Degree Name

Doctor of Philosophy


School of Earth and Environmental Sciences


Fire is a complex natural phenomenon with important social, economic and environmental consequences worldwide. Fuel moisture, acting as a heat sink during the burning process, can have a significant influence on fire activity. Satellite sensors offer spatially and temporally dense data useful to explore fuel moisture dynamics over large areas. The overarching objective of this thesis was to explore the capabilities of Moderate Resolution Imaging Spectroradiometer (MODIS) data to monitor the influence of fuel moisture on fire activity in high biomass (i.e., forested) systems in south-eastern Australia (i.e., the Sydney Basin Bioregion). Specifically, this objective encapsulated two major research threads: (a) to assess the potential of MODIS-based spectral indices to characterise spatio-temporal patterns of live fuel moisture; and (b) to monitor the influence of the spatial continuity of fuels at different moisture conditions on fire activity through a novel and innovative approach which integrated remote sensing data and graph theoretic measures.

A time series (2000-2009) of eight spectral indices were correlated to interannual variations in precipitation patterns (1-month, 3-month and 6-month Standardized Precipitation Index) to test the potential of MODIS data for monitoring drought-related live fuel conditions (i.e., water stress) in high biomass systems. MODIS-based spectral indices were also tested for their suitability to measure seasonal variations in live fuel moisture content (LFMC) of fuel types (i.e., shrubland, heathland and sclerophyll forest) which have not been previously investigated in this context in Australia. The Normalised Difference Infrared Index – band 6 (NDIIb6) showed the highest sensitivity to inter-annual changes in drought conditions and seasonal variations in live fuel moisture content (LFMC).

Moisture maps derived from MODIS-based NDIIb6 time series data (2000- 2010) were integrated with graph theory using an innovative approach to monitor patterns of available (i.e., dry) fuel connectivity. Connectivity of dry fuels was compared to inter-annual and seasonal variations in precipitation patterns, total area burnt each year, size and distribution of fires. Results showed that fuel connectivity is highly dynamic and, under dry conditions, can result in extended arrays of fuels available to burn. Connectivity of dry fuels explained 65% of the variance (P < 0.05) in total area burnt each year and showed a negative relationship with the incidence of small fires. Results also showed that temporal (i.e., monthly) patterns of dry fuel connectivity can be used to monitor the evolution of fire danger.

This research demonstrated that MODIS-based NDIIb6 has strong potential for fire management applications in high biomass systems. NDIIb6 outperforms traditionally used Normalised Difference Vegetation Index (NDVI), and is more suitable for monitoring both inter-annual and seasonal patterns of fuel moisture, and mapping the spatial connectivity of dry fuels across the landscape. NDIIb6-based measures of fuel connectivity are significantly related to fire activity and suitable for spatially-explicit monitoring of fire danger. The implementation of NDIIb6 into existing and future operational fire danger tools is recommended.