Essays on Urban Resilience to Environmental and Health Risks

Abstract

Cities today face growing environmental and health threats due to climate change. Building urban resilience requires understanding the complex interplays between environmental and social systems that account for adaptation dynamics. Using new data, computational tools, and economic analysis, this thesis explores how people and places adapt to environmental risks and the implications for urban policy and infrastructure planning.

Chapter 1 examines how the financing structure of climate resilience infrastructure impacts long-term economic dynamics. Using satellite imagery to develop new performance metrics for U.S. flood protection levees, I find that decentralized financing of infrastructure maintenance creates a feedback loop: lower housing values and property tax revenues reduce fiscal capacity for levee maintenance, which increases levee failure risk and further depresses housing values. These feedback dynamics reinforce under-maintenance and perpetuate spatial inequality.

Chapter 2 analyzes the social cost of behavioral adaptation. Leveraging 27 million fitness app exercise records and quasi-experimental designs, I find that heavy air pollution reduces outdoor exercise likelihood by 28%, with information and risk awareness as key moderators. This behavioral response results in significant health costs often overlooked in environmental health studies.

Chapter 3 explores the role of subjective traits in predicting adaptation behavior. Applying Natural Language Processing to social media posts from 500,000 users, we classify individual fear types and find that pre-pandemic fearfulness strongly predicts social distancing behavior during COVID-19. This project provides a scalable tool for measuring unobserved subjective traits to predict behaviors under risk and target interventions.