Warren Distinguished Lecture Series

Mary A. Armstrong
Women's, Gender & Sexuality Studies
Lafayette College

ABSTRACT:  Everybody knows that diversity matters: it opens opportunities for all, improves teams and outcomes, and serves our highest values. And everybody knows that diversity (however it is defined) is difficult to achieve. If good intentions got good results, inclusivity issues would be a thing of the past. But instead, we often feel “stuck” and/or frustrated with well-meaning but less-than-effective efforts. In this talk, Mary Armstrong shares a three-part schema for thinking about diversity in STEM fields that offers a precise and sophisticated model of “inclusivity work.” This model clarifies differences across types of inclusivity approaches, analyzing the distinguishing characteristics of kinds of diversity work, and examining the strengths and weaknesses of each type. Armstrong concludes with a discussion of how we can recalibrate our thinking around STEM diversity so as to approach inclusivity in new and potentially more effective ways.

BIO:  Mary A. Armstrong is Charles A. Dana Professor of Women's, Gender & Sexuality Studies and English at Lafayette College, where she also chairs the Women's, Gender & Sexuality Studies Program. She earned her Ph.D. in English and Graduate Certification in Women's Studies from Duke University. Her research interests include equity and inclusivity in STEM fields, strategies for institutional transformation in higher education, and inclusive pedagogies. She has been PI on two National Science Foundation ADVANCE grants focused on strategies for institutional change to support underrepresented women in STEM, and she is Director of the Lafayette College Queer Archives Project.

Pierre Gentine
Earth and Environmental Engineering & Earth and Environmental Sciences
Columbia University

ABSTRACT: Over the last couple of years, we have witnessed an explosion in the use of machine learning for Earth system science application ranging from Earth monitoring to modeling. Machine learning has shown tremendous success in emulating complex physics such as atmospheric convection or terrestrial carbon and water fluxes using satellite or high-fidelity simulations in a supervised framework. However, machine learning, especially deep learning, is opaque (the so-called black box issue) and thus a question remains: what (new) understanding have we really developed? 

I will here illustrate the value of lower dimensional, latent, representations to build new physical understanding of complex physical systems using machine learning. I will present several examples where machine learning and physics can advance together our understanding of complex physical systems and highlight the emergent behavior of the system. 

We will start with the example of  convective organization (i.e. the spatial organization of clouds) and their impact on precipitation, and will discuss new strategies for the terrestrial carbon and water cycles, where new physics can be learnt implicitly by building hybrid (machine learning+physics) models.  We will finally show next causal strategies going beyond standard correlations so that we can build more trustworthy and explainable algorithms. 

BIO: Pierre Gentine is the Maurice Ewing and J. Lamar Worzel professor of geophysics in the departments of Earth and Environmental Engineering and Earth and Environmental Sciences at Columbia University. He studies the terrestrial water and carbon cycles and their changes with climate change. Pierre Gentine is recipient of the National Science Foundation (NSF), NASA and Department of energy (DOE) early career awards, as well as the American Geophysical Union Global Environmental Changes Early Career, Macelwane medal and American Meteorological Society Meisinger award. He is the director of the new NSF Science and Technology Center (STC) for Learning the Earth with Artificial intelligence and Physics (LEAP), the largest funding mechanism of the NSF. 

Comparative and Placed-based Hydrology in the Digital Age: Building understanding, promoting inclusivity, and forecasting the future

Laurel Larsen
Geography and Civil and Environmental Engineering
University California Berkeley

