Stockholm Agreement

The problem the Stockholm Agreement was written to solve

A ro-ro passenger ship carries its whole payload of cars, trailers, and coaches on one or two long open decks that run most of the length of the hull. That layout loads and discharges fast, which is the commercial point of the type, and it also removes the transverse watertight subdivision that keeps other ships upright after damage. Water that reaches the vehicle deck at any single point does not stay there. It runs the full breadth and length of the deck, forms a wide free surface high in the ship, and cuts the righting lever within minutes. Two ferry losses in seven years, the Herald of Free Enterprise in 1987 and the Estonia in 1994, showed that the global damage stability rules of the day did not account for this behavior. The Stockholm Agreement is the regional standard that eight north-west European administrations wrote to close that gap.

The gap was specific and quantifiable. The SOLAS 90 deterministic damage stability standard, in force from the early 1990s, requires a ship to retain a defined residual righting-lever curve after an assumed side collision. It says nothing directly about water that then floods the vehicle deck from a breached bow door, a failed ramp seal, or waves washing in over a low damaged freeboard. Analysis after the Estonia loss put a number on the omission: SOLAS 90 implicitly covers the effect of water on the ro-ro deck only for a sea state of roughly 1.5 m significant wave height. Ferries on the Baltic, the North Sea, and the western approaches routinely work in far worse. The Stockholm Agreement raised the assumed sea state and forced the flooded deck into the calculation.

The Herald of Free Enterprise, 6 March 1987

The Herald of Free Enterprise was a Townsend Thoresen roll-on roll-off ferry built for the short Dover to Calais crossing and working that night from Zeebrugge. She left the inner harbor at 18:05 GMT with a crew of 80, 459 passengers, 81 cars, three buses, and 47 trucks. She passed the outer mole at 18:24 and capsized about four minutes later. The assistant boatswain whose job was to close the bow doors was asleep in his cabin; the doors were still open as she left the berth. Seawater poured onto the vehicle deck as the ship built up speed, the free surface rolled her to port, and she came to rest on her side on a sandbank in shallow water. There was no time to send a distress signal, launch boats, or hand out lifejackets. 193 passengers and crew died.

The formal investigation under Mr Justice Sheen, the Admiralty judge, ran the following months and reached past the sleeping crewman. It found the capsize the joint fault of the master, the chief officer, the assistant boatswain, and Townsend Car Ferries at every level of management, and it named a “disease of sloppiness” running through the company. Two design facts mattered for what came later. The master could not see the bow doors from the wheelhouse and had no indicator light to tell him whether they were shut. And the vehicle deck was one contiguous space, so once water was on it, nothing stopped it spreading. A sister ship, the Pride of Free Enterprise, had sailed from Dover with her bow doors open in October 1983 after the same crew failure, and had not capsized; the industry had read that near miss as proof the hazard was survivable.

Herald forced immediate changes. Bow-door indicator lights on the bridge, monitoring of the vehicle deck, and restrictions on undivided decks followed, and the disaster led directly to the creation of the UK Marine Accident Investigation Branch. What it did not immediately produce was a damage stability rule that assumed water would be on the deck. That step waited for a second, larger loss.

The loss of the Estonia, 28 September 1994

The ship and the voyage

The Estonia was built in 1980 at Meyer Werft in Papenburg, launched as Viking Sally, at about 15,600 gross tons. She passed through the names Silja Star and Wasa King before a Swedish-Estonian venture bought her in 1993 and put her on the overnight Tallinn to Stockholm route under the Estonian flag, then the largest ferry serving Estonia. On the night of 27 to 28 September 1994 she ran into a storm in the northern Baltic, with a significant wave height of about 4 m and a strong gale on the bow.

The bow visor failure and the capsize

Shortly after 01:00 the crew heard heavy metallic noises from the bow. The bow visor, the large clamshell structure that hinged up to expose the loading ramp, was working loose under repeated wave impact. Its locking devices and hinges failed, the visor tore away, and as it went it dragged the ramp behind it partly open. The northern Baltic began flooding straight onto the vehicle deck. A mayday went out at 01:22 but did not follow the standard format, which slowed the wider response. The free surface on the car deck rolled the ship hard to starboard within minutes. She lost power, drifted beam-on to the sea, and by about 01:50 had capsized and gone off the radar screens, sinking south of the Finnish island of Uto. Of the 989 people aboard, 137 were rescued. 852 died, most within the first hour, many trapped below decks or lost to the cold Baltic water before help arrived. It stands among the deadliest peacetime maritime disasters in European waters.

