Part 91 vs. Part 135 Turbine Fixed-Wing Accident Data
A comprehensive source compendium — U.S.-registered turbine (turbojet + turboprop), fixed-wing aircraft under 14 CFR Parts 91 & 135. Rotorcraft, piston, Part 121 and Part 137 excluded.
1. Why this cut matters (framing data)
The blended Part 135 accident rate mixes three risk pools: turbine fixed-wing charter, piston air taxi, and helicopters (incl. HEMS). In 2019, of the fatal Part 135 accidents, 7 were helicopters, 4 were in Alaska, 2 in Hawaii, 2 servicing oil platforms (Aviation Week, citing NTSB).
The blended Part 91/GA rate is piston-dominated: piston aircraft flew 16.16M of ~29M GA hours in 2024 (FAA Activity Survey via General Aviation News).
Helicopters run ~30% higher accident rates than fixed-wing overall.
Result: headline "Part 135 ≈ 5x worse than airlines" and "GA ≈ 3.9 accidents/100k" figures say almost nothing about turbine fixed-wing operations under either rule.
2. Published rates — Part 91 turbine fixed-wing
2.1 AOPA Air Safety Institute (2005–2014 decade, noncommercial Part 91, type-certificated, Experimental excluded)
| Aircraft | Total rate | Fatal rate |
|---|---|---|
| Turboprop | 1.54 | 0.55 |
| Jet | 0.51 | 0.09 |
Turboprop rate = ~¼ the overall noncommercial fixed-wing rate; fatal rate less than half.
Source: AOPA, "Accident Analysis: Jets vs. Turboprops" (Aug 2017). aopa.org/news-and-media/all-news/2017/august/pilot/jets-vs-turboprops
2.2 Boyd & Stolzer 2016 (peer-reviewed)
Accident Analysis & Prevention 86:18–24. Part 91 turbine airplanes <12,501 lb, 1–2 engines, 1989–2013. NTSB database; ferry/instructional/aerial-application/etc. excluded.
- 551 accidents in study set; ~87% turboprop, remainder turbojet.
- Part 91 turbine accident rate 4.3 per million hours (= 0.43/100k) in 1989–1993, declining >50% by 2009–2013 (Poisson GLM, p<0.001), adjusted for annual turbine fleet hours.
- Decline partly attributed to 6-fold rise in single-engine turbine share of accidents (7% → 46%).
- Risk factors: "inadequate weather planning" OR 2.22 (p=0.036); "Violation FARs/AIM deviation" OR 2.59 (descent below MDA / failure to execute missed approach = largest sub-category).
- ⚠️ Not directly comparable to §2.1 — different inclusion criteria (weight cap, operation-type exclusions).
pubmed.ncbi.nlm.nih.gov/26590507/ · ERAU Commons full text: commons.erau.edu
2.3 NTSB point estimate (via National Safety Council)
Executive/corporate jet, 2021: ~0.26 total accidents/100k hrs, fatal rate zero.
3. Fixed-wing turbine accident counts by operating rule (AIN / Business Jet Traveler)
3.1 The decade table — U.S. (N-registered), 2010–2020 inclusive
| Segment | Part 91 | Part 135 | Part 91K (fractional) |
|---|---|---|---|
| Business jets — fatal accidents / deaths | 32 / 118 | 3 / 18 | 0 / 0 |
| Business turboprops — fatal accidents / deaths | 82 / 239 | 15 / 43 | 0 reported |
Source: AIN analysis reported via Private Jet Card Comparisons (Oct 2021). privatejetcardcomparisons.com/2021/10/18/are-private-jets-safe/
3.2 Recent years (AIN/BJT annual & interim reports)
2022 (U.S.-registered): Business jets: 26 nonfatal accidents (5 under Part 135). Business turboprops: 37 fatalities.
2023 (U.S.-registered, full year): Business jets: nonfatal accidents fell to 11 (vs 26 in 2022); zero Part 135 nonfatal mishaps; one 91K nonfatal (first fractional accident since Nov 2021). A separate AIN analysis found no fatal U.S. Part 135 jet accidents in 2023. Business turboprops: 10 fatal accidents; all but one under Part 91; the one Part 135 fatal was a Pilatus PC-12 air-medical crash (Feb 2023, 5 deaths). Turboprop fatalities: 25 (down from 37).
