ANALYSIS OF U.S. GENERAL AVIATION ACCIDENT RATES

NLR-TR-2011-236 Executive summary ANALYSIS OF U.S. GENERAL AVIATION ACCIDENT RATES Derivation of a baseline level of safety for a set of UAS categories Problem area The introduction of civil and military Unmanned Aircraft Systems (UAS) in aviation has led to a policy statement from the European Aviation Safety Agency (EASA) that civil UAS must not increase the risk to people or property on the ground compared with manned aircraft of equivalent category. The United States (US) Federal Aviation Administration (FAA) aims to ensure that UAS do no harm to other operators in the US National Aviation System (NAS) and, to the maximum extent possible, the public on the ground. FAA supports the introduction of UAS in Nonsegregated Airspace provided that the risks of flying the unmanned aircraft in the civil airspace can be appropriately mitigated. What are these risks to be considered and what is the level of safety to be achieved? Description of work The objective is to analyse the historical accident rates of US General Aviation (GA), and to provide insight in the risks associated with different flight phases and different flight rules. This would provide a baseline level of safety for UAS categories equivalent to General Aviation aircraft. This is achieved by performing a database analysis of accident flights operated under Federal Aviation Regulations (FAR) Part 91, while focusing on aircraft with a maximum certified gross weight of less than 19000 lb. Three scenarios are considered: Mid-air collisions; Accidents at or near the aerodrome; Accidents outside the aerodrome. Results and conclusions Flights conducted without a flight plan have a higher accident rate for all three accident scenarios. The accident rate at or near aerodromes is higher than outside aerodromes for flights conducted under Instrument Flight Rules (IFR), Visual Flight Rules (VFR), and for flight without a flight plan. For each accident scenario trend data is derived, as well as baseline levels of GA safety. Report no. NLR-TR-2011-236 Author(s) A.D. Balk Report classification UNCLASSIFIED Date Knowledge area(s) Vliegveiligheid (safety & security) Descriptor(s) UAS Safety Accident risk UNCLASSIFIED

NLR-TR-2011-236 Mid-air collisions: Total: 3.95 10-7 IFR: 7.15 10-8 VFR: 1.41 10-7 No flight plan: 1.47 10-6 Accidents at or near the aerodrome: Total: 4.08 10-5 IFR: 1.49 10-5 VFR: 1.40 10-5 No flight plan: 7.41 10-5 Accidents outside the aerodrome: Total: 1.36 10-5 IFR: 5.79 10-6 VFR: 4.80 10-6 No flight plan: 2.41 10-5 Applicability The results of this study may be used by regulators as baseline levels of safety for UAS categories that are equivalent with manned aircraft in category Certification Specifications (CS) 23. If it is shown that proposed operations with such UAS categories do not exceed this safety level, one might argue that such UAS operations may be introduced, without degrading safety levels. NLR Air Transport Safety Institute UNCLASSIFIED Anthony Fokkerweg 2, 1059 CM Amsterdam, P.O. Box 90502, 1006 BM Amsterdam, The Netherlands Telephone +31 20 511 35 00, Fax +31 20 511 32 10, Web site: http://www.nlr-atsi.nl

NLR-TR-2011-236 ANALYSIS OF U.S. GENERAL AVIATION ACCIDENT RATES Derivation of a baseline level of safety for a set of UAS categories A.D. Balk No part of this report may be reproduced and/or disclosed, in any form or by any means without the prior written permission of NLR. Customer National Aerospace Laboratory NLR Contract number ---- Owner Division Distribution Classification of title National Aerospace Laboratory NLR Air Transport Limited Unclassified Approved by: Author Reviewer Managing department

SUMMARY The introduction of Unmanned Aircraft Systems (UAS) in aviation has led to a policy statement of the European Aviation Safety Agency, which states that civil UAS must not increase the risk to people or property on the ground compared with manned aircraft of equivalent category. A similar policy statement exists in the United States (US). The Federal Aviation Administration (FAA) aims to ensure that UAS do no harm to other operators in the US National Aviation System and, to the maximum extent possible, the public on the ground. FAA supports the introduction of UAS in Non-segregated Airspace provided that the risks of flying the unmanned aircraft in the civil airspace can be appropriately mitigated. The objective is to analyse the historical accident rates of US General Aviation (GA) aircraft, and to provide insight in the risks associated with different flight phases and different flight rules. This would provide a baseline level of safety for UAS categories equivalent to GA aircraft. This is achieved by performing a database analysis of accident flights operated under Federal Aviation Regulations (FAR) Part 91, while focusing on aircraft with a maximum certified gross weight of less than 19000 lb. Three scenarios are considered: Mid-air collisions; Accidents at or near the aerodrome; Accidents outside the aerodrome. The following baseline levels of safety are derived for mid-air collisions: Total flights: 3.95 10-7 IFR flights: 7.15 10-8 VFR flights: 1.41 10-7 Flights conducted without flight plan: 1.47 10-6 The following baseline levels of safety are derived for accidents at or near the aerodrome: Total flights: 4.08 10-5 IFR flights: 1.49 10-5 VFR flights: 1.40 10-5 Flights conducted without flight plan: 7.41 10-5 4 NLR-TR-2011-236

The following baseline levels of safety are derived for accidents outside the aerodrome: Total flights: 1.36 10-5 IFR flights: 5.79 10-6 VFR flights: 4.80 10-6 Flights conducted without flight plan: 2.41 10-5 The analysis shows that flights conducted without a flight plan have a higher accident rate for all accident scenarios reviewed. The accident rate at or near aerodromes is higher than outside aerodromes for all flight plan categories. For the mid-air collision accident rate, an upward trend is noticed since the year 2001, which is particularly reflected in the mid-air collision accident rate for flights conducted without a flight plan. The accident rate at or near the aerodrome shows a downward trend for IFR and VFR flights, whereas the accident rate for flights operated without a flight plan has stabilised. The accident rate outside aerodromes shows a decreasing trend, which is particularly reflected in the downward trend in accident rate for flights conducted without a flight plan. NLR-TR-2011-236 5