What the Joint Accident Investigation Commission found

The Joint Accident Investigation Commission of Estonia, Finland, and Sweden reported in 1997. The visor was found on the seabed on 18 October 1994, raised on 18 November, and taken ashore at Hanko in Finland. The commission’s central conclusion was that the visor locking system failed under wave-induced loads it was never designed to take, the visor separated, and the ramp opened enough to admit water to the deck. It found the visor had been under-designed at build, because the yard and the approval process had not treated the visor and its attachments as safety-critical. It criticized the slow crew reaction, the failure to notice deck flooding early, and the thin guidance from the bridge. A survey of the bow structure that regulations should have required, and an exemption that should have been recorded on the certificates, were both missing; had they been in place, the commission judged, the ship would not have been on that route. A later Swedish-led examination of the ramp raised in 2023 and updated structural modeling matched the 1997 findings on the sequence of failure.

The commission’s wider lesson was structural, not personal. A ro-ro passenger ship built to the damage stability of the day had a standing vulnerability to vehicle-deck flooding from the bow, and the sea states the north-west European fleet actually met were beyond what that standard assumed. Fixing that needed a tougher rule, not just better seamanship.

The regulatory response: the SOLAS 95 Conference and Resolution 14

The IMO route to a new global standard was slow. Amending SOLAS across all member states and all ship types takes years, and the ferry states around the Baltic and the North Sea wanted a firmer standard sooner and pitched to their own waters. The 1995 SOLAS Conference in London gave them the mechanism. Among its resolutions, Conference Resolution 14, “Regional agreements on specific stability requirements for ro-ro passenger ships,” adopted on 29 November 1995, let contracting governments agree specific stability requirements for ro-ro passenger ships in a designated area, where they judged that the sea conditions and other local conditions there called for a standard above the SOLAS baseline. Resolution 14 also carried the technical method, including the water-on-deck approach and a model-test alternative, that the regional agreement would adopt.

Resolution 14 is the genuine 1995 precursor to the Stockholm Agreement, and it is a precise instrument, not a general aspiration. It set the legal container: a group of neighboring states could bind ferries on their shared routes to a higher standard, whatever flag those ferries flew, without waiting for the whole IMO membership to agree. That combination, a regional reach plus flag-blind application, is what made the standard workable, because a Baltic ferry route is often served by ships of several flags and a rule that bound only one flag would have been easy to sidestep.

The Stockholm Agreement of 1996

Adoption and legal character

Eight north-west European states concluded the agreement at Stockholm on 28 February 1996, its full name the Agreement Concerning Specific Stability Requirements for Ro-Ro Passenger Ships Undertaking Regular Scheduled International Voyages Between or To or From Designated Ports in North West Europe and the Baltic Sea. Sweden notified it to the IMO on 1 April 1996 under operative paragraph 3 of Resolution 14. It opened for signature at IMO headquarters from 1 July to 30 September 1996 and stayed open for accession after that. It entered into force on 1 April 1997.

Its legal character is worth stating plainly, because the name invites confusion. The Stockholm Agreement is a regional agreement among administrations made under an IMO resolution; it is not itself a SOLAS chapter or an IMO convention. It works by each party applying the standard to the ferries on the covered routes and by the parties recognizing each other’s certificates. That structure let the standard bite quickly on the specific fleet that needed it, without the delay of a global convention amendment.

Scope and the participating administrations

The eight signatory states were Denmark, Finland, Germany, Ireland, the Netherlands, Norway, Sweden, and the United Kingdom. Seven were EU members at the time; Norway was the exception. The covered geography runs across the waters those states share: the English Channel and its western approaches, the Irish Sea, the North Sea, the Skagerrak and Kattegat, the Baltic, and the Norwegian coast. These are the routes that see the Atlantic, North Sea, and Baltic weather the agreement was calibrated against.

The agreement reaches a defined set of ships and voyages, and it is worth being exact about the edges:

• It covers ro-ro passenger ships, as defined in SOLAS Chapter II-1, that carry more than 12 passengers.
• The ship must be on regular scheduled international voyages between, to, or from designated ports of the participating administrations.
• It applies regardless of the ship’s flag, so a ship registered outside the group still meets the standard to work a covered route.

It does not reach pure cargo ro-ro ships with no passengers, cruise ships and other conventional passenger ships without ro-ro spaces, or ferries on ad hoc and one-off charter voyages rather than a scheduled service. Domestic-only ferries sat outside the original agreement, though EU law later drew many of them in by sea area.

