Palair Macedonian Airlines Flight 301

Palair Macedonian Airlines Flight 301 was a scheduled international passenger flight from Skopje to Zurich, operated by Palair Macedonian, the then-flag carrier of Macedonia, now called North Macedonia. On 5 March 1993, the aircraft operating the flight, a Fokker 100, crashed shortly after taking off from Skopje Airport in snowy conditions. Out of the 97 passengers and crew members on board, only 14 survived. At the time, it was the deadliest air disaster in North Macedonia.

The investigation of the disaster concluded that the accident was caused by ice accumulation on the wings. The aircraft had been parked in Skopje in snowy conditions. During the refueling, the ice around the wings' roots had melted due to the temperature of the fuel, while the ice on the tips hadn't. While conducting the pre-takeoff ground inspection, the crew opted not to de-ice the aircraft, thinking that most of the ice had melted and the remaining was safe enough for flying. The aircraft eventually encountered control problems during takeoff, which caused it to crash.

Aircraft
The aircraft involved in the accident was a Fokker 100 with a registration code of PH-KXL and a serial number of 11393. The aircraft was new, built in 1992 and was sent to Palair on 27 January 1993. It was equipped with 2 Rolls Royce Tay 650-15. It had accrued a total of 188 flight hours and 136 flight cycles.

The fuel storage system for the aircraft, according to Fokker, was divided into four distinct chambers. The collector tank (CT) was the compartment with the most inboard position, and the other three compartments—MT1, MT2, and MT3— were located, respectively, from inboard to outboard. The MT3 compartment was placed at the top of the slope-shaped sections and was the most outboard of the others. Each compartment, with the exception of the CT, had ribs to allow fuel to circulate. The CT would be fully fueled during refueling, and any extra fuel would be distributed among the compartments. Less fuel will be present the farther the compartment is from the airframe.

The most outboard fuel compartment, MT3, had the most room for extra fuel due to the way the aircraft was built because it was situated in the most outboard and highest position. As a result, the MT3 tank's fuel level would be lower than that of the other tanks during refueling, which prevented the upper wing skin from making contact with the fuel of the aircraft.

Passengers and crew
Flight 301 carried 92 passengers and 5 crew members. Most of the passengers were Kosovars who were going to work in Switzerland. Dutch newspaper Reformatorisch Dagblad reported that there were also several Dutch and Swiss citizens on board the aircraft.

The flight was flown by members of Aircraft Financing and Trading (AFT), a company based in the Netherlands, based on a leased contract. Both pilots had the rank of captain. The pilot in command (pilot not flying (PNF), training captain) was 49-year-old Peter Bierdrager, a Dutch national. He had accrued a total flying experience of 11,200 hours, of which 1,180 hours were in the Fokker 100. The captain-under-training (pilot flying (PF) and acting captain) was an unnamed 34-year old pilot of Macedonian nationality. He had a total flying experience of 5,580 hours, of which 65 hours were in the Fokker. Before he joined AFT, he was a pilot at Yugoslav's JAT.

Pre-flight
The aircraft arrived in Skopje at 10:40 a.m. after a flight from Frankfurt, Germany. It arrived late as the flight previously had been delayed due to a malfunction on the flap position indication. The aircraft was scheduled to fly to Zurich at 11:50 a.m. with the same crew members. However, due to fears of violating work and rest regulations, a crew switch was conducted and a new set of pilots were brought in. For the flight to Zurich, the aircraft would be piloted by Captain Peter Bierdrager and his Macedonian co-pilot who was undergoing training to transfer from Boeing 737 to Fokker 100.

While the crew from Frankfurt was staying in the hotel, the aircraft was parked in Skopje in snowy conditions. A flight station engineer (FSE) then came to inspect the aircraft. A brief check was conducted on the exterior of the aircraft. During the inspection, at 11:00 a.m. the aircraft was refueled with approximately 2,000 L of fuel. The replacement crew eventually arrived in Skopje approximately 30 minutes after the refueling and immediately boarded the aircraft. Upon arrival in Skopje, the company's dispatcher was requested by Captain Peter to get information regarding the weather. The dispatcher eventually came into the cockpit and briefed the crew about the incoming weather. In response to the information, Captain Peter decided to add an additional 907 kg of fuel. The second refueling was commenced at around 11:40 a.m. and another exterior inspection was conducted.

