DISASTER RISK ANALYSIS LAB
Creating probabilistic methods to make our cities more resilient
Research Overview​
​
We tackle some of the most urgent and complex challenges in disaster risk management and urban resilience. By advancing disaster risk modeling, we aim to deepen the understanding of how natural and human-made hazards affect civil infrastructure. Our mission is to provide innovative engineering insights that help policymakers create more resilient cities.
Our group's research spans three deeply intertwined dimensions (see three axes in the figure below). In the first dimension, we study how multiple hazards (e.g., earthquakes, floods, hurricanes, and cyberattacks) can impact regions across multiple spatiotemporal scales (e.g., from neighborhoods to entire cities and with loads that can last from seconds to days). In the second dimension, we investigate failures of multiple infrastructure systems (e.g., housing, transportation, healthcare, and energy) during extreme events, characterizing how their effects cascade to many communities. In the third dimension, we develop new algorithms (e.g., in uncertainty quantification, stochastic programming, and artificial intelligence) to close fundamental computational gaps that currently hinder our ability to model disaster risks in the largest cities in the world. Some current projects are listed below.
​
Earthquake Resilience of Hospital Systems
​
Hospitals must protect the health and life of communities. Unfortunately, extreme events can easily disrupt them due to infrastructure damage, leaving many communities without access to critical healthcare services for months or even years. Earthquakes can be particularly catastrophic as they also injure thousands of people in seconds, creating high demands for healthcare services. In this project, we create new models to study the impacts of earthquakes on the healthcare system, capturing critical dynamical processes, cascading effects, and interdependencies that emerge in large cities. We have developed large-scale case studies for the Bay Area, US, and Lima, Peru, to elucidate real-world strategies to enhance the seismic resilience and emergency response of hospital systems and improve access to medical services for vulnerable communities.
System-level strategies to mobilize patients, place temporary facilities and reconstruct hospitals are critical for many communities to receive timely (and many times life-saving) access to healthcare services after earthquakes . Images from our articles.
Enhancing ​electricity resilience with clean energy infrastructure
​
Solar generation will become a major and global source of clean energy by 2050. We investigate how solar infrastructure will reshape the resilience of our power grids, which often experience large failures due to extreme events, e.g., Hurricane Ian (2022) with 2.6 M outages, Hurricane Sandy (2012) with 8 M outages. In this project, we create probabilistic frameworks and engineering models to evaluate vulnerability and risk for these new infrastructure due to extreme events, such as hurricanes, at different spatial scales, from entire regions and cities to neighborhoods and households. We also improve models for outage predictions in the traditional grids to create new-generation power grid models that are able to quantify the value of renewables (e.g., rooftop panels) and other new technology (e.g., behind-the-meter batteries, microgrids) for resilience for vulnerable communities. We have developed regional-level case studies in the entire Eastern United States and community-level case studies in New Jersey as testbeds for hurricane risks.
​
Rooftop panels and behind-the-meter batteries can provide households with access to electricity when the main power grid is damaged and non-functional. We evaluate the benefits of these new infrastructures for electricity resilience at different spatial scales. Figures from our articles.
The ​value of urban flood monitoring for risk modeling and management
​
Coastal cities are at exceptionally high flood risk because they can be exposed to high tides, storm surges, high-intensity precipitation, and sea-level rise, and at the same time, they can concentrate millions of people and massive critical infrastructure. Probabilistic risk analysis provides a rigorous quantitative framework to assess multiple possible flood scenarios to inform risk mitigation strategies (e.g., flood protection investments and flood emergency preparedness). However, probabilistic flood methods face fundamental limitations in coastal cities since flood data within coastal-urban systems are exceptionally perishable and scarce; thus, existing runoff models are poorly validated. This collaborative project works with New York City (NYC)’s FloodNet Initiative to evaluate the value of novel sensor networks for hyper-local urban flood monitoring to advance risk modeling and management. Using NYC as a testbed, the project develops a location-agnostic framework to create validated flood models for vulnerable cities.
This project seeks to use sensors that measure local urban floods (upper left) for improving flood risk modeling and mitigation. We have conducted expert elicitation experiments (lower left) to highlight where sensors can help different stakeholders in New York City. Figures from our articles and the FloodNet Initiative.
Funding sources​
​