ANC15FA021

The pilot's decision to continue flight into known icing conditions, contrary to the airplane’s limitations, which resulted in a total loss of engine power due to icing.

The private pilot was conducting a personal cross-country flight in instrument meteorological conditions. The airplane was not certificated for flight into icing. A review of radar data and radio communications recordings revealed that, after an air traffic controller issued the pilot a descent from 10,000 to 8,000 ft mean sea level (msl), the pilot reported that he was having engine trouble and had encountered possible icing. The pilot subsequently declared an emergency, and 3 minutes later, radar contact and communication with the airplane were lost. A day after the accident, search and rescue personnel found the remains of the pilot along the eastern shoreline of Culross Island. The left main landing gear strut and tire were also recovered. The rest of the airplane was not located, and it is presumed to have sunk in Prince William Sound. A review of weather information revealed that marginal visual flight rules conditions prevailed along the flight route due to low ceilings and that occasional instrument meteorological conditions (IMC) prevailed due to low ceilings and visibility in snow showers. Moderate icing was forecast between 4,000 and 10,000 ft above ground level. It is likely that the airplane encountered IMC and potential icing conditions below 10,000 ft msl. A pilot in an airplane about 6 minutes behind the accident airplane reported that, before descending from about 9,500 ft msl, he noticed a slight buildup of ice on the left engine cowling. Given this statement, the weather reports, and the accident pilot's report that he was having engine trouble, it is likely that the airplane flew through an area favorable for the formation of induction or carburetor icing, which resulted in a total loss of engine power. A mechanic reported that the pilot had told him that when he applied carburetor heat, the control felt "mushy." The mechanic instructed the pilot to inspect the control arm for any damage because it could prevent the carburetor heat from operating correctly. Before the accident flight, the mechanic sent a text message to the pilot asking if he had checked the carburetor heat control, and the pilot replied that he had not but would do it that day. Therefore, although it is possible that damage to the carburetor heat control arm may have limited the amount of carburetor heat available to eliminate any possible carburetor ice, it could not be determined if this occurred because the airplane could not be examined.

