Familiar Risks Still Require Careful Decisions

The sixth SACAA General Aviation Accident Reduction Seminar, held on 27 August 2026, examined recurring risks in South African general aviation. Its central message was not that pilots need to relearn familiar theory, but that familiar hazards still have to be managed carefully on every flight.

Density altitude, deteriorating weather, runway limitations and time pressure are not unfamiliar subjects to pilots. They form part of training, flight testing and recurrent safety discussions. Yet the same factors continue to appear in accident and serious-incident investigations, often after several manageable penalties have reduced the available margin.

The practical issue is therefore not whether a pilot can define density altitude or explain the risks of visual flight into instrument meteorological conditions. It is whether that knowledge changes the decision made before and during a particular flight.

What the Occurrence Data Show

Data presented by the SACAA Accident and Incident Investigations Division (AIID) at GAARS indicated that overall accidents fell by 43% from the previous financial year, reaching one of the lowest levels recorded during the preceding five years. However, 2025/26 still included 17 fatal accidents and 24 fatalities.

Between 1 April 2023 and 31 March 2026, AIID recorded 452 accidents and serious incidents. Of these, 275 involved general aviation operations conducted under Parts 91 and 94. These GA occurrences included 35 fatal accidents and 37 fatalities, accounting for 74% of the fatalities represented in the wider dataset.

AIID associated 184 of the GA records with Private Pilot Licence or National Pilot Licence holders. This is an occurrence count, not a risk rate; meaningful comparison between licence categories would also require exposure data such as hours flown.

Of the 275 GA records, 230 occurred during landing or in flight, representing 83.6% of the total. These are phases in which conditions can change quickly and the time available to correct a deteriorating situation may be limited.

AIID’s internal coding attributed 56% of the causal factors in the agreed dataset to flight crew or pilot factors, while 27% related to mechanical or engine factors. These are shares of coded factors, not findings that pilots alone caused the same proportion of accidents. Occurrences can involve several interacting factors.

Experience Does Not Replace a Performance Calculation

A serious incident involving Robinson R44 Raven II ZS-FDV illustrates why familiar performance considerations still require attention.

On 19 February 2026, the helicopter lifted from a stationary mobile trailer platform at a property approximately five nautical miles south-west of Wonderboom Airport. The pilot noticed low engine power and low rotor RPM shortly after lift-off and aborted the departure. During the attempt to return to the platform, the helicopter slipped backwards and its tail boom struck the ground. The aircraft was substantially damaged, but none of the three occupants was injured.

The pilot held a valid Commercial Pilot Licence and had accumulated 1,895.6 flying hours, including 1,751 hours on the R44. A subsequent inspection found no relevant defects.

The environmental and loading figures were central to the investigation. The site elevation was approximately 4,193 ft, the temperature was 32°C and the calculated density altitude was 7,047 ft. The helicopter’s gross weight was 2,342 lb. Under the prevailing conditions, its take-off weight was limited to 2,150 lb, leaving a difference of 192 lb between the actual weight and the weight supported by the applicable performance calculation.

The SACAA report found that the required power exceeded the power available. It identified the loss of rotor RPM during lift-off as the probable cause and the failure to account adequately for high density altitude as a contributing factor.

The value of the case is not in judging the pilot after the event. It demonstrates that experience and familiarity do not alter the available performance. Being within the general weight-and-balance envelope is also not the same as having sufficient take-off, hover or climb performance under the conditions of a particular departure.

When Several Margins Narrow Together

A second case examined the 8 June 2025 accident involving Piper PA-28-181 ZS-KFB. Three aircraft had flown from Wonderboom to Virginia via Ladysmith and Richards Bay the previous day. Strong winds delayed their return departure from Virginia on 8 June.

Their return plan routed via Richards Bay to Ladysmith, where the aircraft were to refuel and one passenger was to disembark. Ulundi was selected as the first alternate and Greytown as the second.

According to the final investigation report, the surviving pilots had not obtained official weather information from the South African Weather Service and instead used an unidentified mobile application. They had also not checked NOTAMs to establish which aerodromes were suitable for night operations and were unaware of restrictions at Greytown despite selecting it as an alternate.

