Notice of Intent to Sole Source - Bayesian Enhancement of Flood Risk Management Depth-Damage Modeling

The U.S. Army Corps of Engineers Humphreys Engineer Center Support Activity (HECSA) has published a Special Notice announcing its intent to award a sole source fixed-price contract to Duke University for "Bayesian Enhancement of Flood Risk Management Depth-Damage Modeling." This notice was posted on August 14, 2025, and serves as a formal announcement under Federal Acquisition Regulation (FAR) subpart 5.2 rather than a competitive solicitation. The contracting agency intends to work with Dr. Mark Borsuk's research team at Duke University, which possesses specialized expertise in artificial intelligence, machine learning, and flood damage analysis. The contract falls under NAICS code 541715 for Research and Development in the Physical, Engineering, and Life Sciences, reflecting its technical and scientific nature. This procurement represents the government's intention to leverage Duke University's unique capabilities in Bayesian statistical modeling for flood risk management applications.

The primary objective of this contract is to develop advanced Bayesian flood damage analysis models that will enhance depth-damage functions used in flood risk management, moving beyond existing deterministic approaches to create probabilistic models with improved accuracy and uncertainty quantification. The 12-month project will involve compiling structure-level flood loss datasets, developing hierarchical Bayesian regression models across at least four U.S. regions, and producing depth-damage functions compatible with USACE's HEC-FDA workflow. Any firm believing it can perform the required services may submit written notice with supporting evidence to the Contracting Officer within 15 days of publication, which establishes a deadline of August 29, 2025. However, responses must demonstrate clear and convincing evidence that competition would be advantageous to the government, as the determination to compete remains solely at the government's discretion. The government emphasizes that acknowledgment of responses will not be provided, nor will respondents be notified of evaluation results, and no payment will be made for preparation of submitted information. Key deliverables include Bayesian model code, probabilistic parcel-level depth-damage functions with credible intervals, technical documentation, and guidance for Project Delivery Team workflows, all designed to improve flood damage estimation accuracy and support decision-making for flood risk management projects.