Can we detect more ephemeral floods with higher density harmonized Landsat Sentinel 2 data compared to Landsat 8 alone?

Spatiotemporal quantification of surface water and flooding is essential given that floods are among the largest natural hazards. Effective disaster response management requires near real-time information on flood extent. Satellite remote sensing only way monitoring these dynamics across vast areas over time. Previous mapping efforts have relied optical time series, despite cloud contamination. This reliance data due to availability systematically acquired easily accessible globally for 40 years. Prior research used either MODIS or Landsat data, trading high temporal density but lower spatial resolution cadence higher resolution. Both pose limitations as can miss ephemeral floods, whereas misses small inaccurately delineates edges. Leveraging frequency 3–4 days existing Landsat-8 (L8) two Sentinel-2 (S2) satellites combined, in this research, we assessed whether increased three sensors improves our ability detect extent compared a single sensor (L8 alone). Our study area was Australia’s Murray-Darling Basin, one world’s dryland basins experiences floods. We applied machine learning NASA’s Harmonized (HLS) Surface Reflectance Product, which combines L8 S2 observations, map dynamics. overall accuracy, estimated from stratified random sample, 99%. user’s producer’s accuracy class 80% (±3.6%, standard error) 76% (±5.8%). focused 2019, most recent years when all HLS operated at full capacity. results show (permanent flooding) identified with greater than by L8, some short-lived events were detected HLS. Comparison (3 m) PlanetScope extensive mixed pixels 30 m resolution, highlighting need improved future work. The has been able cases alone not, 2019 being driest area, few events. dense time-series offered thus critical capturing temporally dynamic phenomena (i.e., drylands), importance harmonized such