Man Overboard Procedures

The sea does not wait for a committee.
It cools a body at its own rate, and the rate is always faster than the response.

Introduction

Every master and OOW has rehearsed the shout. Helm hard over, release the bridge-wing lifebuoy, press the MOB button on the GPS. The checklist sits laminated behind glass. The drill is logged quarterly. And yet, in incident after incident, the time between the person entering the water and the ship arriving back at the datum stretches to fifteen, twenty, thirty minutes — and the casualty is not where the mark says they should be.

The reason is simple. Drills rehearse the mechanical sequence. They rarely rehearse the decisions that matter: which manoeuvre to use and why, when to abandon the initial approach and commit to a search pattern, when to alert the coastguard instead of hoping to handle it alone, and — hardest of all — when to accept that survival time has expired.

This article addresses those decisions. It assumes familiarity with basic bridge team resource management. It is not a primer. It is an attempt to close the gap between the laminated card and the reality of a person in the water at night in a force six.

Contents

• 1. The Arithmetic of Survival
• 2. Immediate Actions — The First Sixty Seconds
• 3. Marker Deployment and Electronic Position Marking
• 4. The Three Recovery Manoeuvres
• 5. AIS-MOB Transmitters and DSC Distress Alerting
• 6. Search Patterns After the Turn
• 7. The Sustained Response and SAR Coordination
• 8. Crew Training and the Drills That Actually Matter
• 9. Closing Reality

1. The Arithmetic of Survival

Everything in a man overboard response is subordinate to one fact: the thermal clock is running from the moment of immersion, and nothing the bridge team does can slow it.

In North Sea winter conditions — water temperatures of 4–8°C — a person without immersion protection loses useful consciousness in fifteen to thirty minutes. In the North Atlantic in February, that window can be under fifteen. Even in tropical waters of 25°C and above, exhaustion drowning without a flotation aid is a realistic outcome inside an hour. These are not textbook abstractions. They are averages, and many individuals fall below the average.

The RNLI’s own data shows that in UK waters, the probability of recovering a person alive drops sharply after thirty minutes of immersion. After sixty minutes in cold water without a lifejacket, recovery alive is an exception rather than a norm.

Every minute spent deciding is a minute of core temperature lost.

This arithmetic must govern the entire response. It means the choice of manoeuvre is not academic. A Williamson turn that takes eight to ten minutes to complete is a fundamentally different proposition from an Anderson turn that takes three. The choice depends on when the casualty went over, how long ago the event was observed, and whether position is known. Get the choice wrong, and the ship arrives at a datum where the person no longer has the capacity to assist in their own rescue — if they are still on the surface at all.

2. Immediate Actions — The First Sixty Seconds

The first sixty seconds are not about the manoeuvre. They are about three things: not losing sight of the casualty, marking the position, and getting flotation into the water.

The OOW’s first action is to put the helm hard over to the side of the casualty. This swings the stern — and the propeller — away from the person. This is instinctive, or it is trained to be instinctive, or it is forgotten in the moment. There is no fourth option.

Simultaneously, the bridge-wing lifebuoy must go over. On most vessels there is a release mechanism that deploys the lifebuoy with its self-igniting light and self-activating smoke marker. This is not primarily a flotation aid for the casualty — though it may serve that purpose if the person can reach it. It is a visual datum. A human head in a moderate sea state is visible from the bridge at perhaps two to four cables. A lifebuoy with an orange smoke trail is visible at a mile.

The MOB function on the GPS or ECDIS must be activated immediately. Every second of delay between the event and the mark represents positional error — the ship has moved, and the mark records where the ship was, not where the person went in. On ECDIS, the MOB waypoint becomes the centre of the return approach, but it must be treated as an approximation.

The mark is where the button was pressed. The person is where the sea has taken them.

The engine room must be informed. If the vessel is on UMS, the engineering officer of the watch needs to know that full manoeuvring capability may be required immediately. Main engine response time matters — a slow-speed diesel on bridge control is one thing; a plant that has been running at steady-state with automation managing load is another. The telegraph should go to standby.

