Probable Cause #29: Freezing Up On Approach
Ice claims another victim as a twin Commander crashes miles short of the runway. But could the accident have been avoided?
The Northeast winter weather was normal for February when the pilot of a Rockwell 500S Shrike Commander departed Naples, Fla., for Newport State Airport (KUUU) in Rhode Island. He had flight planned a stop in Wilmington, N.C., and the first leg of the flight went without incident.Before leaving Naples, the Commander pilot received a thorough telephone weather briefing that covered both legs of the flight from a specialist at the St. Petersburg Automated Flight Service Station (AFSS). He was told to expect moderate rime and mixed icing from the freezing level to 12,000 feet over Connecticut and Rhode Island.The briefer and the pilot had a short discussion about the temperature at Newport. At the time of the briefing, it was 3 C and the pilot made the remark that the freezing level would probably "go up." The briefer agreed saying, "Yeah, it's gonna go up. [But it] probably won't go up much more than five or six thousand, so you're still gonna run into it."During the refueling stop in Wilmington, the Commander pilot contacted the Raleigh AFSS and updated his earlier briefing. He was told that he could expect to see some rain and drizzle in the Hampton, N.Y., area and that it would be moving into the Providence, R.I., area before he arrived. At that time, Providence was reporting 1,600-foot ceilings with light drizzle. The pilot asked if the briefer had any idea what the cloud tops were in that area, but the briefer didn't know.Just before filing his flight plan, the pilot asked about ice. It was in reference to the 9,000-foot level that the briefer had suggested for winds aloft. "I won't get any ice at that level, will I?" the pilot asked. The response was an ambiguous "Yep" and the briefer seemed busy concentrating on NOTAMs. Apparently not sure about the answer, the pilot asked again, "Will I?" This time, the briefer didn't answer, but instead read the NOTAMs. The pilot did not ask the question again and just filed his flight plan.The second leg went smoothly and at 5:11 p.m. local time, when flying over the eastern end of Long Island, the pilot made contact with Providence approach control. The controller told the pilot to proceed directly to KUUU and to begin a descent to 3,000 feet. Asked what kind of approach he would like, the pilot requested the localizer approach to Runway 22 with a circle-to-land to Runway 34.At 5:30 p.m., the controller issued a traffic advisory to the pilot, warning of a Cessna 402 at 10 o'clock and nine miles passing from left to right. The pilot responded, "I'll never see it, sir. My head's so stuck in these clouds."A minute later the controller noticed the airplane was flying a course more easterly than he desired. He asked the pilot to verify his heading. The response was, "I am now gonna move to 050."At 5:35 p.m., the controller instructed the pilot to turn to a 120-degree heading. Shortly after that the controller said, "You're four miles from SMARI [the approach's final approach fix], continue the right turn heading 020. Maintain 2,000 till established on the localizer, cleared localizer Runway 22 approach, Newport. Report established." The pilot acknowledged the clearance.