CONTENTS ABBREVIATIONS 7 1 INTRODUCTION 8 1.1 Background 8 1.2 Objective 9 1.3 Scope 9 1.4 Document setup 9 2 ANALYSIS OF FAR PART 91 ACCIDENTS 10 2.1 Approach 10 2.2 Data inclusion criteria 11 2.3 Data source 12 3 RESULTS 13 3.1 Mid-air collisions 13 3.2 Accidents at or near the aerodrome 18 3.3 Accidents outside the aerodrome 22 3.4 Synthesis of the results 25 4 CONCLUSIONS 26 5 REFERENCES 28 APPENDIX A FLIGHT PHASE DEFINITIONS 29 APPENDIX B FLIGHT EXPOSURE DATA 32 APPENDIX C U.S. GENERAL AVIATION ACCIDENT RATES 33 6 NLR-TR-2011-236

ABBREVIATIONS CS EASA FAA FAR GA ICAO IFR lb NAS NTSB UA UAS US VFR Certification Specifications European Aviation Safety Agency Federal Aviation Administration Federal Aviation Regulations General Aviation International Civil Aviation Organization Instrument Flight Rules Pounds National Aviation System National Transportation Safety Board Unmanned Aircraft Unmanned Aircraft System United States Visual Flight Rules NLR-TR-2011-236 7

1 INTRODUCTION 1.1 BACKGROUND Unmanned Aircraft Systems (UAS) are increasingly being used for either civil and military purposes, or a combination thereof. The European Aviation Safety Agency (EASA) defines UAS as: individual system elements consisting of an unmanned aircraft, the control station and any other system elements necessary to enable flight. The International Civil Aviation Organization (ICAO) definition states: an unmanned aerial vehicle is a pilotless aircraft, in the sense of Article 8 of the Convention on International Civil Aviation [ICAO Convention], which is flown without a pilot-in-command on-board and is either remotely and fully controlled from another place (ground, another aircraft, space) or programmed and fully autonomous [ICAO Doc 9854]. The following ICAO terminology is presently used [ICAO Circular 328]: Unmanned Aircraft (UA). An aircraft which is intended to operate with no pilot on-board; Unmanned Aircraft System (UAS). An aircraft and its associated elements which is operated with no pilot on-board. To ensure that UAS do not increase the risk to people or property, EASA has developed a policy statement which includes the following airworthiness objective: A civil UAS must not increase the risk to people or property on the ground compared with manned aircraft of equivalent category. [EASA, E.Y013-01] A policy statement from the FAA aims to ensure that UAS do no harm to other operators in the US NAS and, to the maximum extent possible, the public on the ground [Sabatini]. FAA supports the introduction of UAS in non-segregated airspace provided that the risks of flying the unmanned aircraft in the civil airspace can be appropriately mitigated [Kalinowski]. In order to determine the maximum acceptable risk for operations with UAS categories equivalent to Certification Specifications (CS)-23 manned aircraft, it is necessary to analyse the accident rates of General Aviation, and to obtain insight in the risks associated with the different flight phases and different flight rules. 8 NLR-TR-2011-236

1.2 OBJECTIVE The objective is to analyse the historical accident rates of US General Aviation aircraft, using a database of accidents with FAR Part 91 operated aircraft that are similar to CS-23 aircraft, and to provide insight in the risks associated with different flight phases and different flight rules. 1.3 SCOPE The EASA policy statement relates to airworthiness and restricts itself to risks to people and property on the ground. There may, however, additional risks be encountered during the actual operation of UAS. Therefore the scope of this study has been broadened to include risks to people both on the ground and in the air. 1.4 DOCUMENT SETUP Chapter 2 of this document describes the approach that has been taken in this study and what data source has been used. Chapter 3 provides the analysis results, from which conclusions are drawn in Chapter 4. NLR-TR-2011-236 9

2 ANALYSIS OF FAR PART 91 ACCIDENTS 2.1 APPROACH In this study, manned aircraft equivalent to UAS are considered to be CS-23 aircraft for Europe. For this category of aircraft, however, there is a lack of detailed accident data. Since the National Transportation Safety Board (NTSB) database provides an extensive source of data for FAR Part 91 operated aircraft, it was chosen to use this database as reference to determine accident rates for manned aircraft equivalent to UAS. There is, however, a difference between CS-23 and FAR Part 91 aircraft. CS-23 describes certification standards for (1) aeroplanes in the normal, utility and aerobatic categories that have a seating configuration, excluding the pilot seat(s), of nine or fewer and a maximum certificated takeoff weight of 5670 kg (12500 lb) or less; and (2) Propeller driven twin-engine aeroplanes in the commuter category that have a seating configuration, excluding the pilot seat(s), of nineteen or fewer and a maximum certificated takeoff weight of 8618 kg (19000 lb) or less. FAR Part 91, on the other hand, prescribes rules governing the operation of aircraft (other than moored balloons, kites, unmanned rockets, unmanned free balloons, ultralight vehicles) within the United States, including the waters within 3 nautical miles of the U.S. coast. In short, CS-23 relates to airworthiness of specific aircraft categories, whereas FAR Part 19 relates to the operation of specific aircraft categories. In the dataset, certain categories of aircraft operated under FAR Part 91 are excluded to best fit the CS-23 aircraft criteria. In order to define the risks to people and properties in the air and on the ground, the following scenarios are considered: Mid-air collisions; Accidents at or near the aerodrome; Accidents outside the aerodrome. 10 NLR-TR-2011-236

