For most of the twentieth century a ship in trouble was heard because another human being was sitting in a radio room with headphones on, waiting for something that almost never came. That arrangement was written into international law after the Titanic sinking, and it was retired on 1 February 1999, when the Global Maritime Distress and Safety System came fully into force under the SOLAS convention. What replaced the listening operator is a set of machines that alert shore rescue authorities directly, identify the vessel by number, and report a position without anyone speaking. The engineering is sound, and the trade it made is worth understanding in detail.

The watch Titanic bought, and why IMO gave it up

The 1912 disaster produced two legal instruments that shaped maritime radio for the next eighty-seven years. In the United States, the Radio Act of 1912 imposed licensing and watch obligations, and internationally the first Safety of Life at Sea convention, signed in 1914, moved toward requiring passenger ships to keep a continuous radio watch. The reasoning was simple enough after the fact, because the nearest ship to Titanic had a radio set and a single operator who had gone to bed, so the apparatus was working and nobody was listening to it.

The watch that emerged was a human one, organized around silence periods. Radio room clocks carried marked sectors on the dial, and twice an hour every station stopped transmitting so that a weak distress signal could be heard through the traffic, with the 500 kHz Morse silence periods falling at a quarter past and a quarter to the hour and the 2182 kHz radiotelephone silence periods falling on the hour and the half hour. A ship that could not staff a twenty-four hour watch carried an auto alarm receiver instead, tuned to 500 kHz and designed to trip bells on the bridge and in the operator’s cabin when it detected the alarm signal, which was a series of long dashes separated by short gaps, deliberately unlike anything in normal Morse traffic. The radiotelephone equivalent on 2182 kHz was a pair of alternating audio tones at 2200 Hz and 1300 Hz.

By the 1970s the International Maritime Organization was looking at a system whose weaknesses were well documented. Morse and MF radiotelephone distress calling was line of sight or groundwave dependent, so a ship in the middle of an ocean was relying on skip propagation and on whoever happened to be listening, and the alert went to other ships rather than to a rescue coordination center ashore. IMO adopted the GMDSS amendments to SOLAS in 1988, phased them in beginning 1 February 1992, and made compliance complete on 1 February 1999. The US Coast Guard ended its own Morse watch several years ahead of the international deadline, in the mid-1990s.

What GMDSS is actually designed to do

The organizing principle of GMDSS is that a distress alert should reach a shore-based search and rescue authority quickly and reliably, along with ships in the vicinity, so that a coordinated rescue can begin with minimum delay. That inverts the older model, in which the alert went out to whoever could hear it and the shore authorities learned about the casualty secondhand. A rescue coordination center has aircraft, cutters, a picture of what merchant traffic is nearby from AIS and reporting systems, and the authority to task all of it, which a passing bulk carrier does not.

SOLAS Chapter IV frames the requirement functionally rather than by naming boxes. A ship must be able to transmit a ship-to-shore distress alert by at least two separate and independent means, receive shore-to-ship distress alerts, transmit and receive ship-to-ship distress alerts, send and receive search and rescue coordinating communications and on-scene communications, transmit and receive signals for locating, receive maritime safety information, and conduct general and bridge-to-bridge communications. The specific equipment a given ship carries falls out of those functions combined with where the ship operates.

The carriage rules apply to SOLAS ships, which in the usual formulation means cargo ships of 300 gross tonnage and upwards on international voyages and all passenger ships on international voyages, though flag administrations extend similar requirements to other categories of vessel and the details vary by country. Fishing vessels, domestic passenger vessels and recreational boats are covered by national rules rather than by SOLAS Chapter IV directly. If you need to know what a particular vessel must carry, that answer comes from the flag state administration, and for US-flag vessels from the Coast Guard together with the FCC rules for maritime services in Part 80 of the FCC’s regulations.

Digital selective calling, the MMSI, and channel 70

Digital selective calling is the piece that most directly replaced the human ear. A DSC-equipped radio sends a short burst of data on a dedicated calling frequency, and receiving stations decode it, alarm, and display who called and what they want. The distress calling channels are VHF channel 70 at 156.525 MHz, 2187.5 kHz on MF, and a set of HF channels at 4207.5, 6312, 8414.5, 12577 and 16804.5 kHz. Nothing transmits voice on those frequencies, because they exist purely to carry the digital call that tells everyone to go somewhere else and start talking.

