Vegmachine

Vegmachine.net, established in July 2016, turns vegetation and rainfall satellite data into an interactive time-series visualisation, enabling farmers and other users to quickly see how all-important ground cover has changed.

One of Vegmachine’s users is the Queensland Department of Agriculture and Fisheries (DAF), which generates Vegmachine ground cover reports for farmers and graziers, such as Mr Burnham and his wife, Emma-Jane Burnham.

The Burnhams asked to receive reports every 12 to 18 months, to see year to year the impact of their land management decisions.

“Vegmachine refines the way we do stuff. It works on land type and you get a benchmark of each paddock so you can see if you’re bettering yourself, if putting in a fence, for example, has worked, and ground cover has improved,” said Mr Burnham.

“We try to keep our ground cover above 80 per cent, and our goal is to get up to 100 per cent. Generally speaking, in previous [Vegmachine] reports, there has been an upward move from between 75– 80% on most paddocks to at least 85% or over 85%. Vegmachine is a massive benchmarking tool.”

Prior to using Vegmachine reports, Mr Burnham relied on what he could see happening on the farm and advice from local farmers. “We can see it with our own eyes but when we get the report it’s hard data that we can store and pull out once we get the next report and see the impact of past decisions. The machine does it for you and it’s a lot clearer and pretty correct as well.

“Our ground cover has come up phenomenally. It’s lifted about 10 per cent and that’s something we can see with the aid of Vegmachine and grazing decisions.”

The Burnhams started receiving Vegmachine reports in 2016 and the reports helped to prove the efficacy of their grazing management when becoming accredited with DAF, Fitzroy Basin Association and AgForce Queensland’s grazing industry Best Management Practice (BMP). One of BMP’s requirements is to have at least 70 per cent ground cover.

Matt Brown, Beef Extension Officer in DAF’s Animal Science division, was one of the staff generating Vegmachine reports for the Burnhams. Vegmachine, he said, is a “great tool” and “really easy” to use. “It only takes a few minutes to produce a report for a property. People have been really happy with the information it provides.

“One of the biggest benefits of Vegmachine is it allows people to look at the degradation of their property, and it benefits the public through outcomes of less runoff to the Great Barrier Reef — people can see the ground cover and plan their strategies to prevent runoff occurring.”

In early 2018, Vegmachine logged its 8,000th user session, and by then the API had extracted 5,000,000 images to generate more than 45,000 ground cover reports, which provide paddock-by-paddock, landtype-by-landtype analysis of ground cover change from 1990 to the present.

Core users include graziers, government agencies, private consultants and natural resource management staff in regional Queensland.

The Web tool is open to public use, does not identify individual users, and provides the full level of service for all of Queensland and a subset of services to a large part of Australia.

QRIScloud and the Joint Remote Sensing Research Program

QRIScloud, QCIF’s cloud compute resource, provides storage for the satellite data powering Vegmachine.

Transferring that data to QRIScloud, and working with its support team, is Australia’s Joint Remote Sensing Research Program (JRSRP), the team behind Vegmachine.

JRSRP is a collaborative effort performing research and development, and specialised training, that supports state government and national land mapping, monitoring and environmental management programs. It also supports research activities aimed at understanding the impacts that humans and climate have on the natural environment.

Established in 2007, JRSRP was originally a collaboration between the University of Queensland (UQ) and the Queensland Department of Science, Information Technology and Innovation. It has since grown to include the University of New South Wales, University of New England, the Brisbane node of Auscover’s Terrestrial Ecosystem Research Network (TERN), and New South Wales’ and Victoria’s environment state departments.

JRSRP is funded by contributions from TERN and the Queensland, NSW and Victorian state governments.

The Tasmanian and Northern Territory governments are also engaged in ongoing projects with JRSRP to build their own satellite environmental measurement and monitoring capabilities.

Dr Peter Scarth, a UQ Senior Research Scientist working on JRSRP, described the team, currently comprising 22 people, as working to “democratise” spatial data access and use.

As well as servicing Vegmachine.net, JRSRP also supports (using QRIScloud for data storage) the national online tool FarmMap 4D (formerly called NRM Spatial Hub), several TERN services and several Queensland Government open data sets.

“QRIScloud is providing the infrastructure to serve this data up to the public,” said Dr Scarth.

Before hearing about QRIScloud from TERN, JRSRP managed a physical server in UQ’s Faculty of Science.

“QRIScloud has given us access to a virtual environment which allows much more flexibility (and has much more processing power and memory) than our previous hardware,” said Dr Scarth.

“QRIScloud has allowed us to scale services as demand has grown; has allowed the use of development servers to test out new processes and software before deployment in production; has allowed a distributed team easy access to the system; and has given us access to far more CPU [Central Processing Units] and memory than we would have gotten elsewhere.”

QRIScloud has enabled JRSRP to deliver data in multiple ways and formats that link to users’ requirements. “Delivery through standard portals would not have allowed the uptake of data that we have seen in our applications,” said Dr Scarth.

QRIScloud provides essential data indexing and delivery services for JRSRP. Many of the satellite image products, particularly the seasonal state mosaics, are pushed on a regular basis across to a Research Data Services (RDS) data allocation. QRIScloud provides HTTP, FTP, THREDDS and Geo server access to these products.

“We also manage several databases to provide faster image data lookup as well as national biomass libraries. More recently we have been experimenting with containerised processing services linking to Digital Earth Australia as well as further processing of these products in preparation for ingest into the Google Earth engine,” said Dr Scarth.

JRSRP is aiming to provide new services to Australia’s grazing industry to enable a better understanding of land conditions. “We are also very keen to further explore new data delivery technology and how we can better link that to the needs of end-users across both the public and private sector.”

Summary

Researcher:
Dr Peter Scarth
Senior Research Scientist
School of Earth and Environmental Sciences
The University of Queensland

Research community:
Agricultural Science, Ecoscience

Resources used:

• QRISdata: TERN AusCover satellite remote sensing data products — 60 TB. Frequent access storage.
• QRIScloud:
  • 80 vCPUs
  • 500 GB volume storage
  • 100 GB object storage.

QCIF advantage:

• Significant storage resources
• Flexible computing environment
• Access to increased processing power and memory.