Floodbase

Floodbase (formerly Cloud to Street) provides satellite-based flood monitoring and powers parametric flood insurance, fusing 35-plus years of satellite imagery, hydrological models and near-real-time observations to map flood extent globally. It underpins index-based insurance products and flood intelligence for re/insurers, brokers, the public sector and corporates, delivering event detection, historical baselines and exposure analytics. Distinctive for addressing historically hard-to-insure flood risk through satellite-derived parametric triggers — a specialist bridging Earth observation and risk transfer.

Vendor methodology

Transparency Score

Beta

Public transparency, not model quality i

Methodology-relevant: vendor websites or external resources explain methodological details.

Methodology-relevant: peer-reviewed publications document the approach reconstructably — the NWM-CNN inundation method (Frame et al., Geophysical Research Letters 2024: 780 training events, 250 m, RMSE 4.58%) and the Nature 2021 satellite flood-tracking science, plus the public Cloud-to-Street global flood dataset and 17 named satellite sources. Core CNN proprietary, so 2 rather than 3.

For further information, see the section 'Transparency Score methodology' in the Readme.

Level 3 signals

Peer-rev. / open source
Peer-reviewed NWM-CNN flood method (Frame et al., GRL 2024): 780 training events, 250 m.
Peer-rev. / open source
Peer-reviewed Nature paper (Tellman et al. 2021): uses Floodbase's Global Flood Database — 913 satellite-observed floods, 2000–2018 — to quantify population exposure.
Peer-rev. / open source
Open-source code for the Global Flood Database: MODIS water-detection algorithms and population-exposure analysis behind the Nature paper.