Informing post-fire recovery planning of northern NSW rainforests

In historic times rainforests have comprised less than 2% of the total forested area of NSW (Baur 1987) yet they have played a pivotal role in the states early timber industry, and the development of conservation policy, and are the only plant community in NSW to have been included in the name of a World Heritage property. The Gondwanan rainforests evolved in an environment of dependable rainfall and low fire frequency. They have persisted through multiple climate oscillations and glacial/inter-glacial cycles in refugia scattered through the Great Escarpment uplift landform. They now represent a collection of relictual species whose origins can be traced to the Eocene Epoch 55-37 million years ago when cool temperate taxa such as Nothofagus first appears in the fossil record.

In northern New South Wales, temperate rainforests are typically considered to be vulnerable to climate-change related natural hazards such as extended spring droughts and wildfires. Numerous reviews predict that extreme climatic events are likely to increase in frequency and/or intensity in coming decades (CSIRO 2019, 2020) and that wildfires will become larger, more intense, and more frequent. The vulnerability of temperate rainforests to fire is assumed to be the result of their limited intrinsic post-fire regenerative capacity, their susceptibility to canopy scorch and bark damage, and extrinsic factors such as their landscape position as patches within extensive areas of elevated fuel hazard in wet sclerophyll forest and woodland plant communities. The impacts of wildfires on temperate rainforests are heightened because wildfires are naturally infrequent, the vegetation is poorly adapted to fire, and fuel biomass is abundant and contiguous in the surrounding landscape.

During the 2019/20 wildfire season, 18.1% of the extent of NSW’s rainforest vegetation classes experienced either low, moderate, high, or extreme severity wildfire, a scale un-anticipated in earlier land management risk planning. While the proportion of each class varied, the greatest impacts occurred in Dry Rainforest and Northern Warm Temperate Rainforest. Across the Gondwana Rainforest of Australia World Heritage property (Gondwana Rainforests), 16 of the 28 reserves in the NSW section saw some fire impacts. This resulted in 8.6% of the 112,145 ha of rainforest in these reserves experiencing either low, moderate, high, or extreme severity wildfire. The extent of the fire impacts was greatest in the Hastings-Macleay, New England, and Washpool-Gibraltar Range sections of the Gondwana Rainforests. The focus of the current study is the Hastings-Macleay section.

The 2019/20 wildfire season coincided with the lowest annual rainfall recorded in the Hastings-Macleay study area since local records commenced in 1959. Soil moisture deficits in rainforest were at their lowest recorded levels at the end of an extended spring drought in northern NSW. An analysis using data from a decade of soil moisture monitoring indicated that when soil moisture in rainforest reached 10-20%, the rainforests become susceptible to fuel ignition and sustaining a wildfire.

In November 2019 multiple ignitions and running wildfires were occurring across northern NSW, and fire-fighting resources were prioritised to life and property protection. On November 9th 2019, an unprecedented seventeen wildfires were at Emergency Warning level on the same day in NSW, fire behaviour was erratic and rainforests were igniting and sustaining running fires. The wildfires impacted directly on long-term research plots established in 1959 in Werrikimbe and Willi National Parks in landscapes with extensive localised climate monitoring equipment. Resources were rapidly deployed within weeks of the wildfire impacting the research infrastructure to capture key elements of fire behaviour and commence post-fire assessments. With the financial support of the Bushfire and Natural Hazards CRC and NSW NPWS, citizen science volunteers began to measure the impacts of the wildfires and describe the recovery trajectory of the rainforest stands within the parks.