The category Accidents at or near the aerodrome comprises all accidents that occurred at the aerodrome, but also accidents that either occurred at the aerodrome or in its vicinity; for example an accident with an aircraft that landed short of the runway. For each scenario, the following data is derived: Total number of accidents; Number of flight hours; Accident rate for total number of accidents; Absolute number of accidents for flights conducted under Instrument Flight Rules (IFR); Number of IFR flight hours; Accident rate for number of accidents for IFR flights; Absolute number of accidents for flights conducted under Visual Flight Rules (VFR); Number of VFR flight hours; Accident rate for number of accidents for VFR flights; Absolute number of accidents for flights conducted without a flight plan; Number of flight hours for flights conducted without a flight plan; Accident rate for number of accidents for flights conducted without a flight plan. 2.2 DATA INCLUSION CRITERIA The following criteria are used to establish the data sample: Only occurrences that are classified as accident are included; Only FAR Part 91 operations are included; Helicopters, gliders, balloons and blimps are excluded; Aircraft with a maximum certified gross weight in excess of 19000 lb are excluded; A mid-air collision between two or more aircraft is considered as a single accident; A collision on the ground between two or more aircraft is considered as a single accident; The time interval is between 1982 and 2010 for mid-air collisions and between 1982 and 2008 for accident at, near or outside the aerodrome (the years 2009 and 2010 are excluded for these scenarios since the flight phase at which the accident occurred is not provided for these years); For the scenario of accidents at or near the aerodrome, the following flight phases are considered applicable: NLR-TR-2011-236 11

Take-off (code 52*) Approach (code 56*) Landing (code 57*) Maneuvering turn to landing area (emergency)(code 583) For the scenario of accidents outside the aerodrome, the following flight phases are considered applicable: Climb (code 53*) Cruise (code 54*) Descent (code 55*) Maneuvering (code 58* excluding 583) Flight phase definitions are provided in Appendix A. 2.3 DATA SOURCE The National Transportation Safety Board (NTSB) determines the probable cause of accidents and issues safety recommendations aimed at preventing future accidents. The NTSB accident/incident database is the official repository of aviation accident data and causal factors. The Database contains data describing the aircraft, operations, personnel, environmental conditions, consequences, the probable cause, and contributing factors of civil aviation accidents within the United States, its territories and possessions, and in international waters. The database comprises three distinct sub-databases, spanning three time periods: 1962 through 1981, 1982, and 1983 to the present. Most data fields, though similar in purpose among the three sub-databases, are incompatible. In the NTSB database, an event is classified as an accident or an incident. "Aircraft accident" means 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. The NTSB defines "Incident" to mean an occurrence other than an accident, associated with the operation of an aircraft, which affects or could affect the safety of operations. The flight exposure data (hours) are obtained from the FAA and included in Appendix B. These data are based on the annual General Aviation and Part 135 Activity Survey conducted by the FAA. 12 NLR-TR-2011-236

3 RESULTS 3.1 MID-AIR COLLISIONS Table 1 shows for the period between 1982 and 2010 the total number of midair collisions, the mid-air collision accident rate, the accident rate for IFR and VFR flights, and the accident rate for flights conducted without a flight plan. Table 1: Mid-air collisions Mid-air collisions Year IFR 1 Rate 2 VFR 1 Rate 2 No flight Rate 2 (acc/flthr) (acc/flthr) plan 1 Unknown Total 3 Rate (acc/flthr) (acc/flthr) 1982 1 1.20E-07 2 3.31E-07 41 3.78E-06 0 24 9.53E-07 1983 0 --- 3 4.36E-07 14 1.14E-06 0 9 3.14E-07 1984 0 --- 2 2.86E-07 34 2.72E-06 0 19 6.53E-07 1985 3 3.21E-07 2 2.94E-07 34 2.79E-06 0 21 7.41E-07 1986 3 3.36E-07 5 7.70E-07 41 3.52E-06 0 26 9.60E-07 1987 2 2.25E-07 4 6.18E-07 29 2.50E-06 0 20 7.42E-07 1988 0 --- 2 3.04E-07 26 2.20E-06 0 15 5.47E-07 1989 3 3.26E-07 0 --- 28 2.33E-06 0 16 5.73E-07 1990 0 --- 4 5.85E-07 24 1.96E-06 0 16 5.61E-07 1991 1 1.09E-07 1 1.51E-07 31 2.60E-06 0 18 6.50E-07 1992 0 --- 4 6.73E-07 16 1.50E-06 0 11 4.44E-07 1993 0 --- 1 1.83E-07 18 1.84E-06 0 11 4.83E-07 1994 0 --- 1 1.87E-07 19 1.99E-06 0 11 4.95E-07 1995 0 --- 1 1.67E-07 26 2.43E-06 0 14 5.62E-07 1996 1 1.22E-07 3 5.02E-07 27 2.52E-06 0 18 7.23E-07 1997 1 1.18E-07 2 3.26E-07 21 1.91E-06 0 13 5.08E-07 1998 1 1.40E-07 3 5.76E-07 20 2.14E-06 0 14 6.45E-07 1999 3 3.66E-07 1 1.68E-07 21 1.96E-06 0 14 5.63E-07 2000 1 1.28E-07 1 1.76E-07 27 2.65E-06 1 18 7.61E-07 2001 0 --- 0 --- 7 6.35E-07 0 5 2.31E-07 2002 0 --- 0 --- 12 1.08E-06 0 7 3.22E-07 2003 0 --- 1 2.15E-07 14 1.29E-06 0 9 4.07E-07 2004 1 2.15E-07 2 3.38E-07 15 1.42E-06 0 9 4.25E-07 2005 1 1.27E-07 0 --- 15 1.95E-06 0 9 4.57E-07 2006 1 1.23E-07 1 2.34E-07 6 7.55E-07 0 6 2.95E-07 2007 1 1.15E-07 0 --- 14 1.87E-06 0 8 3.95E-07 2008 1 1.36E-07 3 6.19E-07 16 2.23E-06 0 11 5.67E-07 2009 0 --- 0 --- 14 2.13E-06 0 8 4.51E-07 2010 0 --- 0 --- 9 1.37E-06 4 7 3.94E-07 1 The number of accidents for the categories IFR, VFR or No flight plan is the number of individual aircraft involved in these collisions; 2 The accident rate is based on the number of individual aircraft involved in these collisions; 3 The total number is not based on the number of individual aircraft but the total number of mid-air collisions. This may involve aircraft of a category that is not in this table; in that case the number of accidents of IFR, VFR and No flight plan does not add up to the Total number of collisions. NLR-TR-2011-236 13