Prior to the inspection, the dispatcher had briefly checked the aircraft for presence of ice and snow. After scraping his hand over the wing leading edges, he discovered that there was neither ice nor snow on the wings and it was particularly wet because all of the ice and snow had melted. This was eventually confirmed again during the second refueling when the FSE and Palair's ground crew members conducted a walk-around to check for the presence of ice and snow on the aircraft. Standing on a baggage cart, the FSE performed a "tactile check" by rubbing his hand over the wings. He noticed that there was some melted snow on the right wing and that the leading edge of the wing was clear of ice and snow. One of the ground crew then tried to confirm it and conducted a tactile check as well, collecting a stroke of snow from the inner flap of the right wing before showing it to the FSE. With the weather condition being taken into account, a discussion on whether the aircraft should be de-iced eventually ensued.

After a brief discussion with the ground crew, the FSE concluded that the aircraft did not need to be de-iced. He went back into the cockpit and told both pilots about his findings. The pilots immediately agreed and continued their routine checks.

Accident
Flight 301 was an international scheduled passenger flight originating from the Macedonian capital of Skopje with a final destination of Zurich, Switzerland. The loadsheet that had been provided by the airport agent listed a total of 91 passengers and 5 crew members. The FSE was later added as one of the passengers.

After having been parked in the snow for more than one and a half hours, the crew contacted Skopje to request start-up clearance. Start-up was conducted soon after, and the aircraft was set for taxiing onto the runway. The engine anti-icing was set to on, the ignition went into automatic, and the flaps were set to 8 degrees. The crew was cleared to line up at Runway 34. Arriving on Runway 34, the crew decided to use the standard VR speed of 134 knots.

At 11:11 a.m., the aircraft was cleared for take-off to Switzerland. At the time, there was moderate snow and visibility was limited to 900 m; observation on the ground confirmed that the visibility was poor enough that the end of the runway could not be seen from Flight 301's position.

Flight 301 became airborne 28 seconds after its takeoff clearance and, according to the ATC who witnessed the take-off, vanished beyond the snow. Just two seconds after taking off from Runway 34, the aircraft began to shake violently. The aircraft was climbing at a rate of 900 ft/min and the crew called out positive climb rate. As the crew was announcing, the aircraft suddenly rolled to the right at an angle of 11 degrees, surprising the crew. The first officer then tried to counteract this by applying a right wing-up aileron input at full force. The wings slowly returned to its previous position and the angle gradually decreased.

Intending to increase the climb, the first officer raised the nose to 10 degrees. The landing gear was retracted, and the aircraft was climbing. However, within a span of three seconds, Flight 301 rolled severely to the left with a bank angle of 50 degrees. Perturbed by the situation, Captain Peter exclaimed "Ah shit!". The crew did not understand the anomalies they were facing at the time. Ailerons and rudder input were immediately applied to correct the attitude of the aircraft. The first officer tried hard to roll the aircraft to the right by making a full right aileron input.

Thinking that the autopilot was at fault, Captain Peter exclaimed "Oh! Deselect!". They tried to turn it off, but then discovered that the autopilot had never been engaged. While applying the full right aileron input, the aircraft immediately rolled to the right at a dangerous angle of 63 degrees. The FSE who was sitting at the cockpit jump seat shouted "Nose up!" as the aircraft plunged towards the ground.

As the aircraft continued to dive, the crew pulled the elevator to almost at its maximum. The angle of attack increased, but the stall warning began to blare inside the cockpit. The aircraft eventually entered a sink rate of 2,000 ft/min and the alarm system warned the crew on the excessive sink rate.

The crew tried to recover the aircraft, and it appeared to be successful as the right angle started to decrease, down to as much as 15 degrees. However, as it approached the ground, the right wing dipped again, reaching 90 degrees. The tip eventually contacted the ground and grazed a road, cutting the airport's perimeter fence and detaching from the aircraft due to the collision. The aircraft, meanwhile, slammed into the ground with its right side and broke up. The cockpit, parts of the front cabin, and the tail surfaces separated from the airframe, while the front portion and the whole right side of the aircraft were immediately destroyed. The resulting collision caused the fuel to ignite and the aircraft burst into flames.

A substantial part of the aft fuselage remained intact after the crash. This was where most survivors were located. Despite this, the aft fuselage had been penetrated by the detached right engine during the crash, causing a fire that eventually destroyed the whole section.

