Background Bushfires pose a serious and increasing threat to Australia. The detection and mapping of burns has many applications to support management of areas impacted by fire. The identification of bushfire burn using Earth Observation is often manual, can come with a significant time delay, and only available at a relatively small scale. This product offers provisional and preliminary change detection using same day satellite data to automatically and rapidly identify burn characteristics. Knowledge about the potential location and extent of fire helps to understand community and ecosystem impacts, enables directed relief and recovery support, and informs planning of mitigation burning for future fire seasons. What this product offers DEA Provisional Burnt Area Characteristic Layers contribute to the understanding of the distribution and frequency of fire in the Australian continent by measuring change in vegetation cover and soil characteristics that may be indicative of fire activity in the landscape. This product contains three layers that each describe change in a specific remote sensing index. Change in each index is measured between a baseline reference image and the most recent observation of Australia from the Sentinel 2 satellite constellation.  The indexes contained in each dataset describe change in a characteristic of the Earth’s surface that may be the result of a burn. The characteristic described are green vegetation cover and the reflective properties of bare soil and of burnt materials. These layers can be used to detect areas that may have been recently burnt, as fire will change the presence of these characteristics in the satellite data. These layers should be used with other information sources to determine if the change is the result of fire or other processes.

The Digital Earth Australia Hotspots web service has been developed as part of the Digital Earth Australia Hotspots national bushfire monitoring system. The service delivers hotspot data derived from (a growing number of) satellite-born instruments that detect light in the thermal wavelengths. The colour of the spot represents the time the Hotspot was last observed by a passing satellite (e.g. 0-2 hours). The colour does not indicate severity. Typically, the satellite data are processed with a specific algorithm that highlights areas with an unusually high temperature. In principle, however, Hotspots may be sourced from non-satellite sources. Lineage (for eCatID 101800 and 101780): The Sentinel Hotspots system was originally developed in 2010. The Sentinel Hotspots webservice was republished in 2016 as part of a platform upgrade. The Digital Earth Australia Hotspots system and webservices was redeveloped in 2019 as part of a platform upgrade.

The Digital Earth Australia Hotspots web service has been developed as part of the Digital Earth Australia Hotspots national bushfire monitoring system. The service delivers hotspot data derived from (a growing number of) satellite-born instruments that detect light in the thermal wavelengths. The colour of the spot represents the time the Hotspot was last observed by a passing satellite (e.g. 0-2 hours). The colour does not indicate severity. Typically, the satellite data are processed with a specific algorithm that highlights areas with an unusually high temperature. In principle, however, Hotspots may be sourced from non-satellite sources.

This dataset contains hotspot point data, derived from satellite-born instruments that detect light in the thermal wavelengths found on the Digital Earth Australia Hotspots application. Typically, satellite data are processed with a specific algorithm that highlights areas with an unusually high temperature. Hotspot sources include the Moderate Resolution Imaging Spectroradiometer (MODIS) sensor aboard the National Aeronautics and Space Administration (NASA) Terra and Aqua satellites, the Advanced Very High Resolution Radiometer (AVHRR) night time imagery from the National Oceanic and Atmospheric Administration (NOAA) satellites, the Visible Infrared Imaging Radiometer Suite (VIIRS) on the Suomi- NPP satellite. Please note: As these data are stored on a Corporate system, we are only able to supply the web services (see download links). email earth.observation@ga.gov.au.

