CHI02FA097

The MU-300 on-demand passenger charter flight sustained substantial damage during a landing overrun on a snow/ice contaminated runway. The captain, who was also the company chief pilot and check airman, was the flying pilot, and the first officer was the non flying pilot. Instrument meteorological conditions prevailed at the time of the accident. Area weather reporting stations reported the presence of freezing rain and snow for a time period beginning several hours before the accident. The captain did not obtain the destination airport weather observation until the flight was approximately 30 nautical miles from the airport. The flight received radar vectors for a instrument landing system approach to runway 30 (5,401 feet by 100 feet, grooved asphalt). The company's training manual states the MU-300's intermediate and final approach speeds as 140 knots indicated airspeed (KIAS) and Vref, respectively. Vref was reported by the flight crew as 106 KIAS. During the approach, the tower controller (LC) gave the option for the flight to circle to land or continue straight in to runway 30. LC advised that the winds were from 050-070 degrees at 10 knots gusting to 20 knots, and runway braking action was reported as fair to poor by a snow plow. Radar data indicates that the airplane had a ground speed in excess of 200 knots between the final approach fix and runway threshold and a full-scale localizer deviation 5.5 nm from the localizer antenna. The company did not have stabilized approach criteria establishing when a missed approach or go-around is to be executed. The captain stated that he was unaware that there was 0.7 percent downslope on runway 30. The company provided a page from their airport directory which did not indicate a slope present for runway 30. The publisher of the airport directory provided a page valid at the time of the accident showing a 0.7 percent runway slope. Runway slope is used in the determination of runway performance for transport category aircraft such as the MU-300. The airplane operating manual states that MU-300 landing performance on ice or snow covered runways has not been determined. The airplane was equipped with a cockpit voice recorder with a remote cockpit erasure control. Readout of the cockpit voice recorder indicated a repetitive thumping noise consistent with manual erasure. No notices to airman pertaining to runway conditions were issued by the airport prior to the accident.

History of flight

On March 25, 2002, at 0901 eastern standard time (EST), a Mitsubishi MU-300, N617BG, operated by Corporate Flight Management, Inc. (CFM), was substantially damaged during a landing overrun on runway 30 at Anderson Municipal-Darlington Field Airport (AID), Anderson, Indiana. Instrument meteorological conditions prevailed at the time of the accident. The 14 CFR Part 135 on-demand air-taxi flight was operating on an instrument rules flight rules flight plan. The two flight crew members and four passengers were uninjured. The flight departed from the Memphis International Airport (MEM), Memphis, Tennessee, and was en route to AID.

At 0430 central standard time (CST), the captain reported receiving a briefing at his residence via telephone from Nashville Flight Service Station for a route from Smyrna Airport (MQY), Smyrna, Tennessee to MEM to AID. He stated that the METAR for AID was not available, but light rain and light snow were forecast for the area. When he arrived at the office, he printed out the weather on DUAT. The captain’s DUAT weather was obtained at 0452 and indicated that no reports were available at the time for AID. During a telephone interview, the captain said that weather information indicated light rain and freeing precipitation. The first officer concurred with this statement. The captain did not go into great deal with the first officer regarding the weather. The first officer said during a telephone interview that he had also checked the weather. It was the first time that the captain had flown to AID.

The captain stated that while en route from MEM to AID, they checked the AID automated surface observing system (ASOS) 100 nautical miles (nm) out, but were unable to receive it until 30 nm out due to what he called "frequency bleed over." He added that they checked the weather at IND when they were unable to receive the AID weather. He stated that, at 0847 EST, the AID weather reported: wind 060 at 9 knots; ceiling 1,300 feet overcast; visibility 2 1/2 statute miles (sm); light snow; temperature -3 degrees Celsius (C); dew point -3 degrees C; and an altimeter setting of 30.10 inches of mercury (Hg).

At 0858:26 EST, the AID air traffic control tower controller (LC) transmitted, "november six one seven bravo gulf runway three zero or circle to land to your choice wind is a zero five zero to zero seven zero at one zero gust two zero and the altimeter three zero zero niner cleared to land."

At 0858:47 EST, N617BG transmitted, "alright cleared to land and ah alright right now we'll plan three zero please."

At 0859:02 EST, LC transmitted, "very well have a slight quartering tailwind on the right hand side."

At 0859:10 EST, LC transmitted, "and the only braking action i have is by a snowplow fair to poor and about a fourth of the runway then plowed down the middle."

At 0859:19 EST, N617BG transmitted, "roger seven bravo gulf."

At 0900:04 EST, N617BG transmitted, "no sir we're off and collapsed a landing gear."

The captain stated the following in a written statement:

"IND approach had vectored us onto the downwind for the ILS, when the ice detection light illuminated. We turned on the remaining anti/de-ice components, and the ice detection light turned off after 1 or 2 minutes. We kept the anti-ice features on. During the vectors onto the approach, we extended flaps to 10 degrees, but the IND approach vectored us through the localizer. We were established prior to intercepting the GS, and lowered the landing gear 1 dot below GS. Just inside the marker flaps 30 degrees was extended, and a stabilized approach commenced.

The tower reported winds 050 at 10 knots and asked if we wanted to do straight in landing or circle to land runway 12. Tower also notified us that the braking action was reported fair to poor by a vehicle. We opted for the straight in since the reported tailwind was minimal, and it appeared that the ceiling and visibility were less than reported. We broke out at 600' AGL and visibility 1 mile in heavy precipitation.

We crossed the runway threshold at 30 feet at Vref (106 KIAS) at which point I closed the power levers. We touched down in a timely manner, without any float. The speed brakes were deployed and the brakes were applied. The FO verified deployment of the speed brakes. The aircraft appeared to be decelerating normal, until approximately the 3,000 foot marker. At that point the aircraft's deceleration slowed down and the aircraft began to skid. It became apparent that we would be unable to stop prior to the end of the runway. As we approached the end of the runway I noticed a drop off at beyond the end of the runway. The visibility was not good enough to determine what was beyond the drop-off. I decided to turn the aircraft to minimize the forward speed after exiting runway and to avoid going down the hill. The aircraft came to rest 30' beyond the departure end to the right of centerline.

We contacted the tower and shut down the aircraft and its systems. We determined that none of the persons on board, including the crew were injured. I noted the time to be 0805 CST.

After exiting the aircraft, I determined that the breaking action was nil, the visibility was approximately 1 mile, the precipitation was a mixture of sleet, snow and ice, and that the ceiling was 600'. Reports received from both the tower controller and the AWOS about weather and airport conditions were not as actually observed by me in flight and after exiting the aircraft. Based on my observations after exiting the aircraft, the airport should have been closed to traffic prior to our arrival due to runway conditions."

The pilot said during an interview that the runway was a sheet of ice with 3/4 of the runway plowed in the center. He did not see any sanding or chemical applicant on the runway.

The first officer stated the following in a written statement:

"The descent took us into icing conditions (as indicated by icing light and visual inspection of the wing). All anti-ice and de-ice was applied. I then observed all ice melt off the wing.

We received poor vectors for the approach-a close downwind to the inbound course and a turn too close to the airport. Controller assigned heading of 300 [degrees] to intercept-this is the inbound course. Because we were vectored so close to the final approach course, out turn to assigned heading took us through the inbound course, and we had to add 30 [degrees] more to re-intercept. Controller then tried to hand us off to Muncie tower. I asked if he meant Anderson tower, and he corrected himself, telling me to contact Anderson tower. We now had flaps 10 [degrees].

At 1 dot below glideslope intercept, gear was extended, ignitions came on, and landing lights came out. At glideslope intercept, 30 [degrees] flaps were applied. Contacting tower, tower advised winds at 050 [degrees] at 10 knots and gusting and gave us the choice of straight-in runway 30 or circle 12. I did not immediately respond so that the captain and I could discuss and decide. We were still mostly IMC, and heavy precipitation limited visibility, so it was clear a circling approach could not be attempted without losing sight of the runway. We decided not to circle, and I advised we would land runway 30. Tower further advised braking action reported fair to poor.

We flew the profile to the runway, decreasing speed to Vref. ...the ALS [airport lighting system] was seen first, followed by the runway about 4 seconds later. I gave standard speed callouts to the runway, finally calling Vref at approximately 20-30 feet on the radar altimeter. The captain quickly and firmly put the airplane on the runway, including the nose wheel, and speed brakes were immediately extended. I would estimate that we landed no more than 800 feet from the end of the runway.

I sensed an initial deceleration, and I could also tell