Technology House, Stratfield Park, Waterlooville, PO7 7XN 023 9223 3611

Flood Early Warning for Critical Infrastructure: How Much Notice Does a Site Really Get?

TL;DR: About a third of England’s water pumping stations and treatment works, and a quarter of its medical facilities and emergency service sites, sit in areas at flood risk. Official warnings are issued for areas, do not cover surface water, and a flood warning usually comes 30 minutes to 2 hours before flooding. Barriers and safe shutdowns can need far longer. A site flood early warning puts level sensors on the water that threatens the site, with alarms set to the site’s own thresholds.

Last updated: 1 October 2026

Key takeaways

  • About 34% of England’s water pumping stations and treatment works are in areas at risk of flooding from rivers, the sea or surface water, along with 26% of emergency service sites and 25.5% of medical facilities.
  • With climate change, those shares rise to about 38%, 31% and 31.5% between 2040 and 2060.
  • The official flood warning service covers rivers, the sea and groundwater. It does not cover surface water flooding.
  • A flood warning usually arrives 30 minutes to 2 hours before flooding, and a flood alert 2 to 12 hours before.
  • Temporary barriers need reliable forecasts and time: the Environment Agency planned to deploy its own anywhere in England within 12 hours.
  • A site flood early warning measures the water that threatens the site and raises alarms at the site’s own thresholds, early enough for the slowest action in its flood plan.

What is flood early warning for critical infrastructure?

Flood early warning for critical infrastructure is a trigger that tells the people responsible for a pumping station, substation, hospital, telephone exchange or data centre that water is rising towards it, early enough to protect it. It combines the official warnings with the site’s own level readings, set against the site’s own thresholds.

The official service answers a question about an area: is flooding possible or expected here? A site needs the answer to a narrower question. When will water reach the door sill, the cable duct or the switchgear plinth, and is there still time to do something about it?

Critical infrastructure flooding also hurts in a different way. A flooded house is a disaster for one household. A flooded pumping station, substation or telephone exchange takes a service away from everyone who depends on it. After the winter floods of 2015 and 2016, the government’s National Flood Resilience Review recorded that communities had lost power, water and telecoms during the flooding, and then faced a lengthy recovery.

This guide sets out how much of England’s infrastructure is exposed, what the official warnings give a site and what they do not, how much notice a critical site really needs, and how a local flood early warning is put together.

How much of England’s critical infrastructure is at flood risk?

The Environment Agency’s 2024 national assessment puts about a third of water pumping stations and treatment works, a quarter of medical facilities and emergency service sites, and more than a fifth of electricity infrastructure in areas at risk of flooding from rivers, the sea or surface water. Climate change raises every one of those figures.

The National assessment of flood and coastal erosion risk in England 2024 gives a single picture of current and future flood risk from rivers, the sea and surface water. It finds around 6.3 million properties in England in areas at risk from one or more of those sources. Its infrastructure tables show how much of the country’s essential estate sits in the same places.

Infrastructure typeAt risk todayShare todayAt risk 2040 to 2060Share 2040 to 2060
Water pumping stations or treatment works12,80034.4%14,20038.2%
Gas infrastructure1,30035.1%1,40037.6%
Emergency services2,00026.0%2,40031.2%
Medical facilities11,80025.5%14,60031.5%
Electricity infrastructure52,60022.1%61,60025.9%
Education facilities30,00021.4%37,00026.4%

Risk is not spread evenly within those shares. Of the water pumping stations and treatment works, 11.4% are at high risk and a further 6.6% at medium risk. Water sites sit low for a reason, close to the rivers they draw from and discharge to, so they will always carry more exposure than a hospital on higher ground.

Bar chart of England's infrastructure in areas at flood risk today and by 2040 to 2060: water pumping stations and treatment works 34.4% to 38.2%, emergency services 26.0% to 31.2%, medical facilities 25.5% to 31.5%, electricity 22.1% to 25.9%
Share of each infrastructure type in areas at flood risk from rivers, the sea or surface water.

Why are official flood warnings not enough on their own?

Official warnings are issued for areas, cover rivers, the sea and groundwater, and do not cover surface water flooding. A flood warning usually comes 30 minutes to 2 hours before flooding. That is enough for a household to move valuables upstairs, but often too little for a site that has to fit barriers or shut plant down safely.

In England, the Environment Agency issues three levels of flood alerts and warnings:

  • Flood alert: flooding is possible, so prepare. It is usually issued between 2 and 12 hours before flooding, and during waking hours where possible.
  • Flood warning: flooding is expected, so act now. It is usually issued 30 minutes to 2 hours before flooding.
  • Severe flood warning: flooding could be a risk to life and cause significant disruption to communities. Among the things that may happen, the guidance lists infrastructure not working, like gas, electricity and water.