The address in that system is the Maritime Mobile Service Identity, a nine-digit number in which the first three digits are the Maritime Identification Digits assigned to a country. The MMSI is how a coast station knows which vessel just alerted, and it is the key into a registration database holding the vessel name, description, and emergency contacts. A distress alert carries the MMSI, the nature of distress if the operator selected one, and the vessel’s position with the time that position was valid, provided the radio has a working GNSS input or somebody entered coordinates by hand.

The procedure is that the DSC alert is the notification and the voice call is the conversation. After sending the alert, the operator shifts to the associated voice frequency, meaning VHF channel 16 or 2182 kHz on MF, and transmits a spoken distress call giving vessel, position, nature of distress and assistance required. A DSC alert that arrives with a position and no voice traffic afterward still gets a response, but the rescue coordination center is working from a single data point and a database entry rather than from a description of what is actually happening on the boat.

VHF channel 16 has not gone away. The SOLAS requirement for ships to maintain a continuous listening watch on 156.8 MHz while at sea remains in force, and proposals to end it have not been carried through. What did change is the coast station side in some countries, since the US Coast Guard announced in 2013 that it would end its guard of 2182 kHz voice and 2187.5 kHz DSC, so if you need to know which frequencies are currently guarded in your area, get that from the Coast Guard rather than from an older reference.

406 MHz: EPIRBs and the Cospas-Sarsat system

The second independent means of alerting on most ships is a beacon transmitting on 406 MHz into the international Cospas-Sarsat satellite system. That program began with a memorandum of understanding signed in 1979 by agencies of Canada, France, the United States and the Soviet Union, and the first rescue credited to it came in 1982. Emergency position indicating radio beacons are the maritime version, emergency locator transmitters are the aviation version, and personal locator beacons are carried by individuals, all using the same 406 MHz alerting and the same ground segment.

The beacon transmits a short digital burst carrying a unique coded identity, repeated at intervals of roughly a minute, and most current beacons also encode a GNSS position taken from an internal receiver. Detection originally worked two ways. Low-earth-orbit satellites determined position by Doppler shift as the satellite passed overhead, which was accurate but could mean waiting for a pass, and geostationary satellites relayed the alert instantly but could give no independent position unless the beacon supplied one. The MEOSAR segment, which places Cospas-Sarsat repeaters on medium-earth-orbit navigation satellites including GPS, Galileo and GLONASS, was built to give both near-immediate detection and independent location from multiple satellites in view. Check the Cospas-Sarsat program’s own site for current system status and capability, because this segment has been evolving for more than a decade.

Shipboard EPIRBs are commonly installed in a float-free bracket with a hydrostatic release, so that a vessel that sinks releases the beacon at shallow depth and it floats up, switches on, and transmits without anyone having touched it. Beacons also carry a low-power 121.5 MHz homing transmitter, which satellites no longer process but which aircraft and rescue boats use for final direction finding once they are in the area. Satellite processing of 121.5 and 243 MHz alerting ended on 1 February 2009, so an old beacon that transmits only on those frequencies will not alert anybody.

Registration is what turns a beacon identity into a rescue. In the United States, 406 MHz beacons are registered with NOAA through the national beacon registration database, and the record holds the vessel or aircraft description and the emergency contacts a rescue coordination center calls first. That phone call is frequently how a false alert is resolved in minutes instead of by launching a helicopter, which is a good argument for keeping the record current on its own.

A1 through A4: carriage that follows geography

GMDSS divides the world into four sea areas, and a ship’s required equipment depends on which ones it operates in. Sea area A1 is the area within VHF coverage of at least one coast station where continuous DSC alerting is available, as declared by the government concerned, which in practice tends to be on the order of twenty to thirty nautical miles from the antenna. Sea area A2 is the area within MF coverage of a coast station with continuous DSC watch, excluding A1, and is often described as extending out to somewhere around a hundred to a hundred and fifty miles. Those distances are typical values rather than parts of the definition, since the definitions are written in terms of declared coverage.

Sea area A3 was originally defined by the coverage of Inmarsat geostationary satellites, excluding A1 and A2, which in practice covered the world between roughly seventy degrees north and seventy degrees south. Sea area A4 is everything left over, meaning the polar regions, where a ship falls back on HF because no geostationary satellite is usefully above the horizon. Ships operating in A4 carry HF DSC and radiotelex equipment for that reason.

The satellite side of this has changed recently and the change matters. IMO recognized Iridium as a GMDSS mobile satellite service provider, its service became operational in 2020, and a modernized version of SOLAS Chapter IV entered into force on 1 January 2024, with the A3 definition rewritten in terms of a recognized mobile satellite service rather than one named operator. The practical effect is that a ship can now satisfy some satellite requirements through a low-earth-orbit constellation with genuine polar coverage, which blurs the old assumption that A4 automatically meant HF only. Because this is an area of active change, confirm the current recognized providers and the current text of Chapter IV through IMO and through your flag administration rather than relying on any secondary summary, including this one.

