Australia — The cutting edge of bushfire surveillance will use drones and social media updates to feed an incredibly sophisticated data engine, making it possible to predict and monitor the unpredictable, writes Andy Park.
The cutting edge of bushfire surveillance will use drones and social media updates to feed an incredibly sophisticated data engine, making it possible to predict and monitor the unpredictable.
Fire authorities are warning this summer’s fire risk is the highest since Black Saturday.
The Victorian Bushfires Royal Commission into the 2009 Black Saturday fires noted there was a need for more forward-looking research.
But since then, the bulk of the work to prevent it happening again has been directed into community engagement schemes.
But one expert says its now technologically possible to predict them in the first place.
Dr Allison Kealy, a senior lecturer at the University of Melbournes Department of Infrastructure Engineering, says that big data can prevent and protect us from bushfires.
Yes. There is enough that we know about fire behaviour and enough that we can measure about the environment in which fires are occurring, its a matter now of making sure we have enough technology in place, she said.
She says the coming fire seasons will see a decentralized information network, being fed with information from Unmanned Aerial Vehicles (UAVs), wireless sensors, improved high-resolution and real-time satellite imagery.
It will even aggregate information from crowd-sourced data, like Twitter and Facebook.
Her team at the University of Melbourne is partnered with the University of California in San Diego, and are working to design and build resilient streaming sensor networks for bushfire response.
Her project aims to add useful information to the currently incomplete data-set, for example, Bureau of Meteorology maps which dont contain detail for high-risk areas.
Slopes and gradients on those maps have micro-climates that have complete different characteristics.
She says its a merge of real-time, event-driven data with pre-existing data.
If we understand how wet the fuel is in the forest, we are able to determine how the fire will decrease in intensity or how its going to propagate through that area, then we can use satellite or live data to model its path.
The use of technology in bushfire communications has improved; for example the CFA FreReady app allows mobile users to share user-generated geo-codded images and view official, real-time warning and advisories.
But new technologies, more specifically mining hugely complex data sets provided by emerging surveillance technologies, wont be used on the front line this summer.
The gap has closed when it comes to the ability to share information this summer, but we need a more detailed and focused approach when it comes to the integrity of data capture, analysis and visualisation, Dr Kealy said.
The failure was in the forest areas.Advertisement
Following a 10-year strategy, ACT fire managers have created a mosaic across the landscape of different fuel levels, burning at every opportunity.
But forests have been too wet to burn this spring and the past two summers.
A network of 500 fire trails and strategic burns along the north-west urban edge, heavy grazing and extra grass slashing will create a fortress for the territory which forecasters say faces a higher than average risk this summer.
After a fire-fuelled tornado in January 2003 killed four Canberrans and frightened thousands more, CSIRO fire expert Phil Cheney told the subsequent inquiry the fire’s penetration into urban areas under extreme conditions did not reflect a failure of fuel management on the urban interface.