UAVSAR Classified Images for the Flooding in Texas July 2025

Dates of Images: July 9, 2025 Date of Next Image: None Expected Summary: These maps show the results of a machine learning classification applied to UAVSAR data to identify open water flooding and inundation under tree canopy.  Colored areas in the image show detected flooding, where the color shows the type of landcover which was flooded (see legend below).  Non-flooded areas are transparent.  The classification uses a U-Net machine learning algorithm which was trained using UAVSAR data from previous flood events.  UAVSAR provides polarimetric synthetic aperture radar data from which different types of scattering mechanisms can be observed, which provides information on the type of flooding.  For example, flooding underneath vegetation produces strong double-bounce scattering from the water surface and tree trunks.  However, strong double-bounce scattering from urban areas aligned with the radar viewing direction, such as in Austin, TX, can produce erroneous classifications.  Non-flooded vegetation is generally dominated by volume scattering from the forest canopy.  In comparison, open water flooding exhibits weak radar returns in all polarizations. Suggested Use: Classified Images: Blue   = Open Water Yellow = Flooded Developed Green  = Flooded Vegetation Orange = Flooded Crop Satellite/Sensor: UAVSAR airborne L-band synthetic aperture radar (SAR) aboard a NASA Gulfstream C-20A jet. Resolution: 10 meters Credits: UAVSAR data courtesy of NASA/JPL-Caltech Esri REST Endpoint: See URL section on right side of the page WMS Endpoint: https://gis.earthdata.nasa.gov/portal/services/DISASTERS_202507_FLOOD_TX/202507_UAVSAR/MapServer/WMSServer Data Download: https://uavsar.jpl.nasa.gov/cgi-bin/download-files.pl

Dates of Images: 10/11/2024, 10/12/2024, 10/13/2024, 10/14/2024 Date of Next Image: None Expected Summary: These UAVSAR false color RGB images provide a unique look at the Earth's surface that can be used for identifying flooding and inundation under tree canopy. The overlay of the intensities of these three polarization channels allows user to visually classify a scene by its backscattering mechanism, such as surface scattering (strong HH and VV return), volume scattering (strong HV return) and double-bounce scattering (strong HH return). Areas dominated by green (HV) intensity are typically vegetated areas. Areas dominated by shades of pink (HH+HV) intensity are typically inundated forests or vegetated fields. Black and dark grey areas are usually smooth surface (roads, open water, smooth bare ground) where there is very little radar backscatter. Suggested Use: The polarimetric color composite is well suited for identifying inundation under tree canopies. This product is especially useful in mapping inundation extent in regions covered with vegetation where the ground is not visible in optical imagery. Classified Images: Blue areas are open water (some roads may also be misclassified as blue) Light Blue/Cyan areas with flooding beneath vegetation and the tree canopy Pink areas are potentially flooded areas with low confidence due to being more developed locations Satellite/Sensor: UAVSAR airborne L-band synthetic aperture radar (SAR) aboard a NASA Gulfstream C-20A jet. Resolution: 10 meters Credits: UAVSAR data courtesy of NASA/JPL-Caltech Esri REST Endpoint: https://gis.earthdata.nasa.gov/portal/rest/services/DISASTERS_202410_HURRICANE_MILTON/2428_uavsar/ImageServer

Dates of Images: 10/11/2024, 10/12/2024, 10/13/2024, 10/14/2024 Date of Next Image: None Expected Summary: These UAVSAR false color RGB images provide a unique look at the Earth's surface that can be used for identifying flooding and inundation under tree canopy. The overlay of the intensities of these three polarization channels allows user to visually classify a scene by its backscattering mechanism, such as surface scattering (strong HH and VV return), volume scattering (strong HV return) and double-bounce scattering (strong HH return). Areas dominated by green (HV) intensity are typically vegetated areas. Areas dominated by shades of pink (HH+HV) intensity are typically inundated forests or vegetated fields. Black and dark grey areas are usually smooth surface (roads, open water, smooth bare ground) where there is very little radar backscatter. Suggested Use: The polarimetric color composite is well suited for identifying inundation under tree canopies. This product is especially useful in mapping inundation extent in regions covered with vegetation where the ground is not visible in optical imagery. RGB Images: Open water can be seen as dark blue/black, areas with inundation under tree canopy can be seen in pink-like colors. Classified Images: Blue areas are open water (some roads may also be misclassified as blue) Light Blue/Cyan areas with flooding beneath vegetation and the tree canopy Pink areas are potentially flooded areas with low confidence due to being more developed locations Satellite/Sensor: UAVSAR airborne L-band synthetic aperture radar (SAR) aboard a NASA Gulfstream C-20A jet. Resolution: 10 meters Credits: UAVSAR data courtesy of NASA/JPL-Caltech Esri REST Endpoint: https://gis.earthdata.nasa.gov/portal/rest/services/DISASTERS_202410_HURRICANE_MILTON/UAVSAR/ImageServer

