Some states have automated or semi-automated methods for rangeland burn extent mapping. However, in forests, burnt area extent and burn severity mapping is currently usually done ad hoc after major events.
Data used in mapping normally include satellite imagery enhanced with on-ground mapping and insights, using mapping techniques that are fine-tuned to suit the characteristics of the event and data. This approach produces appropriate results for the event at hand, but does not produce a longer burn history, which is needed to understand current and future fire risk.
There is a clear need for automated techniques for mapping burnt area extent and severity and fire risk that can be applied anywhere in Australia, including in woody vegetation systems. Continuous mapping of burnt area will also help to inform and attribute land cover change mapping carried out by state and Commonwealth agencies (e.g. NCAS).
ANU and Geoscience Australia together developed an automated workflow to develop a full continental fire history for Australia by processing the enormous volume of Landsat imagery in GA’s Digital Earth Australia data cube. The goal was to map burnt area extent in a manner that is suitable for Australia-wide operationalisation, with a focus on woody vegetation. The method and data were validated against events for which independent spatial data were available.
The algorithm includes a sequence of (i) change detection, (ii) change characterization, (iii) region growing and (iv) attribution steps. You can find the Python code that includes the full workflow in our GitHub repository: https://github.com/ANU-WALD/burn-mapping
The World-Wide Water model (W3) and its Australian version, OzWALD, are two near-identical models for grid-based estimation of daily water balance dynamics and water-related vegetation dynamics. Both are evolutions of the original AWRA-L model. The example Matlab implementation of the models is freely available for download. You can find the code, relevant literature, and some course material via this link or the download button below.
(Note that this is notthe AWRA Community Modelling System (AWRA CMS), which also an evolution of the original AWRA-L model but is maintained by the Bureau of Meteorology and can be found here: https://github.com/awracms/awra_cms)