7%
Increase in rainfall across Ireland comparing 1991–2020 with the 1961–1990 baseline average
EPA Climate Ireland / climateireland.ie
by Kevin Martin, Environmental Services Lead, EMR Integrated Solutions
Flooding doesn’t announce itself. It doesn’t wait for office hours, and it doesn’t care about organisational boundaries. What it does care about is time.
The gap between a threshold being breached and a response team reaching the site is where the real damage is decided. Get that gap right and you protect property, infrastructure, and people. Get it wrong and you’re doing damage limitation instead of prevention.
I’ve spent my career working on the infrastructure that tries to close that gap. Telemetry networks, SCADA systems, sensor deployments across rivers, culverts, and drainage assets. The technology has changed considerably over the years. The fundamental problem hasn’t.
What has changed, and changed significantly, is the nature of the events themselves.
“Flooding doesn’t announce itself. It doesn’t wait for office hours, and it doesn’t care about organisational boundaries. What it does care about is time.”
A different kind of rainfall
From what we’ve seen on the ground, it’s not just that we’re getting more rainfall. We’re getting more intense rainfall over shorter periods.
Events that used to be considered exceptional are becoming much more common. Areas that historically weren’t considered high risk are now experiencing flooding because their drainage infrastructure was never designed for these rainfall patterns.
That matters a great deal for how monitoring systems need to work. A system designed around historical norms is already behind. The thresholds that made sense five years ago may need revisiting. With the increase in housing, industrial and flood prevention projects, the monitoring and control locations within a network may no longer reflect where the vulnerabilities are. And the speed at which data needs to travel, from sensor to operator to response team has to keep pace with how quickly conditions can deteriorate.
This is the environment we work in at EMR, and it’s one that sharpens your thinking about what good monitoring actually means.
Climate data
20% more intense
Human-induced climate change made the heavy storm rainfall that caused flooding across Ireland and the UK between Oct 2023–Mar 2024 approximately 20% more intense
World Weather Attribution / Met Éireann, May 2024
Once every 5 years
What used to be a once-in-50-years storm rainfall event now occurs roughly every five years at current warming levels
World Weather Attribution, 2024
21–31%
Projected increase in frequency of very wet days (>30mm) depending on emissions scenario
EPA Climate Ireland / climateireland.ie
What good monitoring actually involves
A flood monitoring network is not one thing. It’s a collection of instruments, each measuring a different part of the picture. Rain gauges tell you how much is falling and how quickly. River level sensors show how waterways are responding. Trash screen monitors identify blockages before they become critical. CCTV provides visual confirmation of what the numbers are suggesting.
Individually, each instrument tells part of the story. Together, they give operators something to act on. The distinction matters because data alone is not the goal. The goal is getting the right information to the right person early enough for them to do something useful with it.
Take a trash screen on a river. If debris builds up and water can’t pass through, it backs up and breaks the banks. Properties nearby flood. That’s not a dramatic or unusual failure mode; it happens. The monitoring system’s job is to detect the differential level that signals a blockage is developing, alert an engineer, and give that engineer time to act before the situation becomes critical. The CCTV feed adds the context that numbers alone can’t always provide, confirming what’s physically happening at the screen before anyone gets in a van.
River & Culvert Sensors
Real-time level data from vulnerable watercourse locations across the network
Rain gauges
Intensity and accumulation data, tracked per 15-minute period at each site
Trash Screen Monitors
Differential level alerts detect blockages before they cause bank breaches
CCTV Integration
Visual confirmation adds context that sensor data alone cannot provide
Dublin City Council
A case study in long-term partnership
EMR has been maintaining Dublin City Council’s flood management SCADA system for a number of years, and it’s a relationship that illustrates what long-term partnership in critical infrastructure actually looks like in practice.
The council’s network spans scores of outstations across the city and surrounding counties. These include rain gauges, sewer loggers, river level monitors, tide monitors, weather stations, and river trash screen monitoring cameras, all connected to a centralised Ignition SCADA system via a telemetry network.