The data provided in Table 1 are visualised in figures 1 to 4. Figure 1 shows the mid-air collision accident rate, irrespective of type of flight plan. Mid-air collision accident rate 1.00E-06 9.00E-07 8.00E-07 7.00E-07 6.00E-07 5.00E-07 4.00E-07 3.00E-07 2.00E-07 1.00E-07 0.00E+00 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 Figure 1: Mid-air collision accident rate In the period between 1982 and 2010, the mid-air accident rate has decreased from 9.53 10-7 to 3.94 10-7 mid-air collisions per flight hour, with a peak in 1986 and trough in 2001. Although the mid-air collision accident rate has improved after 2001 when compared to the preceding period, the period after 2001 shows an upward trend. This upward trend corresponds with data provided in Appendix C for U.S. General Aviation, although the mid-air collision rates per flight hour slightly differ since all FAR Part 91 operated flights have been included in Appendix C (including helicopters, gliders, balloons and blimps). FAR Part 25 aircraft are also included in Appendix C (e.g. ferry flights), but their number is that low, that they only have a minimal effect on the accident rates provided in Appendix C. Due to the perceived difference in the mid-air collision accident rate before and after the year 2001, it seems reasonable to calculate the average mid-air collision accident rate for the period 2001 to 2010 to determine a baseline level of safety. This equals 3.95 10-7 mid-air collisions per flight hour. 14 NLR-TR-2011-236

Figure 2 shows the mid-air collision accident rate for flights conducted with an IFR flight plan. Mid-air collision accident rate - IFR flights 8.00E-07 7.00E-07 6.00E-07 5.00E-07 4.00E-07 3.00E-07 2.00E-07 1.00E-07 0.00E+00 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 Figure 2: Mid-air collision accident rate IFR flights Figure 2 clearly shows that mid-air collisions with aircraft operated under IFR are at least infrequent. When the average mid-air collision accident rate is calculated for the period between 2001 and 2010, this equals 7.15 10-8 mid-air collisions per IFR flight hour. NLR-TR-2011-236 15

Figure 3 shows the mid-air collision accident rate for aircraft conducted with a VFR flight plan. Mid-air collision accident rate - VFR flights 8.00E-07 7.00E-07 6.00E-07 5.00E-07 4.00E-07 3.00E-07 2.00E-07 1.00E-07 0.00E+00 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 Figure 3: Mid-air collision accident rate VFR flights When compared to IFR flights, the mid-air collision accident rate for VFR flights is higher, although it has to be noted that due to the small number of mid-air collisions, a slight increase in the number of mid-air collisions may cause a large increase in the mid-air collision accident rate. When the average mid-air collision accident rate is calculated for the period between 2001 and 2010, this equals 1.41 10-7 mid-air collisions per VFR flight hour. 16 NLR-TR-2011-236

Figure 4 shows the mid-air collision accident rate for flights conducted without a flight plan. Mid-air collision accident rate - No flight plan 8.00E-06 7.00E-06 6.00E-06 5.00E-06 4.00E-06 3.00E-06 2.00E-06 1.00E-06 0.00E+00 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 Figure 4: Mid-air collision accident rate No flight plan When the period between 1982 and 2010 is considered, figure 4 shows a decreasing trend in the mid-air collision accident rate for flights that are conducted without a flight plan, although an upward trend is noticed after 2001. The average mid-air collision accident rate for the period between 2001 and 2010 is calculated as 1.47 10-6 mid-air collisions per flight hour without a flight plan being issued. NLR-TR-2011-236 17

