Rocky Mountain Low

A cold winter morning in Colorado defeats a Challenger

October 1, 2006

I was the new guy flying for a charter company based out of Teterboro, New Jersey. I had just 70.7 hours of jet time, and the ink was barely dry on my Challenger type rating. If you don't know the Bombardier Challenger personally, it's the business jet that spawned the popular Canadair Regional Jet. It was a perfect winter's day, and we were flying the Ebersol family from Telluride, Colorado, to Oxford, Connecticut. It was January 2, 1999.

That's what I was thinking about nearly six years later as I watched one of the Challengers that I used to fly burning in a Colorado field on television. You may remember the accident, as the chairman of NBC Sports, Dick Ebersol, and two of his children were on board and his wife, actress Susan St. James, had just gotten off the airplane.

Ebersol's 14-year-old son Teddy was killed in the crash that day, as well as Capt. Luis Alberto Polanco-Espiallat, 50, and flight attendant Warren Richardson III, 36. First Officer Eric Wicksell, 30, was severely injured. Ebersol and his 21-year-old son Charlie, who pulled his dad from the flaming wreckage, also were injured. It was November 28, 2004.

It was the same family I had flown, the same company I used to fly for, and one of the same airplanes, N873G, that I used to fly. I can't tell you the feeling of knowing that lives I had flown were lost in an airplane I had flown. And I couldn't help thinking that it could have just as easily been me in that field fighting for my life.

Probable cause

The facts are cold and unforgiving. The NTSB concluded that the probable cause of the accident was the flight crew's failure to ensure that the airplane's wings were free of ice and snow contamination. This failure resulted in an attempted takeoff with upper wing contamination, which induced the subsequent stall and crash. The NTSB also noted the crew's lack of winter flying experience. The captain had more than 12,000 hours of flight time but had rarely flown in snow and ice. Both the captain and the first officer were relatively new to the company.

Given those findings, why did they go, and why didn't they deice?

The day of the crash

Early that morning, N873G left Van Nuys, California, with five passengers for Montrose, Colorado. The airplane landed at Montrose at approximately 9:09 a.m. mountain standard time to drop off two passengers before departing for South Bend, Indiana. Montrose, at 5,759 feet msl, is on the western side of the Rocky Mountains. It's a big airport for such a small town, with two runways — 13/31 (7,500 feet by 100 feet), and 17/35 (10,000 feet by 150 feet) — and significant airline service.

The Montrose AWOS reported at 9:53 a.m. MST calm winds, 1-mile visibility in light snow and mist, few clouds at 500 feet, overcast at 900 feet, minus-1 degrees Celsius temperature and minus-2 degrees C dew point, and an altimeter setting of 29.67. AOPA's Airport Directory recommends Montrose as a good alternative to other less-friendly airports in the region, so the conditions probably weren't any better anywhere else. In fact, the Ebersol family often lands at Telluride but went through Montrose this time because of weather. While the weather always deserves respect — and might ground a less-equipped airplane — there was nothing particularly scary for a Challenger.

Performance

Discussion recorded on the cockpit voice recorder (CVR) indicated the crew established a takeoff weight of 42,000 pounds, resulting in a perfect-day takeoff distance of 7,590 feet. But it was not a perfect day. With the temperature below freezing (minus-1 degree C) and visible moisture present (light snow and mist), engine anti-ice would normally be turned on for taxi and wing anti-ice for takeoff. Anti-ice systems bleed power from the engines, increasing dry-runway takeoff distance in this case to 8,804 feet.

On top of that, there was slush on the runway, and the ramp surface looked grim. For performance purposes, slush is defined as a mixture of snow and water covering greater than 25 percent of the runway surface. With up to one-quarter of an inch of slush on the runway and engine and wing anti-ice on, the takeoff distance at 42,000 pounds in that airplane becomes 12,121 feet, far exceeding the length of any runway at Montrose, and all but a few in North America.

Did the crew intentionally miscalculate takeoff numbers, or was it just ignorance? It's hard to tell. I hadn't seen a Challenger manual in more than six years, and it only took me a few minutes to figure out the takeoff numbers — it was pretty straightforward. In this case, according to the CVR, there appeared to be serious confusion over what the actual takeoff weight was. The crew discussed 40,000 and 41,000 pounds before settling on 42,000 pounds as their takeoff weight.

