Flood Potential Model Using Machine Learning
Status
Not Started.
Description
The Flood Potential Portal, developed by the U.S. Forest Service and Colorado State University, enhances the understanding of flood variability and quantifies design flood discharges and flood frequency relationships. The software assists practitioners with assessments to support infrastructure decisions, including designing road-stream crossings. This project develops and implements new capacities, specifically: (1) advanced ensemble machine learning to improve forecasting of temporal trends in large floods due to changes in the underlying climatic, physiographic, and ecohydrologic conditions; and (2) ensemble modeling for streamgage flood-frequency, to address shortcomings of existing state-of-practice methods (e.g. Bulletin 17C) including assumptions about stationarity and identically distributed observations.
Deliverable
Software update
For More Information
Digital Object Identifier (DOI): Pending
ORCID for the Principal Investigator: Currently Unavailable