3.2 ACCIDENTS AT OR NEAR THE AERODROME Table 2 shows for the period between 1982 and 2008 the total number of accidents at or near the aerodrome, as well as the division of the total number of accidents between IFR and VFR flights, and flights conducted without a flight plan. Table 2: Accidents at or near the aerodrome Accidents at or near the aerodrome Year IFR Rate Rate No flight Rate Rate VFR Unknown Total (acc/flthr) (acc/flthr) plan (acc/flthr) (acc/flthr) 1982 159 1.91E-05 236 3.90E-05 1683 1.55E-04 16 2094 8.31E-05 1983 151 1.60E-05 238 3.46E-05 1487 1.21E-04 17 1893 6.60E-05 1984 154 1.60E-05 189 2.71E-05 1393 1.11E-04 9 1745 6.00E-05 1985 118 1.26E-05 165 2.43E-05 1255 1.03E-04 9 1547 5.46E-05 1986 123 1.38E-05 138 2.12E-05 1181 1.01E-04 2 1444 5.33E-05 1987 82 9.21E-06 165 2.55E-05 1161 1.00E-04 8 1416 5.25E-05 1988 111 1.23E-05 147 2.23E-05 1090 9.24E-05 1 1349 4.92E-05 1989 117 1.27E-05 133 1.98E-05 1112 9.26E-05 5 1367 4.90E-05 1990 95 1.01E-05 132 1.93E-05 1023 8.34E-05 2 1252 4.39E-05 1991 96 1.05E-05 138 2.08E-05 924 7.76E-05 3 1161 4.19E-05 1992 123 1.50E-05 112 1.88E-05 910 8.54E-05 2 1147 4.63E-05 1993 92 1.22E-05 113 2.07E-05 931 9.50E-05 2 1138 4.99E-05 1994 95 1.29E-05 85 1.59E-05 854 8.93E-05 9 1043 4.69E-05 1995 124 1.51E-05 129 2.16E-05 917 8.56E-05 7 1177 4.73E-05 1996 86 1.05E-05 91 1.52E-05 800 7.48E-05 2 979 3.93E-05 1997 77 9.12E-06 110 1.79E-05 813 7.39E-05 9 1009 3.94E-05 1998 48 6.71E-06 61 1.17E-05 539 5.78E-05 3 651 3.00E-05 1999 45 5.49E-06 53 8.88E-06 593 5.55E-05 3 694 2.79E-05 2000 60 7.68E-06 38 6.69E-06 537 5.28E-05 6 641 2.71E-05 2001 96 2.02E-05 88 1.51E-05 806 7.31E-05 9 999 4.62E-05 2002 107 2.24E-05 104 1.77E-05 798 7.21E-05 10 1019 4.69E-05 2003 102 1.54E-05 88 1.90E-05 822 7.59E-05 12 1024 4.63E-05 2004 92 1.98E-05 54 9.12E-06 700 6.62E-05 15 861 4.07E-05 2005 88 1.12E-05 68 1.64E-05 685 8.92E-05 9 850 4.32E-05 2006 111 1.36E-05 41 9.59E-06 514 6.47E-05 14 680 3.34E-05 2007 87 9.99E-06 70 1.73E-05 661 8.82E-05 14 832 4.11E-05 2008 48 6.52E-06 37 7.64E-06 457 6.37E-05 16 558 2.88E-05 The data provided in Table 2 are visualised in figures 5 to 8. 18 NLR-TR-2011-236

Figure 5 shows the accident rate at, or near aerodromes, irrespective of type of flight plan. 9.00E-05 8.00E-05 7.00E-05 6.00E-05 5.00E-05 4.00E-05 3.00E-05 2.00E-05 1.00E-05 0.00E+00 Accident rate at or near aerodromes Figure 5: Accident rate at or near aerodromes 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 This figure clearly shows a decreasing trend in accident rate at, or near, aerodromes, with a temporary rise between the years 2000 and 2001, in which the accident rate increased from 2.71 10-5 to 4.62 10-5 accidents per flight hour. After 2001, the decreasing trend is continued. In order to derive a baseline level of safety, and the perceived difference between the accident rate before and after 2001, it is considered reasonable to calculate the average accident rate for the period between 2001 and 2008. This average accident rate is calculated as 4.08 10-5 accidents per flight hour. NLR-TR-2011-236 19

Figure 6 shows the accident rate at or near aerodromes for flights conducted with an IFR flight plan. 2.00E-04 1.80E-04 1.60E-04 1.40E-04 1.20E-04 1.00E-04 8.00E-05 6.00E-05 4.00E-05 2.00E-05 0.00E+00 Accident rate at or near aerodromes- IFR flights Figure 6: Accident rate at or near aerodromes IFR flights 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 This figure shows a more or less stable accident rate at or near aerodromes for IFR flights, with a peak of 2.24 10-5 in the year 2002. The average accident rate between 2001 and 2008 is calculated as 1.49 10-5 accidents per IFR flight hour. Figure 7 shows the accident rate at or near aerodromes for flights conducted with a VFR flight plan. Accident rate at or near aerodromes- VFR flights 2.00E-04 1.80E-04 1.60E-04 1.40E-04 1.20E-04 1.00E-04 8.00E-05 6.00E-05 4.00E-05 2.00E-05 0.00E+00 20 NLR-TR-2011-236 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 Figure 7: Accident rate at or near aerodromes VFR flights 2000 2001 2002 2003 2004 2005 2006 2007 2008

With regard to flights conducted with a VFR flight plan, a slightly decreasing trend is noticed in figure 7 and the accident rate decreases from 3.90 10-5 in 1982 to 7.64 10-6 in 2008. The average accident rate between 2001 and 2008 is calculated as 1.40 10-5 accidents per VFR flight hour, which is more or less comparable with the average accident rate of IFR flights. Figure 8 shows the accident rate at or near aerodromes for flights conducted without a flight plan. 2.00E-04 1.80E-04 1.60E-04 1.40E-04 1.20E-04 1.00E-04 8.00E-05 6.00E-05 4.00E-05 2.00E-05 0.00E+00 Accident rate at or near aerodromes- No flight plan Figure 8: Accident rate at or near aerodromes No flight plan 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Considering the entire period between 1982 and 2008, a decreasing trend is noticed. However, after the year 2000 the accident rate increased and has more or less stabilised between 2001 and 2008. The average accident rate between 2001 and 2008 is calculated as 7.41 10-5 accidents per flight hour without a flight plan being issued. NLR-TR-2011-236 21