This obviously also contributed to confusion in computing takeoff distance. The crew discussed seven different takeoff distances: 8,000 feet, 7,800 feet, 6,000 feet-something, 7,590 feet, 7,800 feet, 6,875 feet, and 7,000 feet before settling on 7,200 feet. The only number among those that corresponds to their situation was 7,590 feet — the dry-runway takeoff distance at 6,000-feet elevation for a 42,000-pound Challenger — which didn't apply in those conditions. According to the CVR, it appears the crew didn't use wing anti-ice or take the 11-percent takeoff distance penalty for using it. They might or might not have added the 5-percent penalty for engine anti-ice, which they did appear to use.

As for runway contamination, or associated takeoff distance penalties — it appears the crew didn't even discuss it. The CVR makes it sound as if the crew didn't do much of a weight-and-balance or performance calculation either, and comments from the CVR such as, "These numbers are always conservative anyway," while discussing takeoff performance make one wonder.

How did they get this fat?

Going in and out of the Rockies, performance is an issue, so, why weigh any more than necessary? A crew of three, three passengers, luggage and 10,000 pounds of gas to get to South Bend with an alternate and reserves would have put them around 37,500 pounds for takeoff. If more performance is needed, trade fuel for payload and refuel somewhere en route. There is a big difference between 37,500 and 42,000 pounds, and the former would have shortened takeoff distance from 12,121 feet to 9,167 feet. Cargo could only account for a small difference in the weight, so the majority of the difference had to be fuel. Why did they take 14,500 pounds, when 10,000 probably would have been sufficient? Were they tankering fuel? That's not unusual in good weather, but given the marginal conditions, why jeopardize the ability to get out of Montrose just for cheap gas?

But, wait, if they were 42,000 pounds at Montrose and only added 400 gallons of Jet-A there, that means they landed at Montrose just above 39,000 pounds — which is well above the Challenger's maximum landing weight of 36,000. Beyond the obvious violation, this doesn't provide much flexibility for getting out of Montrose, as they were already too heavy to legally get out when they landed.

To deice or not to deice

The focus of the NTSB investigation was the apparent decision by the flight crew not to deice. Why? What reasons would there be not to deice? Was it as simple as a lack of experience in ice and snow, or was it something more?

First, it was supposed to be a drop-and-go at Montrose. The implication being that if you quick turn fast enough and the weather's good enough, you don't need to deice. Unfortunately, the Challenger was on the ground for about 45 minutes, which is probably longer than intended.

Second, is there cost pressure not to deice? Although deicing can be a significant cost, it's normally passed on directly to the customer, which usually removes any economic pressure from the charter operator not to deice. I can't ever recall a client saying "Don't deice, it costs too much."

Third, is there time pressure not to deice? Is the airplane on its way to another charter? Possibly. Are the clients in a hurry? Possibly. Although there are some clients who just can't wait to get going, I don't remember the Ebersols that way — they were simply nice, polite people wanting to safely get where they were going.

So, how does one ultimately decide to deice or not? Go outside and look at the environment, i.e., the weather, the ramp, and most important, the airplane. FBO line personnel said the weather that day was miserable, a moderate, wet snow was falling, and they could only see halfway across the field. Strike 1.

Several witnesses at Montrose observed ice on the wings, but no one saw the crew physically examine the airplane, except for the slush around the nosewheel. That is really hard to believe. Unlike other jets where you need a ladder, the Challenger wing is easy to check by just walking up to it, easier than many general aviation airplanes because it's so close to the ground. It's like looking at a Piper Cherokee.

There was strong potential for icing while the airplane was on the ground in precipitation and freezing temperatures. Add to that the fact that the airplane may have picked up ice on the approach to Montrose and it's hard to justify not deicing. Strike 2.

FAR 135.227 states that a pretakeoff contamination check to ensure the wings and control surfaces are free of frost, ice, and snow must be completed within five minutes before takeoff any time an aircraft is in, or suspects, icing conditions. Given the icing conditions present, the crew should have gone to a passenger window or outside to look at the wings. The CVR indicated that the crew did look at the wings from the cockpit and determined they were clear a little more than 16 minutes before takeoff. Even if the wings were clear then, that's still a violation of FAR 135.227, because it wasn't within five minutes of takeoff. In addition, the NTSB report said Ebersol noticed chunks of slush sliding off the cabin roof and across his window during taxi. This is not what one would expect to hear when describing a contamination-free airplane. Strike 3.