2024 (U.S.-registered, first 9 months): All turbine bizav (jets + TPs): 10 fatal accidents, 26 deaths (vs 12 / 42 same period 2023). Jets: 4 fatal accidents, 11 deaths in 9 months; the only U.S. Part 135 fatal was the Hop-A-Jet (Ace Aviation) Challenger 604, Naples FL highway landing, Feb 9 2024 — both pilots killed, passengers survived. H1: 7 total jet accidents (vs 22 in H1 2023); 3 fatal, all Part 91. Turboprops: 6 fatal accidents / 15 deaths in 9 months (vs 6 / 19 in 2023). Across 2023–24 nine-month windows, 11 of 12 fatal TP accidents were Part 91. 91K fractional: zero accidents/incidents in both 2023 and 2024 nine-month windows. Full-year 2024 context: 21 business jet fatalities, down from 77 in 2019 (resq.com industry summary).
2025 (first half, worldwide): AIN reported 18 fatal turbine business aircraft accidents worldwide in H1 2025 (headline figure; U.S./rule split not captured here — pull the full AIN report for the breakdown).
3.3 Character of the accidents
- Runway excursions = most common event type: 71 turbine bizav excursions in 2023 (39 jets → 16 accidents, 1 fatal non-U.S. charter; 32 TPs → 12 accidents).
- Loss of control in-flight = #1 cause of fatal accidents (NBAA safety committee).
- Nall Report (28th): among commercial fixed-wing, Part 137 ag flying had the majority of accidents; Part 135 accidents concentrated in takeoff (6) and weather (5); "commercial operations delivered a better safety record than their non-commercial fixed-wing counterparts."
4. Denominators — FAA General Aviation & Part 135 Activity Survey
4.1 Hours & fleet by aircraft type (turbine fixed-wing + context)
| Year | Turbojet hours | Turbojet active | Turboprop hours | Turboprop active | Rotorcraft hours | Total GA hours | Total active |
|---|---|---|---|---|---|---|---|
| 2024 | 5.2M | 16,835 | 2.83M | 11,924 | 2.89M | 29.0M+ | 213,756 |
| 2023 | 4.6M | 16,537 | 2.84M | 10,951 | 2.91M | 28.43M | ~214k |
| 2022 | — | — | — | — | — | ~27M | 209,540 |
| 2017 | — | — | — | — | — | 25.2M | 211,756 |
Turbojet hours jumped ~13% 2023→2024.
Sources: FAA GA & Part 135 Activity Survey CY2023/CY2024 (faa.gov/data_research/aviation_data_statistics/general_aviation), summarized in General Aviation News (Feb 2026); GRA/FAA economic-values analysis for 2017/2022.
4.2 The Part 135 hour split — the FAA's own numbers (2022, exact)
Source: FAA Economic Values report, Section 3 (GRA analysis of the 2022 GA & Part 135 Activity Survey), Tables 3-20 to 3-26. faa.gov/regulations_policies/policy_guidance/benefit_cost/econ-value-section-3-capacity.pdf
Hours by operating rule, all survey aircraft, 2022 (Table 3-21):
| Part 91 | Part 125 | Part 133 | Part 135 | Part 137 | Total |
|---|---|---|---|---|---|
| 20,637,763 | 732,187 | 169,280 | 4,407,875 | 1,018,362 | 26,965,466 |
Part 135 composition (2022): air taxi (on-demand pax + cargo) 3.32M + air tours 0.18M + air medical 0.91M = 4.41M hours. "Air taxi" ≠ "Part 135" — the survey tracks the three components separately.
Air taxi hours × aircraft type, 2022 (Table 3-23) — the denominator matrix:
| Aircraft category | Total hours | Air taxi hours | % air taxi |
|---|---|---|---|
| Turbojet/turbofan ≤12,500 lb | 471,547 | 72,671 | 15% |
| Turbojet/turbofan 12,501–65,000 lb | 3,721,624 | 1,580,643 | 42% |
| Turbojet/turbofan >65,000 lb | 1,051,284 | 417,672 | 40% |
| All turbojet | 5,244,455 | 2,070,986 | 39.5% |
| Turboprop single-engine 1–9 seats | 1,011,061 | 20,062 | 2% |
| Turboprop multi-engine 1–9 seats | 89,007 | 7,771 | 9% |
| Turboprop 10–19 seats | 1,589,563 | 494,741 | 31% |
| Turboprop 20+ seats | 291,214 | 139,198 | 48% |
| All turboprop | 2,980,845 | 661,772 | 22.2% |
| Rotorcraft (all) | 2,506,244 | 244,080 | ~10% |
Independent cross-check (Table 3-26, FAA Operating Specification Subsystem, March 2022): 4,235 unique Part 135-certificated jets → est. 1.62M air taxi hrs (2.32M total hours on those tails); 2,188 Part 135 turboprops → est. 0.59M air taxi hrs (1.03M total). All-type OpSpec estimate 2.78M vs 3.32M survey-based; the FAA itself notes "clear uncertainty" in on-demand activity levels — treat the two methods as bracketing the truth.
