Earth Observation (EO) based critical infrastructure exposure models and flood forecasting techniques for risk monitoring and management for the city of Vadodara, India

Abstract:

Every year during monsoon season in India, which spans roughly between April and September, heavy rain and floods affect millions of people in the city of Vadodara and the state of Gujarat. Focused on the city of Vadodara to understand critical infrastructure (CI) exposure from flood risk, this research uses neural network algorithms to identify heavy industrial and power plant locations in and around the city boundaries. The CI exposure outputs will be augmented with existing GIS datasets from the city of Vadodara, and remaining gaps will be filled in using a simulation approach. By collecting a series of Synthetic Aperture Radar (SAR) and optical imagery, flood inundation extents for rainfall events during July-August (2019 and 2020) in Vadodara and its surrounding areas will be established. Event selection will be based on available imagery dates and quality (cloud-free, high resolution for optical), particularly during the monsoon months. The inundation extents will be overlaid on critical infrastructure exposure particularly the heavy industrial areas around Vadodara to examine impacted CI assets. This research is a collaboration between two NASA funded disasters program projects i. “Open Critical Infrastructure Exposure for Disaster Forecasting, Mitigation and Response,” which is investigating novel methods of extracting CI data from Earth Observation (EO) to model the catastrophic impacts of infrastructure disruption, and ii. “Advancing Access to Global Flood Modeling and Alerting”, which is developing a model for integrating flood inundation information from multiple sources into the DisasterAWARE platform, providing a single source of global information on floods that is supported by a common, normalized data model. Using Vadodara as a pilot test, this collaborative research aims to establish the utility of extraction of flood boundaries and dissemination of flood alerts through the DisasterAWARE platform. One of the key expected outcomes is to increase awareness of EO-based CI data in decision making and using flood modeling outputs for DRR/DRM for Vadodara city officials, Gujarat Power Research and Development (GPRD) at IIT-Gandhinagar and Gujarat Urja Vikas Nigam Limited (GUVNL) and other and stakeholders who are partners of the project.