The pilots discussed the fact that daylight would end at 1511Z and that they might not reach Ladysmith in time to land. One aircraft landed successfully. ZS-KFB and ZS-CZU were unable to land after two attempts, by which time it was dark and the unlit runway could no longer be seen. Both aircraft diverted.

ZS-CZU subsequently made an emergency landing near Greytown, with one occupant sustaining minor injuries. Contact with ZS-KFB was lost. Its wreckage was found the following day approximately 3.6 nautical miles north-east of Hylton. All three occupants were fatally injured.

The final report attributed the loss of control to spatial disorientation after external visual references were lost in instrument meteorological conditions. Improper flight planning and a contravention of the regulations governing the privileges of the pilot licence were identified as contributing factors.

The available options did not disappear at once. The margins associated with weather, daylight, aerodrome suitability and diversion planning narrowed progressively. A delay, an unverified alternate, changing weather or another approach may appear manageable in isolation. Together, they can leave fewer viable choices than the original plan assumed.

Complacency Is Better Addressed Before the Flight

Complacency is often used as a broad explanation when a familiar risk is involved. More useful is identifying where familiarity can weaken a decision.

A pilot may have operated from the same airfield many times, but not at the same temperature, weight or wind. A familiar route still requires current destination, NOTAM and weather information. An unidentified mobile application also does not replace authorised information or the approved aircraft flight manual or Pilot’s Operating Handbook.

Experience can improve judgement and handling ability, but it cannot provide performance that is not available or visual references that no longer exist.

The safeguard is a process that makes the decision less dependent on expectation:

• Use current weather, NOTAM and aerodrome information.
• Calculate performance using the actual aircraft weight and prevailing conditions.
• Read the assumptions and limitations attached to the manufacturer’s figures.
• Account for runway surface, slope, wind, obstacles and the condition of the aircraft.
• Decide in advance what indication, position or time will trigger a rejected take-off, go-around or diversion.
• Preserve an alternative while it remains practical, rather than waiting until it becomes necessary.

Weather Is Often Part of a Wider Sequence

AIID recorded 43 weather-related accidents and serious incidents during the three-year period examined. Of these, 36 were accidents, four involved fatalities and 32 occurred in Part 94 operations. Landing and in-flight occurrences together accounted for 77% of the weather-related records.

Wind, crosswind and gusting conditions formed the largest part of the weather-related occurrence profile, with visibility and cloud also contributing. Weather represented only 1% of the internally coded causal factors in the broader dataset, but this should not be read as evidence that it is unimportant. Weather frequently acts as a hazard within a longer operational sequence rather than as the only cause of an occurrence.

South Africa’s summer risks also differ by region. The Highveld interior experiences frequent afternoon thunderstorms, with activity generally peaking between October and March. The east is affected by heat, humidity, summer rainfall, onshore flow and cloud associated with rising terrain. The south and west can still experience frontal weather, strong winds, low cloud, rain and coastal fog.

Pilots therefore need to consider the dominant hazards along the route, at the destination and at each realistic alternate. Weather must also be reassessed as the flight progresses. An acceptable plan at departure can become marginal as temperature rises, thunderstorms develop, wind changes or daylight reduces.

Preserving the Margin

The common thread through the data and case studies is the gradual loss of operational margin. Protecting it requires current information, aircraft-specific calculations and decision points set before workload or time pressure increases.

If a take-off depends on perfect acceleration, precise technique and no further rise in temperature, the available margin may already be too small. If reaching a destination depends on the weather remaining unchanged, the first approach succeeding and an alternate not being required, the plan has little resilience.

SACAA investigations are conducted to identify safety lessons and reduce the likelihood of recurrence, not to assign blame. The recurring nature of these hazards does not make them simple. It makes the discipline of checking them on every flight more important.

Key Points

• Familiar hazards still require aircraft-specific and flight-specific decisions.
• SACAA data shows that landing and in-flight occurrences dominate the GA record examined at GAARS.
• Experience does not replace current performance, weather, NOTAM or aerodrome information.
• Weight-and-balance compliance does not by itself confirm adequate take-off, hover, climb or go-around performance.
• Small penalties can combine until the available options are severely restricted.
• Abort, go-around and diversion triggers are most useful when decided before they are needed.
• Investigation findings should be used to improve future decisions, not to judge those involved after an occurrence.