Assessing Australian aerodrome weather forecasts

Abstract

This thesis investigates issues important to the aviation industry, quantifying the likelihood of flights arriving in unsafe or inefficient circumstances from inaccurate airport weather forecasts (TAF, TTF and TAF3). Key issues include investigating: the likelihood of recurrence for an accident which occurred at Norfolk Island in 2009; the likelihood of adverse situations during aeromedical flights arriving at Australian remote islands and capital city aerodromes; and the potential impact of TAF3 use for major airport arrivals.

Unsafe situations, or ‘misses’, where flights arrive during unplanned unsuitable weather, may result in landings below safe limits, or aircraft safety incidents. Inefficient situations, or ‘false alarms’, where pilots have planned for adverse weather that does not eventuate, have economic and environmental implications due to extra unnecessary fuel being carried. No suitable existing methodology was identified to achieve the research objectives, requiring development of a new utility-based weather forecast verification model. This model compares simulated decisions of historical aerodrome weather forecast use with actual meteorological conditions. The key element of the model is the measurement of likelihood of unsafe or inefficient situations based on the time prior to arrival (‘time-offset’) that forecasts were available to pilots. Results are calculated using a computer-based implementation of this model.

The odds of a ‘miss’ at Norfolk Island were found to be 1 in 859 when simulating the flight plan that was used in the 2009 accident. These odds were predicted to be 1.8 times lower if the ‘time-offset’ between forecast promulgation and planned arrival time was reduced from 5 to 2 hours, the latter time being at the point of last safe diversion. For aeromedical flights to Australian remote islands and to the five busiest capital city aerodromes, strong linear relationships between ‘miss’ likelihood and ‘time-offsets’ between 0 and 12 hours were identified at all locations except for Melbourne and Sydney, with strong second-order relationships identified for ‘false alarms’. ‘False alarms’ are predicted to be considerably more likely at remote islands compared to capital cities, although there was no discernible trend between these groups for ‘misses’. When using all TAF3 information for planning, the replacement of TTF with TAF3 forecasts at Australia’s five busiest major aerodromes is predicted to reduce ‘miss’ arrivals by 34% (enhancing safety), and increase ‘false alarm’ arrivals (when flight have excessive fuel) by 91%. TAF3 PROB segments are predicted to contribute to at least 80% of this change due to the proportion of time these had an operational effect.