The water-on-deck criterion

The core rule

The one idea that defines the Stockholm Agreement is that a damaged ro-ro passenger ship must survive with a body of seawater assumed to be sitting on its vehicle deck, on top of the flooding that SOLAS 90 already assumes inside the hull. The reasoning is direct. After a collision that opens the side, a ro-ro ship with a low damaged freeboard will ship water onto the vehicle deck from waves washing over the damaged edge, and that water then behaves as a free surface across the whole deck. The criterion forces the naval architect to prove the ship keeps a positive righting-lever curve with that deck water present, not just in still water.

Residual freeboard and the 0.5 m height

The height of water assumed on the deck depends first on the residual freeboard, fr, the least distance between the damaged ro-ro deck and the final waterline at the damage location, measured before the deck water itself is added. Low residual freeboard means the deck sits close to the sea, so more water gathers on it. The agreement fixes the relationship between the two bounds and interpolates linearly between them:

So a ferry whose residual freeboard after the design damage is 0.3 m or less must survive 0.5 m, half a meter, of seawater lying across the vehicle deck. Where the residual freeboard reaches 2.0 m or more, no water on deck is assumed, because the deck is judged high enough that waves will not fill it. Between those points the height falls off in a straight line. The height is measured from the lowest point of the flooded deck once trim and heel are accounted for, so a ship that lists after damage sees more water on the low side.

The significant wave height band

The second control is the sea state of the route, expressed as the significant wave height, hs, the mean height of the highest third of the waves, taken as the value not exceeded with more than 10 percent probability over a year. A ship that only ever works sheltered water should not carry the same deck-water penalty as one crossing the open North Sea. The agreement scales the height accordingly, again by linear interpolation between fixed bounds:

At or below 1.5 m significant wave height, no extra water on deck is assumed, which lines up with the sea state that SOLAS 90 already covers implicitly. At or above 4.0 m the full freeboard-based height applies. Between 1.5 and 4.0 m the amount rises in a straight line. The 4.0 m ceiling reflects the working sea states of the north-west European and Baltic routes. This banding is what ties a ship’s certification to a trading area: a ferry proven for a 4.0 m band can work the open routes, while one built to a 2.5 m band is limited to milder areas, and that limit is recorded on its stability documentation.

How the water enters the calculation

Once the height is set, the assumed water is added to the damage stability model as a body of free water on the vehicle deck, with its full free-surface effect included in the righting-lever calculation. The ship must then still meet the survival criteria for the residual righting-lever curve, meaning a positive range, a minimum area under the curve, and a minimum peak lever, with that deck water present in the worst of the damage cases considered. The damage extents themselves stay consistent with the SOLAS Chapter II-1 deterministic assumptions for a side collision. In practice, satisfying the criterion tends to call for higher freeboard at the damaged condition, more subdivision of the vehicle deck, and sometimes added buoyancy in side casings or sponsons.

The model-test alternative

Not every ship that fails the calculation is unsafe, and the agreement recognizes this by allowing a scale-model test in place of the water-on-deck computation. The method sits in the appendix to the agreement’s annex. A model of the damaged ship, ballasted to the worst damage condition, is placed beam-on to long-crested irregular waves generated to a defined spectrum at the significant wave height of the intended area, and run for a period equivalent to 30 minutes at full scale across several wave realizations. The ship passes if the model does not capsize in any run. The test captures dynamic effects the static calculation cannot: how water floods onto and drains off the deck as the model rolls, how the hull form and any deck camber shed water, and how the ship behaves in a real seaway rather than a flat one.

Model testing gives credit that the deterministic calculation withholds, and it has let ships trade that the pure calculation would have excluded, because a hull that sheds deck water quickly can survive a sea state that the frozen 0.5 m assumption would fail. It is a specialist and costly route, run at a hydrodynamics tank, so it is used where a valuable ship sits just the wrong side of the calculated line and the tank time is cheaper than the steelwork a rebuild would need. The results are ship-specific and tied to the tested damage condition and sea state, so they do not transfer to a sister ship on a different route without fresh work.