Initial aftermath
The sound of the crash was loud enough to be heard by a United Nations Peacekeeper Pilot who was walking on a ramp nearby. He immediately offered Skopje Tower via telephone to investigate the source of the noise. Using a Bell 212, the wreckage was immediately found nearly 0.5 km from the airport. The wreckage was spread over an area of 220 x 40 meters. After landing at the crash site, the personnel quickly notified other emergency services and nearby UNPROFOR army base about the accident and transported seven survivors from the area. The Bell 212 returned to the crash site two more times and evacuated more survivors. Shortly after, the Macedonian emergency services arrived at the site and commenced the search and rescue operation. As many as twenty survivors, of whom five were in critical condition, were transported to the hospital in Skopje. However, four survivors would later be pronounced dead on arrival. Several others succumbed to their injuries. Remarkably, the only surviving crew member, a flight attendant, suffered only minor injuries.

Investigation
Following the crash, the Macedonian government immediately set up the State Accident Investigation Commission (SAIC), which was led by Yugoslavia's Ministry of Transport and Communications. Being the state of both manufacturer and registry, the Netherlands was involved in the investigation. Representatives from Rolls-Royce and Swissair were also invited to assist. The French BEA helped encode the FDR and CVR contents.

The result of the flight recorder readout would eventually be presented with a flight simulation, which would take months to complete, the Dutch government noted. The Macedonian government, however, wished to wrap up the investigation as quickly as possible. The final report on the crash was released by the SAIC in May 1993, just one month after the catastrophe. The Dutch government carried out additional investigations after being dissatisfied with the report's final draft. The Dutch government met with Swiss representatives after media reports that the Chairman of SAIC had declared the investigation to be finished and urged the Macedonian government to continue the investigation, which was agreed by the government, albeit reluctantly.

In January 1994, following a meeting that had been convened by representatives of the investigation team, parties involved in the investigation agreed to publish the summary of the investigation and the probable cause of the crash to the press. Substantial disagreements still persisted between both parties and further meetings were conducted. In January 1996, inquiry by the Netherlands regarding the official final report was eventually responded to by Macedonian representatives, who explained that further findings from the Netherlands had been disregarded and added that the May 1993 final report that had been provided by the Macedonian government would be declared the official final report and would be presented to ICAO. The Dutch government ultimately decided to produce a separate investigation report into the crash, and the inquiry would later be the only one to be publicly released.

Icing
The aircraft crashed immediately after takeoff in a violent manner, with its right wing being the first to contact the ground at a perpendicular angle. Account from the surviving flight attendant suggested that the aircraft had suffered control problems during takeoff. The aircraft could be felt shaking while lifting off from Skopje, and seconds later it began to sway violently to the left and right before plummeting towards the ground.

The result of the flight recorder readout confirmed that the aircraft had encountered problems with its controls immediately after takeoff. Just seconds after lifting off, the aircraft began to shake, and the right wing began to drop. After slowly returning to wing level, the wings began to dip again, and the aircraft swayed to the left. The crew tried to recover the aircraft by making a right turn input, but the aircraft suddenly turned towards the right with a dangerous bank angle. The issues that the crew were facing could not be resolved, and they eventually lost control of the aircraft.

Weather data collected by investigators showed that there was light to moderate snow prior to and during the accident. Airport workers reported that the snow melted when it touched the ground. Prior to the accident, there was no visible snow on the runway, taxiway or apron. Investigators confirmed that it was a wet snow, that is, snow mixed with rain. Temperature at the airport was recorded to be at 0 C, while the aircraft itself was experiencing an evaporative cooling effect, demonstrated by the recorded total air temperature value of -2 C. The aircraft was exposed to wet snow and the temperature was lower than the freezing dew point, a conducive condition for airframe icing.

The presence of snow on the wings eventually affected the aerodynamics of the aircraft. The recorded data indicated that due to the contamination, the aircraft's lift capability was limited to half of what it was capable of. The aircraft would not normally stall if the nose angle was below 16.5, but on the day of the accident, the aircraft was already struggling to takeoff when the angle of attack had just reached 10 degrees. When the angle of attack exceeded 10 degrees, the aircraft began to enter a stall and vibrations were noted. Whenever the nose was lowered to below 10 degrees, the aircraft managed to build-up its airflow and recovered from the situation.

