Risk Assessment of Proposed ARFF Standards (2011)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

32 CHAPTER 4 IDENTIFICATION AND REVIEW OF “ACCIDENTS OF INTEREST” Air carrier passenger aircraft accidents were reviewed to determine whether or not if different ARFF standards were in effect at the time of the accident there would have been a different outcome in terms of a reduced number of fatalities and/or serious injuries. Specifically the research scope was directed towards looking at what the differences in fatalities and serious injuries may have been if standards set forth by the International Civil Aviation Organization or the National Fire Protection Association would have been in effect, in lieu of the standards set forth by the Federal Aviation Administration (FAA). In reading the remainder of this report, it is important that the reader keep this study scope in mind. If an aircraft accident did not involve fatalities or serious injuries, it was not of interest to this study. Consequently, there are accidents where an aircraft may have been substantially damaged during the crash sequence or, perhaps, even destroyed by fire. If all the “souls on board” these aircraft were able to evacuate the aircraft without sustaining any serious injuries, the accident was not included in this research effort. The “accident of interest” criteria that are discussed later in this chapter identify other filters that were used to screen accidents and the basis for each of these filters. This effort is being undertaken as legislation is being considered by the U. S. Congress that would require the FAA to adopt consensus ARFF standards for its airport certification program that is set forth in 14 CFR Part 139. If the proposed legislation is passed, it would affect ARFF standards but would not change the airports that are required to have ARFF coverage or the types of aircraft operations for which ARFF services are required to be provided. Consequently, this research study did not include air carrier passenger aircraft accidents if the aircraft had less than 10 passenger seats. It also did not look at accidents involving “all cargo” aircraft. It should be noted that it is not the purpose of this research to recommend whether or not the proposed regulation should be enacted. Rather, it provides technical information and analysis that can be used by others, in conjunction with information from other sources, in formulating policies, regulations, and procedures related to this issue. Definition of Accident, Fatality, and Serious Injury The definition of accident that is contained in 49 CFR Part 830, which governs aircraft accident and incident reporting and investigation in the United States, was used in this research effort. This definition of aircraft accident is: an occurrence associated with the operation of an aircraft which takes place between the time any person boards the aircraft with the intention of flight and all such persons have disembarked, and in which any person suffers death or serious injury, or in which the aircraft receives substantial damage. Since this research study is looking at the reduction of fatalities and serious injuries, it is important to understand the meanings of these two terms as used in the preceding definition. Fatal injury means any injury which results in death within 30 days of the

33 accident. Serious injury means any injury which: (1) Requires hospitalization for more than 48 hours, commencing within 7 days from the date the injury was received; (2) results in a fracture of any bone (except simple fractures of fingers, toes, or nose); (3) causes severe hemorrhages, nerve, muscle, or tendon damage; (4) involves any internal organ; or (5) involves second- or third-degree burns, or any burns affecting more than 5 percent of the body surface. Although these are U.S. definitions, similar definitions are used in ICAO’s Annex 13 and are used by agencies throughout the world that are involved in accident investigation. Consequently, the number of fatalities and serious injuries reported in an accident investigation should be consistent regardless of where the accident took place or what agency was in charge of the investigation. “Accident of Interest” Criteria Using the preceding background, the following criteria were used to identify the accidents that would be analyzed as part of this study. To be considered, the accident had to meet all of the following criteria. The reasoning underlying each of the criteria is set forth after each criterion. The accident must have: 1. Involved an air carrier airplane providing scheduled passenger service with more than nine seats or unscheduled air carrier passenger service with more than 30 seats. Reasoning: The aircraft type, passenger service and seating requirements are based upon the legal authorization that the FAA uses to require airports to obtain airport operating certificates and provide the ARFF service associated with those certificates. At this time there is no indication that ARFF proposals that are under consideration by Congress would change the aircraft type, passenger service and/or seating requirements. 2. Involved a crew member/passenger fatality or serious injury. Reasoning: Objective of this research effort is to see if there may have been reduced fatalities and/or serious injuries if different ARFF requirements had been in effect at the time of the accident. ARFF standards are set by FAA, ICAO, and NFPA based upon aircraft size which is the de facto measure of number of people on the aircraft. Fatalities or serious injuries to people located outside the aircraft such as ground handlers or people living in the vicinity of the airport were not considered since they are not considered in establishing the ARFF standards by FAA, ICAO, or NFPA. However, if the accident involved a collision of two aircraft, fatalities and serious injuries to the crew and passengers on both aircraft were considered as long as one of the aircraft met the first criterion. 3. Occurred on (or the crash sequence started on) airport property or, if adjacent to the airport within 2000 feet (600 meters) of the runway end and 500 feet (150 meters) on either side of the extended runway centerline. Reasoning: This area not only encompasses all the response areas identified by FAA, ICAO, and NFPA standards but goes beyond them to include any accident