Why a shallow layer of water capsizes a ro-ro ship

The water-on-deck criterion is small in absolute terms, half a meter at most, and the reason that half meter matters is the free-surface effect. When water lies free across a wide deck, it runs to the low side as the ship heels, moving its center of gravity outboard and adding to the heeling moment rather than resisting it. The loss of stability scales with the transverse moment of inertia of the free surface, which for a rectangular deck grows with the cube of the breadth. A ro-ro vehicle deck is close to the full breadth of the ship and runs much of its length, so the free surface of even a thin layer produces a large virtual rise in the center of gravity and a matching cut in the metacentric height. This is the same free surface effect that governs slack tanks, magnified because the “tank” is the width of the ship. It is why a car deck with a shallow, wide sheet of water can be more dangerous than a deep, narrow flooded compartment holding the same tonnage of water, and why the Herald and the Estonia both rolled over within minutes rather than settling upright.

SOLAS 90 as the baseline and what Stockholm adds

The Stockholm Agreement does not replace the SOLAS 90 deterministic damage stability standard; it sits on top of it. SOLAS 90 sets the assumed collision damage and the residual righting-lever curve the ship must keep after that damage in still water. The name comes from the 1990 entry into force of the strengthened passenger-ship damage stability rules in SOLAS Chapter II-1. In outline, after the assumed side damage the ship must reach an equilibrium within set heel limits, then show a residual righting-lever curve with a range of at least 15 degrees beyond the angle of equilibrium, a maximum residual lever of at least 0.10 m within that range, and an area under the curve of at least 0.015 metre-radians, with the residual lever taken as the curve less the greatest of the heeling moments from passenger crowding, launching survival craft, or wind. The exact figures and the flooding assumptions are set in SOLAS Chapter II-1, and this deterministic pass-or-fail is the historical baseline against which the Stockholm Agreement was pitched.

The Stockholm Agreement takes those same damage cases and asks the further question of survival with water on the deck. A ferry certified for a covered route therefore carries two linked damage stability results in its approval package: the SOLAS 90 case, and the Stockholm water-on-deck case for its assigned wave-height band. The second is almost always the harder to meet, which is why it drives the design. A useful shorthand within the trade is that SOLAS 90 asks whether the ship survives the hole in her side, while the Stockholm Agreement asks whether she survives the hole in her side plus the sea that then washes onto her car deck.

A qualitative worked case

Take a ferry whose worst damage case leaves a residual freeboard of about 1.0 m and which is certified for a 4.0 m significant wave height route. The freeboard rule interpolates between 0.5 m of water at 0.3 m freeboard and zero at 2.0 m, so at 1.0 m freeboard the assumed water height is $0.5 \times (2.0 - 1.0)/1.7 \approx 0.29$ m. Because the route is a 4.0 m band, no wave-height reduction applies and the full 0.29 m stands. The naval architect must then run the damage case again with roughly 0.29 m of free water spread across the flooded portion of the vehicle deck and confirm the residual righting-lever curve still clears the survival criteria. If it does not, the fix is more residual freeboard, more deck subdivision to shorten the free surface, or added buoyancy, until it does. The same ship assigned instead to a 2.5 m band would see the water height scaled by $(2.5 - 1.5)/2.5 = 0.4$, giving about 0.12 m, an easier case that also limits the ship to the milder area.

EU Directive 2003/25/EC

The Stockholm Agreement bound only its eight parties and their routes, which left a patchwork: a ferry to a Spanish, Italian, or Greek port did not face the standard, even though the physics of a flooded ro-ro deck do not stop at the Skagerrak. The European Union closed that gap with Directive 2003/25/EC of 14 April 2003 on specific stability requirements for ro-ro passenger ships. The directive took the Stockholm water-on-deck method, the residual-freeboard rule and the significant-wave-height banding alike, and made it EU law.

The directive’s reach is wider than the agreement’s in two ways. It applies to every ro-ro passenger ship on regular service to or from a port of an EU member state on an international voyage, whatever flag it flies, so it covers Mediterranean and southern routes that the regional agreement never touched. And it extends the standard to certain ro-ro passenger ships on domestic voyages, keyed to the significant wave height of the sea area they work, so that a domestic ferry in a rough area faces the same deck-water check as an international one. Enforcement runs through the member states’ flag and port state control. Directive 2003/25/EC, rather than the original agreement, is now the instrument under which the water-on-deck standard is applied across most of the covered fleet, and it carries the same 0.5 m ceiling, the same 0.3 m and 2.0 m freeboard bounds, and the same 1.5 m to 4.0 m wave band in its Annex I.