Three limits follow for a critical site. The first is coverage. The flood warnings sign-up page says messages cover flooding from rivers, the sea or groundwater, depending on the area, and that warnings are not available for surface water flooding, often called flash flooding. Yet surface water is part of the risk counted in the national assessment above.

The second is scale. A warning area describes a stretch of river or coast. It cannot know that one site’s cable duct sits lower than the road outside, or that another site’s outfall backs up long before the river reaches its banks.

The third is timing. The notice a warning gives is set by the forecast for the area, not by how long the site needs. When a site’s flood plan needs six hours and the warning gives two, the plan has failed before the water arrives.

None of this is a criticism of the service. It is designed to warn the public across a whole country. A critical site simply needs a second layer, built around its own levels and its own clock.

What the official warning service covers compared with what a critical site has to watch: rivers, sea and groundwater by area, but no warnings for surface water, culverts, outfalls or water inside the site
Official warnings describe an area. A site needs its own levels.

How much lead time does a critical site need?

Work backwards from the slowest action in the site’s flood plan. If fitting barriers, moving mobile pumps and shutting equipment down safely takes six hours, the trigger has to fire more than six hours before water reaches the site. On small, fast-responding rivers, official forecasts rarely give that much notice.

The National Flood Resilience Review is plain about this. Safe deployment of temporary barriers needs reliable flood forecasts far enough ahead to give lead time for deployment, and such forecasts are more often available for the lower reaches of large rivers than for steep, fast-responding upper catchments. Users of temporary defences told the review that enough lead time for transport and deployment was essential and a key challenge.

The numbers in the review show why. The Environment Agency planned its depots so that temporary barriers could be deployed anywhere in England within 12 hours: 3 hours of driving and 9 hours of loading and unloading. The electricity sector held 8.9 km of temporary flood barriers of its own, managed company by company with mutual aid between them. Even so, at the time of the review there were no tested temporary defence products protecting against flooding deeper than 2 metres, and very few tested above 1 metre.

Put a 12 hour deployment window next to a flood warning that arrives 30 minutes to 2 hours ahead, and the gap is obvious. Temporary flood barriers are a good answer for many sites, but only when the decision to fit them is taken early enough.

The practical answer is a written lead-time budget for each site:

  1. List every action in the flood plan: closing flood doors, fitting barriers, raising or moving portable equipment, starting standby pumps, switching supplies, shutting plant down and moving people.
  2. Time each one honestly, at night, in rain, with the people who would actually be on call.
  3. Add the time it takes to get those people to site.
  4. The total is the lead time the trigger has to give. Any official warning that arrives later than that is a confirmation, not a trigger.
Lead times compared: the Environment Agency planned to deploy temporary barriers within 12 hours, a flood alert usually arrives 2 to 12 hours before flooding and a flood warning 30 minutes to 2 hours before
The slowest action in the plan sets the notice a site needs.

Where should water level sensors go to buy lead time?

Put water level sensors on the water that actually floods the site, as far upstream as its travel time allows: on the watercourse above the site, at culvert inlets and screens, in the drainage outfall, and in the sumps and ducts inside. Each point answers a different question and gives a different amount of notice.

Upstream on the watercourse. A sensor on a bridge or bank structure upstream gives notice roughly equal to the time water takes to travel from there to the site. On a small river, that travel time may be the best notice available, which is why the sensor’s position matters more than its specification.

At culvert inlets and trash screens. A screen that blocks with debris raises the water behind it quickly. The level upstream of the screen, and how fast it is rising, is often the first sign that a site is about to flood from a watercourse that looked harmless the day before.

In the drainage outfall. When the receiving river or the tide is high, a site’s surface water outfall cannot discharge, and the site’s own drains back up. Measuring the level in the outfall chamber shows when that is happening, which no area warning can.

Inside the site. Sumps, basements, cable ducts, lift pits and pump rooms are where water arrives first and where it does the most expensive damage. A level sensor in a sump and a leak sensor on a plant room floor give the last and most local warning.

Surveyed levels tie the network together. Each sensor’s thresholds should come from measured heights: the level at which the site floods, the level at which access is lost, and the level at which the slowest action has to start.

Should alerts use level, rate of rise or both?

Use both. A fixed threshold says the water has reached a height that matters. A rate-of-change alarm says it is rising fast enough to get there soon, which on steep catchments often comes first. Persistence rules stop a single wave or a passing branch from waking the on-call engineer at three in the morning.

Fast rises are exactly the case the official system finds hardest. In the Calder Valley on Boxing Day 2015, the National Flood Resilience Review records that Pennine areas had over 60 mm of rain in 24 hours and some places over 100 mm, on catchments already saturated by weeks of rain. The river rose rapidly and many river level stations recorded their highest levels ever. Over 2,200 homes and 1,600 businesses were seriously affected, and a number of substations, schools, roads and bridges were damaged.

