Predictive Deep Learning for Flash Flood Management

This research was completed in tandem as a project funded through MoDOT and the Mid-America Transportation Center. It used deep learning methods, along with weather information from NOAA/National Weather Service and geospatial data from the USGS National Map and other public geospatial data sources, to develop forecasting tools capable of assessing the probability of flash flooding in high risk areas. These tools build on existing models developed by the USGS, FEMA, and others and were used to determine evacuation routing and detours to mitigate the potential for loss of life during flash floods. The project scope included analysis of publicly available data in Greene county in and around Springfield, MO as part of a pilot project in Missouri. This data was then used to determine the probability of flash flooding in order to model evacuation or detour planning modules that can be implemented to assure the safety of the community and highway personnel. These modules used existing rainfall data and weather forecasts in a three-day sliding window to include soil moisture in the flash flood predictions. The transportation safety or disaster planner can use these results to produce planning documents based on geospatial data and information to develop region-specific tools and response methods to potential flash flood events.

Language

  • English

Media Info

  • Media Type: Digital/other
  • Edition: Final Report
  • Features: Figures; Maps; References; Tables;
  • Pagination: 51p

Subject/Index Terms

Filing Info

  • Accession Number: 01762778
  • Record Type: Publication
  • Report/Paper Numbers: cmr 21-001, Project number TR202023
  • Contract Numbers: MoDOT project # TR202023
  • Files: TRIS, ATRI, USDOT, STATEDOT
  • Created Date: Jan 12 2021 1:56PM