In short: A man overboard alarm is built to alert the crew of the boat the person fell from, which is no help to a liveaboard who slips between hull and pontoon at 2am with nobody aboard. We believe marinas should treat the alarm as a site system, with wearable immersion tags, lifebuoy-removal sensors and pontoon CCTV all reporting to the office and the night patrol over a network that reaches the end of every finger.
Key Takeaways
- The first minute decides it — the RNLI's guidance on cold water shock is that its effects peak in the first 60 to 90 seconds, so an alert that takes ten minutes to reach a person on site is too late to matter.
- Boat alarms sound on an empty boat — AIS and wireless kill-cord man overboard devices alert the vessel and nearby ships, whereas a fall in a marina needs the alert at the office, on the patrol phone and on the neighbouring berths.
- The network has to reach the water's edge — wearable tags, lifebuoy sensors and pontoon cameras only work if the far end of every finger pontoon has coverage at 2am in January, which office Wi-Fi rarely provides.
In a nutshell
What is a man overboard alarm, and who is it designed to alert?
A man overboard alarm is a device that detects a person entering the water and raises an alert so that someone can recover them. The market splits into three families, and each was designed with a particular rescuer in mind.
The first is the AIS man overboard beacon, such as the Ocean Signal rescueME MOB1, which clips to a lifejacket, fires when the jacket inflates, and transmits the casualty's position to AIS receivers within a few nautical miles and, via DSC, to the casualty's own boat. The second is the wireless kill cord or crew overboard tag, such as Fell Marine's xEnd MOB+, which stops the engine and sounds an alarm on the helm when a paired tag goes out of range or into the water. The third is the general lone-worker device with a man-down or immersion trigger, which sends an alert over the mobile network to a monitoring centre or a named phone.
The first two families assume a crewed boat at sea. Someone else is aboard to hear the alarm, turn the boat round and get a line to the casualty. That assumption is reasonable on passage. It fails at the berth, where most of the people who fall in are alone, stepping ashore in the dark, often tired, sometimes after a drink, and the boat they have just left has nobody on it to hear anything.
Why are marina falls so dangerous?
The water in a UK marina is cold for most of the year, and cold water shock is what kills in the first minutes. The RNLI's Float to Live guidance is built around this: the gasp reflex and hyperventilation peak in the first 60 to 90 seconds, which is why the advice is to lie back and float rather than swim. The widely taught 1-10-1 rule, from the Canadian physiologist Gordon Giesbrecht, adds that a person has roughly ten minutes of useful movement before their limbs stop working, and perhaps an hour before hypothermia causes unconsciousness.
A pontoon makes that ten minutes harder than it looks. The freeboard of a floating pontoon is typically 400 to 600 mm above the water, and the casualty is often wedged between a hull and the pontoon edge, in the dark, with the fenders moving. Getting out without help is difficult even for a fit person in summer. Getting out with heavy winter clothing soaked through, and in 8°C water, is often impossible.
The national figures put the scale in context. The National Water Safety Forum's Water Incident Database recorded 193 accidental water-related fatalities in the UK in 2024, and the House of Commons Library reports 202 for 2025. The RNLI has said repeatedly that around half of the people who drown in the UK never intended to be in the water. A berth holder who slips on a frosty finger pontoon fits that description.
Who is at risk in a marina at night?
Three groups carry most of the risk, and each one changes what the alarm system has to do.
Liveaboards are the obvious group. Many UK marinas now offer residential berths, and a liveaboard crosses the pontoon several times a day in every season, including late at night. They are also the group most likely to be alone on board.
Marina staff are the second. Night patrols, dock hands taking a late arrival's lines, and engineers fixing a shore-power pillar in the rain all work at the water's edge, often alone. Under section 3 of the Health and Safety at Work etc. Act 1974 the operator also owes a duty to people who are not employees, which includes berth holders and visitors, and a marina's lone-working risk assessment will usually already identify falls into the water as a serious hazard.
Visitors are the third. Crew on a yacht that arrives after the office has closed, or a group walking back from the bar, are the people least likely to know where the nearest ladder or lifebuoy is.
Why don't the usual alarms reach anyone?
The common thread is that each device raises its alert somewhere other than where help is.
An AIS beacon in a marina will put a target on the plotter of any boat with AIS switched on, but at 2am in November most plotters in the marina are off and the boats are empty. The DSC alert goes to the casualty's own VHF, which is on the boat they fell from. HM Coastguard will receive AIS and DSC alerts in some locations, and a coastguard response is the right escalation, but the rescuer who can do the most in the first ten minutes is the person on the pontoon 50 metres away.
