Seeing the Big Picture: Using Deep Learning to Understand the Land Implications of Energy Systems at Scale, CIE Digital Economy and the Environment Series

Estimates of the direct land area required by energy technologies began as rough approximations built on limited data from literature. Methods have rapidly advanced in the past decade—from supervised classification to deep learning of earth observation imagery—enabling a greater understanding of the land implications of energy decisions across spatial and temporal scales. Professor Sarah Jordaan of McGill University will cover this progression, up to new research applying deep learning models to quantify direct land use by solar, wind, hydroelectric and nuclear projects as well as coal- and gas-fired electricity. Deep learning applied to more than 500,000 earth observation images in the western United States produced land inventories for 719 solar photovoltaic facilities, 318 wind farms (15,871 turbines), and more than 100,000 natural gas production wells, complemented by manual delineation of aerial imagery for 594 hydropower plants, 42 coal-fired power plants, 8 concentrated solar plants, and 3 nuclear power projects. Prof. Jordaan will explore what these new techniques mean for understanding the land implications of energy transitions across spatial and temporal scales. Quantified using consistent methods, findings illuminate how new techniques enable us to “see the big picture” in support of a more environmentally sustainable energy future. Dr. Jordaan is an Associate Professor of Industrial Ecology and Life Cycle Assessment in the Department of Civil Engineering and the Trottier Institute for Sustainability in Engineering and Design at McGill University. Her research applies industrial ecology tools and innovation systems to support a sustainable, low-carbon energy future. This lecture is part of the Center for Industrial Ecology's Digital Economy and the Environment Lecture series. Professor Jordaan's presentation will be in person and also available via zoom. It is free and open to the public. For those who are local, lunch will be available in room 319.