ABSTRACT: Hydrologic studies and forecasts have traditionally been place-based, focused on a single watershed or hydrologic region. For this reason, it has often been pointed out that the discipline of hydrology lacks a general theory or set of organizational principles, with a primary disadvantage being difficulty in making predictions for regions without a long-term data record (also known as the PUB, or Prediction in Ungauged Basins, challenge). With the recent proliferation of environmental “big data,” coupled with the rapidly advancing field of data science, opportunities abound to advance the subdiscipline of comparative hydrology—investigations that span many watersheds—and make progress on the challenge of PUB. In this talk I will discuss how the Environmental Systems Dynamics Laboratory is addressing challenges of data curation, using tools from information theory to understand functional classes of behavior that watersheds exhibit, and advancing the integration of physically based and data-driven models to forecast streamflow across the scale of the coterminous United States. Much of the success of this work depends on the Open Science movement and the hydrologic science community’s adoption of principles of transparency, reproducibility, and accessibility. In the second part of the talk, I will explore how these principles are also transforming place-based hydrology, with a case-study focus on the Sacramento—San Joaquin Delta. I will argue that an inclusive shared visioning approach, enabled by technological advances and the Open Science movement, can help transition the science and governance community to one that is well equipped to manage the Delta in the face of rapid climate change. 

BIO: Dr. Laurel Larsen is an Associate Professor at UC Berkeley with appointments in the Departments of Geography and Civil and Environmental Engineering. For 2020-2023, she is on leave to serve as the Delta Lead Scientist, a USGS position housed within the Delta Stewardship Council in Sacramento. Dr. Larsen runs the Environmental Systems Dynamics Laboratory at Berkeley, which has a focus on understanding the interactions and feedback among the physical, biological, and social variables constituting ecosystems, and to apply that understanding to restoration and management challenges. Dr. Larsen has a bachelor's degree from Washington University in St. Louis, with dual majors in Systems Science and Mathematics and Environmental Studies, a master's degree from Washington University in Earth and Planetary Sciences, and a Ph.D. from the University of Colorado Boulder in Civil and Environmental Engineering. She has worked in the Everglades, Chesapeake Bay, coastal Louisiana, and most recently, the Sacramento-San Joaquin Delta. 

Carsten Prasse
Environmental Health and Engineering
Johns Hopkins University

ABSTRACT:  Exposures to anthropogenic chemicals are a key contributor to the human “exposome”, or the sum of environmental stressors that shape and determine health outcomes. In addition to more 85,000 chemicals in commercial use today, we are exposed to thousands of chemicals formed when anthropogenic and natural organic compounds degrade in the environment and/or engineered systems. Frequently it is exposure to a complex mixture of chemicals that results in additive, adverse health effects. However, engineered systems for human and environmental health protection–like drinking water and wastewater treatment–rely on chemical-by-chemical assessments and regulations that rarely consider complex mixtures. Moreover, approaches that help prioritize identification and treatment of the most toxic chemicals are widely missing. As a result, adverse environmental and human health outcomes, unintended consequences of engineered treatment solutions, and inadequate regulations only become evident years after populations have been exposed. If we want to address this issue, we need to develop approaches that help us identify those chemicals that are of highest concern for human health and the environment. In this seminar, I will discuss the development and application of a novel analytical approach, called reactivity-directed analysis (RDA), which can be used to identify and prioritize those compounds that are of particular health concern. RDA combines approaches from analytical chemistry, molecular toxicology, data science, and environmental engineering to detect and identify toxic organic electrophiles, the largest class of known toxicants. RDA provides a new framework for identifying toxic byproducts and their precursors that can be used to optimize engineered treatment systems and minimize risks from toxic byproducts.

BIO:  Carsten Prasse, PhD is an Assistant Professor in the Department of Environmental Health and Engineering in the Johns Hopkins Whitening School of Engineering and the Bloomberg School of Public Health. Carsten received his Master’s degree in Environmental Sciences from the University of Bayreuth, Germany in 2008. In 2012 he obtained his PhD in Chemistry at the Federal Institute of Hydrology in Koblenz, Germany, under the guidance of Dr. Thomas Ternes for which he was honored with the dissertation award of the German Water Chemistry Society. After completing postdoctoral work at the Federal Institute of Hydrology and the University of California at Berkeley (research group of Dr. David Sedlak), he joined Johns Hopkins University in 2018.