CSR
RS Home Features Texas Synergy Research Publications RS Info CSR Home
Floodplain Modeling Based on Data Fusion of Polarimetric SAR, Interferometry, and Laser Altimetry

 


Principal Investigators:

Jakob van Zyl and Yunjin Kim, NASA Jet Propulsion Laboratory

Melba Crawford and Jay Famiglietti, University of Texas at Austin


The human and economic impact of flooding in both inland and coastal lowlands is enormous. Inland riverine watersheds are subject to episodic rainfall events, while low relief coastal watersheds are prone to flooding from both uplands and from the sea.  Hydrologic models can be used to predict the occurrence of flooding provided that the required input fields are accurately described.  Of these inputs, some of the most important include an accurate description of the topography of the floodplain, as well as the state of soil moisture and vegetation cover.  The main objective of this research is to provide these critical input fields for a hydrological model to enable more accurate prediction of flooding events and flood extent in riverine and coastal floodplains.

Recent advances in polarimetric SAR show promise for augmenting the capability of traditional interferometric SAR and laser altimetry.  In particular, a polarimetric topography technique provides useful slope information, and polarimetric interferometry may be used to decompose the topographic response into vegetation and ground surface contributions.  All of these sources of remotely sensed topography, land cover characterization, and even soil moisture provide valuable information for hydrologic mapping for flood prediction and management.  Unfortunately, currently available analysis techniques are typically applied to individual data sources, and the information is later integrated subjectively.  Methodology which is capable of jointly and objectively extracting and using information from a suite of data types in mapping terrestrial features over extensive areas is critically needed, but not yet developed.  Development of an integrated approach that utilizes the combined capability of polarimetric interferometry, polarimetric topographic mapping, and laser altimetry for topographic mapping of flood-prone areas is the primary objective of the proposed research.  Additionally, a secondary goal is to provide soil moisture and land cover characterization inputs from the same remotely sensed data.  As such, new techniques are being developed to exploit the capabilities of remote sensing to provide the required high resolution data sets, with the goal of exploring the utility of these remotely-sensed products for improving the predictive reliability of flood hydrological models.  For this project, data from two primary sites, one on the Finke River of Australia and the other from a riverine system in Texas are being analyzed.