Underlying characteristics of lightning-ignited bushfires in Tasmania, Australia

Extreme wildfires are one of the most unpredictable aspects of the fire environment with significant socioeconomic and environmental impacts. The occurrence and impact of wildfire is the result of a complex interaction between ignition sources, weather conditions, vegetation and topography. A critical component of understanding fire danger is the identification of different ignition sources and their influence on wildfire occurrence.
Lightning-caused wildfires are a major concern for fire management agencies around the world. Lightning fires, unlike other sources of ignition, frequently occur in remote and inaccessible locations, making detection and suppression particularly difficult. Individual lightning storms may also cause a large number of fires to be clustered in space and time, which can overwhelm suppression efforts. In Tasmania, Australia, lightning has always been a cause of fires. However, the intensity and frequency of large-scale landscape fires caused by lightning strikes have increased in recent decades, causing local land and fire management authorities to be particularly concerned.
Understanding cloud-to-ground lightning climatology, identifying its spatial and temporal patterns, and determining how lightning strike characteristics influence the ignition probability of a lightning strike in Tasmania was the first objective of this study. Fire history records from the Tasmania Fire Service and Tasmania Parks and Wildlife Service, as well as the ground positioning and tracking system (GPATS) were used to compile Tasmania's cloud-to-ground lightning inventory for the period January 2011 to June 2019. Individual fires and lightning strikes were systematically linked in order to calculate the probability that a lightning strike ignited a fire. A proximity index analysis was used to categorise lightning strikes in thunderstorms as either potential or actual ignition candidates. The lightning ignition efficiency as a function of different fuel types was also investigated using the currently existing community map (TASVEG). The annual average efficiency of lightning ignition per lightning stroke was 0.24%, with a seasonal maximum during the summer. The lightning ignition efficiency as a function of fuel type also highlighted the role of buttongrass moorland (0.39%) in Tasmanian wildfire incidents.
Identifying whether a lightning strike is associated with wet or dry conditions is one of the most significant challenges of assessing lightning's impact on ignition probability, as "dry lightning" is much more likely to cause fires (lightning strike without significant rainfall). The moisture content of fine surface fuels changes more rapidly during periods of accumulation, making lightning-caused fires more sensitive to changing rainfall amounts prior to lightning strikes. Hence, the second objective of this study was to examine the relationship between rainfall amounts and the period of accumulation before lightning strikes, which control seasonal and regional variability in lightning-ignited wildfires. Hourly rainfall data was derived from the Atmospheric High-Resolution Regional Reanalysis for Tasmania, Australia (BARRA-TA) and lightning observation data was converted into BARRA-TA resolution (1.5km×1.5km×1h). Antecedent Precipitation Index (API) was computed for multiple time windows before lightning strike, and API value was assigned to each grid cells containing lightning strikes. Dry lightning was defined by multiple cumulative rainfall amounts ranging from 0 to 6 mm/season. The analysis was conducted over Tasmania as a broad region, and three smaller regions of Tasmania classified by natural resource management (NRM) sub-regions. The correlation between dry lightning count and lightning wildfires were calculated for each region and season. The findings revealed regional and seasonal differences in correlation and p-values, which could be explained by local factors such as weather, topography, vegetation, and fuel type. Results provide valuable information to fire managers and help them to focus their attention on a reduced sample of lightning strikes associated with higher potential for ignition of wildfires.
Although the lightning ignition probability is relatively independent of fuel moisture, the moisture content influences the risk of sustained ignition. Soil moisture is physically linked to fuel production and live fuel moisture, measured soil moisture data may then improve wildfire probability assessments. In terms of ignition probability and risk assessment, the third primary objective of this study is to investigate and compare the effects and relationships between lightning, different types of vegetation, and soil moisture content in different fuel types associated with lightning-ignited bushfires. We used an advanced soil moisture analysis system recently developed by the Bureau of Meteorology, called the Joint United Kingdom Land Environment Simulator based Soil Moisture Information (JASMIN). JASMIN can estimate soil moisture at 5 km resolution on a daily timestep for the whole of Australia. A conditional probability model was applied to compute the likelihood of ignition within different vegetation types and results were compared with Soil Dryness Index (SDI) method which is used operationally in Tasmania. Outputs of this study include quantitative information on the probability of bushfire ignition that can be used by fire management authorities to assess risk within their jurisdiction.
This thesis contributes to a greater understanding of the impact and interaction of lightning with precipitation and soil moisture that cause wildfires, as well as how wildfires are likely to change over time and space. The findings of this thesis suggest several directions for future research on wildfire risk assessment, as well as providing useful information for operational decision making and improving long-term risk forecasts by identifying the critical conditions that lead to large wildfires.