The master must be called. This is not optional and not a judgment call. The OOW begins the response; the master takes the conn and coordinates the broader operation. If the OOW hesitates to call the master because the casualty appears to be close and recoverable, the OOW has already made the first decision that kills people.

3. Marker Deployment and Electronic Position Marking

Visual markers — the bridge-wing lifebuoy and its smoke float — are the primary short-range datum. In daylight and moderate visibility, the smoke is the single most useful thing in the water. It drifts with the surface current, roughly as the casualty does, giving a better real-time indication of the casualty’s position than a fixed electronic waypoint. At night, the self-igniting light serves the same function, but with greatly reduced effective range.

Additional lifebuoys with lights should be deployed if available and if the vessel’s track allows. The more visual references in the water, the better the OOW’s ability to judge wind-driven drift and current set.

Dan buoys — the tall vertical markers with flag and light — are carried specifically for this purpose. A dan buoy visible above the wave crests is invaluable in sea states where a lifebuoy disappears in the troughs. If the vessel carries one on the bridge wing, it goes over with the lifebuoy. If it is stowed elsewhere, it must be deployed as soon as hands are available.

On the electronic side, the ECDIS MOB mark provides a lat/long reference. This is useful for the return approach and essential if the vessel loses visual contact. But it records the ship’s position at the time of activation, not the casualty’s position. In a beam wind or significant current, the actual point of entry may be fifty metres or more from the mark. This error grows with every minute of delay between the event and the button press.

A GPS waypoint is not a person. It is a guess with six decimal places.

4. The Three Recovery Manoeuvres

Three standard manoeuvres exist. Each serves a different tactical situation. Selecting the wrong one is not a minor error — it is a decision that directly determines whether the vessel arrives at the datum in time.

The Williamson Turn

Helm is put hard over to the side the casualty went over. When the vessel’s heading has changed approximately sixty degrees from the original course, the helm is shifted hard over to the opposite side. The vessel continues turning until it steadies on the reciprocal of the original course. The ship is then running back down its own wake, directly towards the datum.

The Williamson turn is the manoeuvre for a delayed response. If the casualty was not seen to go over — if the absence was discovered at a muster, or reported by a colleague minutes after the event — the Williamson puts the vessel back on its original track line. This is the only manoeuvre that does so reliably. It is the correct choice when the position of entry is uncertain and the best available strategy is to retrace the ship’s path.

It is also the slowest. Depending on the vessel’s turning characteristics, the Williamson takes eight to twelve minutes to complete. In cold water, this is a lifetime — or the end of one.

The Anderson Turn (Single Turn)

Helm is put hard over to the side of the casualty. The vessel completes a tight turn of approximately 270 degrees, arriving back at the original track line from the opposite direction. It is fast. In a manoeuvrable vessel it can be completed in three to five minutes.

The Anderson turn is the manoeuvre for an immediate, witnessed event. The person was seen to fall. The position is known. The bridge team has visual contact or a very recent datum. Speed of return matters more than precision of track retracing, because the casualty’s position is not in serious doubt.

The limitation is that the vessel does not arrive exactly on the reciprocal course — the geometry puts the ship slightly offset from the original track. This matters less when the casualty is visible or the datum is fresh, and matters more when the situation degrades into a search.

The Scharnov Turn

The vessel continues on its current heading for a calculated period, then puts the helm hard over to the opposite side of the original casualty position, turning through approximately 240 degrees to steady on the reciprocal of the original course. This is the manoeuvre designed for a delayed discovery when the time of the event is known but was not witnessed.

The Scharnov is faster than the Williamson and, like the Williamson, brings the vessel back onto the reciprocal course line. Its disadvantage is that it requires accurate knowledge of how long ago the casualty went over, because the initial straight run before the turn must be calculated to ensure the vessel swings back to the correct track segment. If the time estimate is wrong, the vessel arrives on the reciprocal course in the wrong place.

In practice, the Scharnov is the least used of the three. Its reliance on accurate elapsed-time data makes it fragile. If there is genuine uncertainty about when the person went over, the Williamson — slower but more forgiving — is the safer choice.