34 falling just beyond the response areas identified in those standards. Accidents that start on the airport but may end up outside the airport or the areas specified in this criteria were included in the review. For example, an aircraft overrun or veer-off that starts on the runway but where the aircraft ends up outside the area specified within the identified criterion’s boundaries is still included in this effort since the accident sequence started on the airport. 4. Occurred from January 1, 1989, to present. Reasoning: This will provide an approximately 20-year sample. 5. Have an accident report/record available in English. Reasoning: The funds provided in the study were inadequate to retain translation services. Since English is the international language of aviation, many reports are translated into English. These reports usually have a disclaimer up front saying that if there are any discrepancies between the report in English and the one in the native language that the report in the native language governs. Accident Records Search The NTSB’s accident data base was reviewed to identify accidents that occurred in the United States that would be of interest to this study. One of the keys to this research effort was to obtain enough data points to use as input in the risk analysis. For this reason, the search for accidents of interest was expanded to include accidents that occurred in foreign countries. Thanks to the international treaties that govern aviation, the types of data that are recorded for accidents worldwide are basically the same and when reports are written they are normally in the format specified in ICAO’s Annex 13. An Internet search was conducted of government websites containing aircraft accident investigation records for accidents that met the preceding criteria. While some countries like the United States and United Kingdom have posted information on aircraft accidents going back several decades many other countries have posted 10 years or less of data. Some countries like France and Sweden provide English translations for only some of the accidents which occurred within their jurisdiction while others only provide accident reports in their national language. Overall, foreign countries have very few accidents when compared to the United States. The difference in these numbers can be attributed to the difference in the number of aircraft operations. The number of aircraft operations in the United States far exceeds the volume of operations in any other country. An Internet search was also conducted of the aviation accident data bases/reports from several countries to identify accidents that met the criteria for this study. The ability to establish filters to sort events listed in a country’s data base varied from country to country and for most countries was extremely limited. Many of the data bases included both accidents and incidents. The entries for more than 3,000 events were reviewed to identify the ones that met the criteria established for this study. The review of the online accident data bases/reports for several countries revealed that they did not contain any accidents that met the established criteria.

35 The results from the preceding Internet search were crosschecked with accidents that were included in the Flight Safety Foundation (FSF) report entitled Reducing the Risk of on Runway Excursions and the Airport Cooperative Research Program project 04-08, Improved Models for Risk Assessments of Runway Safety Areas. Many of the accidents included in the FSF and ACRP studies did not meet the criteria for inclusion in the study. For example, the accident may not have resulted in any fatalities or serious injuries or there may have been fatalities to the flight crew but the flight was an all-cargo operation rather than an air carrier passenger flight. This crosscheck identified additional accidents of interest---however in many cases there was not enough information to include the accident in the current study effort. The information on many of the international accidents in the FSF report was just an initial notification with few details provided beyond type of aircraft and number of fatalities. Information on the ARFF response, in most cases, was very limited or non-existent. However, we were able to identify some additional accidents with sufficient information for review that were not included in the initial Internet search. The results were also crosschecked with a listing of accidents and incidents involving commercial aircraft that was available on Wikipedia. In two cases, Wikipedia provided a link to the English translation of the accident report for an international accident that was not posted on the country’s website or captured in the previous Internet searches. GENERAL DESCRIPTION AND CHARACTERISTICS OF “ACCIDENTS OF INTEREST” The preceding search effort identified 81 accidents that occurred in 17 countries that met the established criteria. Table 5 provides information on the number of people aboard the aircraft involved in these accidents and the number of fatalities and serious injuries. A breakout is also provided between accidents that occurred at U.S. airports and non- U.S. airports. The 81 accidents included eight that involved a collision between two aircraft. For the collision accidents, Souls on Board (SOB’s) is considered to be the sum of the total number of people on each aircraft. Table 5. Accidents of Interest (1989-2008) Number of Accidents Souls on Board Number of Fatalities Number of Serious Injuries U.S. Airports 51 6881 316 214 Non-U.S. Airports 30 5383 831 352 Total 81 12264 1147 566 Appendix A provides details on each of these accidents. The accidents are listed in chronological order in the appendix. Throughout the remainder of this report references to a particular accident are made by providing the date of the accident and location identifier for the airport where it occurred. Types of Accidents As one would expect the 81 accidents in this study included undershoots, overshoots, veer-offs, and accidents involving aircraft that stopped on the runway during landing or takeoff. In addition there were accidents resulting from runway collisions and accidents

