Why review aerial firefighting occurrences in Australia
On 6 April 2020, the Royal Commission into National Natural Disaster Arrangements (RCNDA) issued a notice to give information to the ATSB. Within this notice, was the requirement to ‘describe any key operational and safety challenges encountered in coordinating and responding to fires associated with the use of aircraft and aerial fire fighting techniques’.
In addition, since 2018 the ATSB has commenced six investigations (two completed and four active) involving aerial firefighting aircraft. This included the high profile investigation (AO-2020-007) into the collision with terrain involving a Lockheed C‑130 near Cooma, NSW on 23 January 2020. This number is about one third of all investigations involving aircraft conducting aerial work commenced by the ATSB since 2018.
As a further basis for conducting this research, there were more occurrences involving aerial firefighting aircraft in Australia between July 2019 and March 2020 than any financial year in the study period (Figure 1). This was more than three times the period’s average. In addition, there were two fatal accidents between August 2018 and March 2020, whereas the previous 17 years only had three fatal accidents. Further, the number of occurrences per financial year has increased steadily since 2016–17. Given the amount of recent bushfire activity in Australia (the 2019–20 season is estimated have around four times the aerial firefighting activity than other recent bushfire seasons), these results could be expected and probably do not indicate a significant increase in the risk per flight. However, it does indicate that the reported occurrences associated with aerial firefighting are probably[2] increasing. A more extensive analysis would require exposure data[3] (hours flown or number of flights) over the full study period.
Figure 1: Number of reported aerial firefighting occurrences per financial year[4]

Data sources: ATSB aviation occurrence database
The Bureau of Infrastructure, Transport and Regional Economics (BITRE) publishes the Australian Aircraft Activity report annually which contains results of its survey of all Australian VH‑registered commercial and general aviation aircraft. Since 2014, this report has contained exposure data for all aerial work activities, including firefighting. Since it only includes VH‑registered aircraft, it is only a sample of the entire aerial firefighting activity, which also includes foreign‑registered aircraft. During the study period, over 75 per cent of reported occurrences involved VH‑registered aircraft. Therefore, the BITRE data probably includes a significant proportion of firefighting activity flown within Australia.
Table 1 displays occurrence rates (per 100,000 hours flown or 100,000 flights) for VH‑registered aerial work activities between 2014 and 2018 (years with available BITRE data). Generally, the rates for incidents, serious incidents and accidents are relatively low for aerial firefighting compared to other aerial work activities.
Aerial firefighting had the second highest average fatal accident rate calculated using hours flown (second to agricultural spreading/spraying) and flights (second to agricultural mustering). However, during this period, there was only one fatal accident involving a VH‑registered aerial firefighting aircraft, therefore there is a high level of statistical uncertainty associated with the aerial firefighting fatal accident rates.
Table 1: Occurrence rates for VH-registered aircraft conducting aerial work (2014–2018)[5]

Data sources: Bureau of Infrastructure, Transport and Regional Economics (BITRE) and ATSB aviation occurrence database. Bars display the average (expected value) of the rates, value ranges (95% confidence interval) calculated using gamma‑Poisson (hours flown) and beta‑binomial distributions (flights)
Where are they happening
Since July 2000, all fatal accidents and around 40 per cent of accidents, serious incidents and incidents occurred in New South Wales (Figure 2). The majority of the remaining occurrences were in Victoria and Western Australia. Additionally, almost half of the occurrences (incidents, serious incidents, accident and fatal accidents) since July 2000 occurred in the last five financial years (2014–15 onwards).
These statistics are probably correlated with the amount of aerial firefighting activity in these states and may not indicate an increased risk to aircrews. Exposure data for aerial firefighting activity within each state was not available to the ATSB for this report.
Figure 2: Number reported aerial firefighting occurrences by State

Data source: ATSB aviation occurrence database
Figure 3 displays the location of reported occurrences involving aerial firefighting aircraft around Australia (July 2000 to March 2020). Indicative of where most bushfire activity is expected to occur, this shows that the majority of aerial firefighting occurrences happened along the eastern coast, stretching from southern Queensland to central Victoria, with additional clusters in Tasmania, and around Perth and Adelaide.
Figure 3: Locations of reported aerial firefighting occurrences

