Global Flood Inundation Calculations in Real-time Compared to Observations

The Global Flood Monitoring System (GFMS) (http://flood.umd.edu) has been developed and used in recent years to provide real-time flood detection, streamflow estimates and inundation calculations. The GFMS is driven by satellite-based precipitation, with the accuracy of the flood estimates being dependent on the accuracy of the precipitation analyses and the models used.  The inundation estimates at 1 km resolution are based on streamflow routing at that resolution and calculations of resulting water depth compared to derived thresholds and taking into account existing water bodies.

Examples of recent GFMS defined flood events during 2016 in the U.S. and outside are compared to validation data, including streamflow observations, and with inundation estimates available from optical and Synthetic Aperture Radar (SAR) satellite sensors. For a case of significant flooding in Texas in May and June along the Brazos River the GFMS calculated streamflow compares favorably with the observed.  Available Landsat-based (May 28) and MODIS-based (June 2) inundation analyses from U. of Colorado shows generally good agreement with the GFMS inundation calculation in most of the area where skies were clear and the optical techniques could be applied.

In a second case in March over the southern U.S. in Mississippi in March both optical and SAR-based inundation analyses are available and are compared to GFMS calculations.  In this case the optical-based inundation underestimates the area as compared to GFMS and the SAR information and the underestimation is validated by aircraft photography.  The comparison of the GFMS inundation in this case improves with spatial averaging as compared to the SAR-based estimates.  The optical-based estimate is shown to be improved by adjusting it based on taking into account elevation information and the GFMS results.