"I'm All Iced Up"
A minute later the controller told the pilot that he appeared to be south of the localizer and asked if he was established on it. The pilot responded, "Negative, sir. I show myself (unintelligible). [It] appears [as] if I've gone through."The controller cancelled the Commander's clearance for the approach at Newport and instructed the pilot to fly a 090 heading and climb to 2,000 feet. His intention was to vector the airplane back to the northeast to begin another approach to KUUU. But soon after issuing the vector, he watched the aircraft's Mode C report change rapidly from 2,000 feet to 1,100 feet. The controller asked about the aircraft's altitude and the pilot responded, "Hold on. I'm having all sorts of problems here." Then, seconds later he added, "Alright. Seventeen-hundred on my way to 2,000."The controller told the pilot to report established on an eastbound heading, but the pilot replied, "Give me a break for a minute." Probably sensing that things weren't going well for the pilot, the controller informed him that Providence (KPVD) and New Bedford (KEWB), Mass., were reporting minimum VFR. The ceiling at KPVD was overcast at 1,200 feet while KEWB was reporting a 1,100-foot overcast. Taking the controller up on his offer, the pilot asked for vectors to Providence.The controller issued a turn to KPVD and told the pilot to climb to 3,000 feet. He provided the complete weather sequence at Providence and told the pilot to expect the ILS Runway 5R approach. Aware of the altitude deviations he had seen earlier, the controller also provided vectors requiring turns of less than 30 degrees. Radar data retrieved after the accident indicates that the airplane's speed and altitude remained constant during these vectors.As the pilot neared KPVD, the controller issued a descent to 2,000 feet and told the pilot there had been wind shear of minus 10 kts reported at 500 feet on the approach into Providence. The pilot responded normally and for the moment it seemed like everything was under control.At 5:48 p.m., the controller said, "Aero Commander 99N, six miles from RENCH [the final approach fix], turn right heading 020, 2,000 until established on the localizer, cleared ILS Runway 5R approach, Providence. Report established." He then called the Providence tower on the landline and advised the controller to "keep an eye on [the aircraft]. He diverted from Newport and gave us a bit of a scare."A minute later the pilot contacted the controller to say he had the localizer. When the controller asked the pilot if he was established, the pilot replied, "I sure hope so." At that point, the controller observed the airplane descending below the glide slope intercept altitude. He instructed the pilot to climb and maintain 2,000 feet until intercepting the glide slope, adding that he was still outside the outer marker, when the pilot came back and said, "Son of a [expletive], I got problems." The controller asked the nature of the problems to which the pilot responded, "I'm all over the place, I have no idea I (unintelligible). I think I'm iced up."The controller then issued a low-altitude advisory, telling the pilot to climb to 3,000 feet as his radar indicated that the airplane had descended to 1,200 feet. At that point the airplane made a tight left turn and began descending again.The controller called the pilot again urging him to climb to 3,000 feet. The pilot responded, "Hey, I'm trying like hell." At that point, Mode C data indicated the aircraft was at 1,000 feet.A few seconds later the Mode C read 800 feet. The controller told the pilot that the Quonset State Airport (KOQU) was off to his right at three miles. The pilot said, "Give me something, would you?" The controller instructed the pilot to fly eastbound. He wasn't sure what heading the airplane was flying because of the turn he had seen it make to the southwest. A few seconds later, the Mode C readout dropped off the radar screen.The controller's supervisor, who had been nearby throughout the incident, called the Quonset tower and asked the controller there to turn the runway lights to full bright and watch for the airplane. At one point, the tower controller thought he saw the airplane in the darkness to the west of the airport, but he later realized that he was looking at the lights of a construction crane.There was no further contact with the Commander. It crashed 11 miles south-southwest of the Providence airport in the town of Exeter. The airplane impacted trees and small boulders before coming to rest upright. The front left portion of the fuselage was crushed, and the pilot, the sole occupant of the aircraft, was killed.
Findings
Investigators sifted through the wreckage for anything that would give them an indication as to what had happened. There was no mention of finding any ice on the aircraft, but it is possible that the NTSB investigators did not arrive on the scene until the following morning.The airplane was equipped with de-ice boots on the wings and tail. The switches for the de-ice boots were found in the "auto" position. The switches for the propeller de-ice and windshield anti-ice systems, which used alcohol, were found in the "off" position. The reservoir for the windshield anti-ice system was empty. The right hand pitot heat switch was in the "on" position, while the left hand pitot heat switch was in the "off" position. It is not known if any of the switch positions were changed due to the ground impact.Investigators examined the aircraft's directional gyro and found that it functioned within tolerances even though it had sustained some minor impact damage. The de-ice distributor valve and boot timers were also examined and found to be operational.The NTSB blamed this accident on the "pilot's failure to maintain control after encountering icing conditions while on approach for landing. Factors in this accident were the night conditions and pilot's failure to select the airplane's propeller de-icing switches to the 'on' position."The pilot held a private pilot certificate with ratings for single-engine aircraft, multiengine aircraft and rotorcraft. The pilot's logbooks were not recovered, but six months prior to the accident, when he renewed his medical, he claimed 860 hours of flight experience and 47 hours in the previous five months. When the pilot obtained his multiengine rating in December 1999, two years and two months before the accident occurred, he reported 656 hours of flight time, 435 hours in airplanes and 142 hours of instrument time. Investigators estimated that he had about 200 hours in the Commander, but his proficiency and currency were not established.There were several witnesses to the crash who observed the airplane at an extremely low altitude "wobbling" from side to side before turning and descending in a left turn into the ground.
What Went Wrong?
We don't know for sure what took place in the Rhode Island sky that evening but there are a few possibilities that we can point to in our efforts to prevent a similar accident from happening again.The reservoir for the airplane's alcohol system was found dry and the switches for the windshield anti-ice and the propeller de-ice systems were in the off position. While it is possible that one or both of the switches were knocked into that position as a result of the impact, let's suppose for a moment that they were not.What if the pilot switched off both of the switches when the alcohol in the tank ran dry? Perhaps turning off the propeller de-ice system was inadvertent, but he probably would have switched off the windshield anti-ice system to save the pump. If he was distracted from the instruments when he attempted to do that he might have hit both switches at the same time.The altitude deviations that were observed on radar appear to be more in line with the pilot losing control of the aircraft as a result of instrument failure or misinterpretation rather than ice building up on the wings and tail. If it was the latter, it is doubtful that he would have been able to climb the aircraft back to 3,000 feet, which he did after the first deviation. So, what if the left pitot heat was off while the aircraft was in flight?The facts seem to back up this theory. If the pitot tube iced up while the pilot was being vectored into Newport he might have been having a hard time controlling the airplane or making sense of what he was seeing on the panel. The temperature on the surface was above freezing, and it is possible that when the aircraft descended the first time to 1,100 feet that the air was warm enough to clear the pitot tube. If that was the case that would explain why the pilot was able to climb to 3,000 feet and fly the assigned headings until it iced up again. Then the trouble repeated itself.There was only one PIREP in the vicinity of the Providence VOR that afternoon and evening. At 3:25 p.m., about 2-1/2 hours before the accident occurred, a regional jet reported a trace of rime ice between 4,000 and 4,500 feet while on descent into Providence. The crew reported that the temperature was 3 C at 3,000 feet.A trace of rime ice on a jet can mean more coverage on a smaller airplane that stays in icing conditions longer. It also is possible that the conditions deteriorated further as time went by so there was more ice in the clouds by the time the pilot arrived in the Providence area.Another possibility is that the pilot's pitot heat did not work. That and the potential icing conditions may have acted in concert and perhaps the pilot never did understand what was happening. A couple of his transmissions could be interpreted to suggest this scenario. Or, it is possible that his comment about being iced up referred to the pitot tube rather than the airframe.The lessons from this accident are clear. Before you begin any IFR flight make sure your equipment is functioning properly. If you have an alcohol anti-icing system, make sure it is full and use it sparingly. Not only do you not want to run out in flight, but alcohol may be hard to get at smaller FBOs. I flew a Baron that had an alcohol propeller and windshield system. Every fall the aircraft owner would order a stock of alcohol that he would keep in his hangar locker, because the FBO he was based with did not sell it.Check your pitot tube to make sure it is heating up before you take off into IFR conditions. Just turning on the switch and looking for a bump in an ammeter might not be sufficient for some systems.Don't forget that you have to maintain currency on partial panel operations. Practice recognizing instrument failures that result from a blocked pitot tube or static system. Know which is which so if you encounter that type of problem you will know which instruments to believe and which should be disregarded.Finally, don't fly in conditions your airplane is not equipped to handle.
More accident analyses are available in AVweb's Probable Cause Index. And for monthly articles about IFR flying including accident reports like this one, subscribe to AVweb's sister publication, IFR Refresher.