Award Abstract # 1760582
RAPID: Fast Reconstruction of Flood Hydrographs in the Houston Metropolitan Area during Hurricane Harvey Based on Image Processing and In-situ Measurements

NSF Org: EAR
Division Of Earth Sciences
Recipient: LOUISIANA STATE UNIVERSITY
Initial Amendment Date: September 20, 2017
Latest Amendment Date: September 20, 2017
Award Number: 1760582
Award Instrument: Standard Grant
Program Manager: Holly Barnard
EAR
 Division Of Earth Sciences
GEO
 Directorate for Geosciences
Start Date: October 1, 2017
End Date: September 30, 2018 (Estimated)
Total Intended Award Amount: $61,167.00
Total Awarded Amount to Date: $61,167.00
Funds Obligated to Date: FY 2017 = $61,167.00
History of Investigator:
  • Navid Jafari (Principal Investigator)
    njafari@lsu.edu
  • Xin Li (Co-Principal Investigator)
  • Qin Chen (Co-Principal Investigator)
Recipient Sponsored Research Office: Louisiana State University
202 HIMES HALL
BATON ROUGE
LA  US  70803-0001
(225)578-2760
Sponsor Congressional District: 06
Primary Place of Performance: Louisiana State University & Agricultural and Mechanical College
LA  US  70803-2701
Primary Place of Performance
Congressional District:
06
Unique Entity Identifier (UEI): ECQEYCHRNKJ4
Parent UEI:
NSF Program(s): Hurricane Harvey 2017
Primary Program Source: 01001718DB NSF RESEARCH & RELATED ACTIVIT
Program Reference Code(s): 7914, 9150
Program Element Code(s): 071Y00
Award Agency Code: 4900
Fund Agency Code: 4900
Assistance Listing Number(s): 47.050

ABSTRACT

The ability to construct flood hydrographs in urban areas in real-time during flash floods, hurricanes, and other extreme weather events is difficult because of the low spatial density of water level measurements and the complex interactions of built infrastructure, ground topography, and natural landscape with flowing water. The goal of this RAPID project is to leverage perishable images and video footage from traffic intersection and interstate highway cameras, major news media outlets, and social media along with reference objects/points. Subsequent photo image processing, scaled to the reference objects, will enable development of a more continuous, accurate hydrograph in the Houston metropolitan area. By reconstructing the flood hydrographs at a large number of locations in flooded highways, streets and residential subdivisions, high-resolution, process-based urban inundation modeling from hurricane-generated surge and rainfall will become significantly more accurate. For example, it will facilitate a better understanding of transport of sediments and pollutants in and out of Houston during Hurricane Harvey. Such a model validated by the reconstructed hydrographs will also aid state and local governments in making timely evacuation decisions for low-lying areas to mitigate the impact of similar hurricane-induced hazards.

This RAPID project based on reconstruction of flood hydrographs in Houston using image processing and in-situ measurements has significant intellectual merit: (1) The proposed methodology is innovative and creative because it does not employ any traditional stream gages. Instead, it relies on a unique form of existing data employed for a new application. (2) The reconstructed flood hydrographs will significantly improve the understanding of the hydrological processes of this unprecedented flood event caused by the extreme rainfall of Hurricane Harvey. (3) The data will benefit the development of a new flood model for Houston. (4) The developed algorithms and software for processing the traffic image data, which will be available on the NHERI DesignSafe-CI platform, can be readily applied to many other flood-prone urban centers, such as New York City, New Orleans, and Miami.

PROJECT OUTCOMES REPORT

Disclaimer

This Project Outcomes Report for the General Public is displayed verbatim as submitted by the Principal Investigator (PI) for this award. Any opinions, findings, and conclusions or recommendations expressed in this Report are those of the PI and do not necessarily reflect the views of the National Science Foundation; NSF has not approved or endorsed its content.

Hurricane Harvey led to unprecedented flooding in the Houston and Beaumont metropolitan areas, Texas. The ability to construct flood hydrographs in urban areas in real-time during flash floods, hurricanes, and other extreme weather events is difficult because of the low spatial density of water level measurements and the complex interactions of built infrastructure, ground topography, and natural landscape with flowing water. This RAPID project aimed to leverage traffic cameras and webcam videos along with photo image processing to advance flood inundation mapping in urban areas because time-lapse cameras are continuously photographing the rising and falling water levels.

The Intellectual Merit of this RAPID project includes (1) creating a procedure to download image and video data from traffic cameras and internet webcams and (2) using computer vision and machine learning techniques to analyze images for estimating flood hydrographs. For outcome (1), we downloaded approximately 14.5 GB of image and video data from Hurricane Nate (October 7-8, 2017), Tropical Storm Gordon (September 4-6, 2018), Hurricane Florence (September 14-19, 2018), and Hurricane Michael (October 9-11, 2018). This dataset is archived in the NHERI Designsafe-CI portal so other researchers can use it to train their computer vision and machine learning methods. For outcome (2), we developed a deep-learning based segmentation algorithm to automatically detect and label the water and reference objects, and estimate flood elevation from a time-lapse video of Buffalo Bayou in downtown Houston, Texas during Hurricane Harvey. The algorithm- and program-estimated water elevations were recorded as hydrographs and compared to in-situ measurements and nearby stream gages. (The results demonstrate desirable estimation consistency in these experiments).

The Broader Impact of this RAPID project includes improving community resilience to floods caused by hurricanes and other extreme weather events. The results are leading to the ability to better model urban flooding and transport of pollutants from overland fluid flow during these events. This includes the planning and design of residential, commercial and industrial areas, evacuation shelters, schools, hospitals, roadways etc., all of which may unavoidably be located in inundation zones. Improved understanding of urban flooding as well as city and regional planning and design can increase life-safety and decrease property loss and environmental contamination. The developed methodology can be used to monitor flood at traffic stops in real time, which can facilitate emergency response and rescue efforts. Three graduate students and an undergraduate student participated in this research effort.

 


Last Modified: 02/13/2019
Modified by: Navid Jafari

Please report errors in award information by writing to: awardsearch@nsf.gov.

Print this page

Back to Top of page