The telemetry network evolves with technology. Critical sites have dual redundant communication options. Pre-configured alarms at each outstation alert engineering teams around the clock, giving them advance warning of conditions that could lead to localised flooding.
Our role covers far more than the initial deployment. We provide the managed service that keeps the system available when it matters most, respond to communications faults and equipment issues, and give DCC’s engineers on-demand access to our SCADA and telemetry specialists. We also develop the system continuously; most recently, we’ve been building new reporting around rainfall intensity, tracking hourly data on a monthly basis across all sites so that threshold levels can be reviewed against how rainfall patterns are actually evolving.
That continuity of relationship is something we take seriously. A system like this isn’t commissioned and handed over; it’s a live thing that evolves continuously with engineering input from both sides, new requirements, new sensors, and new technologies absorbed over time.
For 45 years EMR has been working in critical national infrastructure, giving us a very clear sense of what it takes to deliver and support systems that people depend on when environmental conditions are at their worst.
69
Outstations across city & surrounding counties
24/7
Alarm monitoring & escalation via Ignition SCADA
EMR has a full team of field and SCADA engineers that support the DCC regional SCADA system through associated site work, installations, repairs and calibrations to instruments.”
Alan Vickers · Chief Technician, Flood Projects & Water Framework Directive, Dublin City Council
“A system like this isn’t commissioned and handed over; it’s a live thing that evolves continuously with engineering input from both sides, new requirements, new sensors, and new technologies absorbed over time.”
Kevin Martin, Environmental Services Lead, EMR Integrated Solutions
The data question
The Dublin Regional Flood Management system has gathered very useful data for many years from weather stations, rainfall measurements, river levels, tide readings, trash screens, equipment status, river water quality, alarm events and all of it collected and stored on SQL open database platforms.
We’ve recently automated the transfer of historical data going back several years directly to DCC’s own servers, alongside a mechanism that keeps that data current.
That’s useful, but it prompts a more interesting question. What are the next steps? Predictive analyses? AI? Machine learning? What are we not asking of that data?
The historical records could help us understand how flood events develop, identify locations that are becoming increasingly vulnerable, and improve prediction models. We can start asking not just what happened, but why it happened and what conditions typically preceded it. That shift, from retrospective analysis to genuine foresight, is where the real value lies.
The challenge is that the data we hold doesn’t exist in isolation. Weather forecasting data, hydraulic models, GIS information about drainage infrastructure, gully monitoring data, even combined sewer overflow records. These data silos sit in different systems, owned by different organisations, in different formats. As a systems integrator, our data integration expertise spans many utilities in the UK and Ireland. It’s what we do.
“The challenge is that the data we hold doesn’t exist in isolation. Integration is the hard part, not the collection.”
The case for shared intelligence
Flooding doesn’t recognise organisational boundaries. A sewer managed by Uisce Éireann can overflow and cause a flood event in an area where a local authority carries responsibility for roads and drainage. Environmental data held by one local authority could enhance the reliability of flood prediction on networks that are the responsibility of other organisations.
Bringing more information onto a single EMR Platform will result in more reliable decision-making. When you combine data from multiple sites, from different sensor types and from different network databases, we can more confidently trigger “call for specific action” alarms. These actions could include to clean gullies on a certain road section, clear a trash screen or close flood gates in anticipation of extreme weather events.
Years of sensor data and local operational knowledge, when processed through AI, can do something that manual analysis never could: surface a live picture of risk status across every point on the network simultaneously. The output is a dynamic flood risk map, with boundaries that shift in real time as conditions change. Operators see not just where thresholds are being approached, but how risk is moving across a catchment. Asset management decisions follow from that picture rather than from instinct or incomplete data.
This is the direction the industry needs to move in, and it requires open platforms, genuine data sharing between authorities, and engineering experience in integrating systems that weren’t originally designed to talk to each other.
These are capabilities EMR has been building across multiple sectors over the last 45 years. Smart decisions made at the start of this project has meant the data is still useable and very important to develop trends that can allow us predict future events and intensity and what we need to protect people and property.
