The impact of climate change can be felt around the world and in a variety of ways – from longer, hotter summers to increased ocean acidity and the disappearance of sea ice – but perhaps the most concerning consequence is the major impact on our global food and water supply. Agriculture accounts for about 70 percent of the total global freshwater consumption, but fluctuations in temperature and rainfall around the world are exacerbating the potential for drought and floods around the world. Even the most minor of environmental changes can be felt across the food and agriculture supply chain.
North Carolina State University’s (NCSU) research program, in partnership with Lenovo, is actively invested in addressing the effects of climate change on the agricultural ecosystem.
To help minimize the disruption to food production, it’s critical that farmers are able to prepare for anticipated regions that will experience drought (or flooding) which will negatively impact crop growth. Led by Dr. Ranga Raju Vatsavai, Associate Professor in the computer science department and associate director of Center for Geospatial Analytics, NCSU’s research team is leveraging innovative Geospatial Image Analysis technology to map and monitor croplands and preemptively identify local areas that will be affected by flooding or droughts.
Dr. Vatsavai and his team start by collecting and processing big geospatial data from satellites and sensors. Very high-resolution imagery at global scale constitutes more than 500 trillion multidimensional observations (pixels) that then need to be sorted and analyzed so the team can gain insights about crop growing patterns. The computational problem is further compounded by the need to analyze petabytes of climate data in order to understand the climate change impacts (e.g., floods, droughts) on food and water systems. The research team’s data analysis workflow demands high-end computing resources due to algorithmic complexity of artificial intelligence methods such as multiple instance learning and deep learning, and I/O needs stemming from the big data.
Coarse resolution, medium resolution and high-resolution examples of remote sensing data
The university needs powerful underlying technical infrastructure to support the deep learning algorithms that process this massive amount of geospatial data very quickly. The lack of large scale computing power at the Spatio-Temporal Analytics and Computing Laband the Center for Geospatial Analytics at NCSU has limited the researchers’ and students’ ability to leverage advanced machine learning problems for spatio-temporal big data.
To address this problem, Lenovo is offering access to one of the recently opened Lenovo AI Innovation Centers in Morrisville, North Carolina. The Innovation Center will provide AI-optimized infrastructure to crunch billions of data points in hours that would otherwise have taken NCSU about a year and significant operating expenses to establish and maintain.
“Without a means to process our geospatial big data, we are unable to advance our research methods to predict agricultural trends and adverse impacts of droughts and other climate change impacts. Our partnership with Lenovo brings meaning to this data with advanced computing power that lets us use deep learning, machine learning and AI to prevent negative impacts to the food supply chain, all without worrying about the reliability and performance of the hardware equipment on the back end,” says NCSU Professor Dr. Ranga Raju Vatsavai.
Over the next 12 months, Lenovo’s AI Innovation Center will provide NCSU with access to industry-leading computing and expert consulting resources that can turn their geospatial data into actionable insights that they can then share with local farmers and agricultural experts in order to predict drought areas and appropriately prepare for any disruption.
Together, using technologies like Geospatial Image Analysis to predict future trends, Lenovo and NCSU can help limit the impact of climate change on our food and water supply.