36 that occurred during apron and taxiing operations. Finally there were 11 accidents that were categorized as “Other.” Each of these accident types is discussed further in this section. Table 6 provides a breakout for 44 accidents that occurred during landing or take off. The last column of the table shows accidents where an aircraft in either the landing or takeoff phase of operation stopped on a runway or exit taxiway when the captain believed there was something amiss with the operation. This column also includes accidents where the captain decided to reject the takeoff and brought the aircraft to a stop on the runway or exit taxiway. Table 6. Number of Accidents by Accident Type (1989-2008) Phase of Operation Undershoot Veer-off Overrun Remained on R/w or T/w Landing 2 8 8 13 Takeoff - 3 6 4 Total 2 11 14 17 Some of the accidents involved the captain declaring an in-flight emergency. Of the 81 accidents, the ARFF units were pre-deployed prior to aircraft touching down in 11 of them. (They were not pre-deployed in 59 of the accidents. The pre-deployment status could not be determined for 11 of the accidents.) In addition to the 44 accidents shown in the table there were seven runway collisions and 11 accidents that were labeled as “Other.” These “Other” accidents included situations where the aircraft stalled during landing or takeoff and proceeded to crash on the airport. “Other” events also occurred when the flight crew was unable to control the aircraft due to mechanical problems resulting in a crash occurring on airport property. There were nine accidents that occurred on the apron and 10 that occurred on taxiways during taxiing. One of the accidents that occurred on the apron involved a collision between two aircraft. To the extent information was available, the location of each accident for all the accidents except “taxiway” and “apron” accidents is graphically depicted relative to a generic runway in Appendix B. Accidents Involving Fire Only 35 of the 81 accidents involved an actual fire. For the remaining 46 accidents, there were several accidents where the flight crew suspected there was a fire on board the aircraft. These included such things as faulty fire sensors in the engines or cargo compartments and reports of fire from parties outside the aircraft. During the accident investigation, it was found that there was never an actual fire in these instances. Accidents Where All Fatalities and Serious Injuries Occurred during the Evacuation There were 38 accidents where all the fatalities and serious injuries occurred during the evacuation of the aircraft. This represents 47% of the 81 “accidents of interest” in this study.

37 There was only one accident that involved a fatality resulting from the evacuation of the aircraft (11/20/2000-KMIA). This accident is somewhat of an anomaly since the fatality involved a flight attendant who was ejected from the aircraft onto the apron when he opened a door to start the evacuation while the aircraft was still pressurized. However, there were three serious injuries that also occurred during the evacuation of that aircraft. The serious injuries that were sustained in the 10 accidents that occurred on taxiways during the taxi phase of operation and all except one of the nine accidents that occurred on the apron resulted from evacuation of the aircraft. Table 7 provides a comparison of these “evacuation” injury accidents to all the accidents. Table 7. Comparison of Evacuation Injury Accidents to All Accidents (1989-2008) Number of Accidents Accidents with All Serious Injuries Due Only to Evacuations Percent of Evacuation- related Accidents to Number of Accidents Number of Accidents Souls On Board Serious Injuries U.S. Accidents 51 29 3102 33 57 Non-U.S. Accidents 30 9 2435 32 30 Total Accidents 81 38 5537 65 47 These accidents included such things as people fracturing limbs from falling off the slides or breaking legs or ankles from jumping off wings. In the case of the accidents that had suspected but not actual fires, many of the aircraft did not meet the “substantial damage” criteria in the definition of aircraft accident. The event was classified as an accident due to the serious injury that occurred during the evacuation. Events involving evacuations of the aircraft merit some additional discussion. Generally, the decision on whether or not to evacuate an air carrier aircraft is made by the captain. This normally involves the flight deck crew notifying the cabin crew and then proceeding with the actions listed on an emergency evacuation checklist such as shutting down the engines. However, most airlines allow the flight attendants to initiate the evacuation based on the immediate circumstances of the moment that they may be privy to. In the review of accidents there were some events where the passengers saw or heard something that they deem to be unsafe and self-initiated the evacuation. This usually catches the flight and cabin crews by surprise. Any aircraft evacuation, regardless of who initiates it, can result in serious injuries. It is important to note that in many of the 38 accidents in this study only one or two people were seriously injured on an aircraft that may have had more than 100 souls on board. These injuries are normally not life-threatening—they often involve fractures such as a broken wrist, ankle, or leg. To keep things in the proper perspective, it should be noted

