Detection · Risk Scoring · Continuous Monitoring

Every culvert.
Scored. Prioritised.

We map every culvert in your authority area, score each one for flood risk and blockage likelihood, and turn it into a prioritised register your team can work down — so you know where to look before the next storm, not after.

See your authority area How it works

Risk-scored culvert map — real data, live platform

27,000+road–watercourse crossings mapped nationally
4-tierrisk classification, every culvert
152Lead Local Flood Authorities in England
The Problem

Small culverts.
Big consequences.

Most culvert blockages aren't found until water is already on the road. Inspections happen on a two-yearly reactive cycle — if at all — and there's no continuous record between visits. When a major event hits, Section 19 of the Flood and Water Management Act 2010 obliges Lead Local Flood Authorities to investigate; without a register of known assets, that investigation starts from scratch. Storm Babet in October 2023 made the gap visible across Yorkshire, Lincolnshire and the East Midlands.

S.19

Required after significant flood events. Continuous data supports your statutory duty.

2-yearly

Typical reactive inspection cycle. Blocked inlets fill silently between visits.

£2,500

Typical cost of one gully clean and road closure after a culvert blockage.

152

Lead Local Flood Authorities in England — each with a statutory duty for ordinary watercourses and culverts.

The Register

Your culvert register.
Pre-built and risk-scored.

Two tiers. A national open-data baseline you can view today, and a deeper, enhanced pass when you work with us.

National baseline

Available today

27,000+ road–watercourse crossings already identified across England from OS and Environment Agency open data, each risk-scored by flood probability, infrastructure proximity, and flood history. Your authority area is ready to view today.

Detailed analysis

When you work with us

We run your area through the full detection pipeline using the OS NGD Water Network — including the buried and culverted sections free data misses. Pilot councils typically see significantly more crossings detected, each classified as culvert, bridge or ford from OS survey data.

How crossings are classified
Confirmed culvert

OS NGD Underground classification + surveyed entrance points.

Likely culvert

NGD Underground crossing without surveyed points.

Bridge or ford

NGD Surface Level crossing.

Surveyed buried

OS-surveyed culvert entrance with no road–water intersection — urban and Victorian assets that intersection-based detection misses.

See the register for your authority area
Risk Scoring

Every culvert classified.
Likelihood × Consequence.

Every culvert in your register is scored on two independent axes — how likely it is to fail, and how bad the consequences would be if it did — then combined into a single number you can sort and prioritise by.

Blockage likelihood

How likely to fail

We analyse upstream land cover, channel gradient, vegetation canopy, inlet aspect, watercourse size, pipe gradient, and historical flood evidence to estimate how likely each culvert is to block.

Flood consequence

How bad if it fails

EA surface water flood probability, proximity to housing, schools, hospitals, emergency services, road classification, and flood history — combined to estimate the impact of a failure.

Likelihood × Consequence = a risk classification for every culvert — Critical, High, Moderate, or Low — with the contributing factors visible so your team understands exactly why each culvert was flagged.

Critical
Risk classification

Likelihood: High  ×  Consequence: High

Woodland upstream Steep gradient South-facing inlet Housing ≤ 50m Surface water: High
Terrain Screening

LiDAR terrain analysis
at every crossing.

Using Environment Agency 1m LiDAR data, we analyse the terrain at each culvert crossing to estimate channel depth, embankment height, and pipe gradient — identifying potential structural issues without a site visit.

Longitudinal cross-section of Macon Way / unnamed watercourse, Cheshire East — elevation profile sampled along the NGD waterlink centreline at 1m intervals over 60m, with red upstream and green downstream segments and a grey road-crossing zone at d=0.

Channel depression depthEstimates how deep the watercourse cuts below the surrounding terrain.

Embankment heightMeasures how high the road sits above the channel — approximating the available space for a culvert pipe.

Adverse gradient detectionFlags culverts where the outlet appears higher than the inlet, indicating siltation risk or structural settlement.

Ponding indicatorsIdentifies upstream depressions where water may pool against the road embankment.

Terrain screening is indicative, not measured — it identifies sites worth investigating. Field verification confirms the findings.

Continuous Monitoring

For your highest-priority culverts,
add continuous monitoring.

For your highest-risk sites, add a water-level sensor at the inlet headwall. No mains power, no cellular dependency — LoRaWAN-connected, and designed to alert your team before overtopping or blockage develops.

ALERT THRESHOLD SENSOR GATEWAY UPSTREAM · backing up DOWNSTREAM · flowing
How It Works

Four parts.
One early warning.

Standard off-the-shelf hardware. No proprietary lock-in, no cellular dependency, no in-field mains power.