Analysis made by investigators suggested that the ice contamination was likely to be located on the front part of the wing, particularly close to the leading edges of the aircraft. This was indicated by the massive increase in drag and loss of control on roll damping and roll control when the angle of attack exceeded 10 degrees. The analysis further suggested that the outer parts of the wings were probably the ones that got contaminated, even though it did not exclude the possibility of contamination on the inner parts of the wings.

Estimation conducted by investigators showed that the aircraft had been exposed with light precipitation, which later increased to moderate, for 1 hour and 15 minutes, with an unspecified amount of thickness. For another 15–20 minutes after the walk-around check, the aircraft was further exposed by snow, which presumably worsened the wing contamination. Due to the prevailing weather condition, investigators believed that the aircraft's wing upper surface had been covered with a thin layer of snow and possibly frozen to the skin.

Based on the available data, there was no indication that a technical malfunction had occurred on board. Many people had managed to observe the aircraft and stated that there was no damage on the wings, and the wreckage indicated that there was no pre-existing damage. The data, however, supported the theory that there was wing contamination on board. The issue could not have appeared had the crew de-iced the aircraft prior to takeoff.

De-icing
The investigation board opined that based on the weather data and aircraft information, a de-icing procedure was warranted for the aircraft. Prior to the de-icing, the manual stated that one of the crew members, either the FSE or the captain, would be the one who conducted the external inspection of the aircraft. The decision to de-ice the aircraft however would be given to the captain as the commander of the flight.

The aircraft had been checked for signs of snow and ice by the crew at least twice prior to takeoff. The first inspection was conducted during the first refueling, and based on the findings, it was concluded that no de-icing was required for the aircraft. Following the second refueling, another check was conducted, with the same conclusion that there was no need for de-icing.

The inspection was carried out by the FSE who was accompanied by three ground crews. According to reports from the ground crew, rather than being distributed uniformly, the inspection seems to have been focused mostly on the inner part of the wings, which was located near the wing roots, particularly on the right side. The FSE initially checked the wing by walking from front to back before using a baggage cart to get a closer look at the wing's surface. Upon touching his hands with the wing, he discovered that the snow had melted and felt that his hands were wet. After finishing the brief check, he then discussed the possibility of using de-icing on the aircraft. At the time, another ground crew took slush from one of the flaps and showed it to the FSE. Thinking that the aircraft was clean enough from ice and that the remaining ice would sledge off during the takeoff, he decided that the aircraft did not need to be de-iced.

According to the investigation, it would not be possible for the FSE to check the wing thoroughly for ice as the FSE's position during the external inspection with the baggage cart was near the wing roots. His decision was weighted based upon his findings during the check on the wing roots and the inner part of the wings instead of the outer parts, which likely had been contaminated by ice.

Following his decision to not de-ice the aircraft, the FSE then went into the cockpit to inform the crew, who immediately agreed without any questions. Apparently, due to his highly regarded skills and the fact that he was the representative of Swissair, which was reputable for its safety, the crew decided to immediately trust his judgement without any criticism. The crew did not leave the cockpit to confirm the result of the check and instead focused their attention on the instruments and other preparations for the takeoff.

The investigation also opined that both the crew and the FSE might have misinterpreted the weather conditions at the time. Snow had just started falling by the time they arrived at the airport via car. During the trip, it was snowing lightly, and most of it immediately melted upon touching the ground. After arriving at the aircraft, the pilots immediately went into the cockpit to prepare for the flight and did not pay much attention to the weather. The FSE, who had just checked the wings and concluded that there was only melted snow on the wings, further reinforced the crew's belief that the weather condition was harmless. During the taxi, while the crew noticed that the snowing had intensified, they thought that it was still safe enough to fly as the snow immediately melted upon contacting the ground. They probably did not think that the falling snow could freeze into the wings.

Eventually, due to the belief of both the FSE and the flight crew that the snow was not dangerous for the operation of the flight, the aircraft was not de-iced.

Fuel distribution
The crew decided not to conduct the de-icing procedure after apparently thinking that most of the snow on the aircraft had melted and that all that was left was slush of snow, which was expected to slide off during the takeoff and was not dangerous for flight. Even though the ice near the wing roots had indeed melted, the rest, which was located on the outer parts, had not. This was probably caused by the mechanism of the Fokker 100's fuel distribution system.