Data source: ATSB aviation occurrence database. Map source: Bing
Registration type
Over the study period, around three quarters of aerial firefighting occurrences involved an Australian VH‑registered aircraft. Foreign‑registered aircraft made up the bulk of the remaining occurrences (16%); followed by aircraft where the registration details were unknown (7.5%) and a significantly smaller proportion (0.6%) involved remotely piloted aircraft systems (RPAS).
Figure 4 shows that the proportion of occurrences involving foreign‑registered aircraft increased significantly (p<0.01)[6] from July 2000. In the financial year 2019–20, foreign‑registered aircraft were involved in two thirds of more severe occurrences (serious incidents, accidents and fatal accidents). This is probably correlated with an increased use of foreign‑registered aircraft conducting aerial firefighting within Australia.
Figure 4: Proportion of aerial firefighting occurrences involving Australian VH‑registered and foreign‑registered aircraft

Data source: ATSB aviation occurrence database
Aircraft age
As of March 2020, specific details for every aircraft conducting aerial firefighting within Australia were unknown. Therefore, as a proxy, the average age[7] was calculated from aircraft that were involved in a reported occurrence with a known year of manufacture (around 73% of occurrence aircraft).
At the time of the occurrence, the average age of an aircraft conducting aerial firefighting involved in a reported occurrence was 23 ± 1 years. The average age of an Australian VH‑registered aerial firefighting aircraft was 23 ± 1 (131 aircraft, 90% with known aircraft age) years and the average age of a foreign‑registered aircraft was 35 ± 3 years (28 aircraft, 32% with known aircraft age).
Considering aircraft type, the average age of an aeroplane conducting aerial firefighting involved in a reported occurrence was 22 ± 2 years (72 aircraft, 82% with a known aircraft age). For helicopters, the average age was 25 ± 2 years (98 aircraft, 70% with known aircraft age).
Figure 5 appears to indicate that the average age of an aerial firefighting aircraft involved in a reported occurrence is greater than the average for all VH‑registered aerial firefighting aircraft. However, the high level of statistical uncertainty (p=0.35) associated with this comparison makes this level of inference about the data imprecise.
Further, the occurrence data indicates that it is highly likely (p=0.01) the average age of an aircraft conducting aerial firefighting involved in a reported occurrence increased over the study period. However, the BITRE data does not indicate a significant increase (p=0.22) in the average age of VH‑registered firefighting aircraft between 2014 and 2018.
Figure 5: Average age of aerial firefighting aircraft involved in a reported occurrence compared (Jul 2000–Mar 2020) with VH‑registered aerial firefighting aircraft (2014–2018)

Data sources: Bureau of Infrastructure, Transport and Regional Economics (BITRE) and ATSB aviation occurrence database. Error bars display the standard error
Aircraft size
Another potential indicator of a change in the safety risk associated with aerial firefighting is the average aircraft size (presented by the aircraft’s maximum take-off weight – MTOW). As discussed above relating to the average aircraft age, at the time of writing there was limited data available to the ATSB concerning the entire fleet of aerial firefighting aircraft operating in Australia. Instead, the MTOW of aircraft involved in a reported occurrence was used.
Over the period, the average MTOW of an aerial firefighting aircraft involved in a reported occurrence, where the MTOW was known, was 6,600 ± 1,100 kg. The average for VH‑registered aircraft was 3,400 ± 200 kg (131 aircraft, all with known MTOWs) and the average for a foreign‑registered aircraft was 32,000 ± 7,000 kg (28 aircraft, 61% with known MTOWs). This indicates that average MTOW of a foreign‑registered aircraft conducting aerial firefighting involved in an occurrence is probably around 10 times greater than VH‑registered aircraft.
Considering aircraft type, the average MTOW of a helicopter involved in a reported occurrence was 4,000 ± 400 kg (98 aircraft, 86% with known MTOWs) whereas the average aeroplane MTOW was 10,000 ± 2,400 kg (72 aircraft, 92 % with known MTOWs).
Figure 6 indicates that it is highly likely (p=0.01) that the average MTOW of a firefighting aircraft involved in an occurrence increased over the study period. Excluding 2019–20 as an outlier,[8] it is still highly probable (p=0.049) that the average MTOW increased.
Figure 6: Average maximum take-off weight of aerial firefight aircraft involved in a reported occurrence