Runway selection

The crew originally planned Runway 35 for departure, but they discovered it was being plowed and went to the shorter Runway 31 instead. In fact, both runways were being plowed, so either way, they had to wait for plows to vacate. This also indicates that the runways were still at least partially contaminated, and it is consistent with the NTSB report, which said the passengers estimated 1 to 2 inches of slush on the ground. Bombardier does not recommend taking off with more than one-quarter of an inch of slush on the runway, and this appears to be far worse than that. Bombardier also cautions that the sudden reduction in drag upon rotating off a contaminated runway can cause an overrotation, which if not corrected in a timely manner could lead to a stall.

If they had departed on the long runway would the accident have been avoided? Would the crew have recognized a problem and aborted while they still had the runway remaining to do it? If they had decided to wait for Runway 35 to be plowed, would that have been enough to tip the balance in favor of heading back to the ramp and deicing? As in, "I was willing to go right now without deicing, but I'm not gonna wait and pick up more snow and ice." Perhaps they just didn't recognize how much contamination was on the airplane and on the ground, and that it was getting worse.

Let's review

So far, we know it's possible the Challenger landed overweight in Montrose from tankering fuel. The crew seemed unsure about how much the aircraft weighed and how much runway it needed for takeoff. If the crew had calculated their takeoff performance correctly, the runway required would have increased by more than 50 percent — from 7,590 feet to 12,121 feet.

This meant the Challenger couldn't legally or safely depart from any runway at Montrose at the time they departed. Runway 31, which they used, was far too short at 7,500 feet long. And at 42,000 pounds they couldn't depart from the 10,000-foot-long Runway 35 unless it was plowed and free of contamination — allowing them to use the dry runway takeoff distance of 8,804 feet.

We know the crew didn't examine most of the aircraft surfaces, including the wings, for ice. We know the crew didn't deice, even though they should have. We know they didn't do the pretakeoff contamination check within the timeline required by FAR 135.227.

The result

At 9:54 a.m. MST, N873G contacted Denver Center, ready for takeoff. Passengers reported they reached 20 to 50 feet above the ground when the left wing dropped, then the right wing, and finally the left wing before it struck the ground. The airplane slid approximately 1,400 feet through the airport fence and across a road before stopping next to a dairy farm.

The NTSB expressed frustration that repeated warnings during the past 15 years that even trace amounts of ice on wings can bring down aircraft have largely been ignored. The board recommended better training to identify when ice is present on aircraft surfaces. The crew did not demonstrate a great understanding of the effects of contamination on either the airplane or the runway and missed several winter performance issues.

The NTSB found no significant safety problems at the charter company and said that its training is above what is required. Notwithstanding, the first officer, Eric Wicksell, has filed a lawsuit alleging he didn't get adequate training for the flying conditions that day — even though his training met FAR Part 135 requirements.

Because of several crew resource management deficiencies, the NTSB also recommended CRM training programs for Part 135 operators (mostly charter), similar to what is already required for Part 121 operators (airlines).

What is there to be learned?

There's always cause and effect, physics, and the NTSB to help us figure out what happened and perhaps even rationalize such a tragedy.

At the time I flew the Ebersols I was just as green, if not greener than the first officer on that flight. I was lucky; even though it was the same airplane and roughly the same winter trip, it was a great day and I was flying with Dave, one of the best pilots I've ever flown with. However, reviewing the conversations on the Challenger's CVR, I can see how easily it could have all gone wrong.

What have I learned since then? Much of what I am as a pilot. After a while, we all develop a sense of what we should or shouldn't be doing in an airplane. In general, we call it experience. It doesn't mean we still don't make mistakes or use bad judgment, just that our propensity to do it goes down. When we're just starting out, we often depend on others — friends, flight instructors, other pilots, and crewmembers — to look out for us. We trust them to have the knowledge to keep us out of trouble while we figure things out for ourselves. For the most part, it's a culture that works. But every once in a while it doesn't, and we find ourselves in a field somewhere wondering what went wrong.


