27 March 2018

It is quite unusual to have extreme fire weather in autumn in Australia which has caused widespread bushfires (e.g. fires in SW Victoria and fires in the Alpine National Park) and devastating property damage.

We developed and contributed three information products in response to the disaster in Tathra, a picturesque town at the Sapphire Coast (Southern NSW, Australia). The very fast-moving fires on Sunday 18 March 2018 destroyed 65 houses and 35 caravans / cabins, and damaged another 48 buildings (source: NSW RFS).

As the main information products have been shared on social media (especially on Twitter), we only make links here.

(With the developed cloud-based geospatial big data and analytics platform, we are now able to efficiently produce a number of similar information products for future major / extreme events in Australia and around the world.)

Level 1: Earth Observation & Imagery Analysis

We have used the Sentinel-2 satellite imagery (11 March 2018, immediately before the fire) to classify bushland. The Sentinel-2 imagery (26 March 2018, soon after the fire) was also processed to visualise the full extent of burnt area. The before/after imagery is shown below.

Level 2: Integrated Exposure Location Profile Report

This comprehensive location profile report (Version 4.2, 03/2018) features an exposure estimation component we have recently finished. The number of addresses in the vicinity of bushland was estimated. (If the exposure data is in another form, e.g. building stocks, infrastructure assets or sum insured, we are able to produce reports similarly.)

Level 3: Bushfire-prone Areas (animation)

In this case, the freshly classified bushland was used to identify bushfire-prone areas. In 2017, we finished a national study for this type of application (link).

Note: For future events and reporting, we will follow the above format and try to deliver relevant information products in a timely manner. We constantly enhance the cloud-based platform by including new content, analytics and delivery means.