3.3 ACCIDENTS OUTSIDE THE AERODROME Table 3 shows for the period between 1982 and 2008 the total number of accidents outside the aerodrome, as well as the division of the total number of accidents between IFR and VFR flights, and flights conducted without a flight plan. Table 3: Accidents outside aerodromes Accidents outside aerodromes Year IFR Rate Rate No flight Rate Rate VFR Unknown Total (acc/flthr) (acc/flthr) plan (acc/flthr) (acc/flthr) 1982 40 4.81E-06 42 6.95E-06 393 3.63E-05 2 477 1.89E-05 1983 48 5.07E-06 57 8.28E-06 429 3.48E-05 5 539 1.88E-05 1984 64 6.66E-06 57 8.16E-06 549 4.39E-05 4 674 2.32E-05 1985 56 5.99E-06 55 8.09E-06 625 5.13E-05 1 737 2.60E-05 1986 64 7.16E-06 52 8.00E-06 526 4.52E-05 0 642 2.37E-05 1987 59 6.63E-06 53 8.19E-06 483 4.16E-05 2 597 2.21E-05 1988 47 5.19E-06 42 6.38E-06 488 4.13E-05 0 577 2.10E-05 1989 40 4.34E-06 36 5.37E-06 367 3.06E-05 1 444 1.59E-05 1990 44 4.68E-06 63 9.21E-06 394 3.21E-05 2 503 1.76E-05 1991 50 5.47E-06 65 9.79E-06 501 4.21E-05 2 618 2.23E-05 1992 70 8.56E-06 61 1.03E-05 408 3.83E-05 2 541 2.18E-05 1993 45 5.98E-06 53 9.69E-06 417 4.25E-05 2 517 2.27E-05 1994 63 8.59E-06 58 1.09E-05 400 4.18E-05 1 522 2.35E-05 1995 49 5.96E-06 49 8.20E-06 363 3.39E-05 2 463 1.86E-05 1996 39 4.75E-06 29 4.86E-06 253 2.36E-05 3 324 1.30E-05 1997 45 5.33E-06 43 7.00E-06 330 3.00E-05 3 421 1.65E-05 1998 29 4.05E-06 48 9.22E-06 248 2.66E-05 1 326 1.50E-05 1999 27 3.29E-06 38 6.37E-06 262 2.45E-05 2 329 1.32E-05 2000 35 4.48E-06 27 4.75E-06 248 2.44E-05 2 312 1.32E-05 2001 36 7.57E-06 33 5.65E-06 278 2.52E-05 9 356 1.65E-05 2002 30 6.28E-06 26 4.43E-06 247 2.23E-05 7 310 1.43E-05 2003 38 5.73E-06 32 6.90E-06 256 2.36E-05 3 329 1.49E-05 2004 47 1.01E-05 17 2.87E-06 218 2.06E-05 2 284 1.34E-05 2005 30 3.81E-06 17 4.11E-06 212 2.76E-05 2 261 1.33E-05 2006 44 5.40E-06 20 4.68E-06 200 2.52E-05 1 265 1.30E-05 2007 33 3.79E-06 32 7.90E-06 199 2.66E-05 3 267 1.32E-05 2008 27 3.67E-06 9 1.86E-06 154 2.15E-05 4 194 1.00E-05 The data provided in Table 3 are visualised in figures 9 to 12. 22 NLR-TR-2011-236

Figure 9 shows the accident rate outside aerodromes, irrespective of type of flight plan. 9.00E-05 8.00E-05 7.00E-05 6.00E-05 5.00E-05 4.00E-05 3.00E-05 2.00E-05 1.00E-05 0.00E+00 Figure 9: Accident rate outside aerodromes Accident rate outside aerodromes 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Figure 9 shows a decreasing trend in accident rate outside aerodromes, ranging from the highest rate of 2.60 10-5 in 1985 to the lowest rate of 1.00 10-5 in 2008. The average accident rate outside aerodromes for the period between 2001 and 2008 is calculated as 1.36 10-5 accidents per flight hour. Figure 10 shows the accident rate outside aerodromes for flights conducted with an IFR flight plan. 2.00E-04 1.80E-04 1.60E-04 1.40E-04 1.20E-04 1.00E-04 8.00E-05 6.00E-05 4.00E-05 2.00E-05 0.00E+00 Accident rate outside aerodromes- IFR flights Figure 10: Accident rate outside aerodromes IFR flights 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 NLR-TR-2011-236 23

The accident rate outside aerodromes for aircraft operated with an IFR flight plan shows little variation in the period between 1982 and 2008. The average accident rate between 2001 and 2008 is calculated as 5.79 10-6 accidents per IFR flight hour. Figure 11 shows the accident rate outside aerodromes for flights conducted with a VFR flight plan. 2.00E-04 1.80E-04 1.60E-04 1.40E-04 1.20E-04 1.00E-04 8.00E-05 6.00E-05 4.00E-05 2.00E-05 0.00E+00 Accident rate outside aerodromes- VFR flights Figure 11: Accident rate outside aerodromes VFR flights 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Similarly as with IFR flights, the accident rate outside aerodromes for aircraft operated with a VFR flight plan shows little variation in the period between 1982 and 2008. The average accident rate between 2001 and 2008 is calculated as 4.80 10-6 accidents per VFR flight hour, which is a little less than for IFR flights. 24 NLR-TR-2011-236

Figure 12 shows the accident rate outside aerodromes for flights conducted without a flight plan. 2.00E-04 1.80E-04 1.60E-04 1.40E-04 1.20E-04 1.00E-04 8.00E-05 6.00E-05 4.00E-05 2.00E-05 0.00E+00 Accident rate outside aerodromes- No flight plan Figure 12: Accident rate outside aerodromes No flight plan 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Figure 12 shows a decreasing trend in the accident rate outside aerodromes for flights conducted without a flight plan, ranging from a peak of 5.13 0-5 in 1985 to a trough of 2.06 10-5 in 2004. In 2008 the accident rate is a little higher at 2.15 10-5 accidents per flight hour without a flight plan being issued. The average accident rate between 2001 and 2008 is calculated as 2.41 10-5 accidents per flight hour without a flight plan being issued. 3.4 SYNTHESIS OF THE RESULTS Table 4 summarises the calculated accident rates for the three scenarios. Table 4: Average accident rates by scenario Scenario Mid-air collision (2001-2010) Accident at or near aerodrome (2001-2008) Accident outside aerodrome (2001-2008) Average accident rates (accidents per flight hour) Total IFR flights VFR flights No flight plan 3.95 10-7 7.15 10-8 1.41 10-7 1.47 10-6 4.08 10-5 1.49 10-5 1.40 10-5 7.41 10-5 1.36 10-5 5.79 10-6 4.80 10-6 2.41 10-5 NLR-TR-2011-236 25

