*Article* **Mapping Flood Extent and Frequency from Sentinel-1 Imagery during the Extremely Warm Winter of 2020 in Boreal Floodplains and Forests**

**Liis Sipelgas \*, Age Aavaste and Rivo Uiboupin**

Department of Marine Systems, School of Science, Tallinn University of Technology, Akadeemia tee 15a, 12618 Tallinn, Estonia; msi@taltech.ee or age.aavaste@taltech.ee (A.A.); rivo.uiboupin@taltech.ee (R.U.) **\*** Correspondence: liis.sipelgas@taltech.ee

**Abstract:** The current study presents a methodology for water mapping from Sentinel-1 (S1) data and a flood extent analysis of the three largest floodplains in Estonia. The automatic processing scheme of S1 data was set up for the mapping of open-water flooding (OWF) and flooding under vegetation (FUV). The extremely mild winter of 2019/2020 resulted in several large floods at floodplains that were detected from S1 imagery with a maximal OWF extent up to 5000 ha and maximal FUV extent up to 4500 ha. A significant correlation (r<sup>2</sup> > 0.6) between the OWF extent and the closest gauge data was obtained for inland riverbank floodplains. The outcome enabled us to define the water level at which the water exceeds the shoreline and flooding starts. However, for a coastal river delta floodplain, a lower correlation (r<sup>2</sup> < 0.34) with gauge data was obtained, and the excess of river coastline could not be related to a certain water level. At inland riverbank floodplains, the extent of FUV was three times larger compared to that of OWF. The correlation between the water level and FUV was <0.51, indicating that the river water level at these test sites can be used as a proxy for forest floods. Relating conventional gauge data to S1 time series data contributes to flood risk mitigation.

**Keywords:** Sentinel-1; flood; climate change
