bwi

With the release of v2.14, Blue Water Intelligence continues to strengthen the performance and reach of its hydrological forecasting service. This update focuses on a key improvement: significantly faster access to forecasts, alongside expanded geographic coverage and ongoing model enhancements.

BWI powers continental river flow forecasting delving on EU Data Foundation as many of the data sources BWI uses come from Copernicus, EU’s Earth Observation Programme, which offers information services that draw from satellite Earth Observation and in-situ data.

We’re pleased to welcome Hanbing Peng, Precise Orbit Determination Specialist at BWI. With a strong background in geoscience and satellite navigation, Hanbing contributes to the REVALTO satellite mission by ensuring highly accurate orbit determination – a key element for reliable Earth observation data.

A posteriori analysis of the 16 November 2025 flood event on the Ardèche river, reported by the Service de Prévision des Crues du Grand Delta (SPC Grand Delta)

We’re pleased to welcome Bassam Chehade, Signal Processing Engineer at BWI. With a PhD in SAR imaging and extensive experience in remote sensing and radar signal processing, Bassam contributes to the REVALTO satellite mission by transforming radar measurements into meaningful geophysical parameters used to monitor inland waters.

BWI participated in the first edition of the Eau & IA Congress in Grenoble, organized by the Société Hydrotechnique de France. Three days of exchanges at the intersection of water, data and artificial intelligence.

We’re pleased to welcome François Verdeil, Signal Processing Engineer at BWI. With around ten years of experience in signal processing and a recent specialization in radar signal processing, François contributes to the REVALTO satellite mission by transforming raw radar measurements into usable data for hydrological analysis.

We’re pleased to welcome Thibault Lebeke, Industrial Manager at BWI. With over 10 years of experience in spacecraft engineering and a strong background in flight dynamics and system coordination, Thibault contributes to the REVALTO mission by managing industrial partners and technical interfaces – ensuring high-quality space data to power advanced hydrological forecasting.

We’re pleased to welcome Roxane Blattes, Project Management Officer at BWI. With a PhD in Biology, an MBA, and extensive experiences in program governance, Roxane supports the REVALTO satellite mission by overseeing planning, risks and cost management – contributing to the delivery of high-revisit hydrological data for improved water monitoring.

Climate volatility makes accurate river flow forecasting mission-critical for hydropower operators, irrigation planners, and basin authorities. Blue Water Intelligence (BWI) has developed an AI engine that transforms sparse satellite, EU Earth observation, and meteorological data into precise probabilistic forecasts up to 10 days ahead—even for ungauged rivers.

BWI’s technology uses a two-tier modeling architecture balancing speed and physical realism. The first tier, a lumped model for rapid deployment, aggregates each basin into a single “super-cell.” Recurrent neural networks capture temporal dynamics from precipitation inputs to outflow, with training in just minutes per basin. It achieves RMSE under 20% for daily discharges on gauged catchments, while physical constraints enforce mass balance and non-negative storage to prevent unrealistic drift and enable day-one production readiness.

The second tier, a semi-distributed hybrid for higher precision, discretizes basins into hydrological response units (~1–5 km²). Each unit employs RNN or LSTM layers to evolve states like surface runoff, subsurface flow, and root-zone storage.