Decoding Biodiversity Across Scales: Harnessing AI & Big Data to Preserve Ecosystems Under Global Change

Wednesday, March 29  |  10:15 a.m.

Plant Sciences Building  |  Seminar Rooms A&B

Sergio Marconi

Sergio Marconi is a post-doctoral associate at University of Florida. He earned his PhD in Interdisciplinary Ecology at University of Florida, where he focused on building algorithms for estimating plant properties from either airborne remote sensing or big field datasets. His research stands at the intersection between terrestrial ecology and artificial intelligence (AI). His recent work focused on: (a) building AI architectures for surveying species, physical and chemical attributes for millions of individual plants; and (b) using these surveys for understanding the effects of climate and topography on the distribution of functional biodiversity across scales. 

Along with his collaborators he contributed to open scientific software and datasets for plant detection and characterization from RGB and Hyperspectral imagery. His work has been published in high-impact journals, including Remote Sensing of Environment, Ecological Applications, and Environmental Research Letters. He is a University of Florida Biodiversity and Informatics Institute fellow and US-Italy Fulbright scholar.

Research Seminar:

As global temperatures continue to rise and habitats are altered, understanding how species will plastically adjust their functional traits and how their distribution will shift over time is key for conservation planning. Traditionally, models used to investigate these relationships have limitations such as oversimplified assumptions, sparse data, and a focus on few species. With advancements in technology, such as low-cost hyperspectral remote sensing and ecological networks, it is now possible to overcome these limitations by building advanced Artificial Intelligence (AI) applications to generate inventories of species and physiological attributes at an unprecedented scale. In this talk, I will present my research on building AI pipelines to generate inventories of species and traits, for millions of individual trees across the United States. Then, I will present how I used this data to disentangle the effect of environmental and evolutionary drivers on the distribution of functional traits, within and across species, from local to continental scale.