38 that there were also several other accidents (far more than 38) involving emergency evacuations that were initially reviewed but not included in this study effort since the accident did not involve any serious injuries. Extrication of Occupants There were nine accidents where ARFF entered the aircraft and extricated people who were trapped in the aircraft or needed assistance to exit. Six of these accidents occurred at U.S. airports. (See Table 8 below.) Table 8. Accidents Where Occupants Needed to be Extricated from Aircraft by ARFF (1989-2008) Accident Date Airport LOC ID Fire Description 7/19/89 KSUX Yes* Extricated flight crew that was trapped in deformed cockpit. 9/27/89 KGCN Yes* Extricated several passengers that were trapped in deformed fuselage. 2/1/91 KLAX Yes Rescued first officer from cockpit as flames approached. 7/2/94 KCLT Yes* Rescued several passengers from three rows of seats after extinguishing fire. 12/16/97 CYCF No Extricated seven passengers from deformed fuselage. 6/1/99 KLIT Yes Extricated first officer from the cockpit and assisted some passengers from first class in exiting aircraft. 8/22/99 HKG Yes Assisted passengers in evacuating the cabin. 7/9/06 IRK Yes Rescued 11 passengers from the cabin. 8/27/06 KLEX Yes Extricated first officer from the cockpit. * Fire was either not present in area of rescue or had been extinguished prior to rescue. Collision Accidents The 81 accidents included 8 where two aircraft were involved in a collision. One of these occurred on an apron when the captain of a DC-9 that lost hydraulics taxied into the wing of an A-319 (5/10/05, KMSP). The captain of the DC-9 suffered serious injuries. The remaining seven collisions all occurred on the runway with at least one of the aircraft being on takeoff roll or landing rollout. All seven accidents resulted in fatalities. Table 9 provides a summary of the numbers.

39 Table 9. Accidents Involving Runway Collisions (1989-2008) Accident Date Airport LOC ID Aircraft Involved Souls on Board* Fatalities* Serious Injuries* 1/18/90 KATL Boeing 727 Beech King Air 159 1 1 12/3/90 KDTW Boeing 727 DC-9 198 8 10 2/1/91 KLAX Boeing 737 Fairchild Metroliner 101 34 13 11/22/94 KSTL MD-82 Cessna 441 142 2 0 11/19/96 KUIN Beech 1900 Beech A90 14 14 0 5/25/00 CDG MD-83 Shorts 330 159 1 1 10/08/01 LIN Boeing MD-87 Cessna 525A 114 114 0 * Sum from both aircraft Response Times As stated in Chapter 3, the FAA, ICAO, and NFPA all have standards that require a demonstration that an ARFF vehicle can reach a designated point on the airport in a designated time frame. Under the three organizations’ standards these demonstrations are all done in ideal conditions, e.g., daylight, good visibility, and dry pavement. Unfortunately most accidents do not occur in ideal conditions. As was pointed out in Chapter 2, an FAA Technical Center study looked at the time that elapsed from when the aircraft stops until the firefighters are in a position to fight the fire. For 50 per cent of the occasions firefighters were in place within 4 minutes and for 90 per cent of the occasions they were there within 12 minutes. Although the time period measured in the study differs from the “demonstrated” time period, one can still conclude that actual response times are greater than demonstrated times. In the 81 accidents in this study, there were 12 accidents where the ARFF response was delayed. There were five accidents where the delay could be attributed to weather conditions, i.e., the ARFF units had trouble locating the aircraft because of the environmental conditions. In the runway collision at Detroit (12/3/90, KDTW) the ARFF knew the runway on which the collision had occurred but because of the low visibility had to drive slowly to the site as they navigated around aircraft that were holding on taxiways. In the Linate (10/08/01, LIN) runway collision, the responding ARFF units initially thought that they were responding to a building fire that was called in by a security guard. It was only when they arrived on site did they find that an aircraft had crashed into the building. Approximately 18 minutes elapsed before an aircraft returning to its parking position revealed that there was another aircraft located on the runway that was also ablaze. Communications Posed Similar Challenges In one accident (9/27/89, KGCN), the accident aircraft cut the electric line that provides power to the airport including the radio communications system. In another accident

