Most of the scientists attending the workshop “Sustainable Computing for Sustainability,” hosted by the National Science Foundation last spring, were computer scientists developing algorithms for more nuanced sustainability and climate-change models.
Rina Ghose, professor, industrial engineering, was one of the minority from another discipline serving as an invited panelist at the workshop. She brought a vital piece to the event’s two broad topics: making the computing industry greener and addressing data and analytical needs for predicting climate change impacts.
With a background in climatology and data analysis, Ghose was able to share her expertise on the use and interpretation of spatial data, a topic that few computer scientists know much about.
That interdisciplinary perspective was one of the NSF’s aims for the workshop, Ghose said.
“That’s important because the multitude of data that is used in climate-change modeling comes from remote sensing,” she said. “Spatial data collection and analysis using Geographical Information Systems and remote sensing is fundamental to modeling climate change.”
Large forest fires, ocean temperatures and hurricane flooding are examples of the data that can be mapped from space, she said. And most of the data is free through federal agencies.
Among the wider findings of the workshop were:
- Industries need to immediately prioritize climate change in their strategic planning. For instance, industries could stop locating data centers in dry, hot locations.
- More attention should be directed at recycling in the computer industry, which currently generates a lot of waste.
- More education and research about sustainable computing is needed. Ghose called out the growing number of data centers needed to make artificial intelligence possible. And yet, data centers are incredibly energy intensive.
She also participated in discussions on a variety of topics, including disaster mitigation plans, sustainable agriculture and the United Nations’ Sustainable Development Goals.
An early release version of the report documenting the workshop’s findings and recommendations can be found at https://arxiv.org/abs/2407.06119.