Skip to the cards
Filed
Source
Tim Barnwell
Loans
747

Dublin City Council to Test AI Flood Alerts Through Early 2027

Dublin’s pilot will analyse existing rainfall, river and tide sensor data, with past trials indicating 13 to 21 hours’ warning for some events.

Dublin City Council to Test AI Flood Alerts Through Early 2027

Dublin City Council plans to test whether artificial intelligence can give its teams more time to prepare for serious flooding. The pilot, developed with EMR Integrated Solutions, is scheduled to run through autumn and winter in late 2026 and early 2027. Its model will analyse data already collected around Dublin and nearby counties, looking for conditions that may precede major flooding.

The practical question is whether those signals can produce useful alerts early enough to change how response teams prepare. EMR says trials using historical events showed potential warning times of 13 to 21 hours for some mountain rain-driven events. That result is limited to some of the events examined; it does not establish a general lead time for every flood risk or confirm how the system will perform during the pilot.

GreenFutures initiative to bring AI and climate education to secondary schools
A bird's eye view of Dublin with the river Liffey in the centre and a blue sky with some small clouds above.

Image: © Cristi/Stock.adobe.com

Circle K and Tesco to boost Irish EV charging capacity

Existing monitoring provides the data

EMR has operated Dublin City Council’s regional flood monitoring system for nine years. Its network includes outstations in the city and surrounding counties, where instruments monitor rainfall, river levels, tide readings and other conditions. The new model is intended to assess accumulated readings across those categories and identify combinations or patterns associated with conditions that could lead to serious flooding.

Irish co-founded Pretxt raises $5m seed to supervise AI agents

That approach makes the quality and coverage of the existing monitoring network central to the pilot. The model depends on data arriving from sensors and on those readings being useful for identifying developing risks. EMR also provides engineering support for the council’s flood management work, including site work, installations, repairs and instrument calibration. That operational role connects the monitoring equipment with the maintenance needed to keep it working.

EU leaders set for Dublin artificial intelligence summit

For IT and operational teams, the description points to an analytical layer being added to an established monitoring setup. The available details do not specify the model architecture, data-refresh frequency, alert thresholds or how alerts will reach staff and residents. Those details will matter when assessing how the system fits into existing monitoring and response procedures.

Dublin’s Akara raises $5.5m to accelerate hospital AI tech

Staff retain control of decisions

EMR says a central principle of the pilot is that qualified council staff will make all decisions. The AI model is expected to generate alerts from its analysis of signals and patterns; staff will review and interpret those alerts. That division places the system in a support role, where its output can inform preparation without being treated as an automatic instruction to act.

Five Irish research teams eye €1m decarbonisation award

This distinction matters because a warning is useful only when teams can judge its relevance and respond in time. Dublin City Council assistant chief executive for climate and urban resilience Dennis Keeley said operational teams need enough time to deploy resources, protect critical infrastructure and respond to emerging risks. He said the technology could help reduce disruption to communities, limit damage to public assets and improve public safety during extreme weather.

Council chief executive Richard Shakespeare said the pilot would examine how technology and data could support decision-making and improve flood preparedness. EMR chief commercial officer Mark O’Connor described additional preparation time as the benefit the system is intended to offer: rain and flooding cannot always be prevented, but earlier notice may give residents and council teams more opportunity to prepare and act.

What the trial needs to establish

The historical trials provide an initial indication, rather than a performance guarantee. The 13-to-21-hour window applies to some mountain rain-driven events, as reported by EMR. The pilot will need to show how well alerts work in live conditions and how consistently the model can identify situations that warrant attention. The supplied details do not give a target accuracy rate, a false-alert rate or a formal measure of success.

There is also a difference between producing an early signal and delivering a useful warning. Council staff need enough context to interpret an alert, while residents who may be affected need information that is timely and relevant. EMR says the project aims to support earlier warnings for significant flooding across the capital and improve the information provided to residents. The precise communication channels and alert process have not been described.

For teams responsible for the system, the trial will therefore put both data and operations under scrutiny. Sensor readings must be available for analysis; the model’s alerts must be understandable to the staff assessing them; and those staff need a workable route from alert review to preparation. The pilot’s stated human review principle provides a decision point, while the trial period should show whether the additional analysis supplies enough lead time to influence real response planning.

EMR, founded in the 1980s and based in Dunboyne, Co Meath, supplies communications, supervisory control and data acquisition (SCADA), and instrumentation solutions. Its work spans water, power, utilities, corporate business, broadcasting and public safety organisations. For Dublin, the upcoming test builds on a monitoring network already in operation. The result to watch is whether analysis of those existing rainfall, river and tide readings can produce alerts that council staff can act on and residents can use.

Source: Tim Barnwell · www.siliconrepublic.com

Filed alongside