How big data is providing pearls of wisdom for Tasmanian oyster …

ByBrad Howarth australia, computing, crcsi, csiro, data, environmental, infrastructure, research, sense, tasmania, tickle, university Comments Off

It takes human senses to appreciate the wonders of Australia’s natural environment. But it is machine intelligence that may prove critical to our ability to manage it.

Numerous projects around Australia are now using advanced data analytics techniques (also known as big data) to better understand and manage the complexity of natural systems. Modern analytical techniques – particularly the capability to analyse and compare vast sets of data simultaneously – are delivering new capabilities to understand how natural systems work.

The hope is that through better understanding the interplay of different systems through data it may be possible to manage natural resources in a way that is both more efficient and more sustainable.

This is the intention of Sense-T, a sensing and data analysis project currently being developed in Tasmania by the CSIRO, University of Tasmania and numerous supporting parties including environmental groups and farmers.

It is using vast volumes of data already collected by the state to better manage its natural resources, while augmenting this with new sensor-based data sets to provide specific details.

The four projects currently underway cover beef and dairy production, viticulture, aquaculture and water management, and have even seen researchers attaching sensors to oysters to collect data for farm management.

Sense-T’s director, Ros Harvey, says where possible the projects will re-purpose existing data sources for new purposes, such as using environmental monitoring data to improve farming methods.

“And we are looking at repurposing production optimisation data from our agricultural projects and repurposing that for environmental reporting,” Harvey says. “If we repurpose individual farm-level data and aggregate it up, can it become a really important source of environmental monitoring that can be used for public policy purposes?”

Sense-T’s oyster farming project for instance is taking existing data on inflows into water catchments and using this to help farmers determine when environmental pollutants will move in and out of their farming areas, helping them better determine the optimal times for harvesting.

In Tasmania’s north-east, a water catchment management program is providing farmers with real-time information on the health of waterways to reduce the number of times a year that farmers are issued with orders to stop drawing water for environmental reasons.

“We were working with the community to develop a dashboard that allows that community to actually understand what is happening with the environmental flows in the river that they are drawing irrigation from,” Harvey says.

Vital to the success of Sense-T is a partnership with Sirca, a not-for-profit organisation focused on financial research and innovation, but whose big data crunching capabilities have proven adaptable to the needs of Sense-T.

The concept of using big data for environmental purposes is only just beginning to be understood within the public sector. While the Australian public service big data strategy does make reference to the use of analytics to monitor environmental issues, activity today is being driven by individual departments or research projects.

But nationwide examples do exist. One of the biggest users of big data for environmental management is the Bureau of Meteorology (BoM), which also makes available large data sets for external analysis. This includes past weather and climate information from its weather stations, with many popular data sets available free of charge.

The BoM has also been a key contributor to CRCSI (Cooperative Research Centre for Spatial Information), and specifically to a project that is using archival data from the Landsat satellites. This includes information on characteristics of the earth’s surface taken at a resolution of 25 metres and refreshed every 16 days stretching back to 1987.

According to CRCSI’s business development and research manager, Philip Tickle, this information is vital in helping land managers understand changes in ground cover and water.

“A landholder will be able to bring in a boundary of a paddock or farm and get the average ground cover for the last 30 years and how it compares to neighbours,” Tickle says.

What makes this possible is use of the National Computing Infrastructure at the Australian National University, which provides the computational grunt to crunch the petabytes of data that the Landsat archives contain.

“Something that might have taken three to five years’ worth of computing two or three years ago can basically be run overnight now,” Tickle says.