The long-term research plots provide a baseline to understanding pre-fire rainforest stand dynamics and condition. The 60 years of accumulated data highlighted the high biomass, slow growth rates, and highly episodic nature of tree regeneration in these rainforests. Annualized tree mortality was less than 1%, and canopy dieback was associated with previous selective harvesting in the 1960s. Fine fuel biomass is low and inputs are strongly seasonal. Basal fire scars were present on trees across the stands surveyed prior to the 2019/20 wildfire season, indicating past wildfires had occurred in these rainforests during the past century, and it appears the majority of established trees were resilient to their effects. Evidence from char heights indicates the predominant fire behaviour experienced by the larger rainforest trees is low-intensity litter fires that progress slowly through the rainforest stands. This was corroborated in discussions with fire managers who had witnessed fires across multiple northern NSW rainforest reserves. Evidence from archival fire reporting and mapping suggests that the typical fire season in the 1950s peaked in November and now peaks two months earlier. These rainforests are approaching the driest months in early spring when lightning ignitions are also at their seasonal maximum.

Wildfires entered rainforest from front or flank spread leading to attrition of trees on rainforest margins. Monitoring dominant trees for 12 years in smaller rainforest stands demonstrated that they can tolerate multiple fires until their structural integrity weakens and fails. In 2019/20 ember attack was widespread in rainforest, but generally led to the ignition of small, localised areas (<100 m2) or of individual trees. Mortality rates associated with the wildfires were 20-30% across all tree species, with the proportion increasing with the severity of the wildfire event. This figure may prove to be an underestimate as lagged tree mortality effects from wildfires have been reported elsewhere for rainforests. The majority of burnt trees regenerated with basal coppicing. Seedling regeneration of tree species was infrequent due to the lack of a soil or canopy seed bank. Coppicing was more abundant where fire severity was low. Large trees with well-developed buttress hollows and scars were more likely to ignite due to the accumulation of dry wood and fuel at their base compared to smaller trees. Larger, older trees that had more pre-existing fire scars were more likely to die and collapse after burning. This is potentially an age effect where older and larger (diameter and height) trees have been subjected to more fire-induced wounds and hollows.

The temperate rainforests of northern NSW have been subject to wildfires in the past, and have a level of resilience to them. However, the susceptibility of the canopy dominant rainforest trees to basal wounding and collapse is a concern due to the longevity and age of these trees, their role in maintaining the rainforest microclimate, and their importance as habitat to a very diverse cargo of epiphytic bryophytes. Options to manage wildfire risk using conventional zoning and fuel hazard management may be limited. Fuel hazard was not particularly high during the 2019/20 wildfire season in the Hastings-Macleay section of Gondwana Rainforests, with most areas burnt in 2019 also having been burnt in 2013. Regardless, landscape-level fuel hazard management will not address the risk of individual mature rainforest tree ignition during conditions of significant ember attack when rainforest fine fuel biomass is naturally low.

A range of proposals are made to improve the knowledge base to better inform the spatially explicit wildfire risk assessment process trialled following the 2019/20 wildfire season. The central recommendation is the scoping of a systematic Monitoring, Evaluation and Reporting (MER) system for the Gondwana Rainforests of Australia World Heritage property (Gondwana Rainforests) that focuses on establishing a baseline to track trends in the nominated World Heritage outstanding universal values or criteria. The system will require an integrated remote sensing and ground-based plot monitoring approach stratified by rainforest type and location. The monitoring plots established for this study can be readily integrated into the proposed MER system, and can be used to test the significance of ongoing or lagged tree mortality and to validate models of post-wildfire tree mortality and regeneration.

The current study has informed post-fire recovery planning by assessing the significance of the impacts of the 2019/20 wildfire season on rainforest stand structure, composition, tree mortality, and regeneration strategies. While the field measurements were in progress, recovery planning workshops hosted by the Commonwealth and NSW governments were attended and preliminary results were provided. The availability of existing long-term rainforest monitoring experiments dating from the 1950s, sections of which were burnt in November 2019, was critical to providing preliminary results within months of the fire event (Peacock 2020) to inform post-fire recovery planning. This outcome alone underscores the critical role long-term ecological monitoring systems have in rapidly supporting land managers information needs when responding to widespread natural hazards such as landscape-scale wildfires.