Soil and fuel moisture precursors of fire activity during the 2019-20 fire season, in comparison to previous seasons

Australia experienced unprecedented bushfires during the 2019-20 fire season. Millions of hectares were burned, almost 3,000 homes were destroyed, there were 26 fatalities, and communities were exposed to smoke for extended periods.  Low soil and vegetation moisture content due to antecedent dry conditions were a key driver of fire activity.

This report extends the work in two Bushfire and Natural Hazards CRC projects to examine the 2019-20 fire season:

• Mitigating the effects of severe fires, floods and heatwaves through the improvements of land dryness measures and forecasts: this project was conducted by the Bureau of Meteorology and examined soil moisture using a system called JASMIN. The research team produced soil moisture datasets at a range of depth levels that are consistent with Australia's weather and climate prediction system.
• Mapping bushfire hazards and impacts: this project was conducted by Australian National University and examined vegetation moisture using the Australian Flammability Monitoring System (AFMS). The project has produced live fuel moisture mapping from satellite data.

The national spatial datasets produced by these projects are consistent and regularly updated, separately examining soil and vegetation moisture variability, as both quantities influence fuel availability and therefore fire behaviour.

This project explores the overlaps between these data sets, as well as comparing other available data (including the Australian Landscape Water Balance) to identify atmospheric, soil and fuel moisture characteristics that contributed to fire risk during Black Summer.

Spatially coherent historical and regularly updated soil and fuel moisture datasets are critically important to fire managers so that an accurate, quantifiable assessment of potential fire activity in the landscape can be made, and this can be benchmarked against historical conditions. This information is crucial to making accurate assessments of risk and identifying situations that fall outside of historical records. One of the valuable aspects of JASMIN is that it can also be used in a predictive way and is therefore useful in anticipating soil moisture and therefore fire risk into the future in a quantitative manner on weather (short-term) and seasonal timescales. Research work in the 'Land Dryness' project also demonstrated the capacity of soil moisture to predict future vegetation moisture content and thus more directly anticipate fire potential.

This project examines the interaction of atmospheric parameters with soil and fuel moisture content over scales of days, weeks to months, and years. Such an integration has not been undertaken previously. Consideration of each of these timeframes is important when making accurate assessments of fire risk. The interaction between them is also critical, as a compounding of the individual processes at the different scales occurs and this was seen in the cumulative effects of antecedent dry years, low winter rainfall and heatwaves during the Black Summer. 

This project examined six fire events that were nominated for investigation by the relevant jurisdictions. The fires are:

• Badja Forest Rd, Courtegany- New South Wales
• Stanthorpe – Queensland
• Corryong – Victoria
• Kangaroo Island – South Australia
• Yanchep – Western Australia
• Orroral Valley – Australian Capital Territory

The scope of the project included examining the following:

• multi-annual to annual variation of rainfall distributions
• multi-week influence of soil moisture on live fuel moisture content
• the multi-day influence of heatwaves on vegetation moisture
• soil and fuel moisture properties that distinguished 2019-20 from the previous two seasons.

Key findings of the project include: 

• Antecedent rainfall over most firegrounds was exceptionally low on timescales from months to years.
• High seasonal temperatures set a broad scene for shorter timescale heatwaves.
• Hot conditions were also characterised by anomalously low atmospheric moisture close to the time of several fires (Stanthorpe, Corryong and Orroral Valley).
• Heatwaves (measured by 'Excess Heat Factor') were associated in most fire locations with a perceptible decrease in live fuel moisture content immediately prior to fire occurrence.
• Strong correlation is seen between the two soil moisture metrics AWRA-L and JASMIN over the domains in which fires occurred (although the metrics have different underpinning physical basis), confirming similar findings during the 'Land Dryness' project.
• Spatial monitoring of soil and fuel moisture presents an accurate mechanism for quantifying potential fire risk.
• Drought Factor has limited representation of moisture deficit in dry conditions.
• Rapid falls in live fuel moisture can occur with the onset of warmer weather in dry conditions.
• The magnitude of simulated soil moisture content is highly land surface model dependent, so further refinement of numerical weather prediction model land surface schemes will benefit fire management.
• The 2018-19 and 2017-18 seasons were dry and hot, but not as extreme as 2019-20.
• The dominant drivers are not the same at all fires. For example, rainfall deficits were historically low at Stanthorpe, but not at Kangaroo Island, indicating that conditions could have been worse in the latter event.  

