This article has been reviewed according to Science X's editorial process and policies. Editors have highlighted the following attributes while ensuring the content's credibility: For insurers and operators alike, what grows under a solar array is almost as important as what sits on top of it. Grass height, moisture and biomass all influence how a site is insured, maintained and grazed.
At ACEN Australia's New England Solar site in New South Wales, a collaboration with CSIRO and Macquarie University is using satellite-based remote sensing and artificial intelligence (AI) to bring new precision to vegetation management, fire preparedness and day-to-day operations. As CSIRO senior photovoltaic scientist Kenrick Anderson explained, today's solar farms must constantly balance two competing demands. "On one hand, you have solar farms wanting to keep grass levels down to reduce fire risk and keep insurers happy," he said.
"On the other hand, the farmer, who leases the land to the solar operator, prefers higher grass levels for grazing purposes." Higher grass levels are also preferred by farmers to mitigate the ingestion risk of barber's pole worm—a fatal, blood-sucking parasite common in Queensland and the northern half of NSW. "Industries are trying to find a 'happy medium' in grass length: short enough for insurers and fire services, long enough for graziers' livestock health and productivity—all in a climate where droughts and extreme rainfall events are becoming more frequent and severe," Anderson said. Until now, grass conditions have largely been determined through manual processes.
"We've worked with technicians who'll drive by a site and go '1, 2, 3, 4' to describe the grass length, with '1' being the lowest and '4' being the highest," CSIRO principal research scientist Dr. "This provides limited insight, is subjective and introduces the chance for human error." Manual assessments are also time-consuming. "A solar farm like ACEN's spans 2,000 hectares," Anderson said.
"To traverse the site and inspect everything is extremely labor intensive—a technician could cover hundreds of kilometers in the space of a few days." "We're hoping this research will unlock a new quantitative, repeatable system using satellites, airborne data and field measurements to map vegetation fire risk across the entire site—and update it regularly." Nona Sepahrom, a CSIRO Industry Ph.D. student from Macquarie University, has been tasked with the project. She is using hyperspectral imagery from Germany's EnMAP satellite to measure vegetation down to a biochemical element. "For utility-scale solar operators, biochemical detail translates to something practical: the ability to distinguish green, moist biomass from dry, flammable fuel and see how that fuel is distributed around high-value assets such as transformers and cabling," Ong said.
"This is how we determine fire risk in a quantitative form. It can also help farmers better manage their grazing patterns and work collaboratively with the solar farm operator to balance both the fire risk and yield of a farm." Using field samples and satellite imagery, Sepahrom is building a larger, more crystallized picture of ACEN's New England Solar site. "Once I've rationalized lab and satellite information, I will use specific spectral features learned from the lab and satellite to classify the solar farm into distinct vegetation characteristics," Sepahrom said.
Extract — continue reading at the source.
