This report presents the results of the annual product and service quality assessment of the Global Flood Monitoring (GFM) product of the Copernicus Emergency Management Service (CEMS), which was performed during 2025. This report is the fifth in a series, following the GFM annual quality assessment reports for 2024 (Seewald et al., 2025), 2023 (Seewald et al., 2024), 2022 (Seewald et al., 2023a) and the GFM pre-operational quality assessment report (Seewald et al., 2023b). The main part of the 2025 GFM product and service quality assessment was the thematic accuracy assessment of the GFM output layers Observed Flood Extent and Reference Water Mask, for 12 selected Use Cases of worldwide flood events (described in Section 3.2 of this report). Ten of the Use Cases were floods that occurred in 2025, which were mapped by the operational, near real-time GFM product. Two Use Cases were floods that occurred during 2017-2020, which are included in the GFM re-processed Sentinel-1 archive of worldwide observed flood events and water extent. As required by the Technical Specifications (European Commission, 2020), the GFM product should have a thematic accuracy of at least 70%, computed based on the Critical Success Index (CSI). The main results of the GFM thematic accuracy assessment, which are presented in detail in Section 4.1 of this report, and discussed in detail for each Use Case in Section 4.2, are summarised below: - For the GFM output layer Observed Flood Extent, ten Use Cases had CSI values ranging from good to very good (i.e. 72.7 - 89.1%), while two Use Cases — i.e. Coastal Ecuador (February 2025) and Central Ireland (November 2025) — had CSI values just below the target value of 70% (i.e. 63.4% and 67.0%, respectively). - For the GFM Reference Water Mask, nine Use Cases had CSI values ranging from good to very good (i.e. 72.8 - 92.8%), while three Use Cases — i.e. Northern Togo (October 2020), Eastern India (August 2017), and Northern Thailand (July 2025) — had CSI values ranging from poor to fair (i.e. 31.3%, 53.1%, and 38.9 %, respectively). Note that the thematic accuracies of the seasonal (i.e. monthly) water of the GFM Reference Water Mask are not presented, since, as highlighted in previous GFM annual quality assessment reports, this is generally very low, due to major discrepancies between the seasonal water extent in the GFM product and the reference datasets, as explained in Section 4.1 of this report. The new post-processing approach that was developed and tested during 2024 (Seewald et al., 2025), was implemented in the new GFM version (v4), released in April 2025. For Use Cases 1, 2, and 3, the thematic accuracy assessment was carried out using the previous GFM version (v3.2), while for Use Cases 4 to 12, the thematic accuracy assessment was carried out using the latest GFM version (v4). For three Use Cases, the thematic accuracy results for GFM v4 were compared with those for GFM v3.2 and were found to be significantly improved (Use Cases 3 and 5) or unchanged (Use Case 2), as described in Section 4.2 of this report. Finally, the results of the assessment of GFM product timeliness, service availability and user uptake in 2025, and the general plausibility of the GFM Exclusion Mask, are also presented. GFM service availability and product timeliness in 2025 generally met or exceeded the expected targets, and user uptake and experience improved compared to 2024.

SCHERF Nadine;  PASIK Adam;  GRUBER Christian;  INNERBICHLER Florian;  RIFFLER Michael;  REIMER Christoph;  STACHL Tobias;  KIDD Richard;  MCCORMICK Niall;  SALAMON Peter; 

2026-07-13

Publications Office of the European Union

JRC146857

978-92-68-41007-3 (online),   

1831-9424 (online),   

EUR 40765,    OP KJ-01-26-272-EN-N (online),   

https://publications.jrc.ec.europa.eu/repository/handle/JRC146857,   

10.2760/9393706 (online),