Implementation

Design consequences for ferries

A designer building a ro-ro passenger ship for covered service works the water-on-deck case from the first weight-and-buoyancy study, because it shapes the hull. The common answers are a higher freeboard at the damaged condition than SOLAS 90 alone would need, transverse and longitudinal subdivision that breaks the vehicle deck into shorter watertight sections up to a defined height, stronger bow and stern doors with redundant locking and bridge monitoring, deck drainage that clears water before it builds, and, where the numbers are tight, added buoyancy in the form of side casings or sponsons that lift the righting lever in the damaged condition. Each of these buys back part of the margin the assumed deck water takes away.

Intact stability and the loaded condition

Damage stability is only half the picture, because a ship that meets the water-on-deck case on paper still depends on the master loading her to the condition the calculation assumed. The Stockholm result is worked for defined loading conditions, and each carries a required metacentric height and a limiting position of the center of gravity that the ship must be kept within in service. That link runs back to intact stability : the damaged righting-lever curve starts from the intact condition, so a ship trimmed by the head, loaded high with light freight on the upper vehicle deck, or run with slack ballast tanks reaches the damage case with less margin than the approval assumed. The loading computer on board checks each sailing condition against the limiting curve, and the stability booklet records the conditions the ship was approved for.

This is why the Estonia and Herald losses read as system failures rather than isolated equipment faults. On both ships the watertight boundary that the stability case assumed, the closed bow, was not there when it mattered: the Herald sailed with her doors open, and the Estonia’s visor tore off. No damage stability standard survives an assumption that the shell is intact when it is not, which is why the Stockholm Agreement travels with door design, monitoring, and watertight-integrity discipline rather than standing alone. The naval architect provides the margin; the operator has to keep the conditions that make the margin real, sailing after sailing, in the weather the wave-height band assumed.

Classification societies and flag states

The classification society that approves a ro-ro passenger ship for covered service reviews the Stockholm water-on-deck calculation alongside the SOLAS 90 case, the vehicle-deck subdivision arrangement, and the bow, stern, and ramp designs, then the flag state or its recognized organization confirms compliance before the ship enters service. The stability documentation carries the assigned significant wave height band, which fixes the areas the ship may work. A change of route to a rougher band forces a fresh assessment, because a ship proven for 2.5 m is not proven for 4.0 m.

Existing ships, phase-in, and retrofit

The agreement reached the ships already trading, not just newbuildings. Existing ferries on covered routes were phased in on a schedule tied to how far each already met the standard: the ships that fell furthest short of the water-on-deck survival line had to comply earliest, and those close to it were given longer. The staged approach was deliberate risk-ranking: a ship’s compliance date was set by a value measuring how close it already came to the water-on-deck survival standard, so the least survivable ships in the fleet were fixed or removed first and the marginal ones last. Owners met the deadlines in different ways. Many rebuilt, adding subdivision to the vehicle deck, reworking bow and stern door arrangements, and adjusting ballast and buoyancy, work often combined with a scheduled dry-docking to hold down the off-hire cost. Some ships were withdrawn from the region rather than converted, because the rebuild cost more than the remaining hull was worth, and moved to routes outside the covered area where the standard did not apply. That migration of older, weaker tonnage to less-regulated waters is itself a limitation of a regional rule, and it is examined below.

Later IMO work and the relationship to SOLAS 2009

The IMO kept moving toward a tougher global standard, and the relationship between the regional rule and the global one has three later milestones. First, the harmonized probabilistic subdivision under Resolution MSC.216(82) entered force on 1 January 2009 and changed the whole method of damage stability assessment. It replaced the deterministic pass-or-fail of SOLAS 90 with an attained subdivision index A, the sum over flooding cases of the probability of each case times the probability of surviving it, which must reach a required index R that grows with ship length. This is the probabilistic damage stability framework of the current SOLAS Chapter II-1 .

Second, the Safe Return to Port rules brought passenger ships whose keels were laid on or after 1 July 2010 under a philosophy that the ship itself is its own best lifeboat: a large passenger ship, generally 120 m or longer or with three or more main vertical zones, must keep defined systems working and return to port after a casualty below a set threshold. Third, the required subdivision index R for passenger ships was raised by amendments in force on 1 January 2020 under Resolution MSC.421(98).

That 2020 increase is the clearest line from the Stockholm Agreement into current global rules, because it rests on research aimed squarely at the ro-ro passenger ship survivability problem the agreement first named. A run of EU and EMSA-funded projects, among them GOALDS on goal-based damage stability, FLOODSTAND on flooding progression, and the later eSAFE work on survivability after a flooding event, rebuilt the probability-of-survival factor for passenger ships from ship-model and full-scale flooding data. EMSA’s own study on the specific damage stability parameters of ro-ro passenger ships fed the case that the probabilistic s-factor understated how quickly these ships lose stability once the deck floods, and the raised R for passenger ships was the regulatory result. The probabilistic method reached, by a different road, the same conclusion the water-on-deck rule had reached deterministically two decades earlier: a ro-ro passenger ship needs more survival margin than a general damage stability standard gives it.

