DEEP LEARNING-Based DISASTER MANAGEMENT PLANNING and RISK ANALYSIS of FLASH FLOOD-PRONE REGIONS
Abstract
An improved ability to predict flood events reduces risk to life and property. This research focuses on the use of deep learning algorithms to increase the accuracy and timeliness of flash flood predictions. Historical rainfall and Geographic Information System (GIS) data are used as inputs to a set of deep learning models. These models are then trained using historic flash flood event data to capture relationships between the weather and geographic data. Greene County, Missouri is used for this study as it encounters several weather events that have at times led to flash flood events. A risk analysis study is performed using this data to advance the current flash flood management strategies for the region. The data-driven approach is applied to publicly available data sourced from the United States Geological Survey (USGS), National Oceanic and Atmospheric Administration (NOAA), and National Weather Service (NWS).
Recommended Citation
B. Kanwar and S. Corns, "DEEP LEARNING-Based DISASTER MANAGEMENT PLANNING and RISK ANALYSIS of FLASH FLOOD-PRONE REGIONS," 2021 ASEM Virtual International Annual Conference "Engineering Management and The New Normal", pp. 360 - 368, American Society of Engineering Management, Jan 2021.
Department(s)
Engineering Management and Systems Engineering
Keywords and Phrases
Deep learning; Flash floods; Flood risk; Missouri; Neural networks
Document Type
Article - Conference proceedings
Document Version
Citation
File Type
text
Language(s)
English
Rights
© 2024 American Society of Engineering Management, All rights reserved.
Publication Date
01 Jan 2021
Comments
Missouri Department of Transportation, Grant TR202023