Date of Images: 5/5/2024 Date of Next Image: N/A Summary: Scientists at NASA's Marshall Space Flight Center created these water extents on May 5, 2024 using PlanetScope imagery. These images can be used to see where open water is visible at the time of the satellite overpass. This product shows all water detected and differentiates between normal water areas and some flooded areas. This product was classified using WorldCover. It's important to note that all flooded areas may not be captured due to the sensors limitations of not being able to "see" through vegetation and buildings. To determine where additional flooding may have occurred, combine this layer with other data sets. Suggested Use: This product shows water that is detected by the sensor with different colors indicating different land cover/land use classifications from WorldCover that appear to have water and are potentially flooded. Blue (1): Known Water Red (2): Flooded Developed Green (3): Flooded Vegetation Orange (4): Flooded Cropland/Grassland Gray (5): Clouds/Cloud Shadow (0): No Data Satellite/Sensor: PlanetScope Resolution: 3 meters Credits: NASA Disasters Program, Includes copyrighted material of Planet Labs PBC. All rights reserved. Service URL: https://gis.earthdata.nasa.gov/portal/home/item.html?id=24843d7f43694de290c000b0aacae919 WMS Endpoint: https://gis.earthdata.nasa.gov/portal/services/DISASTERS_202405_FLOOD_TX/2408_planet_water/ImageServer/WMSServer

Dates of Images: Sept 17, 2018 (1700), Sept 20, 2018 (1410), Sept 22, 2018 (1705), Sept 23, 2018 (1545), all times are UTC Date of Next Image: None Summary: UAVSAR (Uninhabited Aerial Vehicle Synthetic Aperture Radar), a NASA facility instrument mounted on a Gulfstream-III aircraft, is an airborne radar system capable of monitoring rapidly evolving flood events regardless of cloud cover. It can support daily flood extent maps that track flood cresting and recession following hurricane landfall or periods of heavy precipitation.  The UAVSAR radar produces an L-band signal that can detect open water as well as water under canopies. This makes it especially useful for identifying flooding in places where it may not be visible in other aerial or satellite images. The flood extend map described here identifies the following classes: open water, flooded vegetation, not flooded, and no data.Repeated UAVSAR acquisitions fill gaps between satellite passes and support both:Response phase: providing high-resolution detection of active flooding, identifying flood-crest locations, and extending situational awareness in areas not covered by stream, tide, or rain-gauge networks. Recovery phase: documenting peak flood extent and areas of sustained inundation, supporting damage assessment, access planning, and post-event impact analyses. The map legend has been reconciled with the OPERA Dynamic Surface Water Extent product from Sentinel-1 (DSWx-S1), specifically the WTR layer. The product specification document for OPERA can be found here: https://www.jpl.nasa.gov/go/opera/products/dswx-product-suite/ UAVSAR-derived flood inundation products were used to characterize both storm-surge and riverine flooding in North Carolina following Hurricane Florence, which made landfall on September 14, 2018. The Gulfstream-III aircraft flew over impacted areas on September 17, 20, 22, and 23, capturing the progression of floodwaters as they receded. Suggested Use: Legend Classifications: Blue – Open Water: Areas that are completely covered by water and clearly visible from above. This means the sensor has a direct view of the water surface, with no blockage from trees, buildings, or terrain. This class includes lakes, rivers, and flooded areas where land is completely covered by water. Green – Flooded Vegetation: Areas where water is present but partially hidden by vegetation. This includes places like flooded forests (water under trees), agricultural fields, wetlands, golf courses, and urban green spaces with standing water. White – Not Flooded: Areas without flooding. This includes both vegetated land (like forests or grass) and non-vegetated surfaces (such as bare soil, roads, and parking lots). Transparent – No Data: Areas where no information is available. This happens either because the sensor did not capture data there or those areas were excluded from the current map. Things to keep in mind: Users are advised to examine the spatial context and multiple observation dates to infer flooding in roads. While SAR-based flood classification typically achieves high accuracy (>80%), two persistent challenges remain: (1) distinguishing open water from other low-backscatter surfaces (e.g., roads), and (2) separating flooded vegetation from built-up areas, which generate similar double-bounce scattering responses. Additional documentation on this data can be found here. Data can be downloaded here. Satellite/Sensor/Resolution: Instrument: UAVSAR L-band Synthetic Aperture Radar. Model: this classification followed Denbina et al. (2020) DOI: 10.1109/IGARSS39084.2020.9324379.  Spatial grain: The pixel posting is approximately 30 meters (0.0002778 degrees) and the original UAVSAR images used to make this map have approximately 6 meters posting. Credits: This map was created by Richard Chen and Annemarie Peacock (NASA/JPL) with imagery from the UAVSAR sensor (https://uavsar.jpl.nasa.gov). For questions, please contact UAVSAR Project Manager, Yunling Lou (ylou@jpl.nasa.gov).

Dates of Images: 10/11/2024, 10/12/2024, 10/13/2024, 10/14/2024 Date of Next Image: None Expected Summary: These UAVSAR false color RGB images provide a unique look at the Earth's surface that can be used for identifying flooding and inundation under tree canopy. The overlay of the intensities of these three polarization channels allows user to visually classify a scene by its backscattering mechanism, such as surface scattering (strong HH and VV return), volume scattering (strong HV return) and double-bounce scattering (strong HH return). Areas dominated by green (HV) intensity are typically vegetated areas. Areas dominated by shades of pink (HH+HV) intensity are typically inundated forests or vegetated fields. Black and dark grey areas are usually smooth surface (roads, open water, smooth bare ground) where there is very little radar backscatter. Suggested Use: The polarimetric color composite is well suited for identifying inundation under tree canopies. This product is especially useful in mapping inundation extent in regions covered with vegetation where the ground is not visible in optical imagery. RGB Images: Open water can be seen as dark blue/black, areas with inundation under tree canopy can be seen in pink-like colors. Satellite/Sensor: UAVSAR airborne L-band synthetic aperture radar (SAR) aboard a NASA Gulfstream C-20A jet. Resolution: 10 meters Credits: UAVSAR data courtesy of NASA/JPL-Caltech Esri REST Endpoint: