instruction, again with no response. At this point, the airplane is over Hollywood. The turn would have taken the pilot just south of the Hollywood sign. The airplane is heading northeast at about 150kts (170 mph), covering almost 3 miles every minute.

15:24:50Z – the controller is very aware there is a problem. The pilot is now almost 1 minute late making his turn. “Skylane 3BW, SoCal Approach, Radio Check”

15:25:00Z – “3BW. How do you hear?”. I’ve heard several cases where controllers and pilots couldn’t hear each other, even though both were successfully transmitting, if transmissions are weak or there is terrain in the way. If I were a pilot in the air at this time and 3BW was responding, I would jump in and attempt to relay communications. The only fault I can find with the controller is that he never specifically asks nearby aircraft to try and reach 3BW or confirm if they can hear him. I think it unlikely that another pilot hears 3BW, but fails to let controllers know.

15:25:30Z – “N133BW SoCal Approach”.

15:25:40Z – “3BW, SoCal Approach. Low altitude alert. Check your altitude immediately. Minimum vectoring altitude in your area is 4,300, increasing to 6000.” The controller is now warning the pilot that he is crossing into an area of rising terrain. The MVA is the lowest altitude a controller would be allowed to assign an airplane in that area; that altitude would give sufficient clearance over local terrain (that a pilot can’t see in the clouds). He is approximately over Glendale at this point, heading northeast.

15:25:50Z – “3BW, SoCal Approach”. A different voice is now heard. This may be a different controller or supervisor trying on other frequencies to see if he can reach the pilot. The controllers clearly understand the seriousness of the situation and the need to get the pilot turned around.

15:26:00Z – “N133BW, Radio check. How do you hear?” I know the controllers sound repetitive, but there are specific communication phrases that if a pilot hears are meant to get their attention.

15:26:10Z – “…hear this transmission, ident”. The controllers know that the airplane’s transponder is still operational, providing altitude information. The controllers are attempting to determine if the pilot can hear them, but is unable to transmit voice over the radio. “Ident” means the pilot would push a button on their transponder. Pushing that button would allow the controllers to see on their radar that the button had been pushed. If that worked, controllers would know the pilot could hear them, but was unable to transmit voice. That level of two way communication would be enough to give instructions and ask yes/no questions. It apparently was also unsuccessful. At this point, the airplane is approximately over the 134/2 interchange.

15:26:45Z – there is an exchange that is not clear on LiveATC at this point “… been trying to get ahold of you. Turn left immediately and climb maintain 5.. 6000”. FAA tapes will help answer whether there was momentary response from the pilot. However, there is not a change in the airplane’s course or altitude to indicate that pilot heard or understood the controller’s instruction. At this point, the airplane is in the area of the Verdugo Hills and may be passing less than 1000 feet over terrain in that area – well below the altitude criteria for IFR.  The highest point in the Verdugos is charted as 3160′ MSL and the airplane is at 4000′ MSL.

15:26:55Z – “3BW, SoCal Approach. Low altitude alert. Check your altitude immediately. Climb and maintain 6000 immediately. and turn left. Left turn 210.”

15:27:10Z – “3BW, SoCal Approach. Radio check, how do you hear?”

15:27:20Z – “…SoCal approach, how do you hear?”

15:27:30Z – “3BW, if you hear this transmission, low altitude alert. Climb immediately to 6000 and turn left to heading 210 immediately. You’re going into a 7,000′ MVA, increasing to 7,700”. The airplane is now almost directly over JPL (NASA’s Jet Propulsion Laboratory) and heading towards 6000+ foot peaks in the San Gabriel mountains north of Altadena.

15:28:00Z – “3BW, SoCal on 124.6. How do you hear?”. 124.6 is not the frequency the pilot had been assigned. However 135.05 and 124.6 are two frequencies that are often handled by the same controller for the area generally south and east of Burbank.

15:28:15Z – “..33BW, SoCal, if you hear this transmission, ident, Turn left heading 210 immediately, climb and maintain 8000”

15:28:35Z – “3BW, SoCal Approach. Turn right. Right turn heading 180. Climb and maintain 8000 immediately.” At this point. while climbing will help, only a turn has a chance of keeping the airplane from impacting terrain. In the best of conditions, that 4000′ climb would take 3-4 minutes. The airplane is well into the foothills of the San Gabriel Mountains, in the area of Brown Mountain, a peak at nearly 4500′ MSL, well above the airplane’s altitude.

For the next several minutes, the controller makes many more attempts to contact the pilot and eventually asks several other aircraft in the area if they can hear an ELT. An Emergency Locator Transmitter is a piece of equipment designed to begin transmitting in the event of a crash. It transmits a series of tones on the frequency 121.5 MHz which can be received by airplanes tuning in 121.5. No pilots report hearing an ELT. After 15:29:30Z (8 minutes after the last apparent communication), WebTrak and FlightAware show no more radar information for the airplane and it is almost certain the airplane has crashed, destroying the ELT in the process. From what I know of timing, it is clear that controllers very quickly alert rescue personnel with a last known position and altitude; the stream of fire and sheriff vehicles I saw begins. With the treacherous terrain and low clouds and visibility in the area, a location of the wreckage will take well into the afternoon at which point news report a single fatality in the crash.

May 18 Update

News reports have now identified the pilot as a 57 year old doctor from the San Diego area. They note he had flown 20 hours in the previous month and had been a member of the plus one flying club for several years.

April 4, 2017 Update

By sheer coincidence, my wife texted me about a large helicopter that was repeatedly taking off and landing near our house in Altadena. I happened to be up flying on my own soon after that and flying over the area. I flew over my house and I noticed that the helicopter seemed to be working to recover the wreckage on Brown Mountain. IF you look carefully, you can see people standing near the wreckage.

As of today, the NTSB has still not published much information:

On May 15, 2016, about 0829 Pacific daylight time, a Cessna R182, N133BW, was destroyed when it impacted terrain during cruise flight near Altadena, California. The airplane was registered to San Diego Skylane LLC., and operated by the pilot under the provisions of Title 14 Code of Federal Regulations Part 91. The private pilot, sole occupant of the airplane, was fatally injured. Instrument meteorological conditions prevailed and an instrument flight rules (IFR) flight plan was filed for the personal flight. The cross-country flight originated from Montgomery Field, San Diego, California, at 0737, with an intended destination of the Santa Monica Airport (SMO), Santa Monica, California.

Preliminary information provided by the Federal Aviation Administration (FAA) indicated that the pilot was being vectored for an instrument approach to SMO. The pilot established radio communication with the controller and subsequently acknowledged obtaining weather information at the destination airport. About 2 minutes, 26 seconds later, the controller issued the pilot a heading change to 290 degrees and a descent clearance to 3,000 feet for vectors to final approach. However, the controller received no response from the pilot despite multiple attempts over the course of about 2 minutes. The pilot then transmitted that he was on a 030 degree heading. The controller continued to issue vectors away from rising terrain and made several attempts to communicate with the pilot; however, no further radio communication from the pilot were heard. Radar contact with the airplane was subsequently lost and an alert notice (ALNOT) was issued by the FAA. The wreckage was located later that evening by a Los Angeles County Sheriff Office air unit. The wreckage was located within mountainous terrain near Brown Mountain, about 2 miles north, northwest of Altadena. Law enforcement personnel and initial responders reported that the airplane was mostly consumed by a post impact fire. Recovery of the wreckage is currently pending.


May 2018 Update

A comment elsewhere alerted me to the fact that the NTSB has published their “Probable Cause” analysis on this accident. It’s coincidental that this comes exactly 2 years after the accident (though it was actually published last month). A probable cause is the National Transportation Safety Board’s attempt to explain what happened. If you read through the probable cause, as well as the Full Narrative, you get an idea of the amount of effort that goes into analyzing some accidents. Some will find this information disturbing – it includes rather graphic descriptions of the after effects of impact at 150 kts. This still just scratches the surface of the information that was created and examined over the last two years. If you want more details, you can look at the Full Docket (I find that links to the docket don’t always work, so you can go to the Docket Search and enter “WPR16FA111” for the NTSB accident ID and search). There are currently 17 different files stored there – many of these are not for the faint of heart.

So, what else have they learned? The reality is not a lot more. The analysis does confirm there was a single transmission from the pilot while the controllers were frantically trying to contact him and get him turned (at 15:26:45Z in my transcript). The pilot never acknowledged any instructions nor indicated he could hear controllers. The NTSB seems to believe the pilot was not incapacitated, as a result of that transmission and everything sounding normal previously. There is no indication of any mechanical failures, so they are left with “The pilot’s failure to maintain clearance from rising mountainous terrain while flying in instrument meteorological conditions. Contributing to the accident was the loss of radio communications for a reason that could not be determined because of the extensive impact and thermal damage to the airplane.”… he flew a plane into the side of the mountain he couldn’t see at high speed because he didn’t hear/do what the controllers told him to do. The NTSB does indicate some concerning medical issues – serious heart disease, as well as indications of use of prescription pain killers, but stops short of blaming those.

We are left where we were before – either an incapacitated (maybe partially?) pilot or a pilot who lost situational awareness and failed to implement lost communication actions. In either case, a very sad outcome. RIP.

Fly safe.

23 thoughts on “Terrain and IFR

  1. Very nice analysis of the crash. Very worthy of the “Aftermath” columns in Flying Magazine. Thank you

  2. David

    Thank you for your thoughtful article. I was on the ground at Palm Springs looking to fly my 182 into SMO that day, and decided to stay put as it wasn’t a day for VFR flight, so had a particular interest in this accident.

    It seems that the pilot was either incapacitated, distracted by a loss of communications, or dealing with some other emergency such as an electrical fire? In the latter two scenarios, had he had situational awareness, he would have known to make a turn, while he dealt with whatever was going on. I wonder if his aircraft was equipped with GPS?

    • I have read some other comments that indicate it had a Garmin 530 (though not specific to whether it was a 530W – that does matter for terrain).

    • There’s a post on reddit that has a picture of the panel, confirming a Garmin 530, Aspen PFD, and STEC 55X autopilot

    • I fear you’re right. I guess it is possible there is still some piece of information that we don’t yet know that will shed some light on it. I have a hard time imagining them getting much from the wreckage.

  3. This aircraft had an Aspen 1000, Garmin 530, and a very loud terrain warning system. I just flew this approach in my Grumman last week. I was happy to be above the marine layer until the final intercept turn. This pilot would have been in the soup for 15 minutes at least, once he descended below 4500′. My impression of his voice during the transmissions I have heard, is that he sounds “behind the plane”, and not quite sharp and anticipating what’s about to happen during the approach procedure. I believe he becomes distracted by a loss of comm’s, might be troubleshooting the comm panel, his headset, the frequency he should be on or something that takes his attention away from the approach. At that speed, things unwind fast, and he becomes confused as to what his lost comm procedure should be in the busy airspace. Indecision at 3 miles a minute.

  4. Do you know what the link is to the Live ATC file that preceeds the handoff to his final approach controller?

  5. Great point about the media. I have extensive experience where they get the facts wrong either by intentionally lying or just making assumptions to “fill in the blanks.” There’s not much of a difference in my opinion.

    Now, without rushing to judgement, just knowing the statistics of accidents involving CFIT we can probably all agree this will likely be a result of pilot error. If that turns out to be the case it is very sad that every brand name hanging on or in that airplane will be invariably sued by the family of the pilot, and GA will continue to get more expensive mostly due to product liability.

    • Indeed – these almost always end up in litigation, even in cases where it is very clearly bad pilot decisions.

  6. I’ve listened to the transmissions he was making with the LAX controller, and I’ll retract my previous statement about him “sounding behind the plane” Hearing the sequence of transmissions, it actually sounds like he’s on top of it..responding quickly and accurately, except for one incorrect heading readback. Interestingly, the previous controller calls for him, 6 minutes after passing him onto the final approach controller. ATC was trying everything to reach him.

  7. Regarding your comment that at 15:21:30Z the controller is acknowledging a check-in. This is not necessarily true. Had the controller not received a check-in from the pilot, the controller would likely have issued the instructions we hear, assuming that the pilot was on the frequency and either hadn’t checked in yet, or that the controller had missed the check-in. Without hearing the check-in, it remains possible that the pilot dialed in the wrong freq (despite having read back the correct one) or that the pilot’s radio was inop on the new freq, and that in either case, he didn’t recognize the problem in time to take corrective action.

    In addition, there is a third possibility: that the pilot became incapacitated just after acknowledging the freq change. An article in the Pasadena Star-News states the following:

    “An autopsy determined [the pilot] died from ‘multiple blunt traumatic injuries,’ and the death was determined to be accidental, according to coroner’s records. Atherosclerotic cardiovascular cisease [sic] was listed as a significant factor in the death.”

    I’m not sure how to interpret that last sentence, but perhaps it suggests that the pilot may have suffered a heart attack while being vectored for the approach.

    • Thanks for the note on autopsy. Pasadena-Star News added that, but didn’t change the publication/update date on the article. That would add some credence to the incapacitation theory.
      I agree I can’t confirm a check-in, but if the controller hadn’t heard a check in, but thought he was on frequency, the phrasing is usually different. I’d expect a “did you check in” or altimeter setting + say altitude, but that should be easy for the FAA/NTSB to sort out.

Leave a Reply

Your email address will not be published. Required fields are marked *