<div>A package of deliverables for the Australian Research Data Commons (ARDC), Bushfire History Data Project, Work Package 5. If you require further information or access, please contact earth.observation@ga.gov.au</div><div><br></div><div>Outputs generated for this Project are interim and represent a snapshot of work to date, as of September 2023. Deliverables are developmental in nature and are under further advancement. Datasets or visualisations should not be treated as endorsed, authoritative, or quality assured; and should not be used for anything other than a minimal viable product, especially not for safety of life decisions. The eventual purpose of this information is for strategic decisions, rather than tactical decisions. For local data, updates and alerts, please refer to your State or Territory emergency or fire service.</div><div><br></div><div>The purpose of this Project (WP5) was to generate fire history products from Earth observation (EO) data available from the Landsat and Sentinel-2 satellites. WP5 aimed to implement a suite of automated EO-based algorithms currently in use by State and Territory agencies, to produce National-scale data products describing the timing, location, and extent of bushfires across Australia. WP5 outputs are published here as a “deliverable package”, listed as documents, datasets and Jupyter notebooks.&nbsp;</div><div><br></div><div>Burnt area data demonstrators were produced to a Minimum Viable Product level. Four burnt area detection methods were investigated: </div><div>* BurnCube (Geoscience Australia, ANU, (Renzullo et al. 2019)),</div><div>* Burnt Area Characteristics (Geoscience Australia, unpublished methodology),</div><div>* A version of the Victoria’s Random Forest (Victorian, Tasmanian and New South Wales Governments). Based on method as described in Collins et al. (2018), and</div><div>* Queensland’s RapidFire (Queensland Government, (Van den Berg et al. 2021). Please note that demonstrator burnt area data from the Queensland method was only investigated for the Queensland location. Data were sourced from Terrestrial Ecosystem Research Network (TERN) infrastructure, which is enabled by the Australian Government National Collaborative Research Infrastructure Strategy (NCRIS). </div><div><br></div><div>In addition demonstrator products that examine the use of Near Real Time satellite data to map burnt area, data quality and data uncertainty were delivered. </div><div><br></div><div>The algorithms were tested on several study sites:</div><div>* Eastern Victoria,</div><div>* Cooktown QLD,</div><div>* Kangaroo Island SA,</div><div>* Port Hedland WA, and</div><div>* Esperance WA.</div><div><br></div><div>The BurnCube (Renzullo et al. 2019) method was implemented at a national-scale using the Historic Burnt Area Processing Pipeline documented below “GA-ARDC-DataProcessingPipeline.pdf”. Continental-scale interim summary results were generated for both 2020 Calendar Year and 2020 Financial Year. Results were based upon both Landsat 8 and Sentinel-2 (combined 2a and 2b) satellite outputs, producing four separate interim products:&nbsp;</div><div>* Landsat 8, 2020 Calendar Year, BurnCube Summary (ga_ls8c_nbart_bc_cyear_3),</div><div>* Landsat 8, 2020 Financial Year, BurnCube Summary (ga_ls8c_nbart_bc_fyear_3),</div><div>* Sentinel 2a and 2b, 2020 Calendar Year, BurnCube Summary (ga_s2_ard_bc_cyear_3),</div><div>* Sentinel 2a and 2b, 2020 Financial Year, BurnCube Summary (ga_s2_ard_bc_fyear_3).</div><div>&nbsp;</div><div>The other methods have sample products for the study sites, as discussed in the "lineage" section. </div><div><br></div><div>The Earth observation approach has several limitations, leading to errors of omission and commission (ie under estimation and over estimation of the burnt area). Omission errors can result from: lack of visibility due to clouds; small or patchy fires; rapid vegetation regrowth between fire and satellite observation; cool understorey burns being hidden by the vegetation canopy. Commission errors can result from: incorrect cloud or cloud-shadow masking; high-intensity land-use changes (such as cropping); areas of inundation. In addition cloud and shadow masking lead to differences in elapsed time between reference imagery and observations of change resulting in differences in burn area detection. For more information on data caveats please see Section 7.6 of DRAFT-ARDC-WP5-HistoricBurntArea.</div><div><br></div><div>The official Project title is: The Australian Research Data Commons (ARDC), Bushfire Data Challenges Program; Project Stream 1: the ARDC Bushfire History Data Project; Work Package 5 (WP5): National burnt area products analysed from Landsat and Sentinel 2 satellite imagery.</div><div><br></div><div>We thank the Mindaroo Foundation and ARDC for their support in this work.</div>

The datasets created to produce the emergency mapping support products which contributed to fulfilling GA's arrangements in supporting the outcomes sought by the Australian Government during disaster events.