Projects

Research Problem

Wildfires increasingly threaten people, the built environment, and ecosystems as climate conditions, human activities at the wildland-urban interface, and land-use patterns continue to change. Existing wildfire risk-management strategies often rely on static, outdated, and coarse-scale risk estimates. They also provide limited opportunities for sustained citizen engagement, even though individual mitigation decisions can affect neighboring properties and the broader community. Limited information exchange between governments and citizens creates a barrier to informed and coordinated wildfire-risk reduction.

SHAPE Lab Contributions

• Characterize interdependent wildfire-mitigation decisions between governments and citizens using evolutionary game theory.
• Develop a dynamic wildfire risk-assessment model that incorporates ongoing data collection and processing.
• Examine how citizen engagement and multidirectional information exchange influence individual and collective mitigation decisions.
• Integrate decision and risk models into a prototype next-generation Community Wildfire Protection Plan platform.
• Evaluate the technical feasibility and cost-effectiveness of the prototype platform while incorporating user feedback.
• Translate project findings into educational courses, practitioner training, and public outreach materials.

Expected Impact

The project seeks to advance wildfire planning from traditional static assessments toward a dynamic and interactive process. The resulting approach will help governments and citizens exchange information, coordinate mitigation actions, and engage in collective, risk-informed decision-making. By strengthening collaboration between local governments and residents, the project can help expand responsibility for wildfire-risk reduction from emergency personnel to the broader community.

Research Problem

Wildfire-related power outages can significantly affect people, businesses, and communities in different ways. This project brings together key partners, including utility companies, ratepayers, regulators, local agencies, and health and energy organizations, through a community collaborative. We study wildfire-related outages, develop actionable solutions, and establish a more effective and balanced resilience planning framework.

SHAPE Lab Contributions

• Model wildfire ignitions across diverse landscapes, community settings, and ignition sources.
• Build an advanced wildfire simulator for regional-scale risk assessment.
• Create a cluster-based active-learning scenario reduction framework for probabilistic wildfire risk assessment.
• Predict wildfire-related power outages using satellite imagery.
• Evaluate community-centric solutions using an interactive dashboard currently under development, which will display the spatial distribution of benefits, costs, and burdens associated with different resource-allocation choices across the service areas of our utility partners.

Broader Project Collaboration

The broader multidisciplinary project team contributes complementary expertise in wildfire modeling, power grid resilience planning, public health, and community engagement.

Expected Impact

The project will help utility and community partners compare resilience strategies and understand how the benefits, costs, and burdens associated with different resource-allocation decisions are distributed across utility service areas. The resulting framework will support technically sound and community-informed planning for wildfire-related power outages.

Research Problem

Wildfire evacuations are becoming increasingly difficult as fires become faster and less predictable. Existing evacuation guidance often provides standardized instructions without accounting for household-specific needs, road capacity, visitor populations, transportation limitations, livestock, mobility challenges, or traffic congestion. As a result, current approaches do not adequately balance individual constraints with broader community safety goals. This project addresses that gap through a customizable, LLM-based co-planning tool that grounds personalized recommendations in local wildfire and transportation data and iteratively balances them with network capacity and community-level safety.

SHAPE Lab Contributions

• Develop an LLM-driven interface that helps residents create personalized evacuation plans based on their household needs and constraints.
• Use retrieval-augmented generation to ground AI recommendations in local road networks, parcel information, wildfire data, refuge areas, and emergency-planning protocols.
• Adapt EVAC-Advisor responses to different levels of user trust, technical knowledge, and skepticism toward AI-generated guidance.
• Develop a simulation-based optimization engine that balances personalized departure times, routes, destinations, and shared-mobility needs with community-level evacuation goals.

Broader Project Collaboration

The multidisciplinary project team combines wildfire modeling, transportation simulation, optimization, artificial intelligence, emergency planning, and social-science methods. Focus groups and cognitive interviews with residents, fire professionals, emergency managers, land-use planners, and other local stakeholders will support the co-development and evaluation of EVAC-Advisor.

Expected Impact

EVAC-Advisor will help residents and emergency managers establish personalized and coordinated evacuation plans before wildfire emergencies. By balancing household needs with network capacity, shared mobility, and clearance time, the system aims to improve community evacuation efficiency and safety.