4 CONCLUSIONS The database analysis of accidents with FAR Part 91 operated aircraft considered three scenarios: Mid-air collisions; Accidents at or near the aerodrome; Accidents outside the aerodrome. From the calculation of accident risks of the three scenarios, the following conclusions are drawn. Flights operated without a flight plan have a higher accident rate than IFR and VFR flights for all accident scenarios reviewed. The accident rate at or near aerodromes is higher than outside aerodromes for all flight plan categories. The following baseline levels of safety are derived for mid-air collisions: Total flights: 3.95 10-7 IFR flights: 7.15 10-8 VFR flights: 1.41 10-7 Flights conducted without flight plan: 1.47 10-6 The following baseline levels of safety are derived for accidents at or near the aerodrome: Total flights: 4.08 10-5 IFR flights: 1.49 10-5 VFR flights: 1.40 10-5 Flights conducted without flight plan: 7.41 10-5 The following baseline levels of safety are derived for accidents outside the aerodrome: Total flights: 1.36 10-5 IFR flights: 5.79 10-6 VFR flights: 4.80 10-6 Flights conducted without flight plan: 2.41 10-5 26 NLR-TR-2011-236

The mid-air collision accident rate has decreased since 1982. However, after a significant drop in the year 2001, an upward trend is noticed. This is particularly reflected in the mid-air collision accident rate for flights conducted without a flight plan. IFR flights have the lowest mid-air collision accident rate. Flights conducted without a flight plan have the highest mid-air collision accident rate. The accident rate at or near the aerodrome has decreased until the year 2000. After peaks in 2002 and 2003, a downward trend is noticed for IFR and VFR flights. For flights conducted without a flight plan, the accident rate has more or less stabilised. VFR flights have the lowest accident rate at or near the aerodrome. Flights conducted without a flight plan have the highest accident rate at or near the aerodrome. The accident rate outside aerodromes shows a decreasing trend, which is particularly reflected in the downward trend in accident rate for flights conducted without a flight plan. VFR flights have the lowest accident rate outside the aerodrome. Flights conducted without a flight plan have the highest accident rate outside the aerodrome. NLR-TR-2011-236 27

5 REFERENCES CICTT EASA E.Y013-01 FAR Part 91 ICAO Convention ICAO Circular 328 ICAO Doc 9854 Sabatini Kalinowski CAST/ICAO Common Taxonomy Team; Phase of Flight Definitions and Usage Notes,, Version 1.0.3, April 2011 EASA; Rulemaking Directorate Policy Statement Airworthiness Certification of Unmanned Aircraft Systems (UAS), European Aviation Safety Agency, 2009 FAA; Code of Federal Regulations (CFR) 14 Aeronautics and Space, Part 91 General Operating and Flight Rules ICAO; Convention on International Civil Aviation ICAO; Unmanned Aircraft Systems (UAS), Circular 328- AN/190, ISBN 978-92-9231-751-5 ICAO; Global Air Traffic Management Operational Concept, Doc 9854 N.A. Sabatini, FAA Associate Administrator for Aviation Safety; Statement before the House Committee on Transportation and Infrastructure, Subcommittee on Aviation on UAS Activities, 29 March 2006 N. Kalinowski, FAA Vice President for System Operations Services; Statement on the Role of Unmanned Aerial Systems on Border Security, 15 July 2010 28 NLR-TR-2011-236

Appendix A FLIGHT PHASE DEFINITIONS The following flight phase definitions are provided by the Commercial Aviation Safety Team/International Civil Aviation Organization Common Taxonomy Team (CICTT). Take-off From the application of takeoff power, through rotation and to an altitude of 35 feet above runway elevation. This phase of flight includes the following sub-phases: Takeoff. From the application of takeoff power, through rotation and to an altitude of 35 feet above runway elevation or until gear-up selection, whichever comes first. Rejected Takeoff. During Takeoff, from the point where the decision to abort has been taken until the aircraft begins to taxi from the runway. Initial climb. From the end of the Takeoff sub-phase to the first prescribed power reduction, or until reaching 1000 feet above runway elevation or the VFR pattern, whichever comes first. Climb IFR: From completion of Initial Climb to arrival at initial assigned cruise altitude. VFR: From completion of Initial Climb to initial cruise altitude. Cruise Any level flight segment after arrival at initial cruise altitude until the start of descent to the destination. Descent IFR: Descent from cruise to either Initial Approach Fix (IAF) or VFR pattern entry. VFR: Descent from cruise to the VFR pattern entry or 1000 feet above the runway elevation, whichever comes first. Approach IFR: From the Initial Approach Fix (IAF) to the beginning of the landing flare. VFR: From the point of VFR pattern entry, or 1000 feet above the runway elevation, to the beginning of the landing flare. NLR-TR-2011-236 29