A site ladder that mirrors the three official levels keeps the response familiar to everyone involved:

Site stageTypical triggerWhat happens
WatchRate of rise above normal, or an official flood alertThe duty engineer checks readings and the forecast
PrepareUpstream level passes the first thresholdThe team is mobilised, barriers and pumps are checked
ActLevel passes the action threshold from the lead-time budgetBarriers are fitted, plant is isolated, people are moved

The thresholds are the site’s own, taken from its survey and its lead-time budget. The official warnings still count. They become one input among several instead of the only trigger.

How do the alerts reach the people who act?

Through a monitoring platform that sends alarms to named people and escalates when needed, over telemetry that keeps working when site power and local networks do not. A flood early warning is only as good as the moment a person with the authority to act reads it.

Three details matter more at a critical site than anywhere else:

  • Power. The sensor should not depend on the supply it is trying to protect. A battery-powered level sensor keeps reporting when the site’s power fails.
  • Communications. The review recorded communities losing telecoms as well as power and water. Sensors that store readings locally and resend them when the link returns leave no gap in the record, even if the network drops during the event.
  • People. Alarms should go to a named duty contact first and escalate to the next person, rather than to a shared inbox nobody watches at night.

The record matters after the event as well. Every reading and every alarm, with the time it was sent, becomes evidence for the post-incident review: when the water started rising, when the site knew, and whether the lead-time budget held.

What did the National Flood Resilience Review ask infrastructure owners to do?

The 2016 review tested key local infrastructure against extreme flood scenarios and, setting health aside, found around 530 sites still vulnerable. It asked the water and telecoms sectors to plan temporary resilience measures by Christmas 2016, in line with the electricity industry, and asked the utilities to plan permanent improvements.

It also committed £12.5 million to increase the Environment Agency’s stock of temporary flood defences and incident response equipment, and noted that the agency could deploy four times more temporary barriers that winter than the one before.

Hospitals were treated differently. The review found that temporary barriers are not the preferred option there, because encircling a hospital can stop vehicles getting in or out, although barriers can protect specific departments where access is not affected. For health estates, early warning is less about walling off a site and more about protecting plant rooms and the routes in and out.

The lesson for any owner of critical sites is that every temporary measure in the review depends on a trigger. A stockpile of barriers is only useful if someone knows, early enough, that it is time to use it.

What does a site flood early warning system include?

Water level sensors at the points that give notice, telemetry that survives the event, thresholds taken from surveyed levels, alarms that escalate to named people, and a record of every reading. Most sites also add water sensors inside plant rooms and keep the official warnings as one of the inputs.

The AQUAIOT Radar reads water level from above using 60 GHz radar, accurate to within 2 mm on still surfaces, without touching the water, so debris and silt cannot foul it. It connects over LoRaWAN or LTE Cat-1, with battery life of up to 7 years on LoRaWAN and up to 5 years on cellular, and it is rated IP68.

Our radar level monitoring service sets high and low thresholds, rate-of-change alarms for sudden rises, and persistence rules that filter out short spikes, with escalation notifications to configured contacts. On LoRa models, up to 2,000 readings are stored on the sensor and resent when the connection returns, so a network gap does not become a data gap.

AQUAIOT Radar level sensor with the layers of a site flood early warning system: upstream level, culvert screen, outfall chamber, plant room floor and escalating alarms
One network, from the river upstream to the plant room floor.

Frequently asked questions

Can you get flood warnings for surface water flooding?

No. The official service in England sends warnings about flooding from rivers, the sea or groundwater, depending on the area, and does not issue warnings for surface water flooding. Sites exposed to surface water need their own monitoring on drains, culverts and outfalls, alongside Met Office weather warnings.

How much notice does a flood warning give?

The Environment Agency usually issues a flood warning 30 minutes to 2 hours before flooding, and a flood alert between 2 and 12 hours before. A critical site should compare those times with the lead time its own flood plan needs.

What is the best sensor for flood early warning?

For most outdoor points, a non-contact radar level sensor. It is mounted above the water, so debris and silt cannot foul it, turbulence does not affect the reading, and it measures the level to within millimetres. Inside buildings, a floor-level water sensor adds the last warning before equipment is reached.

Who needs a flood early warning system?

Any site whose loss would take a service away from many people: water and wastewater sites, substations, telephone exchanges, data centres, hospitals and emergency service stations. The 2024 national assessment puts between a fifth and a third of these infrastructure types in areas at flood risk.

Where to start

Start with one site and one number: how many hours the slowest action in its flood plan really takes. Then find the water that threatens the site, place water level sensors where they give that much notice, and set the thresholds from surveyed levels.

Our guide to choosing a flood monitoring system covers the wider options, and the guide to river level monitoring covers watercourse sensors in more detail. To plan a flood early warning for a critical site, tell us about the site and a UK engineer will come back within one working day.

Previous Post
Newer Post
Shopping Cart (0 items)