None of these repealed the Stockholm Agreement. The regional water-on-deck requirement, kept in force by Directive 2003/25/EC, continues to apply to covered ferries in parallel with whichever SOLAS regime a given ship was built to. A modern ro-ro passenger ship built to the 2020 probabilistic standard is a stronger ship than the water-on-deck rule alone would produce, and the practical gap between the global standard and the regional one has narrowed with each amendment, but the two remain separate legal requirements rather than one having swallowed the other.

Operational and commercial consequences

Compliance is a standing discipline, not a one-time approval. Class certificates must keep reflecting the current condition, damage stability is re-verified when the operating profile changes, and the route, speed, and weather limits tied to the assigned wave-height band are documented and enforced on board. Crews drill the damage scenarios the standard assumes, and vehicle-deck monitoring and bow-door inspection are routine, sharpened by the two disasters that produced the rule. Compliance also sits inside the commercial arrangements around a covered ferry: it is a baseline the north-west European ferry resale market expects, and a ship that lacks it is limited to markets outside the covered area, where it usually fetches a lower price.

Ferry losses outside the covered area

The Stockholm standard is regional by design, and its limits show in casualties beyond the covered waters, where comparable water-on-deck rules did not apply. The clearest is the loss of the Sewol off South Korea on 16 April 2014, in which 304 people died, most of them secondary-school students, after a ro-ro passenger ferry with an altered loading condition and poorly secured cargo took a hard turn, heeled, and could not recover. The Sewol is a different failure mode from the Estonia, cargo shift and stability mismanagement rather than a bow-visor failure, but it makes the same underlying point: a ro-ro passenger ship pushed past its stability limits capsizes fast and kills at scale. Since the Stockholm Agreement took effect, no large north-west European ro-ro passenger ferry has been lost with an Estonia-scale death toll, which is the outcome the standard was written to prevent. That record is not proof the rule alone is responsible, since door design, monitoring, weather routing, and crew training all improved over the same years, but the water-on-deck margin removed the specific failure path that took the Estonia.

Limitations

The Stockholm Agreement is a specific instrument with specific edges, and a practitioner should hold several caveats in view.

It is regional. Its direct legal force runs only to its parties and their designated routes, and its EU extension runs only to EU ports and EU sea areas. A ferry that leaves the covered area for the Mediterranean beyond EU reach, or for routes in south and east Asia, escapes the standard, and ships have moved for exactly that reason. The rule protects a fleet, not a ship type everywhere.

It addresses one failure mode. The water-on-deck criterion targets water accumulating on the vehicle deck after collision damage in a seaway. It is not a general answer to every ro-ro casualty. It does not by itself prevent the cargo-shift and loading-condition failures that sank the Sewol, nor fire, nor grounding, and it assumes the bow and stern doors hold, which is precisely what failed on both the Herald and the Estonia. Watertight-integrity discipline, door design, and deck monitoring sit alongside the stability rule, not inside it.

The wave-height banding is a statistical assumption. The significant wave height that defines an area is the value not exceeded with more than 10 percent yearly probability, so a ship certified for a band can still meet worse than its design sea state on a bad day, and a shift in the storm climate of a route can change the band a ship needs. The banding is an engineering simplification of a variable sea, not a guarantee for any single voyage.

Finally, the numbers here are the standard’s stated design assumptions, not a substitute for the approved stability documentation of a particular ship. The residual freeboard, the assigned wave-height band, and the resulting water-on-deck height for a given ferry come from its own class-approved damage stability calculation or model test. Any real compliance question is answered from that ship’s papers and the current text of Directive 2003/25/EC and SOLAS Chapter II-1, read with the flag administration.

Related Articles

• MS Estonia 1994 disaster
• Ro-ro vessel
• Passenger ship
• Damage stability
• Probabilistic damage stability
• SOLAS Chapter II-1: construction, subdivision, and stability
• Subdivision and floodable length
• Intact stability
• Free surface effect
• Freeboard and reserve buoyancy
• GZ curve and righting arm
• Cruise and passenger operations
• SOLAS Convention
• Classification society