My perspective of the talk
There are 15-20 researchers at UNSW involved in working out the dynamics of bushfire behaviour. The traditional opinion has been that of a quasi steady state with independent foci. There are many examples where this assumption is violated.
Junction Fires - vortex point allows joining of fires to rapidly advance.
Eruptive Fires - with a plume attachment. Where the slope is greater than 25 degrees there is an extra dynamic where there's an interaction between terrain and atmosphere and the plume instead of rising vertically lies over the ground to cause acceleration and this can be irrespective of the prevailing wind.
Vorticity driven lateral spread (VLS) Fires - fires normally travel around 4-5 km/hour whereas it can now travel up a Lee slope greater than 20 degrees across the wind of up to 40% change, at 7 km/hour. These are like a whirlwind and as they reach the crest the fire generated can act like corkscrews going left and right. They are developed where there is a heavy fuel load like in forest, not in grasslands (ie unlikely to develop in Hawker). Part of this complex is Deep Flaming, where spot fires merge leading to violent pyroconvection and firestorm development that the reach higher into the atmosphere with more erratic and unpredictable wind mixing with dry upper atmosphere air and creating down draughts and hence increased ember transport.
The Australian Standard 3959 on bush fires is fundamentally flawed as it doesn't encompass these realities and is being rewritten. It is predicated on radiation causing loss, not encompassing ember attack and pyroconvection.
Although not the province of our speaker, one way of protecting one's house could be to install sprinklers on the roof with independent access to power and water.