The Application of Artificial Intelligence in Flood Risk Assessment and Mitigation in the State of Texas

Project summary:
The purpose of this project is to determine the applicability of artificial intelligence (AI) and machine learning (ML) technologies to enhance flood risk assessment and mitigation in Texas. The guidelines, recommendations, and workflows developed in this project will result in a roadmap for TWDB to leverage advanced AI models to optimize current modeling efforts and processes, such as Base Level Engineering (BLE), coastal modeling, and flood risk mitigation.
As part of this project, Texas A&M University Engineering Experiment Station (TEES) will collaborate with the University of Texas at Austin Computational Hydraulic Group (UT-CHG) to provide an early implementation test case for the roadmap’s recommended AI governance workflow. This effort aligns with the roadmap’s Compound Coastal Flood Mapping use case and will leverage ADCIRC simulation outputs generated by UT-CHG to develop and evaluate an ML model for characterizing compound flooding along the Texas coast. The work will analyze a range of coastal flooding scenarios to estimate flood occurrence, identify primary drivers of flood (including storm surge, rainfall-runoff, and the timing of their interaction), and quantify the degree to which coastal flooding results from multiple interacting processes. The trained model may serve as a decision-support workflow that produces flood classifications, identifies the dominant drivers of flooding, and produces spatial insights and feature-importance metrics, providing practical insights for coastal planning.
Project deliverable(s):
  • A technical memorandum presenting key insights from a comprehensive literature review on the capabilities and limitations of various AI techniques for flood risk assessment and mitigation.
  • A technical memorandum identifying flood-related datasets which can be leveraged to develop and implement AI applications in flood risk assessment and mitigation.
  • A strategic roadmap for applying AI in flood assessment and mitigation across Texas.
  • Guidelines for implementing machine learning workflows in Snowflake using various flood-related datasets.
  • A technical memorandum identifying potential concerns related to data governance and risk reduction for AI applications in flood risk assessment and mitigation.
  • Develop, validate, and package an interpretable machine learning model of compound coastal flooding in Texas using ADCIRC simulation outputs.
Contractor (and Principal Investigator, if appropriate):
Texas A&M University Engineering Experiment Station, Dr. Ali Mostafavi
Contract amount:
$155,000
Project lead:
Krutikkumar Patel, Ph.D., CTCM, and Kimia Karimi, Ph.D.
Project timeline:
March 2025 - May 2027