Key correction to earlier working assumptions: jets carry the higher Part 135 share (~40% of jet hours), not turboprops (~22%).
Use-category → rule mapping (FAA/GRA): Part 91 = personal + business + instructional + corporate + sightseeing + work; Part 135 = air taxi + air tours + air medical; Part 137 = aerial application; Part 125 = ≥20-seat airplanes outside 121/135/137; Part 133 = rotorcraft external load.
2017 edition comparison: ~2.4M air taxi / ~3.5M Part 135 hours → the Part 135 pool grew ~26% by 2022 (charter boom). 1999: air taxi ~2.4M hrs (~10% of GA).
Per-year raw cross-tabs: Survey Table 3.2 (hours by actual use × aircraft type, annual Excel); Table 3.3 = fractional hours; Table 1.4 = use trend 2012–2023.
4.3 Survey caveats (use these when defending the numbers)
- Voluntary, sample-based; overall response rate 37.1% (2023). Twin-turboprop response among the lowest (~26%); piston rotorcraft similar. Turbine rotorcraft, gliders ~50%.
- No hours published for 2011 (GA & Part 135) — rates can't be computed that year.
- FAA notes uncertainty in air-taxi hour reporting; aircraft owners (not operators) respond.
5. Blended headline rates (context / the numbers your cut corrects)
| Category | Total rate | Fatal rate | Year / note |
|---|---|---|---|
| Part 121 scheduled | 0.188 | ~0 (1 fatality) | 2019: 36 accidents / 19M+ hrs |
| Part 121 scheduled | 0.133 | 0 | 2021 (NSC) |
| Part 121 scheduled | 0.23 | — | 2020 |
| Part 121 long-run | 0.181 avg | ~0.006 (≈1 per 16.3M hrs) | 2001–2017; zero-fatal years: 2002, '11, '12, '14–'17, '20, '21, '23 |
| Part 135 on-demand (blended, incl. helicopters & piston) | 2.036 | — | 2000 |
| Part 135 on-demand (blended) | 2.494 (period high) | — | 2003 |
| Part 135 on-demand (blended) | 0.903 (period low) | 32 fatalities | 2019: 34 accidents / ~3.8M hrs; sawtooth pattern, no clean trend |
| Part 135 on-demand (blended) | 0.76 | 0.201 | 2021 (NSC) |
| GA / Part 91 (blended, piston-dominated) | 3.86 | 0.65 | 2023, 28.4M hrs |
| GA / Part 91 (blended) | 5.91 | — | 2020 |
| GA fatal trend | — | 1.27 (2001) → 0.87 (2016 low) → 0.65 (2023) | NTSB |
| U.S. civil aviation overall | 5.23 | 0.94 | 2011–2020 average |
| Accident share by rule (10-yr, 11,859 total accidents) | Part 91 >93% of accidents; Part 135 = 3.75% of accidents but 6.73% of fatalities | Kryder analysis of NTSB data | |
| Worldwide 2023 (per million flights, IATA-style) | Turboprop hull loss 0.57 / fatality 0.31; Jet hull loss 0 / fatality 0.01 | context only — different denominator | |
6. Crewing effect within turbine (single-pilot vs. two-crew)
Breiling & Associates (2010 report): turbine aircraft certified for single-pilot ops had 3.4x the accident rate and 13x the fatal accident rate of aircraft requiring two pilots (figure propagated widely, incl. MITRE "digital copilot" patents).
AOPA (2005–2014, Part 91):
- Jets: 35 single-pilot accidents, 11 fatal (31% lethality) vs 12% lethality in two-crew jet accidents → death ~2.5x as common in single-pilot jet accidents. Single-pilot = 25% of Part 91 jet accidents.
- Turboprops: 220 single-pilot accidents (79% of all Part 91 TP accidents), 82 fatal (37%) vs 30% in two-pilot TP accidents (+25% excess).
Interpretation for the 91-vs-135 gap: Part 135 pax ops structurally filter toward two-crew, trained, dispatched operations; Part 91 turbine contains the owner-flown single-pilot segment — the highest-risk turbine flying.
