A 10 Gigabit per second AARNet link is enabling animal scientists based at the CSIRO FD McMaster Laboratory near Armidale to share research data more efficiently with experts at other CSIRO facilities and partner institutions across Australia.

Electronic sensors and sensor networks used in agricultural research generate very large data sets. In the current research environment, members of large multidisciplinary research teams can be located hundreds or thousands of kilometres apart. The ability to transfer very large data files quickly and efficiently between team members at various locations is becoming increasingly necessary.

“Under the previous eight to ten megabit per second internet system, a video file of an animal’s behaviour used to annotate sensor data took one hour and forty minutes to transfer from Armidale to Brisbane. This can now be done over AARNet in less than five minutes, a whopping twenty-fold increase in speed,” said CSIRO’s David Paull.

He says AARNet’s high-speed file transfer capacity will lift the capability of this regional research centre by allowing researchers to share large data files in real time. Access to the network enables data analysis to be performed by staff with specific statistical expertise at different locations within a very short period of time.

For example, a new CSIRO and NSW Department of Primary Industries (DPI) research project aimed at estimating feed intake of individual grazing animals uses audio and motion sensors for capturing data related to feeding behaviours.

The sensors generate very large data sets that require a high level of annotation to deliver precise and accurate estimates of pasture intake for individual animals within a herd.

These data sets are transferred over the AARNet connection from where the experimental animals are located at CSIRO’s FD McMaster Laboratory to scientists with specific statistical expertise located in Brisbane. The scientists in Brisbane then are also able to work with the raw data as an integral part of the research team to develop methods and prediction algorithms for the pasture intake of individuals within the herd.

Team leader, Paul Greenwood from NSW DPI says, new sensor technologies have the potential to become very useful research tools for future studies in animal science, particularly for grazing livestock.

“Sensor technologies give researchers the ability to record information from animals without constraining the expression of their normal behavioural patterns. Capturing this information with a high level of precision and repeatability is extremely valuable for developing new breeding and precision management programs,” he says.

The aim is to enable enhanced monitoring of livestock health and welfare in differing landscapes, and to make Australia’s livestock industries more efficient and better adapted to our ever-changing environment.