Extreme precipitation events are intensifying due to a warming climate, which, in some cases, is leading to increases in flooding. Detection of flood extent is essential for flood disaster management and prevention. However, it is challenging to delineate inundated areas through most publicly available optical and short-wavelength radar data, as neither can “see” through dense forest canopies. The 2018 Hurricane Florence produced heavy rainfall and subsequent record-setting riverine flooding in North Carolina, USA. NASA/JPL collected daily high-resolution full-polarized L-band Uninhabited Aerial Vehicle Synthetic Aperture Radar (UAVSAR) data between September 18th and 23rd. Here, we use UAVSAR data to construct a flood inundation detection framework through a combination of polarimetric decomposition methods and a Random Forest classifier. Validation of the established models with compiled ground references shows that the incorporation of linear polarizations with polarimetric decomposition and terrain variables significantly enhances the accuracy of inundation classification, and the Kappa statistic increases to 91.4% from 64.3% with linear polarizations alone. We show that floods receded faster near the upper reaches of the Neuse, Cape Fear, and Lumbee Rivers. Meanwhile, along the flat terrain close to the lower reaches of the Cape Fear River, the flood wave traveled downstream during the observation period, resulting in the flood extent expanding 16.1% during the observation period. In addition to revealing flood inundation changes spatially, flood maps such as those produced here have great potential for assessing flood damages, supporting disaster relief, and assisting hydrodynamic modeling to achieve flood-resilience goals.

Improved monitoring of inundation area variations in lakes and reservoirs is crucial for assessing surface water resources in a growing population and a changing climate. Although long-record optical satellites, such as Landsat missions, provide sub-monthly observations at fairly fine spatial resolution, cloud contamination often poses a major challenge for producing temporally continuous time series. We here proposed a novel method to improve the temporal frequency of usable Landsat observations for mapping lakes and reservoirs, by effectively recovering inundation areas from contaminated images. This method automated three primary steps on the cloud-based platform Google Earth Engine. It first leveraged multiple spectral indices to optimize water mapping from archival Landsat images acquired since 1992. Errors induced by minor contaminations were next corrected by the topology of isobaths extracted from nearly cloud-free images. The isobaths were then used to recover water areas under major contaminations through an efficient vector-based interpolation. We validated this method on 428 lakes/reservoirs worldwide that range from ~2 km2 to ~82,000 km2 with time-variable levels measured by satellite altimeters. The recovered water areas show a relative root-mean-squared error of 2.2%, and the errors for over 95% of the lakes/reservoirs below 6.0%. The produced area time series, combining those from cloud-free images and recovered from contaminated images, exhibit strong correlations with altimetry levels (Spearman’s rho mostly ~0.8 or larger) and extended the hypsometric (area-level) ranges revealed by cloud-free images alone. The combined time series also improved the monthly coverage by an average of 43%, resulting in a bi-monthly water area record during the satellite altimetry era thus far (1992–2018). Given such performance and a generic nature of this method, we foresee its potential applications to assisting water area recovery for other optical and SAR sensors (e.g., Sentinel-2 and SWOT), and to estimating lake/reservoir storage variations in conjunction with altimetry sensors.