The accident took place about 1330 AKD and originally the search was focused on the east side of Perry Island, as that is where the last radar return was observed. About 1440, the search was refocused to the east side of Culross Island, about 7 nautical miles west of Perry Island after receiving the last known coordinates from the Spidertracks data. The conditions of Prince William Sound at the time and location of the accident was water temperature 42 degrees F with 1-4 foot seas. When the pilot was located on Culross Island, it was noted that he was not wearing a life jacket or other personal floatation device. A February 2008 NATO Research and Technology Organization publication titled Survival at Sea for Mariners, Aviators and Search and Rescue Personnel described the four stages of cold water immersion as: Stage 1: Initial immersion responses or cold shock (3-5 min.) Stage 2: Short-term immersion or swimming failure (5-30 min.) Stage 3: Hypothermia (= 30 min.) Stage 4: Post-rescue collapse or circum rescue collapse In this publication, it is stated that during stage 1, "death from cold shock is not uncommon," and it takes place "within 3-5 minutes of immersion." It further states that swimming in "cold, dense water" is very dangerous (stage 2). As the body becomes exhausted, the person transitions to a more vertical position before total submersion. A US Coast Guard article from January 6, 2015, titled A Lifejacket Buys You Time, states that it takes at least an hour for the full effects of hypothermia to set in and another hour after loss of consciousness for the heart to stop. It further states that without a lifejacket or other flotation device, drowning will occur prior to death from hypothermia. A new Ameri-King Corporation AK-451 406MHz emergency locator transmitter (ELT) was installed in the airplane on February 04, 2015. The ELT is designed to transmit GPS/NAV data immediately on 406MHz, 243MHz and 121.5MHz frequencies. The 406MHz is stated to last for 24 hours with the 243MHz and 121.5MHz lasting a total of at least 78 hours. Shortly after the last communication was received from N9247C, an ELT began broadcasting a signal on 121.5MHz. Based on the Spidertracks data, the ELT was activated in the air with the signal broadcasting for about 20 seconds before it was no longer received. The Alaska Rescue Coordination Center (AKRCC) reported never receiving a 406MHz signal. Currently, ELTs are not certified, nor required to be certified, for operation during or after submersion. A post accident review of archived FAA radar data and radio communication recordings revealed that, about 1315, the on-duty Anchorage (ZAN) Air Route Traffic Control Center (ARTCC) radar controller cleared the airplane for the LDA/DME H instrument approach to the Valdez Airport (VDZ). The airplane was about 60 miles southwest of VDZ, at an altitude of about 10,000 feet mean sea level (msl). Shortly after the pilot began a descent from 10,000 feet msl to 8,000 feet msl, he advised the ARTCC controller of an engine problem. As a result of static and background noise on the air traffic recording, a variety of filters were applied to clarify the audio by the NTSB Vehicle Recorders Laboratory in Washington, DC. The transcription from that recording can be found in the public docket for this accident. At 1316, after determining from the pilot that he did not have the VDZ weather, the ZAN controller issued the 1256 weather observation. One minute later, he cleared N9247C for the LDA/DME-H approach to VDZ with a restriction to cross the PEPPI intersection at or above 8000 feet. At 1318, the pilot reported leaving 10,000 feet for 8,000 feet. At 1319, the pilot of N9247C stated that he was having trouble with his engine, but the transmission was blocked by other aircraft on the frequency. The controller acknowledged N9247C's descent from 10,000 to 8,000 feet, but did not acknowledge the pilot's report of engine trouble. At 1321, the pilot transmitted, "Four seven charlie is declaring an emergency I have a problem with my engine I think I've been through the light stuff (unintelligible) descending for PEPPI for four thousand." The controller acknowledged the emergency call, then asked the pilot for the number of souls on board and if his engine was out or if he was able to continue. The pilot responded, "…I have windmilling power and (unintelligible) I'm in pretty poor shape. I still maintain forward airspeed. Descending for PEPPI at four thousand." The controller replied, "…Roger, you're coming in a little bit scratchy right now. Right now I am showing you over an island you're about to head out over a little bit of water uh are you going to try to continue to Valdez or somewhere else along your route?" The pilot responded that he was going to try to get below the cloud deck at his current location, circle over the island, and see if he had room to land on the beach. The controller acknowledged and asked again how many souls were on board the aircraft. The pilot responded that one soul was on board. At 1323, the controller advised the pilot that he was no longer in radar contact but that he was last observed northeast bound with the island behind him. The pilot acknowledged with "four seven charlie copy." There were no further communications between N9247C and air traffic control (ATC). At 1325, Empire 922, an ATR-42-300, heard transmissions from N9247C. They relayed to the ZAN controller that the accident airplane reported that he was at 5500, had not broken out of the clouds and the prop was wingmilling, having not fired yet. The controller advised Empire 922 that he could not hear N9247C and asked the crew to continue relaying N9247C's transmissions. They agreed to do so. At 1325, the ZAN controller asked Yukla 02, an E3 airborne early warning and control system (AWACS) aircraft, if they could help locate a mode 3 transponder code of 2221, the transponder code assigned to N9247C, at their 4 o'clock position and 40 miles at low level. Yukla 02 advised that they would help out. At 1330, the pilot of Empire 922 advised the controller that he was receiving a strong emergency locator transmitter (ELT) signal on frequency 121.5. The ZAN controller acknowledged Empire 922 and relayed to Juneau (JNU) FSS that N9247C had "crash landed." The ZAN controller asked the JNU controller if JNU was in contact with any other aircraft or helicopters in the area that could assist, but there were none. At 1331, the controller asked Empire 922 to attempt to establish communications with the pilot of N9247C via the emergency frequency [121.5]. At 1332, Empire 922 advised ATC that they had not re-established communications with N9247C and that they were no longer hearing the ELT. Two minutes later, Empire 922 advised ATC that they had tried to reach N9247C on frequency 121.5 without success. No further radio transmissions were received from the airplane and an emergency locator transmitter (ELT) signal was broadcasting shortly thereafter for about 20 seconds. A detailed air traffic aircraft accident package is located in the public docket for this accident. A NTSB detailed air traffic control study is also located in the public docket for this accident. A pilot in an airplane about 6 minutes behind the accident airplane stated that while at 9,500 feet msl, in-flight visibility varied between 1 and 10 miles and was restricted by suspended ice crystals. The temperature was about 0°F. The pilot stated that prior to descending from 9,500 feet msl, he noticed a slight build-up of ice on the left engine cowling. The FAA Civil Aeromedical Institute performed toxicology examinations for the pilot on May 08, 2015, which were negative for alcohol and drugs. A postmortem examination conducted by the Alaska State Medical Examiners Office attributed the cause of death to be drowning with co