This spatio-temporal composite flood extent product was produced using 34 flood extents derived from SAR and Multispectral imagery acquired over the course of 11 days (Aug 26th – September 5th, 2017) by NASA MSFC SPoRT, Copernicus EMS, MDA Systems and ARIA NASA JPL/Caltech throughout Hurricane Harvey using a wide range of sensors and sensor types as well as different derivation methodologies.
Layer names and descriptions are as follows:
|Layer Name||Description||File Name|
|Coverage||All AOIs/footprints merged and summed together to indicate overlapping coverage throughout the event. The pixel value indicates the number of times the pixel was sampled by a SAR- or MS- sensor acquisition used in the analysis.||Harvey_Coverage.tif|
|Flood Extents Combined||All flood extents merged and summed together to indicate the number of times a pixel was identified as flooded throughout the event. The pixel value indicates the number of times the pixel was identified as “flooded” in a SAR- or MS- derived flood extent.||Harvey_FloodExtentsCombined.tif|
|Flood Detection %||All flood extents merged and summed together, divided by coverage, to indicate water detection percentage throughout the event. The pixel values range from 0% – 100% and equates to the number of times the pixel was identified as “flooded” in a SAR- or MS- based flood extent, divided by the number of times the pixel was sampled.||Harvey_FloodDetection.tif|
All flood extents were resampled to 30m spatial resolution.
Access & Use Information
Public: This dataset is intended for public access and use.
Downloads & Resources
In the Flood Detection % Raster, some pixels that were identified as “flooded” but only imaged once (ex. 1/1=100%) will have an equal detection % as pixels that were imaged several times and identified “flooded” each time (ex. 5/5=100%). It is best to view both the Coverage and the Flood Detection layers together to better understand how coverage affects the detection percentage.
Some features with smooth surfaces (such as roads and airports) have similar backscatter signatures to water and may be falsely detected as flooded areas. In addition, water detection in urban areas is limited due to spatial resolution and building-related shadow effects on radar signal.
The flood detection % product weighs all input flood extent products equally despite differences in production methodology, spatial resolution, date/time of acquisition, and spatial coverage.
Credit goes to the following who provided source data, data processing services, and derived products used for this map product:
SPoRT NASA MSFC
ARIA NASA JPL/Caltech
NGA for the Northernview Agreement (provides access to MDA Systems, Ltd.)
University of Alaska Fairbanks
Sentinel-1: Contains modified Copernicus Sentinel data (2017), processed by ESA
ALOS-2: Derived from ALOS-2 Products, (c) JAXA, METI (2017)
COSMO-SkyMed: Derived from COSMO-SkyMed Products, (c) ASI (2017)
RADARSAT-2 Derived from RADARSAT-2 Products, (c) MacDONALD, DETTWILER AND ASSOCIATES LTD. (2017)
TerraSAR-X: Derived from TerraSAR-X Products, (c) DLR (2017)