7. Helicopter data (what the exclusion removes)
- 2024 U.S. helicopters (all rules): 0.44 fatal/100k (13 fatal accidents — lowest count in 25 years); fatality rate 1.02/100k. 2020 fatal rate was 0.79.
- 2024 split: Part 135 helicopter fatal rate 0.46 vs Part 91 helicopter 0.34 — commercial exceeding private, called "pretty remarkable" (FAA/USHST via Vertical Aviation Intl).
- Part 135 helicopter ops average ~2 fatal unintended-IMC accidents per year, every year (NTSB, Island Express/Kobe Bryant investigation).
- Rotorcraft flew 2.89M hours in 2024 — nearly as many as all turboprops — which is why leaving them in wrecks the blended Part 135 rate.
8. Part 91K fractional (bonus rule, strongest proof anchor)
- Zero fatal accidents, 2010–2020 (AIN).
- Zero accidents/incidents in the 2023 and 2024 nine-month windows; single nonfatal event in 2023 was the first fractional accident since Nov 2021.
- Fractional hours by aircraft type are published annually in Activity Survey Table 3.3.
9. Derived rates — Part 91 vs Part 135 turbine fixed-wing (2010–2020) — v2, FAA-anchored
Method: AIN fatal-accident counts (§3.1) ÷ cumulative 11-year hours, with Part 135 shares anchored to the FAA's actual 2022 matrix (§4.2). Scenario envelope: jet hours 3.7–4.3M/yr; TP hours 2.5–2.9M/yr; jet Part 135 share 30–42% (2022 actual 39.5%; earlier decade lower); TP Part 135 share 20–30% (2022 air taxi 22.2% + fixed-wing medevac allowance); 0.5–0.8M/yr turbine-TP Part 137 ag hours removed from the Part 91 pool. Poisson 95% CIs (Garwood) at central denominators.
| Fatal rate /100k hrs, 2010–2020 | Part 91 (incl. 91K) | Part 135 | Ratio 91:135 (central) |
|---|---|---|---|
| Jets | 0.10–0.14 (central ~0.11; Poisson CI 0.08–0.16) | 0.015–0.025 (central ~0.02; Poisson CI 0.004–0.055) | ~6x |
| Turboprops | 0.41–0.79 (central ~0.54; Poisson CI 0.43–0.67) | 0.16–0.27 (central ~0.20; Poisson CI 0.11–0.33) | ~2.7x |
| Blended turbine FW | 0.22–0.34 (central ~0.26) | 0.06–0.10 (central ~0.08) | ~3.4x |
Validation: central Part 91 TP = 0.54 vs AOPA's independently published 0.55 (2005–14); Part 91 jet 0.10–0.14 vs AOPA's 0.09 (adjacent window). The model reproduces published numbers where they exist — the case for trusting it where they don't.
Statistical robustness: the Part 91 and Part 135 Poisson CIs do not overlap for jets (0.08–0.16 vs 0.004–0.055) or turboprops (0.43–0.67 vs 0.11–0.33). The gap is not a small-numbers artifact.
Direction-of-bias notes (all conservative for the "135 is safer" finding): air-medical fixed-wing fatals (e.g., the Feb 2023 PC-12) sit in the AIN Part 135 numerator while medevac hours are only partially in the denominator band → Part 135 rates biased high, gap understated. Certificated-135 operators' Part 91 positioning-leg crashes land in the Part 91 column, slightly inflating it with commercial-adjacent ops.
What changed vs v1: the FAA matrix flipped the assumed shares (jets ~40% under 135, TPs ~22% — v1 assumed the reverse). Net effect: jet gap widened (~6x, was ~4x), TP gap narrowed (~2.7x, was ~4x), blended ~3.4x essentially unchanged. Superseded v1 figures: 135 TP 0.10–0.14 → 0.16–0.27; 91 blended 0.21–0.25 → 0.22–0.34.
Known gap: total-accident (fatal + nonfatal) rates by rule still need an NTSB CAROL pull — AIN reports nonfatal counts only patchily by rule.
Comparison ladder (fatal/100k): Part 121 ~0.006 ≪ Part 135 jet ~0.02 < Part 135 turbine FW blended ~0.08 < Part 91 jet ~0.11 < Part 135 TP ~0.20 < Part 91 turbine FW blended ~0.26 < Part 91 TP ~0.54 < GA blended 0.65. (Note the nuance: a Part 91 jet outperforms a Part 135 turboprop per hour.)
