Date of Award

8-2026

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

College/School

College of Science and Mathematics

Department/Program

Earth and Environmental Studies

Thesis Sponsor/Dissertation Chair/Project Chair

Aparna Varde

Committee Member

Clement Alo

Committee Member

Pankaj Lal

Committee Member

Weitian Wang

Committee Member

Jan Polzer

Abstract

The global energy transition is accelerating, yet its benefits are not reaching all communities equally. Low-income households, communities of color, and those living in aging housing stock face disproportionate energy costs, limited access to clean technologies, and greater exposure to climate and transportation-related pollution — burdens that existing policy frameworks have struggled to address. Drawing on data spanning U.S. power plants and New Jersey census tracts, this dissertation investigated how climate variability, socioeconomic structure, and transportation electrification intersect to produce and compound energy inequity across the United States and New Jersey, using artificial intelligence and data-driven modeling to reveal where disparities exist, why they persist, and how interventions can be better targeted. A hybrid CNN-LSTM model demonstrated that temperature nonlinearly shapes electricity generation across U.S. power plants, with climate sensitivity varying significantly by region and infrastructure type. Building on this foundation, interpretable machine learning and public discourse analysis revealed that energy burden in New Jersey is structurally driven by educational attainment, income, and housing conditions, and that community-level concerns about affordability are largely absent from formal policy processes. These findings were operationalized through NJ-EQUIP, a visualization portal designed to make energy equity evidence accessible to planners and policymakers. Extending the analysis to transportation, the same communities identified as most energy-burdened proved least positioned to benefit from electric vehicle infrastructure, with environmental justice designation carrying no weight in observed charging deployment patterns. Across all three analyses, explainable AI methods are used not only to generate predictions but to produce interpretable, auditable accounts of the structural drivers behind observed patterns, moving beyond description toward evidence that planners and policymakers can act on. Taken together, these findings demonstrate that energy inequity compounds across residential and transportation systems in predictable, structurally rooted ways. As decarbonization timelines accelerate, the risk that clean energy benefits bypass the communities most in need of them is not hypothetical but empirically documented and spatially addressable, and integrated, equity-centered analytical frameworks are essential for ensuring the clean energy transition reaches the communities that need it most.

File Format

PDF

Download

Available for download on Saturday, August 21, 2027

Share

COinS