EARTH SYSTEM SCIENCE

SEMINAR SERIES

May 22, 2025

12:00 - 1:00 PM

Huang 300 Mackenzie

Sparkle L. Malone, Ph.D.

Assistant Professor of Natural Carbon Capture
Yale School of the Environment
 

“Advancing Ecological Understanding Through The Convergence of Machine Learning and Environmental Infrastructure: Understanding Carbon Exchanges Rates”

Abstract

In ecological systems, nonlinearities emerge from complex interactions and feedback loops across multiple scales. Capturing resilient and emergent patterns is challenging, yet essential for thriving under uncertain conditions. As global change accelerates, the adaptive management of complex systems necessitates advancements in ecology, network infrastructure, and machine learning. There is an urgent need to establish robust physical and analytical research infrastructure, as well as to develop convergence research paradigms that integrate ecology and machine learning. These efforts are crucial for understanding the complex environmental processes governing the exchange of greenhouse gases (CO2, H2O, and CH4) between the Earth’s surface and atmosphere. Currently, our understanding of the factors influencing exchange rates is limited, particularly regarding spatial and temporal variability. The intricate feedbacks between biological processes, climate, and land cover changes further complicate our understanding of surface–atmosphere exchange rates across diverse ecosystems.

To enhance our understanding of greenhouse gasses from natural ecosystems, leveraging existing infrastructure and data streams from the National Ecological Observatory Network (NEON), AmeriFlux, and FLUXNET is required. The integration of network science and machine learning is essential to overcome biases in data collection and measuring functional relationships between flux observations (responses) and their environmental drivers (meteorological and land surface) to elucidate the spatio-temporal variability of flux responses across ecosystems. Flux data products can be improved by accounting for surface heterogeneity and evolving sample characteristics over time. This work represents a convergence of physical and analytical infrastructure to deepen our understanding of ecosystems' effect on the atmosphere.

Biography

Dr. Sparkle L. Malone obtained her PhD from the University of Alabama in 2014, where she studied the carbon fluxes in subtropical wetland ecosystems. She then worked for the USDA Forest Service at Rocky Mountain Research Station from 2014-2017 and at Florida International University (FIU) from 2017- 2022. At FIU, she established the Malone Disturbance Ecology lab. Her primary research focused on improving our understanding of how climate and disturbance regimes influence spatial and temporal variability in ecosystem structure and function. She explores questions related to ecosystem conditions, sustainability, and vulnerability to climate extremes using remote sensing and eddy covariance. In the fall of 2022, she joined the Yale School of the Environment (YSE) and the Yale Center for Natural Carbon Capture (YCNCC).

https://www.malonelab.org/