10. Source registry
Primary / official
- NTSB CAROL database — every accident record; filter: airplane + turbine + FAR part. data.ntsb.gov/carol-main-public/basic-search
- NTSB Aviation Accident Statistics (annual tables) — blended rates by Part, per year (incl. on-demand 135 series). ntsb.gov/safety/data
- NTSB Annual Review of U.S. Civil Aviation Accidents (e.g., ARA1201 for CY2010) — splits Part 135 fixed-wing vs helicopter accident counts per year. ntsb.gov/safety/data/Documents/ARA1201.pdf
- FAA GA & Part 135 Activity Survey — CY pages 2001–2024; Tables 1.3, 1.4, 2.1, 3.2 (hours by actual use × aircraft type — the denominator cross-tab), 3.3 (fractional). faa.gov/data_research/aviation_data_statistics/general_aviation
- FAA Economic Values / Section 3 (GRA analysis) — maps survey use categories → FAR parts; source of the 2.4M air-taxi / 3.5M Part 135 hour figures. faa.gov/regulations_policies/policy_guidance/benefit_cost/econ-value-section-3-capacity.pdf
- BTS General Aviation Profile — long-run hour/fleet series. bts.gov/content/general-aviation-profile
Peer-reviewed
- Boyd & Stolzer (2016), AA&P 86:18–24 — Part 91 turbine rates & precipitating factors.
- Boyd (2020), "A Review of General Aviation Safety (1984–2017)," Aerosp Med Hum Perform 91(2):86–90.
- Samuels & Takawira (2025), systematic review of GA accident factors, J. Air Transport Management — factor themes, no rule-split rates.
- Li & Baker (2007), JAMA 297:1596–98 — foundational GA crash-risk framing.
Industry / trade (the only sources with the 91-vs-135 turbine split)
- AIN annual & interim business-aviation safety reports (ainonline.com) + Business Jet Traveler mirrors — counts by rule, jets vs TPs, U.S. vs non-U.S.
- Private Jet Card Comparisons summary of the AIN 2010–2020 decade analysis. privatejetcardcomparisons.com/2021/10/18/are-private-jets-safe/
- Robert E. Breiling Associates annual turbine business-aircraft accident analyses — single-pilot vs two-crew multipliers.
- NBAA Business Aviation Accident & Incident Analysis dashboard — bizjet/TP by operation type (91/91K/135).
- AOPA ASI Nall/McSpadden Reports + "Jets vs Turboprops" (2017) — published Part 91 turbine rates; commercial fixed-wing chapter.
- Vertical Aviation Intl / USHST — helicopter rates (for the exclusion argument). verticalavi.org/magazine-articles/the-safety-dividend/
- Aviation Week, "Why All the Interest in Part 135 Safety?" (2021) — blended-135 sawtooth + 2019 fatal-accident composition. aviationweek.com
11. Known gaps & honesty notes
- No agency publishes a turbine-fixed-wing rate by operating rule. Every figure in §9 is derived; §2 rates are Part 91 only.
- Denominator uncertainty: survey-based hours, low twin-TP response rate, 2011 hours missing entirely.
- AIN "business jet/turboprop" scope ≠ all turbine fixed-wing (excludes e.g. some cargo, ag-turbine, government ops); NTSB rule classification itself is fuzzy at the edges — certificated 135 operators crash on Part 91 positioning/maintenance legs, which shifts events into the Part 91 column (see Mondor's 2019–2024 dataset work).
- Part 91 column includes owner-flown single-pilot ops and professionally-crewed corporate flight departments — two very different risk pools sharing one rule. Corporate/two-crew Part 91 likely performs near Part 135 levels; the survey's corporate vs business use split (Table 3.2) is the lever to separate them.
- AOPA decade (2005–2014) and AIN decade (2010–2020) overlap but don't match; Boyd covers 1989–2013 with different inclusion rules. Don't mix rate figures across them without flagging the windows.
- Fatal-rate comparisons are robust; total-accident-rate comparisons by rule need a CAROL pull.
12. Next-step queries (to harden the numbers)
- DONE (2022): rule-by-type hour matrix obtained from FAA Economic Values Section 3 (§4.2). Remaining: pull Survey Table 3.2 Excel files for CY2015–CY2024 to build the per-year share series — shares drift, and the 2022 jet share (39.5%) likely overstates the decade average.
- CAROL query: EventType=ACC, AircraftCategory=Airplane, EngineType=Turbofan|Turbojet|Turboprop, FAR=091|135, 2000–2024 → numerators (fatal + total) per year per rule.
- AIN year-end reports 2021–2025 full-year (behind the interim figures above) → complete the annual count series.
- Optional: NBAA dashboard export for an independent count check.