Carriage rules also cover the problem of what happens when the equipment breaks at sea. SOLAS requires that availability be maintained by one or more of three approaches, which are duplication of equipment, shore-based maintenance, and at-sea maintenance capability, and ships operating in the offshore areas are required to use at least two of the three. That requirement exists because GMDSS assumed the radio officer who used to repair equipment underway would no longer be aboard.

Radar SARTs, AIS-SART, and the safety information broadcasts

Alerting tells the rescue coordination center that something has happened and roughly where, and the separate problem of putting a searching vessel or aircraft alongside a liferaft is handled by dedicated equipment. The radar search and rescue transponder works in the 9 GHz band, responds when it is illuminated by a ship’s X-band radar, and paints a line of twelve dots on the searching vessel’s radar display extending outward from the transponder’s position along its bearing. As the searcher closes, those dots widen into arcs, which gives the radar operator an unambiguous target that cannot be confused with sea clutter.

The AIS-SART does the same job differently, transmitting position reports on the AIS channels so that the survivor shows up as a distinct target on any AIS display, including those on aircraft and small boats that have no X-band radar. Either type satisfies the locating requirement, and each has a different failure mode, since the radar SART depends on someone running X-band radar and looking at the screen, while the AIS-SART depends on an AIS receiver within VHF range and on the GNSS position it carries being correct.

The receiving half of GMDSS is maritime safety information, meaning navigational warnings, meteorological warnings and forecasts, and urgent shore-to-ship messages. NAVTEX carries this automatically on 518 kHz for international English broadcasts and 490 kHz for national language broadcasts, printing or displaying messages aboard without the crew tuning anything, with a message numbering scheme that lets the receiver suppress bulletins it has already shown. Outside NAVTEX coverage the same traffic arrives by satellite through the enhanced group call service, addressed to ships by geographic area. Both are one-way broadcasts, which is precisely the point, because the alerting side of a shore-to-ship distress relay uses the same infrastructure to reach every ship in a defined box without calling them individually.

False alerts and unregistered beacons

Every automated alerting system generates false alerts, and GMDSS generated a great many of them, particularly in its first years. The causes are the ordinary ones. Somebody presses the distress button on a DSC radio while cleaning the bridge, a beacon is activated in a storage locker or during a battery replacement, equipment is tested improperly, or a vessel is sold and scrapped with a registered beacon still aboard. IMO and Cospas-Sarsat have both published guidance aimed at reducing the rate, and current false alert statistics should come from Cospas-Sarsat’s own published figures rather than from an older secondary number, because the rate has moved considerably over the years and different sources count different things.

Two design decisions reduce the damage. DSC distress buttons on type-approved equipment are recessed under a cover and require a deliberate press and hold, with an audible countdown, so an accidental brush does not alert. Registration is the other, because a rescue coordination center holding a beacon identity that resolves to a vessel and an emergency contact can make a phone call, establish that the boat is on a trailer in a driveway, and close the case without moving an asset. An unregistered beacon offers none of that, so the response has to assume the alert is real.

What you must not do is silently switch off a beacon or a radio after an accidental alert. A false alert that is cancelled properly costs a phone call, while one that is switched off without notification produces a search for a vessel that has apparently stopped transmitting, which is the signature of a sinking. The correct action after an accidental DSC alert is to transmit a cancellation on the associated voice frequency, identifying the vessel and the MMSI and stating that the alert was sent in error, and to contact the Coast Guard or the appropriate rescue authority directly. For an accidental beacon activation, turn it off and call the rescue coordination center or the national beacon authority immediately.

What automation cost, and who to ask about current rules

The gains from GMDSS are real and measurable. An alert now goes directly to a shore authority with rescue assets, it carries an identity that resolves to a vessel description and contact information, it usually carries a position good to a few tens of meters, and it happens automatically if the ship sinks fast enough that nobody reaches the radio. None of those properties existed under the old regime, in which the alert went to whoever was listening, the position was whatever the operator could state in the time available, and a vessel that capsized without transmitting simply disappeared.

The cost sits on the ship. The radio officer as a distinct professional aboard merchant vessels largely disappeared, and the radio duties moved to deck officers who hold a General Operator’s Certificate or a Restricted Operator’s Certificate under the STCW convention and who are also navigating, handling cargo documentation and standing bridge watches. Radio is now one competency among many rather than one person’s entire job, and the practical consequence shows up in the details that nobody notices until an emergency, such as an MMSI that was never registered, a GNSS feed to the DSC radio that has been dead for months, or a crew that has never actually sent a DSC test call. Equipment maintenance moved off the ship too, which is why SOLAS had to write the duplication and shore-based maintenance requirement into the rules as an explicit substitute.

