Windshear – Lessons Learned
The National Transportation Safety Board cited the lack of specific training, policies, and procedures for avoiding and escaping low-altitude windshear, as well as the lack of onboard microburst detection equipment as major causes. The radar equipment onboard most airliners at the time was capable of detecting rain showers and thunderstorms, but it was unable to detect windshear.
Today, windshear and microbursts are a part of every airline pilot’s initial and recurrent training. Also, ground based warning systems are installed at major airports around the world and even the aircraft themselves are capable of predicting windshear and microburst events before they are encountered.
Fast forward 31 years…
Nearing the top of descent, I was the pilot flying as we neared the end of a 5-hour flight from Panama City, Panama to Dallas, Texas. The sun had long since dipped below the horizon and the night was dark. With an undercast below and no moon above, the only visible light outside the cockpit came from the glow of city lights from beneath the clouds and an occasional pop of lightning from storms that were well past our destination, but that appeared much closer. The Captain ripped a weather report off the ACARS printer and read the important details aloud. Everything on the report was basically in-line with the forecasted weather conditions except for the wind, which was stronger than expected, blowing from 310 degrees at 12 knots gusting to 26 knots.
Runways 35R, 35C, and 36L were in use for arriving aircraft. The runway heading for all three runways is 356 degrees, which, taking the gusts into account, translated to approximately a 20 knot crosswind. The airport has 2 additional runways that were almost directly aligned with the wind, but, for various reasons, those runways are rarely used except when crosswinds are out of limits for landing and departing aircraft. That certainly wasn’t the case on this flight, but it wasn’t the 20 knot crosswind that raised the red flag, it was the 14 knot gust. The maximum demonstrated crosswind of our Boeing 737-800 is 33 knots, well below the 20 knots reported at the airport, but the 14 knot gust foreshadowed a turbulent approach.
I gave my standard approach briefing to the Captain and set up for a visual approach with a flaps 30 landing. Coming in from the southeast, we would most likely be assigned Runway 35C; but, our gate was on the west side of the airport and I was feeling optimistic, so we set up for Runway 36L. I wasn’t overly concerned about the wind. It had been a long day and I had assumed my late night arrival would end with an easy, smooth air approach and landing, but that was really the extent of my concern. Even so, the turbulent conditions ahead must have been on my mind as I ended my brief by saying, “If we don’t like what we see, we’ll go-around and reevaluate.” “Sounds good to me,” said the Captain.
I turned the radar on as we passed through 20,000 feet. It was clear that we would enter the cloud tops soon and I have learned (the hard way, I might add) that you should always have the radar on when you don’t have a clear view of what lies ahead. There weren’t any returns indicated on the radar, so we pressed on and got a much smoother ride through the clouds than I expected. As we descended through the bases of the clouds at approximately 10,000 feet, I could tell I wasn’t going to need the radar and turned it off.
After descending another couple thousand feet, I told the Captain that I had the runway in sight and he relayed this information to the approach controller. We were vectored onto the localizer for Runway 36L, informed of traffic ahead, and cleared for a visual approach. The ride had deteriorated at this point, and as I intercepted the glideslope, the airspeed was fluctuating by 8 to 10 knots. The air was quite a bit rougher than I expected it would be and the airplane was bouncing around enough that I had trouble focusing on the instruments. I did my best to bracket the airspeed between too fast and too slow.
When instructed to do so, the Captain switched over to the tower frequency just in time to hear the aircraft ahead, a Bombardier CRJ700, announce that he was going around. No explanation…just going around. I was a little task saturated at that moment, and considering the fact that I may have missed something, I voiced the thought as it entered my mind: “They’re going around? Did they say why?” Maybe the pilot was uncomfortable with the crosswind. Maybe he was unstable on the approach. Could it be a mechanical problem? That was wishful thinking. Maybe it was windshear. This situation was beginning to feel like a “what would you do now” question in an airline interview.
What would you do now?
As I write this, I remember the 1999 job interview hosted by my current employer. They had me in a tiny room, just big enough for three chairs and a small table. I sat across from two pilots, a Captain and a First Officer, both line pilots at the airline. I was surprisingly comfortable in my grey suit and red tie, and I wasn’t particularly surprised by any of the questions they had to ask. I was well prepared and specifically remember thinking that the interview was going well. It would have been much easier to explain the correct course of action for my current situation while sitting in that interview than it was to process and act on the same information in the heat of battle. But that’s exactly why windshear escape maneuvers are an important part of every recurrent training cycle.
