Authors: Anupal Baruah – University of Alabama
Title: Forecast-Horizon Based Probabilistic Approach for Communicating Uncertainty in Operational Flood Inundation Mapping
Presentation Type: Poster Presentation
Abstract: The NOAA-OWP operational Flood Inundation Mapping (FIM) services currently rely on the Height Above Nearest Drainage (HAND)-Synthetic Rating Curve (SRC) framework. This framework uses the hydrologic forecasts from the National Water Model (NWM) and River Forecast Centers to produce deterministic, binary flood inundation maps at the HUC-8 scale. While computationally efficient and operationally scalable, the current framework provides limited capability to communicate uncertainty associated with hydrologic predictability. Recent efforts within NOAA-OWP emphasize probabilistic flood mapping through distribution fitting to long-term NWM hydrological simulations and terrain parameters (Slope and roughness).
The proposed project develops a complementary, forecast-centered Research-to-Operations (R2O) pathway to generate probabilistic flood inundation maps utilizing NWM forecasts within an operational forecast horizon at a given reference time and assigns weights based on: (1) ensemble mean of forecast FIMs, (2) simple and exponential moving average of forecast FIMs, (3) non-linear scaling relationships between forecast lead time and maximum inundated area, (4) temporal weighting schemes derived from probability distribution functions across short- and medium range forecast, (5) Bayesian approach and (6) Multi-Model Streamflow Forcing. The proposed approach is explicitly designed for real-time operations and supports uncertainty communications through different tiers of probabilistic FIMs that evolve with successive forecast updates. A large-scale benchmark FIM dataset, FIMbench, will be used to evaluate the probabilistic maps using the FIMEval framework.