Speed of return and accuracy of return are in tension. The situation decides which matters more.

Selection Logic

The decision tree is simpler than the textbooks make it appear:

• Person seen to fall, position known, visual contact likely: Anderson turn.
• Person not seen to fall, time of event uncertain, position uncertain: Williamson turn.
• Person not seen to fall, time of event reasonably known: Scharnov turn — but only if the elapsed time calculation can be trusted. If in doubt, Williamson.

In rough weather or at night, the bias should always shift towards the Williamson. The reciprocal-course approach gives the best chance of running back through the debris field — the lifebuoy, the smoke, the dan buoy — and using those markers to refine the search datum. In calm daylight with a witnessed fall, the Anderson’s speed advantage is decisive.

5. AIS-MOB Transmitters and DSC Distress Alerting

AIS-MOB devices — personal transmitters that activate on immersion and broadcast an AIS position — have changed the geometry of the problem, but not as much as their marketing suggests.

When they work, they provide a continuously updating position for the casualty, displayed on the ECDIS or radar overlay of every AIS-equipped vessel in range. This is transformative. It converts a search problem into an approach problem. The vessel no longer needs to retrace its track hoping to find a head in the waves. It has a target.

When they do not work — and the failure modes are numerous: not carried, not armed, not registered, battery expired, antenna submerged because the casualty is face-down, signal blocked by wave crests in heavy weather — the operation reverts to the traditional datum-and-search model. Reliance on the device as a substitute for immediate visual tracking is a trap. The lifebuoy, the smoke, and the lookout’s pointed arm remain the primary references. The AIS-MOB is a supplement, not a replacement.

DSC distress alerting is a different function with a different purpose. The vessel’s GMDSS console allows immediate broadcast of a distress alert with position and nature of distress. For a man overboard, this should be initiated early — far earlier than most bridge teams are comfortable with.

The reluctance to press the distress button is well documented. Masters hesitate because a false alarm brings consequences, because activating SAR feels like an admission of failure, because there is a belief that the ship can handle it alone. This hesitation costs lives. The RCC can always stand down assets that are not needed. It cannot conjure a helicopter out of nothing thirty minutes into an event when the bridge team finally admits they have lost the casualty.

Alerting the coastguard is not an escalation. It is a parallel action. It buys time that the thermal clock is consuming.

The DSC alert should go out as soon as it is clear that immediate recovery alongside is not going to happen within minutes. That threshold is lower than most people think.

6. Search Patterns After the Turn

If the initial manoeuvre does not result in visual reacquisition of the casualty, the operation transitions from recovery to search. This transition must be recognised explicitly. It changes the psychology, the communications, and the strategy.

The standard search patterns are defined in IAMSAR Volume III. For a single vessel, the expanding square search is the default when the datum is reasonably well known. The vessel starts at the datum and runs outward in a square spiral, each leg slightly longer than the last, expanding the searched area systematically.

If the datum is uncertain — a delayed discovery with no AIS-MOB and a questionable last-known position — the sector search (also known as the Victor Sierra pattern) is more appropriate. The vessel runs outward from the datum along a bearing, turns and returns through the datum on a different bearing, and repeats, covering a circular area in overlapping passes.

For multiple vessels, the parallel track search (coordinated by the OSC or RCC) divides the probable area into lanes. Each vessel sweeps its lane at a spacing calculated from the prevailing visibility and sea state.

In all patterns, the critical variable is sweep width — the effective distance at which a lookout can reliably detect a person in the water. In a calm sea with good visibility, this might be several hundred metres. In a sea state of four or five with reduced visibility, it collapses to tens of metres. The IAMSAR tables provide theoretical sweep widths; the reality on the day is invariably worse.

Track spacing must be set conservatively. Covering more area faster at wide spacing sounds efficient. It means the probability of detection in each pass is low. Tight spacing with high probability of detection is slow but more likely to find the casualty while they are still alive.

Searching fast is not the same as searching well.