📡
Sensor Node →
📶
LoRa Gateway →
☁️
Network Server →
🔔
Alert

Select a component to explore

Sensor Node

A waterproof ultrasonic sensor mounts above the inlet and points down at the water. It fires a 40kHz pulse every few minutes, measures echo return time, transmits the reading over LoRa, then sleeps. Battery-powered, no mains, no SIM.

ESP32 microcontroller Solar option available
LoRaWAN Gateway

A single gateway, a small weatherproof box mounted on a rooftop or utility pole, receives transmissions from sensor nodes up to 2km away in urban areas and further in open terrain. One gateway can serve dozens of sensor nodes across an area. It connects to the internet over ethernet or WiFi and forwards readings to the server. If LoRa coverage is unavailable at a site, a mobile data connection can be used as a fallback.

~2km urban · 10km+ line of sight One gateway, many sensors No cellular required NB-IoT fallback available
Network Server

Once the gateway receives a reading from a sensor node, it passes the data to a server that decides what to do with it. This is where the alert rules live: if the water level crosses a set threshold, an alert fires. The server is privately operated, meaning your council's data stays under your council's control. Over time, readings are correlated against Met Office rainfall data to build a clearer picture of which culverts are most at risk and when.

Council-operated or hosted Threshold + rate-of-rise logic Met Office API integration
Alert

When a threshold is breached, a webhook fires within seconds, delivering an alert to email, SMS, or an operations dashboard. The early warning gives highways and drainage teams time to inspect, close roads proactively, or deploy temporary measures. All data is logged, providing the continuous sensor record that supports Section 19 flood investigations.

Email / SMS / dashboard Configurable early warning < 2 min alert latency S.19 investigation evidence
What Can Be Detected

More than just
rising water.

One upstream sensor, two alert types from the same hardware: flood early warning, and blockage detection.

Flood early warning
Water level rising with rainfall

During heavy rainfall, upstream water backs up as flow exceeds the culvert's capacity. The sensor catches the rising surface and fires an alert before overtopping — buying highways and drainage teams time to respond before road damage or property flooding.

Threshold alert Rate-of-rise alert Weather correlated
Blockage detection
Water rising faster than rain explains

When debris blocks the inlet, water backs up faster than rainfall alone explains. If the sensor sits directly over the inlet, floating debris also reflects the pulse early — both signals trigger an alert. Caught in dry weather, it's a routine clean. Caught during a storm, it's a Section 19 investigation.

Anomalous rate-of-rise Debris surface reflection Dry-weather trigger
Optional extension: upstream + downstream sensors

Adding a second sensor downstream allows direct flow differential measurement. If upstream is rising but downstream is not, the culvert is blocked. This gives maintenance teams a precise, actionable signal even without rainfall context.

Benefits
  • Unambiguous blockage confirmation
  • Flow rate estimation across culvert
  • Richer dataset for ML model training
  • Useful even in dry conditions
Tradeoffs
  • Two installation points per site
  • Two potential failure points to maintain
  • Downstream access can be difficult
Return on Investment

The cost of one flood event
dwarfs a decade of monitoring.

One culvert blockage at Eel Brook, Church Minshull generated £476,000 in insurance claims, triggered an MP inquiry and a formal Section 19 investigation. A year of CulvertSense across an entire authority is a fraction of the cost of one event like this.

£476k

insurance claims from a single blocked culvert (Eel Brook, Church Minshull)

839,000

homes at high risk of surface water flooding in England (NHF/EA, 2026)

Pilot Programme

Low risk.
Measurable outcomes.

We're working with our first cohort of Lead Local Flood Authorities. Three stages, each evidence-gated before you commit to the next, and no operational dependency at any point.

1 Register
Detect & prioritise

Your authority's full culvert register — every crossing detected, classified, risk-scored 0–100, and mapped. Review with your team, prioritise, and layer in your local knowledge. Browser-based, nothing to install.

2 Monitoring
Continuous water level

Add continuous water-level monitoring at your highest-priority sites. Half-day install per sensor, no mains power, and a shared dashboard with configurable alerts.

3 Predictive
Forecast-correlated alerts

As your sensor network matures, we correlate water-level trends with Met Office rainfall forecasts to alert teams before water starts to rise. Rolled out to pilot partners as it develops.

Request a pilot

Tell us about your council and the culverts that worry you most. We'll confirm whether it's a good fit and walk you through what comes next. No commitment.

Or email directly: james@culvertsense.com

Built on Ordnance Survey NGD Water Network · Environment Agency LiDAR & flood risk data · OpenStreetMap