Sacramento-San Joaquin Levee Health Assessment Utilizing NASA UAVSAR Data

EarthzineDEVELOP Virtual Poster Session, Original

Images A and B depict levee seeps on June 16, 2011, at 4:14 a.m. and 7:00 a.m. local time, respectively. Image C depicts the change in classed pixels from the two seep detection products. Image D depicts the frequency of seeps along the levee over the 32 UAVSAR flights. Image Credit: DEVELOP JPL Team.
Images A and B depict levee seeps on June 16, 2011, at 4:14 a.m. and 7:00 a.m. local time, respectively. Image C depicts the change in classed pixels from the two seep detection products. Image D depicts the frequency of seeps along the levee over the 32 UAVSAR flights. Image Credit: DEVELOP JPL Team.

Images A and B depict levee seeps on June 16, 2011, at 4:14 a.m. and 7:00 a.m. local time, respectively. Image C depicts the change in classed pixels from the two seep detection products. Image D depicts the frequency of seeps along the levee over the 32 UAVSAR flights. Image Credit: DEVELOP JPL Team.

Authors: Austin Madson, Robin Sehler, Scott Barron

Mentors/Advisors (affiliation): Benjamin Holt, (NASA Jet Propulsion Laboratory), Dr. Cathleen Jones, (NASA Jet Propulsion Laboratory)

Team Location: Jet Propulsion Laboratory, Pasadena, California

Abstract: In previous DEVELOP projects, the team has studied methods to accurately determine flood extent and levee seep locations using UAVSAR data collected along the Mississippi River during the Spring 2011 floods. This project applied previously developed techniques to the levees in the Sacramento Delta and nearby urban areas, as well as new techniques created with the Delta specifically in mind. Past DEVELOP projects have shown the value of masking the data sets to isolate areas along the levees, then applying algorithms to the reduced data set to detect areas of change. The first phase of the project applied levee masks to several islands (Sherman, Twitchell, Bouldin, Bradford, Jersey, Jones Tract) in the west and central delta where ground validation data sets are available, and used the masked data sets to identify areas along the levees where the most change has occurred. After application of the levee mask, the seep detection algorithm developed in spring 2012 was tested on datasets that were collected during high water events. This project partnered with both the California Department of Water Resources, whose interest is confined to the legal boundaries of the Delta, and the U.S. Army Corp of Engineers, who has responsibility for levees extending into local urban areas, including the cities of Stockton and Davis. Having the levee masks implemented in a standard analysis package enabled rapid production of GIS products showing levee leak locations during an emergency. The team also created a high-water disaster model, which highlights the levees most likely to break or leak the most, during a high water event. This project provides usable products for response crews during a future flood emergency.

Transcript available here.