7. The Sustained Response and SAR Coordination

Beyond the first fifteen to twenty minutes, the operation is no longer a bridge team exercise. It is a coordinated SAR event.

The master must communicate with the MRCC. The vessel’s position, the nature of the event, the number of persons in the water, the sea state, water temperature, and whether the casualty was wearing a lifejacket — all of this must be transmitted clearly. The MRCC will allocate an on-scene coordinator if multiple assets respond. If the ship is the only asset, the master is the OSC by default.

Helicopter response, where available, is the most effective search platform. A helicopter covers area at a rate that no surface vessel can match, and its elevated vantage point multiplies sweep width. But helicopter availability depends on range, weather, and tasking. In many ocean areas, there is no helicopter response. The ship is alone.

Other merchant vessels in the area should be contacted via VHF. The obligation to render assistance to persons in danger at sea is absolute under SOLAS and the law of the sea. A second vessel significantly improves the probability of detection in a search pattern.

As time passes, the search area expands. Current drift, leeway (wind effect on a person in the water, with or without flotation), and elapsed time combine to enlarge the probable position from a point to a circle to an ellipse that may cover several square miles. IAMSAR Volume II provides drift calculation methods. The ECDIS can overlay tidal data. But these are estimates layered on estimates, and the uncertainty grows faster than most people expect.

The decision to suspend a search is the master’s responsibility, in consultation with the RCC. It is not taken lightly. But it must be taken when the evidence — water temperature, elapsed time, sea conditions, absence of flotation — makes survival implausible. Continuing a search beyond the point of reasonable hope exposes the searching vessel and crew to fatigue and risk without corresponding benefit.

8. Crew Training and the Drills That Actually Matter

The quarterly MOB drill on most vessels consists of throwing a dummy over the side in calm weather during daylight, executing a Williamson turn because that is the one everyone remembers, and recovering the dummy using the rescue boat. The drill is logged, the box is ticked, and no learning of consequence occurs.

Effective MOB training requires discomfort. It requires drills at night. It requires drills where the OOW is not told in advance. It requires the master to stand back and let the OOW make the initial decisions — and then debrief honestly about what went wrong.

It requires practising the Anderson turn, not just the Williamson. It requires understanding the vessel’s actual turning characteristics — the tactical diameter, the advance, the transfer — in the current loading condition, not the one in the manoeuvring booklet from sea trials twenty years ago.

It requires testing AIS-MOB devices. Not reading the manual. Activating them in the water and confirming they appear on the ECDIS. Discovering, before the real event, that the battery is dead or the device is not registered in the AIS transponder’s list.

It requires the engine room to participate. A fast manoeuvre means rapid telegraph changes. The main engine must respond. If the plant needs time to come to manoeuvring status from a steady-state sea passage configuration, that time must be known and planned for.

A drill that does not create doubt is a drill that has taught nothing.

Above all, it requires confronting the survival statistics honestly. The bridge team must understand that in winter North Atlantic conditions, a person without a lifejacket who is not recovered within twenty minutes is very likely dead. This is not pessimism. It is the planning assumption that should drive every decision towards speed, towards early alerting, towards not waiting to see if the situation resolves itself.

9. Closing Reality

A man overboard event is the collision of seamanship with thermodynamics. The seamanship is trainable. The thermodynamics are not negotiable.

The manoeuvres exist for a reason. The Anderson turn is fast and suits a witnessed event. The Williamson retraces the track and suits a delayed discovery. The Scharnov is a compromise that demands accurate time data. Choosing between them is not an academic exercise — it determines whether the vessel arrives at the datum in time for the recovery to mean anything.

Markers go over immediately. The DSC alert goes out early. AIS-MOB devices are checked and carried, not stored in a locker. Search patterns are selected for detection probability, not area coverage speed. The coastguard is called before the bridge team runs out of ideas, not after.

Everything comes back to the thermal clock. It does not pause for indecision. It does not pause for a muster to confirm who is missing. It does not pause while the OOW debates whether to wake the master.

In cold water, hesitation is the mechanism of death. Procedure exists to eliminate it.