Thanks in part to the lessons learned from Delta 191, my 737 is equipped with a windshear detection system. Activation of this system is automatic and invisible to me as the pilot. Even with the weather radar system turned off, windshear detection is automatically activated as I descend through 2,300 feet above the touchdown zone. Additional WARNINGS and CAUTIONS are enabled as I passed through 1,200 feet. The system scanned the airspace ahead and would issue a windshear WARNING if a windshear event was detected within .25 NM of the longitudinal axis of the aircraft and within 30 degrees of the aircraft heading. A windshear CAUTION would be generated if a windshear event was detected within 30 degrees of the aircraft heading and less than 3 NM from the aircraft.
Automation in the cockpit is a fantastic thing that I’ve learned to trust and appreciate, but there’s also some old fashioned pilotage at work. If the windshear detection system sounded an alarm, we would certainly execute a missed approach. However, in addition to an aircraft alerted windshear, there is a specific list of uncontrolled changes from normal, steady state flight conditions that would necessitate a go-around.
- 15 knot change in indicated airspeed
- 500 fpm change in vertical speed
- 5 degrees change in pitch attitude
- 1 dot displacement from the glide slope
- Unusual thrust lever positions for a significant period of time
These metrics are so ingrained in my thought process that it wasn’t necessary for me to list them, one by one, in my head as I flew the approach. The air was turbulent and the airplane was bouncing around enough that we were close to meeting several, if not all, of those conditions from the time we intercepted the glide slope.
It bears mentioning that rough approaches aren’t as uncommon as you might think. Large airspeed fluctuations and rough air is common, especially at certain times of the year, at certain airports, and when flying in and around certain weather patterns. The ride was uncomfortable and I was working hard to keep the airplane under control; but, other than the unexplained go-around ahead, I still didn’t see a valid excuse to abandon the approach. Flashing back to my 1999 interview one more time, my answer hadn’t changed: I’d continue the approach.
At about 500 feet above the ground, the aircraft sounded the alert. “Go around! Windshear ahead!” I called out the standard litany I’d practiced so many times in the simulator, “Go around, flaps 15, positive rate, gear up.”
We didn’t actually fly through the windshear. That’s an important distinction because it determined how I executed the go-around. If we had flown through an actual windshear, standard procedure would have dictated a few very important differences from a normal go-around. These differences are designed to emphasize the severity of the situation and maximize aircraft performance. First of all, I would have announced “escape” instead of “go-around” so that we were all on the same page regarding the procedure that we were performing. Then, I would have simultaneously disconnected the autopilot and autothrottle, pushed the TO/GA switch, and “aggressively” applied full thrust. Also, I would have rolled the wings level to get as much lift as possible working the correct direction and I would have pulled the nose of the aircraft up to an initial pitch attitude of 15 degrees. Changing the aircraft’s configuration (For example, raising the landing gear or retracting the flaps) is prohibited until the windshear event is over, but it is vitally important to ensure that the speed brakes are retracted.
Our go-around, and the one from the CRJ ahead of us on the approach, were the only two that occurred. There was a Boeing 737 on the approach behind us that elected to continue and landed without incident. By the time we came around for another attempt at the approach, 5 other aircraft had successfully flown the approach. It was just as bumpy the second time around, but I managed to squeak out an especially smooth landing. I’ve always believed that there’s a little bit of luck involved with any good landing, especially when the conditions are poor. Besides, as the old saying goes, I’d rather be lucky than good.
As we taxied off the runway, I transferred control to the Captain and began my after landing duties. I noticed my hands were shaking a little, as they do sometimes when my adrenaline really gets pumping. The Captain asked me if I had noticed the windshear depicted on the radar screen at the time of the windshear warning. I had not. As I mentioned before, even with the radar system turned off, the windshear alerting system had begun to scan the airspace in front of the aircraft automatically. It’s something the system does on every takeoff and landing regardless of the anticipated conditions.
Even with the weather radar turned off, the windshear system automatically activated and displayed the location of the windshear. Apparently, when it popped up on the display, the windshear was slightly ahead and to the left of our track. We flew right past and probably a little above it as we executed the go-around. I had my hands full of airplane as I was struggling to see and make sense of fluctuations in airspeed, altitude, and attitude. This was a procedure I rarely perform outside of the simulator-training environment. I don’t believe my course of action would have been different if I had seen the windshear on the display, but this was definitely one of those times that justifies having two pilots in the cockpit. I was task saturated and missed something potentially important. Fortunately, the Captain saw the windshear and would have shared that information if he had deemed it relevant
The event we avoided that night was, almost certainly, nowhere near as severe as the conditions encountered by the pilots of Delta 191. But it is a reminder that many of the rules, regulations, procedures, and safety equipment that we rely on today have often come about as a direct result of the lessons we learn from past mistakes and lives lost.
Special thanks to my friend and editor, Ryan @theAviationGeek