Researchers recognised for AI use in stormwater breakthroughs

Two PhD researchers from Griffith University have been recognised for their research into green stormwater infrastructure, having been awarded an artificial intelligence ‘Al for Earth’ Azure award grant from Microsoft.

The ‘AI for Earth’ award grantees apply computer vision, machine learning, deep learning, and other AI disciplines to conservation challenges around the world. The grants provide access to AI, training and educational offerings, and investments in scalable, innovative solutions, and, to date, Microsoft has awarded 139 grants to projects in 45 countries.

Emad Kavehei is in his final year at Griffith’s Cities Research Institute (CRI), supervised by Dr Graham Jenkins, Dr Fernanda Adame and Professor Charles Lemckert. Abdul Karim is in his first year at Griffith’s Institute for Integrated and Intelligent Systems, supervised by Professor Abdul Sattar.

“Traditionally, we monitor a few influencing factors of soil carbon accumulation while we know that there are more factors involved,” Mr Kavehei said.

“Microsoft is supporting this project, which builds on my PhD studies, to create and deploy an open source model to predict the below-ground carbon content based on above ground image data.”

Terrestrial carbon storage in soil accounts for more than 65 per cent of global carbon which makes it a recognised strategy for climate change mitigation.

However, there is a little understanding of the carbon storage in urban areas. Green stormwater infrastructure such as bioretention basins are normal features of an urban landscape.

These aquatic and terrestrial ecosystems have great potential to capture and sequester carbon within below-ground media.

Mr Kavehei and Mr Karim will access Microsoft’s data to create computer vision and Artificial Intelligence-based solutions for analysing and predicting the carbon sequestration potential of such bioretention basins.

“We believe integrating different levels of relevant data with the state-of-the-art AI methods would better enable us to predict the urban carbon budget,” Mr Karim said.

They intend for their work, as part of CRI’s Digital Earth and Green Infrastructure Group, to have immediate application in predicting and quantifying the benefit of similar green infrastructure installations, within sandy loam soils, for capturing atmospheric carbon.