There is also a subtler loss in the ship-to-ship layer. A DSC alert produces an alarm on a bridge and a line on a display, and the response depends on a busy officer of the watch understanding what it means and acting on it, whereas a Morse operator listening to a distress call heard the sending fist deteriorate and knew immediately that something was badly wrong. IMO addressed part of this by keeping the channel 16 watch requirement in place rather than retiring it as originally contemplated, which means a human being on every SOLAS ship is still supposed to be listening to a voice frequency at sea.

For anyone whose job touches this, the authorities are clear. SOLAS Chapter IV and the associated IMO performance standards define the system, national administrations define what specific vessels must carry, the FCC’s Part 80 rules and its commercial operator licenses cover the US regulatory side including the GMDSS operator and maintainer licenses, the Coast Guard operates the alerting infrastructure and publishes current guarded frequencies and coverage, and Cospas-Sarsat publishes the beacon and satellite system documentation. Anything you need to rely on professionally should come from one of those, in its current edition.

What to do at your agency

• Have the officer responsible for your marine unit confirm, on each department vessel, that the DSC radio has an MMSI programmed, that the MMSI is registered to a record listing your communications center’s twenty-four hour number rather than a member’s personal phone, and that the radio display is showing a live position from its GNSS input while under way.
• Pull every 406 MHz beacon your agency owns, whether it is an EPIRB on a boat or a personal locator beacon in a technical rescue cache, and record the battery expiration date and the NOAA registration status for each one on the inventory sheet you already maintain.
• Ask your Coast Guard sector liaison for the current list of distress frequencies guarded in your area and the approximate limit of VHF DSC coverage on your water, and file that answer in the marine annex of your communications plan with the date you received it.
• Add one paragraph to your existing marine or water rescue SOP telling telecommunicators exactly what to do when the Coast Guard relays a DSC alert or a beacon hit inside your response area, naming the sector command center contact already in the call-taker reference material.
• Put a ten-minute block into the next scheduled marine operations drill in which crews send a DSC test call to a station that will answer it and read back the received identity, because a test that has never been performed is an assumption rather than a capability.
• Have your radio technician or shop supervisor verify that handheld marine radios carried by crews have not been programmed with a factory default or duplicated MMSI, and correct any that have before the next boating season.

Takeaways

• GMDSS was adopted by IMO as amendments to the SOLAS convention in 1988, phased in from 1 February 1992, and became fully mandatory on 1 February 1999, ending the continuous human Morse watch that had been an international obligation since the years following the Titanic sinking.
• The core design principle is that a distress alert should reach a shore-based rescue coordination center directly and quickly, rather than depending on whichever ship happened to be listening on the right frequency.
• SOLAS Chapter IV states requirements as functions, including the ability to send a ship-to-shore distress alert by at least two separate and independent means, and the specific equipment follows from those functions plus the sea areas the vessel operates in.
• Digital selective calling carries a short data burst on dedicated frequencies, including VHF channel 70 at 156.525 MHz and 2187.5 kHz on MF, containing the vessel’s nine-digit MMSI and, if the equipment is fed a position, coordinates and a time.
• Cospas-Sarsat beacons on 406 MHz provide the independent second alerting path, satellite processing of the old 121.5 and 243 MHz alerting ended on 1 February 2009, and an unregistered beacon deprives rescue authorities of the phone call that resolves most false alerts.
• Sea areas A1 through A4 are defined by declared coast station and satellite coverage rather than by fixed distances, and a modernized Chapter IV that took effect on 1 January 2024 redefines A3 in terms of any recognized mobile satellite service, so older summaries of the sea area rules are now out of date.
• Radar SARTs on 9 GHz and AIS-SARTs solve the separate problem of locating survivors once searchers are in the area, and NAVTEX on 518 and 490 kHz plus satellite group calls deliver maritime safety information without the crew tuning anything.
• The automation traded a dedicated radio professional aboard ship for equipment that alerts without human action, which is why registration records, GNSS connections and maintenance arrangements now carry the reliability that a watchstander used to carry, and why the VHF channel 16 listening watch was kept rather than retired.
• Carriage requirements for any specific vessel come from the flag administration, and in the United States from the Coast Guard and the FCC Part 80 maritime rules, which is where any operational decision should be verified.