Extreme weather events are testing the limits of how cities and communities prepare for, respond to, and recover from disasters. In this session, presenters showcase how high-resolution imaging, UAV data collection, and AI-powered analytics are transforming emergency management—from rapid post-event assessment to long-term resilience planning. Through case studies spanning tornado recovery in St. Louis and hurricane resilience efforts in The Bahamas, attendees will see how advanced aerial imaging, feature extraction, and digital twin technologies enable faster situational awareness and data-driven recovery. Presentations will also explore how integrating UAV photogrammetry with AI-based terrain modeling supports real-time risk mapping, community engagement, and local capacity building.
Together, these projects illustrate how geospatial innovation is closing the gap between crisis response and continuity planning—empowering municipalities and coastal communities alike to adapt, rebuild, and thrive in a changing climate.
The following presentations will be shared in this session:
From Crisis to Continuity: High-Resolution Emergency Response and Digital Twin Integration for Municipal Resilience
Presented by Jamie Reford, Phase One, and Scott Merritt, Surdex Corporation
The May 16, 2025 tornado that struck St. Louis created an unprecedented opportunity to demonstrate how cutting-edge aerial imaging technology, AI-powered feature extraction, and comprehensive digital twin planning can transform emergency response and long-term municipal resilience. This presentation showcases the complete workflow from immediate crisis response to comprehensive urban digital modeling, highlighting how Phase One’s imaging systems and Surdex’s geospatial expertise delivered actionable intelligence within 48 hours and established the foundation for a city-wide digital twin.
This session presents a real-world case study of the complete emergency response-to-recovery pipeline, demonstrating how high-resolution imaging capabilities enable both immediate damage assessment and long-term municipal planning through integrated AI/ML workflows and digital twin implementation.
Integrating ArcGIS Pro’s MaxEnt and Forest-Based Machine Learning Tools for Rapid Post-Hurricane Helene Landslide Susceptibility Mapping
Presented by Grace Braver, East Tennessee State University
Landslides triggered by extreme rainfall events pose increasing threats to infrastructure, transportation networks, and communities across the Southern Appalachian region. This project focuses on the Nolichucky headwaters in East Tennessee, which experienced widespread slope failures following Hurricane Helene (2024). Leveraging ArcGIS Pro’s integrated machine learning environment, this study compares two predictive modeling approaches: Maximum Entropy (MaxEnt) and Forest-based & Boosted Classification and Regression (FBCR) to map landslide susceptibility and assess hazard potential across the area. Both models were developed entirely within ArcGIS Pro, eliminating the need for external coding environments and enabling rapid model iteration. Environmental predictor variables included lidar-derived terrain indices (slope, curvature, roughness, elevation, aspect), soil erodibility (K-factor), geologic units, and NDVI from post-Helene Sentinel-2 imagery. Model performance was evaluated using AUC-ROC, and confusion matrix outputs to assess predictive accuracy and interpret variable importance.
The resulting susceptibility maps identify high-risk slope zones and infrastructure corridors vulnerable to debris flows during extreme rainfall. By comparing MaxEnt’s presence-only predictive framework with the ensemble capabilities of FBCR, this work demonstrates how GIS-integrated machine learning tools can provide fast, data-driven risk assessments for real-world applications in hazard mitigation, infrastructure planning, and emergency management.
This work highlights the growing potential of ArcGIS Pro as a full-stack geospatial modeling platform, enabling agencies and private industry partners to operationalize machine learning and remote sensing data for resilient terrain and infrastructure management across mountainous regions. Results demonstrate that both the MaxEnt and FBCR offer reliable, scalable approaches for assessing landslide risk in data-limited, geologically complex terrains. These tools support future efforts in predictive risk mapping, emergency response planning, and climate-resilient infrastructure development in the southern Appalachians.
Additional presentations to be announced.