After being exposed to low temperatures at very high altitudes, the fuel in the tanks became colder than the ambient temperature. When they arrived at the airport, the temperature was below freezing. This enabled the falling snow to freeze into the wings. During the refueling, the newly added fuels were warm enough to melt the ice that had been collected during the flight as the wing's skin temperature was lower than the fuel. The surface of the CT and MT 1 responded rapidly to the added fuel and ice immediately melted from the skin. However, due to the distribution of the fuel, the warming effect did not reach the outer parts of the wings.

The fuel compartments were designed in slope-like chambers, with the MT3 being the furthest from the main airframe, located near the wing tip. Each compartment was equipped with stringers that would limit the movement of the fuel. Fuel would progressively fill each tank, but none would receive equal fuel. Each compartment would have a different fuel level due to the design of the fuel tanks. Accordingly, each compartment would have a different warming effect. This was confirmed by a test that had been conducted by Fokker. The test, called the "fuel temperature test" by Fokker, suggested that the further the compartments were from the airframe, the slower the temperature would rise. The compartments near the wing roots would be warmer than the compartments located further in the wing tips, which were not affected significantly by the added fuel. The ice in this area eventually did not melt.

The warming effect from the first refueling lasted until the second refueling and this caused ice in the wing roots to continuously melt from the aircraft. This might have caused observers to think that most of the ice had melted from the aircraft, even though the ice on the outer part of the wings did not benefit from the warming effect due to the flawed fuel distribution system. It was noted that the melting of the snow on the inboard wing would anyhow disappear due to the activation of the booster pumps later on.

Lack of attention
Fokker conducted multiple studies regarding the effects of ice on its aircraft. Using 1-2 mm of ice spread over the entire wing, an analysis regarding its effect on the aircraft's lift was carried out. The result indicated that even with such thin ice, the aircraft lost a total of 35% of its maximum obtainable wing lift, and the nose could never exceed 5–6 degrees. If the same amount of ice were distributed near the wings' leading edges and outer parts, the effects would be even more severe.

Unfortunately, the devastating effects of icing on Fokker aircraft were already well known in the aviation industry. The design of both the Fokker F100 and its sister craft, the Fokker F28 Fellowship, made both aircraft more vulnerable to icing than any other type, as relatively small amounts of ice could produce extreme effects on the controllability of the aircraft. The aircraft was more likely to lose more lift due to its heavily swept wings that were not equipped with leading edge slats, causing it to be more susceptible to smaller amounts of ice. The issue had caused multiple deadly accidents in the past. Notable accidents that were cited in the final report included the crashes of Air Ontario Flight 1363 near Dryden in 1989 and USAir Flight 405 near LaGuardia Airport in 1992. The latest crash resulted in the release of a personal briefing on recent studies regarding the effects of contaminated wings by Fokker, emphasizing the dangerous effect that could be inflicted by a small amount of ice on the controllability of their aircraft.

The investigation eventually revealed that the operator of the aircraft, AFT, did not include de-icing procedures in adverse weather conditions in their Fokker F100 manual. The operator's Standard Operating Procedure (SOP) and operating manual also did not publish any de-icing procedures that were applicable for ground operations during bad weather conditions. Additionally, neither the operator's Standard Operating Procedure (SOP) nor operating handbook published any de-icing techniques that might be used for ground operations in inclement weather. The manual provided generic explanations for the meaning of icing and other relevant concepts, but it lacked specific guidance on when anti-icing should be applied. The board ruled that each operator was required to publish particular policies or procedures regarding aircraft operations in cold weather and submit them in advance of the operation.

Conclusion
The investigation board concluded the cause of the crash as follow: "The Board determined that impact with the ground in a steep right bank shortly after liftoff was caused by loss of roll controllability due to contamination of the wings with ice. This situation resulted from an omission to carry out spraying of the aircraft with deicing or anti-icing fluid in meteorological conditions conducive to icing, due to a lack of ice-awareness of the flight crew and the Flying Station Engineer. Contributing factors were a lack of common background and procedures in a difficult multi-sources operational environment.'"

A total of six recommendations were published by the board, covering issues regarding the training of crew and inclusion of more specific de-icing procedures, review of the existing regulations regarding air operation, and further procedural clean-up. In the wake of the crash, Dutch investigation agency RLD issued an airworthiness directive that stated that every operator of Fokker 100 was mandated to conduct physical hands-on check for possible ice contamination on the wing leading edge and wing upper surfaces.