Where AI fits and where it doesn’t
There is considerable enthusiasm in the sector at the moment for AI as a solution to flood management challenges. Some of it is well-founded. A great deal of it is not.
AI has genuine potential in identifying patterns across large datasets, improving predictive capabilities, and reducing false alarms that lead to alarm fatigue among operators. When you have sensor readings, weather data, hydraulic model outputs, and historical flood event records all feeding into a system, the volume of information quickly exceeds what any human analyst could process in real time. That’s where machine learning can add real value, finding precursors to events that experience alone wouldn’t uncover.
But AI should not be mistaken for a replacement for engineering expertise. My colleague Brian Power, EMR’s Head of Engineering, has written about this directly: responsible AI in critical infrastructure starts with a well-defined engineering problem, keeps qualified engineers in the decision loop at every stage, and uses domain knowledge to validate what the system surfaces. That principle applies with particular force in flood management, where the consequences of acting on flawed information, or failing to act on correct information, are borne by real people and real communities.
Confidence in a system comes from transparency. When an operator understands why an alert was generated and can interrogate the reasoning, trust is maintained. When the system becomes a black box, trust erodes, and with it the willingness to act decisively when it matters most. That’s a dangerous outcome in a safety-critical environment, and it’s one that the responsible deployment of AI has to actively avoid.
I see AI as another tool in the belt in assisting us with everything we’ve been doing since the first iteration of these monitoring networks were installed. Using the technologies available to us to give operators better information faster. The aim doesn’t change. The methods evolve.
“If you do not have deep engineering judgement and critical thinking, you can go in completely the wrong direction with AI output. The tool is only as good as the person interpreting it.”
Brian Power, Head of of Engineering, EMR Integrated Solutions
Key Takeaway
Flood monitoring is not fundamentally about sensors or software. It’s about preparedness. The technology provides visibility. Successful outcomes depend on the processes, the people, and the response plans that sit around it.
What the next five years look like
You can’t manage what you can’t see. The most important improvements in flood early warning capability over the next five years will come from two directions:
- Better network coverage in areas that are currently under-monitored.
- Making better use of historical data with the introduction of new technologies.
The future of flood management lies in better-connected ecosystems, where telemetry, weather forecasting, hydraulic models, drainage infrastructural data and real-time sensor readings are brought together to give operators accurate and timely data presented on a real-time, user-friendly platform.
AI will play a significant role in processing larger volumes of information both real and historical to provide operators with more accurate alarms and lightly outcomes. It will do so in support of experienced engineers and operators.
Monitoring and physical flood alleviation infrastructure are not separate investments; they are complementary ones. Monitoring tells you where the problems are and helps validate whether alleviation measures are actually working. Without data, you’re investing blind.
Predictive capability improves with depth of data. The further back the historical record goes, the more confident you can be about what conditions typically precede a flood event and how quickly one develops once those conditions are in place.
What I’d want any local authority or utility thinking about flood risk to understand is this: flood monitoring is not fundamentally about sensors or software. It’s about preparedness. The technology provides visibility. Successful outcomes depend on the processes, the people, and the response plans that sit around it.
At EMR, we’ve spent 45 years building and maintaining systems that people depend on in precisely those moments when things go wrong. The tools are evolving. The responsibility is the same as it’s always been.
Kevin Martin is Environmental Services Lead at EMR Integrated Solutions, with responsibility for flood monitoring, environmental telemetry, and managed services for local authorities and water utilities across Ireland. EMR has delivered critical national infrastructure solutions for 45 years, serving water utilities, renewable energy providers, and national transport and maritime bodies across Ireland, the UK, and Europe.
Latest News
EMR to exhibit at Solar Ireland Annual Conference in JuneApril 1, 2026 - 10:08 am
EMR to Exhibit at SWIG Continuous Water Quality Monitoring ConferenceFebruary 25, 2026 - 11:51 am
Same Team, New Mill: EMR’s UK Team Gets a New HomeFebruary 18, 2026 - 5:28 pm