Data source: ATSB aviation occurrence database. Error bars display the standard error
Aircraft type
BITRE publishes exposure data (hours flown and flights) for commercial and general aviation activities involving Australian-registered aircraft within its annual Australian Aircraft Activity report. As of March 2020, exposure data was available for (VH-registered) aeroplanes and helicopters conducting aerial firefighting (Figure 7).
From this data, between 2014 and 2018, around two thirds of firefighting activity was conducted using helicopters. However, while helicopter activity remained relatively constant, it is very likely that the hours flown (p=0.03) and number of flights (p=0.09) conducted by firefighting aeroplanes increased between 2014 and 2018.
Figure 7: Hours flown and number of flights for VH-registered aerial firefighting aeroplanes and helicopters (2014–2018)

Data source: Bureau of Infrastructure, Transport and Regional Economics (BITRE)
Figure 8 displays the proportion and number of aerial firefighting occurrences involving aeroplanes and helicopters per financial year. It is likely (p=0.06) that the number of occurrences involving firefighting helicopters increased over the study period, in contrast there was no significant trend detected from the number of occurrences involving aeroplanes.
Over the 20-year study period, helicopters and aeroplanes were each involved in 49 per cent of incidents. The remaining consisted of occurrences where the aircraft type was unknown (1.2%) and RPAS (0.6%). Concerning more severe occurrences (serious incident, accidents and fatal accidents), around 72 per cent involved helicopters with aeroplanes making up the remaining 28 per cent. However, this largely reflects the higher activity levels of helicopters and the rates of reported occurrences per 100,000 hours flown and flights (for 2014–18) were consistent between VH‑registered aeroplanes and helicopters (Figure 9).[9]
Figure 8: Proportion of aerial fighting occurrences involving aeroplanes and helicopters

Data source: ATSB aviation occurrence database
Figure 9: Rate of reported occurrence for VH-registered aeroplanes and helicopters conducting aerial firefighting (2014–2018)

Data sources: Bureau of Infrastructure, Transport and Regional Economics (BITRE) and ATSB aviation occurrence database. Bars display the average (expected value) of the rates, error bars (95% confidence interval) calculated using gamma‑Poisson (hours flown) and beta‑binomial distributions (flights)
Aircraft model
In Figure 10, for brevity and ease of comparison, aircraft models that were variants within the same family were grouped. Bell 212/412, UH‑1 and 204/205 aircraft were separated due to their different operational characteristics. Appendix C contains a list of the groupings of each specific aircraft model.
Figure 10 should not be used to compare the relative safety risk between aircraft models. A more extensive analysis would incorporate exposure data (hours flown or number of flights) for each aircraft model to utilise occurrence rates for the comparison.
Over the study period, the Airbus AS350 family of helicopters were involved in more reported occurrences (13%) than any other aircraft model (Figure 10). This was followed by the Bell 206 family helicopters (9%) and the PZL‑Mielec M18 Dromader family of aeroplanes (8.5%).
Airbus AS350 family helicopters were involved in 18 per cent of more severe occurrences (serious incidents, accidents and fatal accidents); this was followed by Bell 212/412 family helicopters (14%) then the M18 family aeroplanes and Bell UH-1 family helicopters, both accounting for 12 per cent of more severe occurrences.
Figure 10: Number of reported aerial firefighting occurrences per aircraft model

Data source: ATSB aviation occurrence database
Engine type
Over the study period, almost half of more severe occurrences involved a helicopter with a turboshaft engine (Figure 11). However, when incorporating exposure data (hours flown and flights) for VH‑registered aircraft between 2014 and 2018 (Figure 12), the occurrences rates are consistent with the other aircraft/engine types. Helicopters with piston engines had almost double the rate of more severe occurrences than turboshaft helicopters.
A more extensive analysis would incorporate exposure data of all aircraft conducting aerial firefighting aircraft in Australia.
Figure 11: Number of reported aerial firefighting occurrences per engine type

Data source: ATSB aviation occurrence database
Figure 12: Rate of reported occurrences for VH-registered aeroplanes and helicopters conducting aerial firefighting per engine type (2014–2018)

