Google Flood Hub — Bangladesh Water Development Board's AI-Powered River Flood Alert System
Bangladesh
Since 2020, Google Research's machine-learning flood model — which forecasts river levels even in ungauged basins — has combined with …
Uganda · Kampala · See the Uganda profile · See the Kampala profile
Top 46% 60/100 · Ask Evidence Copilot about this practice
Uganda launched an AI flash-flood nowcasting system in Kampala (April 2026), blending satellite rainfall, radar and hydrodynamic modelling. It is the first live site in a WMO-backed five-city East African pilot, with expansion to four more Ugandan towns planned.
Uganda faces recurring urban flash flooding in Kampala and other towns, with warning systems that have historically been limited. In March-April 2026, the Ministry of Water and Environment's Department of Meteorology, working with the World Meteorological Organization (WMO) and Dutch hydrology specialists HKV and Deltares, launched the Urban Flash Flood Forecasting System (UFFFS) in Kampala — the first live deployment of a five-city pilot also covering Addis Ababa, Juba, Khartoum and Kigali, funded by the Netherlands Ministry of Foreign Affairs under its "Water at the Heart of Climate Action" and "Early Warnings for All" programmes.
UFFFS combines the PySTEPS machine-learning nowcasting algorithm with satellite rainfall estimates, automatic weather-station and radar data, and SFINCS hydrodynamic flood modelling to generate impact-based warnings that account for drainage and terrain, not just rainfall volume. A technical workshop held in Kampala on 23-25 March 2026 trained more than 20 meteorological and hydrological staff from the five countries, and the system was launched for operational use on 8 April 2026.
As of the sources reviewed, no forecast-accuracy, lead-time or avoided-damage figures have yet been published; the system is newly operational rather than independently evaluated. Ugandan officials themselves flagged limited monitoring infrastructure and inadequate equipment as ongoing risks to effectiveness. Uganda's Ministry has announced a roadmap to extend the system next to Fort Portal, Kasese, Gulu and Mbale.
Read the full analysis: https://www.monitor.co.ug/uganda/news/national/govt-rolls-out-real-time-flood-detection-app-to-boost-early-warnings-5417390
Implementation detail (cost, timeline, staffing, conditions for success) is not yet available for this practice.
Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.
Where this practice's information was retrieved from, and when.
Bangladesh
Since 2020, Google Research's machine-learning flood model — which forecasts river levels even in ungauged basins — has combined with …
Samoa
Samoa's national weather service now issues impact-based forecasts, not just hazard alerts, for cyclones, coastal flooding and storm surge, under …
Mozambique
Mozambique's INGD and INAM are piloting machine-learning forecasts atop a multi-hazard early-warning system that helped cut Cyclone Freddy's 2023 death …
United Kingdom
ECMWF made its AIFS machine-learning weather model operational in February 2025, with a probabilistic version following in July — cutting …
Open full copilot Grounded in cited practices — always check the sources.