This phase of flight includes the following sub-phases: Initial Approach (IFR): From the IAF to the Final Approach Fix (FAF). Final Approach (IFR): From the FAF to the beginning of the landing flare. Circuit Pattern Downwind (VFR): A flight path (normally 1,000 feet above the runway) which commences abeam the departure end of the runway and runs parallel to the runway in the direction opposite to landing, and terminates upon initiating the turn to base leg. Circuit Pattern Base (VFR): From start of turn at end of downwind leg until the start of the turn for final. Circuit Pattern - Final (VFR): From the start of the turn to intercept the extended runway centerline, normally at the end of base leg, to the beginning of the landing flare. Includes VFR straight-in approaches. Circuit Pattern Crosswind (VFR): A flight path of the VFR traffic pattern, which is perpendicular to the landing runway, crosses the departure end of the runway, and connects with the downwind leg. Missed Approach/Go-Around: From the first application of power after the crew elects to execute a missed approach or go-around until the aircraft reenters the sequence for a VFR pattern (go-around) or until the aircraft reaches the IAF for another approach (IFR) Landing From the beginning of the landing flare until aircraft exits the landing runway, comes to a stop on the runway, or when power is applied for takeoff in the case of a touch-and-go landing. This phase of flight includes the following sub-phases: Flare: Transition from nose-low to nose-up attitude just before landing until touchdown. Landing Roll: After touchdown until aircraft exits the landing runway or comes to a stop, whichever occurs first. Aborted Landing After Touchdown: When an attempt is made to get airborne after touchdown (successful or not). This does not include the take-off portion of a touch-and-go. Manoeuvring Low altitude/aerobatic flight operations. This phase of flight includes the following sub-phases: 30 NLR-TR-2011-236

Aerobatics: Any intentional manoeuvring that exceeds 30 degrees of pitch attitude or 60 degrees of bank, or both, or abnormal acceleration (usually associated with air shows and military flight, or with related training flights). Low Flying: Intentional low-altitude flight not connected with a landing or takeoff, usually in preparation for or during observation work, demonstration, photography work, aerial application, training, sight seeing, ostentatious display, or other similar activity. For rotorcraft, this also includes hovering (not associated with landing or takeoff) and handling external loads. NLR-TR-2011-236 31

Appendix B FLIGHT EXPOSURE DATA Year All hours IFR flight hours VFR flight hours No flight plan flight hours 1982 25194000 8314020 6046560 10833420 1983 28673000 9462090 6881520 12329390 1984 29099000 9602670 6983760 12512570 1985 28322000 9346260 6797280 12178460 1986 27073000 8934090 6497520 11641390 1987 26972000 8900760 6473280 11597960 1988 27446000 9057180 6587040 11801780 1989 27920000 9213600 6700800 12005600 1990 28510000 9408300 6842400 12259300 1991 27678000 9133740 6642720 11901540 1992 24780000 8177400 5947200 10655400 1993 22796000 7522680 5471040 9802280 1994 22235000 7337550 5336400 9561050 1995 24906000 8218980 5977440 10709580 1996 24881000 8210730 5971440 10698830 1997 25591000 8445030 6141840 11004130 1998 21689450 7157519 5205468 9326464 1999 24859100 8203503 5966184 10689413 2000 23662300 7808559 5678952 10174789 2001 21615500 4755410 5836185 11023905 2002 21713250 4776915 5862578 11073758 2003 22097450 6629235 4640465 10827751 2004 21154800 4654056 5923344 10577400 2005 19692555 7877022 4135437 7680096 2006 20368495 8147398 4277384 7943713 2007 20245867 8705723 4049173 7490971 2008 19383950 7365901 4845988 7172062 2009 17732586 6738383 4433147 6561057 2010 17765000 6750700 4441250 6573050 Data includes General Aviation FAR Part 91 operated aircraft, but excludes helicopters and gliders. 32 NLR-TR-2011-236

Appendix C U.S. GENERAL AVIATION ACCIDENT RATES U.S. General Aviation Accident Rates Year Flight Hours Accidents Accident rate per flight hour Mid-air collisions 1 Mid-air collision rate per flight hour 1982 29,640,000 3,233 1.1E-04 27 9.1E-07 1983 28,673,000 3,075 1.1E-04 12 4.2E-07 1984 29,099,000 3,017 1.0E-04 21 7.2E-07 1985 28,322,000 2,739 9.7E-05 22 7.8E-07 1986 27,073,000 2,581 9.5E-05 28 1.0E-06 1987 26,972,000 2,494 9.2E-05 24 8.9E-07 1988 27,446,000 2,388 8.7E-05 18 6.6E-07 1989 27,920,000 2,242 8.0E-05 19 6.8E-07 1990 28,510,000 2,242 7.9E-05 20 7.0E-07 1991 27,678,000 2,197 7.9E-05 21 7.6E-07 1992 24,780,000 2,110 8.5E-05 13 5.2E-07 1993 22,796,000 2,064 9.1E-05 12 5.3E-07 1994 22,235,000 2,021 9.1E-05 12 5.4E-07 1995 24,906,000 2,055 8.3E-05 15 6.0E-07 1996 24,881,000 1,908 7.7E-05 18 7.2E-07 1997 25,591,000 1,840 7.2E-05 15 5.9E-07 1998 25,518,000 1,902 7.5E-05 15 5.9E-07 1999 29,246,000 1,905 6.5E-05 16 5.5E-07 2000 27,838,000 1,837 6.6E-05 18 6.5E-07 2001 25,431,000 1,727 6.8E-05 5 2.0E-07 2002 25,545,000 1,715 6.7E-05 8 3.1E-07 2003 25,998,000 1,741 6.7E-05 11 4.2E-07 2004 24,888,000 1,617 6.5E-05 10 4.0E-07 2005 23,168,000 1,671 7.2E-05 10 4.3E-07 2006 23,963,000 1,523 6.4E-05 8 3.3E-07 2007 23,819,000 1,651 6.9E-05 9 3.8E-07 2008 22,805,000 1,569 6.9E-05 12 5.3E-07 2009 20,862,000 1,480 7.1E-05 9 4.3E-07 2010 20,900,000 1,435 6.9E-05 10 4.8E-07 1 Collision counts for one event FAR Part 25 aircraft are also included (e.g. ferry flights), but their number is that low, that they only have a minimal effect on the accident rates NLR-TR-2011-236 33