40 (3/9/97, KRN), the air traffic controllers could not contact ARFF because of the firefighters’ failure to reset the alarm from an earlier notification. It is not unusual for the response time to be affected by a combination of communication and weather issues. In the Little Rock accident (6/1/99, KLIT) the controller initially informed the ARFF units that he had an aircraft down on Runway 4R. In limited visibility conditions due to a driving rain storm, the ARFF units proceeded to the approach end of Runway 4R to search for the aircraft. The controller then informed the ARFF units that he had last seen the aircraft rolling out past the mid-point of Runway 4R. When ARFF arrived at the departure end of Runway 4R, they saw where the aircraft was located off the end of the runway but several feet below the runway elevation. The ARFF units had to backtrack to a maintenance roadway that provided access to the accident site. A similar situation happened in Charlotte (7/2/94, KCLT). The controllers knew the aircraft had crashed but did not know its location. With that knowledge ARFF units began to drive on the airfield in search of the aircraft. When the weather improved, the controllers were able to provide the ARFF units with the location of a large plume of smoke. This occurred about the same time the ARFF units heard over the city radio frequency that there was an aircraft crash in the vicinity of a certain highway intersection. One of the best examples of the differences between “demonstrated” and “actual” response times is illustrated by an accident that occurred at Denver International Airport (12/20/08, KDEN). During takeoff in the early evening hours, a Boeing 737 veered off the runway and came to rest approximately 100 to 200 yards behind an ARFF (see photo). The controller initially provided the wrong information on the accident location and failed to properly reset the notification system when he attempted to correct the location information. The vehicle from one of the ARFF stations exited the station towards the initially reported location without seeing the accident located right behind the station. By the time the ARFF units received the correct location information and reached the aircraft, it was ablaze. Fortunately, everyone on board had already evacuated. Incidentally, ARFF units from another station on the airport were the first to reach the accident site.

41 Photo courtesy of Denver International Airport Rapid Response Area The Rapid Response Area (RRA) is a much larger area than either the FAA or ICAO have in their standards. For this reason the research team looked at how many accidents were located beyond the runway end that were outside the FAA’s runway safety area which generally extends 1000 feet beyond the runway end and 250 feet either side of the extended runway centerline. The maximum RRA extends 1650 feet beyond the runway end and 500 feet either side of the extended runway centerline. The actual RRA for a runway extends from the runway end to the airport property line or 1650 feet, whichever comes first. In the approach area there was one aircraft that touched down at 1080 feet from the runway threshold, i.e., 80 feet prior to the runway safety area (1/17/08, LHR). However, the aircraft came to rest 100 feet beyond the runway threshold off to the side of the runway. In the departure area there were two aircraft that went beyond the runway safety area and ended up in the RRA. The one aircraft came to rest 1020 feet beyond the runway end and 100 feet to the right of centerline (7/9/06, IKT). The other aircraft which was involved in a runway collision became airborne and crashed into a building located 1520 feet beyond the runway end and 165 feet to the right of centerline (10/08/01, LIN). While the runway safety area extends 250 feet each side of the runway centerline abeam the runway, the rapid response area extends to 500 feet on each side of the runway centerline abeam the runway. There were seven accidents that were in the RRA but not in the runway safety area. They are listed in Table 10. ARFF Station

42 Table 10. Accidents Abeam the Runway Outside the Runway Safety Area but in the Rapid Response Area (1989-2008) Date Airport ID Type of Accident Distance Off Centerline (ft) Left or Right of Centerline 03/22/1992 KLGA Other 370 ft Left 07/30/1992 KJFK Takeoff-VO 290 ft Left 04/14/1993 KDFW Landing-VO 325 ft Right 04/26/1994 RNGO Other 400 ft Right 06/18/98 CYMX Other 325 ft Left 08/22/1999 HKG Landing-VO 325 ft Right 11/30/2001 ESNS Landing-VO 400 ft Left