Data sources: Bureau of Infrastructure, Transport and Regional Economics (BITRE) and ATSB aviation occurrence database. Bars display the average (expected value) of the rates, error bars (95% confidence interval) calculated using gamma‑Poisson (hours flown) and beta‑binomial distributions (flights)
Occurrence types
Occurrences are often the result of a complex set of circumstances, involving multiple events and conditions. The ATSB categorises each reported occurrence into one or more occurrence types to identify what happened, and how the sequence of events developed to lead to an occurrence. Classifying occurrences in this way helps to understand what types of occurrences have taken place, and to identify potential areas for safety improvement and communication.
Occurrence types do not explain why an occurrence happened; they are generally a description of what occurred. The ATSB uses a three-level hierarchical structure to classify occurrence types. There are broad occurrence type categories (level 1). These are:
- airspace-related
- infrastructure-related
- environment-related
- operational-related
- technical-related.
Consequential events that happen as the result of an occurrence, for example forced and precautionary landings, emergency descents, rejected take-offs, evacuations and fuel dumps to reduce landing weight, are also recorded.
The five level 1 occurrence types are broken down further into different level 2 occurrence types, which are further broken down into level 3 occurrence types. These are detailed in the ATSB’s SIIMS Occurrence Type Coding Manual. The ATSB records one or more occurrence types for all aircraft involved in each occurrence. More severe occurrences (serious incidents, accidents and fatal accidents) generally have more occurrence types coded than incidents, as they are more likely to be investigated, and their severity usually means that there is a greater amount of information to draw upon for analysis and coding.
The frequency of a particular occurrence type does not necessarily reflect its importance or safety risk. For example, fuel-related events may be relatively rare, when compared with fumes events, but fuel starvation is more likely to lead to an accident. Many fuel starvation events result in an attempt at an emergency landing, and potential aircraft damage and injury to people on board or outside the aircraft. In comparison, most fumes-related events are minor in nature, and do not affect the safety of flight, or result in any injuries.
Figure 13 displays the number of reported occurrences per level 1 and level 2 occurrence type; Table 2 also includes level 3 occurrences types.
Note: Occurrences can have more than one associated occurrence type. Figure 13 and Table 2 display the number of occurrences and not the number of aircraft involved in an occurrence, if there were more than one firefighting aircraft involved in an occurrence (for example aircraft separation), it would count as one occurrence. In addition, only occurrence types involving one or more aerial firefighting aircraft are included.
Over the study period, over half of reported occurrences, all fatal accidents and four fifths of more severe occurrences were operational in nature. Terrain collisions (level 2) accounted for around half of the operational occurrences and 71 per cent of the more severe occurrences. Additionally, around one quarter of operational more severe occurrences involved aircraft control (level 2).
Around one quarter of more severe occurrences involved technical issues (level 1). Over half of these were powerplant/propulsion related (level 2) typically engine failure or malfunction (level 3). A further 23 per cent of occurrences were airspace-related; the majority of these were aircraft separation (level 2), of which 28 per cent were near collisions (level 3).
Around one quarter of occurrences and 30 per cent of more severe occurrences resulted in a consequential event, generally forced/precautionary landings (level 2).
Figure 13: Number of reported aerial firefighting occurrences per occurrence type (level 1 and 2)

Data source: ATSB aviation occurrence database
Table 2: Number of reported aerial firefighting occurrences per occurrence type (all levels)

Data source: ATSB aviation occurrence database
Comparison with other low‑level aerial work activities
To examine the risks associated with aerial firefighting in addition to those inherent with low‑level flying,[10] the occurrence rates for each associated occurrence type[11] between VH‑registered aerial firefighting aeroplanes and helicopters[12] and the combined results for other VH‑registered low‑level flying aerial work (agricultural mustering, agricultural spreading/spraying and survey/photographic) aeroplanes and helicopters were compared between 2014 and 2018.
It is highly likely (greater than 95%) that the rate of reported occurrences involving a VH‑registered aerial firefighting aircraft, compared to other low‑level flying aerial work activities, was greater for:
- communications‑related occurrences
- encounters with remotely piloted aircraft (RPA)
- airframe‑related technical issues
- flight preparation/navigation operational occurrences
- aircraft separation occurrences
- operational non‑compliance occurrences.
It is highly likely that the rate of reported terrain collisions was lower for VH‑registered aerial firefighting aircraft than other low‑level flying aerial work activities. Further, it is also likely that aerial firefighting aircraft had a lower rate of reported occurrences associated with aircraft control.
Figure 14: Rate of reported aerial firefighting occurrences per occurrence type for VH‑registered low‑level aerial work activities (2014–2018)

Data sources: Bureau of Infrastructure, Transport and Regional Economics (BITRE) and ATSB aviation occurrence database. Bars display the average (expected value) of the rates, error bars (95% confidence interval) calculated using gamma‑Poisson (hours flown) and beta‑binomial distributions (flights)
__________