The findings confirm impressions of fire practitioners regarding the extremity of conditions antecedent to the 2019-20 fire season. Importantly, the datasets used in the project present a measurable and spatially coherent approach to estimating fire risk from observed and modelled soil and fuel moisture. Operational application of the datasets and approaches used here will assist in producing accurate soil and vegetation moisture forecasts for prediction of fire risk in the future.

Some immediate benefits for fire managers of this work include:

• AWRA-L provides comparably valuable root-zone soil moisture information to JASMIN. AWRA-L is currently available on a daily basis (e.g. at http://www.bom.gov.au/water/landscape/#/sm/Actual/day/-28.4/130.4/3/Point////2021/6/14/). In particular, relative values of AWRA-L soil moisture provide an historical context for values on several timescales, allowing a qualitative assessment of fire risk in the environment.
• On an annual and area-integrated basis, mean levels of fuel moisture content can change substantially, especially in forested areas of eastern Australia. By implication, successive years of dry conditions can lead to dramatic downward changes in fuel moisture content, and thus heightened fire risk.
• Related to these points, and as noted in the Key Findings above: if soils are very dry in the cool season, even if fuel moisture content is close to normal at that time, the fuel moisture can rapidly decrease coinciding with the onset of much warmer conditions. This observation can act as a useful warning indicator for fire managers of potential rapid onset of increased fire risk.

Each of the meteorological and moisture variables investigated in this study contributed useful information to an understanding of increased fire risk at each site during the 2019-20 Australian summer. Fuel moisture content provided perhaps the most immediate indicator of present fire risk. Soil moisture content permitted an assessment of future changes in fuel moisture content. Both of these integrated changes in meteorological parameters. As such, the meteorological parameters (temperature, precipitation, atmospheric moisture represented by vapour pressure) contributed an understanding of why soil and fuel moisture changed in the ways that they did and offered information on how they would change in the future. Monitoring these quantities can help fire and land managers understand how and why fire risk changes across a landscape. 

Future work

This project was conducted over a short time period with one objective being to establish a pathway for future research and operational application.

The following opportunities for future research with likely operational benefits have been identified:

• The project results indicate that most of the studied variables interact over a range of timescales. Further investigation to understand the nature and quantity of these interactions will be of benefit to better estimate fire risk. 
• The project focused on spatial data of live fuel moisture. An analogous spatial approach to assess dead fuel moisture in very dry conditions is required to develop a complete assessment of fuel availability.
• Detailed examination of the fire intensity against spatial heterogeneity of soil and fuel moisture variability within each fireground would be valuable, to better predict likely fire intensity resulting from soil and fuel moisture variations.
• The project briefly explored the connection between heatwaves and live fuel moisture. Appropriate datasets were not readily available, however further investigation of the links between heatwaves and fire risk resulting from changes in fuel moisture content is critical to understanding, for example, overnight fire activity in a changing climate.
• The soil moisture model parameterisations in land surface schemes can be further developed to improve soil moisture estimation. Developments will have positive flow-on effects to a range of weather forecasting applications, including seasonal fire risk predictions, and thus providing fire managers with enhanced predictive capability for a range of important activities.
• Inclusion of ground water may improve the soil moisture models and benefit seasonal predictions.
• Unexpected interactions and connections that emerged between the strands of this project should be further explored to improve future services. For example, it seems likely that live fuel moisture content can constrain the range of possible values of sub-surface water content.

In summary, better understanding and spatial mapping of the influence of atmospheric, soil and fuel moisture will assist fire and land managers predict fire risk, which is critical for strategic and tactical planning. Many of the recommendations could be inclusions to the Australian Fire Danger Rating System (AFDRS), and therefore of immediate benefit to fire and land managers following the introduction of the AFDRS to operational use over the next two fire seasons.