Uncertainty Quantification (UQ) is a critical component of digital twin systems, as it enhances confidence in actionable information and builds trust among end users. Using water quality, flooding, and agriculture as representative application domains, this project focuses on improving prediction reliability through explicit representation and communication of uncertainty in interconnected environmental and agricultural systems. The project’s primary objective is to develop a unified UQ framework that systematically quantifies, propagates, and communicates predictive uncertainties across water quality, flooding, and agricultural models to support risk-informed decision-making. All models will generate probabilistic forecasts with confidence intervals and uncertainty bounds, enabling decision-makers to assess forecast reliability and quantify associated risks. The framework will be demonstrated through three representative use cases: (1) chlorophyll-a gap filling using satellite observations, (2) river gauge–based flood forecasting, and (3) climate-driven agricultural yield prediction. Project deliverables include operational forecasting systems that provide uncertainty-aware outputs to support water quality management, flood risk mitigation, and agricultural planning.
