Fire intensity effects on post-fire fuel recovery in Eucalyptus open forests of south-eastern Australia.

This is a study of the re-accumulation of bushfire fuels following both prescribed fire of low fireline intensity (<700 kW m) and wildfire of high intensity (>10,000 kW m) in Australian Eucalyptus open forests of differing annual rainfall. Repeated measurements over 5 to 7 years of litter, elevated fuels, coarse woody debris, and bark revealed more rapid fuel recovery in higher rainfall forests compared with lower rainfall forests, following prescribed fire. In prescribed-burnt forests with mean annual rainfall 900-950 mm all fuel categories recovered to very high within seven years, with elevated fuels exceeding pre-fire loads by up to 200%. No fuels in prescribed-burnt forests with mean annual rainfall 600-650 mm recovered to pre-fire loads after six years suggesting that rainfall is an important driver of the rate of fuels recovery. High intensity wildfire in lower rainfall forests (600-650 mm) stimulated the rapid recovery of elevated fuels to over 600% of pre-fire loads - effectively transforming open forest formations into shrublands over the 6 years after fire. The recovery of elevated fuels following both prescribed fire in high rainfall forests and wildfire in low rainfall forests did not follow a gradual negative exponential increase often approximated by an Olson curve, but peaked early after fires. This suggests that the Olson recovery function, the default for predicting loads for these fuels in the operational fire behaviour models in use in south-eastern Australia, may not be appropriate in all cases. Fire simulations were run for forests burnt in wildfires using default (forest) and observed (shrubland) vegetation types. Under weather conditions similar to the previous wildfire, predictions for fireline intensities and the rate of spread would be at least 50% greater in transitional shrubland than forest, emphasizing the importance of accounting for vegetation dynamics for safe response management.