A crew overboard tag sounds on the helm of the casualty's boat. Again, nobody is there.
A lone-worker device with an immersion trigger does send the alert off-site, usually over the public 4G network, to a monitoring centre that then phones the marina. That chain can work well for a road worker. In a marina it adds minutes the casualty does not have, and it depends on mobile coverage at the water's edge, which in many UK marinas is weak. Steel hulls, aluminium masts and the low position of a person in the water all attenuate the signal, so coverage at the end of a long pontoon is often materially worse than at the marina office.
What does a marina-wide man overboard alarm system look like?
We think the useful design is a site system, owned by the marina, that brings several inputs into one alert and sends it to the people who can respond fastest. In practice that means (i) wearable immersion tags for staff and for berth holders who opt in, (ii) sensors on lifebuoy and throwline stations that report when one is removed, (iii) pontoon-edge CCTV with analytics that flag a person entering the water, and (iv) a single alerting layer that rings the night patrol's handset, sounds a local alarm on the affected pontoon, notifies the office and, where agreed, sends a message to the liveaboards on adjacent berths.
The lifebuoy sensor is the cheapest part of the system and among the most useful. A lifebuoy lifted from its stand at 2am is either theft, vandalism or an emergency, and all three deserve a camera check. The Yacht Harbour Association's Code of Practice for marinas already expects ladders, lifebuoys and lighting along the pontoons; adding a sensor to equipment that is already there costs tens of pounds per station.
Location matters as much as the alert itself. "Person in water, C pontoon, finger 14" gets the patrol to the right place; "immersion alert, tag 07" sends them to the office to look up a map. A system that knows which access point, small cell or gateway heard the tag, combined with the camera covering that stretch, can give a berth-level location without GPS, which performs poorly at water level between hulls anyway.
There are honest limits. No alarm prevents the fall, and the basics still matter more than the technology: ladders at regular intervals, lighting, non-slip decking, clear access in frost. Immersion sensors produce false alarms from rain, hosing down the deck and dropped devices, and they need tuning before staff trust them. Marinas also cannot make berth holders wear a tag, though a liveaboard agreement can reasonably offer one.
What network does the alarm need?
Every input in that design has to reach the alerting layer within seconds, from the far end of every finger pontoon, on the worst night of the year. That is a coverage and reliability requirement before it is a bandwidth one.
Marina Wi-Fi is usually designed to give berth holders internet access. Access points sit on poles at the head of each pontoon, shared with the berth-holder service, and the coverage is tuned for a laptop in a saloon rather than a tag on a person in the water. Wi-Fi can carry safety alerts if it is designed for them, with access points down every pontoon, a separate SSID or VLAN for safety devices, and monitoring that tells someone when an access point drops. Many marina networks were not built that way.
LoRaWAN is a good fit for the lifebuoy sensors and other low-data, battery-powered devices. A single gateway can cover a large marina, the sensors last years on a battery, and the latency of a few seconds is acceptable for a "lifebuoy removed" event. It cannot carry CCTV or voice.
A private 4G or 5G network, of the kind we run at our Bath site, covers the whole basin from a few small cells, gives each device a SIM that the marina controls, and lets staff handsets, immersion tags, pontoon cameras and berth-holder services share one managed network with traffic priority for safety alerts. The trade-off is cost and complexity: it is more infrastructure than a marina needs for alarms alone, and it makes sense when the same network also carries CCTV, card payments at the fuel berth, shore-power metering and the berth-holder internet service the marina already sells.
The right answer depends on the site. A small inland marina with a dozen liveaboards may be well served by a properly designed Wi-Fi network and a LoRaWAN gateway. A 400-berth coastal marina with a night patrol, a fuel berth and a camera network is a strong candidate for a private network that does all of it.
Where should a marina start?
We would start with the night. Walk the pontoons at 2am with a phone and a lone-worker device, test the signal at the end of each finger, and time how long an alert takes to reach the person on patrol. Most marinas that try this find at least one stretch of pontoon where nothing gets through.
Then fit sensors to the lifebuoy stations, issue immersion tags to the night patrol and anyone working alone at the water's edge, and offer them to liveaboards. Route every alert to the patrol handset and a local pontoon alarm first, and to a monitoring centre or the coastguard second. Add CCTV analytics on the pontoon edges where liveaboards are concentrated, and review every alarm, real or false, monthly for the first six months.
If the coverage test shows gaps, fix the network before buying more devices. We are happy to run that 2am walk with any UK marina that wants to know where its pontoons go quiet.