Marc K. Henegar, of Bend, Oregon, flies a Boeing 737.


Links to additional information about the Challenger accident may be found on AOPA Online.


ASF Online Safety Center

Accident-avoidance help is free on the AOPA Air Safety Foundation Online Safety Center.

  • Mountain Flying is an interactive online course that teaches pilots about the challenges of high density altitude operations, flight planning, performance considerations, and mountain weather.
  • Cold Facts is a short publication that discusses preflighting steps to avoid wing contamination of snow, frost, and ice.
  • Runway Safety is an interactive online course designed to help pilots avoid and prevent runway incursions by studying the various factors involved.

My Flight Time in N873G

I had a lot of memorable moments in the Challenger, and specifically N873G. Even though I was far from home, the people I worked with made it a great experience. It was my first jet job, and I was like a sponge, soaking up the details that formed the basis for the commercial pilot I would become.

The Challenger was my first type rating, and I loved the airplane — and still do. It's incredibly comfortable, with one of the widest business jet cabins anywhere — wider even than the Gulfstream. It performs well, even though it will float for an hour on landing if you're anything above your approach speed — I've watched many a black runway-remaining sign whiz by waiting for the wheels to touch down. It also has a nasty tendency to pop a wheelie if you throw in too much reverse thrust when you're at light weights.

As I recall, N873G was one of three former GE corporate Challengers, N606BA, N875G, and N873G, which the charter company I worked for purchased from Bombardier in early 1999. The company took delivery of N873G a few days after my trip to Telluride to pick up the Ebersol family, and my logbook records my first flight on January 30, 1999. I had a total of 179.6 hours in N873G during the year I was with that company. Working on this article, I went through my logbook and found a lot of memories, many of which I've written about in past pages of AOPA Pilot.

In fact, the opening of my very first story for AOPA Pilot (" Reverse CRM," March 2002 Pilot) was about a jet headed for Teterboro, New Jersey, that got a last-minute hold sprung on the crew that they had a little trouble figuring out. The two pilots and the airplane weren't identified in my original article. Well, the airplane was N873G, it was August 14, 1999, and I bet you can guess who was part of the crew. I learned not to just say "Roger" and assume I could figure things out later.

I almost got hit in N873G at Belmar, New Jersey's Monmouth Executive one night by a radio-silent Cessna 172 doing an opposite-direction departure. It's a little unnerving to see runway lights, VASI, and then a huge landing light in your face on short final. All landing lights look huge when they're in your face — whether it's a 172 or a 747. I learned that just because the CTAF is silent, it doesn't mean there's nothing going on.

My worst duty day ever was in N873G. We left Teterboro at 3 p.m. and went to San Diego, San Francisco, Las Vegas, and Pontiac, Michigan (near Detroit) before finally landing in Boston just after daybreak the next day. We ended up far beyond maximum duty and flight times, which was never part of the plan. We had a well-meaning dispatcher who had overcommitted us, and we did him a favor. I can still remember making that approach into the sun on Runway 9 at Boston using two-by-fours to keep my eyes open. Dumb, dumb, dumb. Not doing that again. Like the drug commercials, I learned to "just say no" when required.

And finally, during the very last trip I flew in N873G I lined up to do a visual approach to Runway 31 at Hailey, Idaho's Friedman Memorial Airport at night. I learned the valuable lesson that you need to be really sure you are looking at the correct runway, and by extension, the actual airport, before you accept a visual approach. I saw a rotating beacon and a bunch of lights in a straight line, and that was enough for me. Cleared for the visual, contact advisory, have a nice night. It was only as we got closer that I noticed the runway seemed to be moving. It occurred to me that we would have much better success if we didn't land on Idaho Highway 75, the road we were lined up on, and tried the runway instead, which ran right next to said highway (see " Hard Knocks — There's No Reasoning With Terrain," November 2004 Pilot). I learned to be a little more careful before calling the airport and runway in sight in unfamiliar territory.

The trip to Sun Valley was my last flight in the airplane. However, I always assumed it would be my last flight in N873G because I left the company, not because the airplane was destroyed. — MKH