The Scandinavian Star Incident

A Case Study

by Alan Robinson, BSc, MSc, PhD - Partner, Dr J.H. Burgoyne & Partners

At 21.45 hours on 6 April 1990, the motorised ferry Scandinavian Star left Oslo bound for Frederikshaven. Several hours later while the ship was crossing to Denmark a fire broke out that was to claim the lives of 158 people on-board. This paper is about the reasons for that tragedy.

1. The Ship and Her History

Deck 1 Engine room and tanks
Deck 2 Engine room, other machinery spaces refrigerated stores and crew cabins
Decks 3&4 Car deck with passenger cabins outboard
Deck 5 Passenger cabins
Deck 6 Lounge, restaurant and shops
Deck 7 Shops, officers cabins, embarkation deck for the lifeboats
Deck 8 Wheel house, officers cabins and disco
Deck 9 Open air deck

The ship was built to SOLAS 1960 requirements and Method I was used with respect to fire protection. This means that fire resistant materials were used for internal bulkheads. For the accommodation decks, the bulkheads were constructed of 30mm asbestos silicate covered generally with a 1.5mm layer of laminated plastic, although in some areas there were 2 layers of plastic covering. The ceilings were formed of 10mm asbestos silicate, again generally covered with 1.5 mm layer of laminated plastic but for 90 cabins amidship on deck 5 the asbestos silicate ceiling was covered with 4mm PVC. The "A" class bulkheads and deck insulation consisted of 25mm rockwool and the cabins were fitted with B-15 fire doors.

The ship was divided into three main vertical fire zones which were numbered 1 to 3 from aft to forward. On Decks 3-5, there were four stairways to port and four to starboard. They were numbered from aft and stairways 3P&S and 4P&S extended down to Deck 2 and 3P&S also to deck 1. There were some internal stairways, also on the upper decks, but they were generally not in the same positions as those on the lower decks. The doors to stairways were usually fitted with A-60 self closing fire doors.

As Method I was used, the ship generally was not fitted with either an automatic fire detection system or an automatic fire fighting system, although some spaces, such as the engine room, were fitted with this type of equipment. In the event of a fire, the alarm was sounded manually by pressing activation buttons on the Bridge.

2. Events Prior to Disaster Voyage

Until 1987 the ship had been classed with Bureau Veritas but thereafter it was classed with Lloyds Register and carried the Bahamian flag.

The ship had previously been operating out of Miami on short cruises but was purchased in March 1990 by the V R Dano Group to replace the Holger Danske on the run between Frederikshaven in Denmark and Oslo, Norway. The ship was quickly brought into service by the new owners, making its first run on the new service on 1 April. Only nine original members of the crew remained with the ship and they comprised mainly engineering crew including the chief engineer. The rest of the crew were either previous members of the crew of Holger Danske, consisting of deck officers and catering officers, or they were recruited new to the ship. This latter group were mainly "hotel staff' of Portuguese nationality.

3. The Disaster Voyage

The Scandinavian Star came into service on the new run on 1 April. It appears to have operated without incident until the tragic voyage on 6 April. The ship left Oslo at 21.45 hours that day under the command of Captain Hugo Larsen and with a crew of 99 and 383 passengers.

Between 01.45 hours and 02.00 hours on 7 April, a small fire was discovered in a pile of bed clothes outside Cabin No.416 on the port side of Deck 4. This fire was quickly extinguished. However, a little after 02.00 hours, a second fire started in the aft section of the starboard corridor of Deck 3 near to staircase 2S. The accommodation on this deck was not in use. The fire was not extinguished quickly and at 02.24 hours the ship sent a mayday message giving its position. Subsequently at 03.20 hours, the captain decided that it was not possible to extinguish the fire and the decision was taken to abandon ship.

I first became aware of the incident at about 08.00 hours on the morning of 7 April when I heard about the fire on the BBC news. The report said at that time that four people were believed to have died in the incident. Shortly afterwards I received instructions to investigate the cause of the fire on behalf of one of the parties and I left later that morning for Sweden.

At 11.55 hours the ship was taken under tow to the small town of Lysekil in Sweden where she arrived at 21.17 hours. I also arrived there shortly afterwards. At this time there was a small amount of smoke coming from the ship and externally there appeared little evidence of a major fire. However, during the night whilst the fire brigade was trying to extinguish the fire, the fire developed and spread significantly causing extensive damage to most decks.

The fire was eventually extinguished at 16.00 hours on Sunday 8 April. During the next week, the emergency services were employed in removing the bodies from the ship. It was eventually found that 158 people had died in the tragedy; 156 were passengers and two were crew. Four bodies had been carried out onto the after part of Deck 5 prior to the ship being abandoned. All except one of the bodies that were recovered from inside the ship in Lysekil were found in the passenger accommodation Decks 4 and 5. The one exception was recovered from the restaurant on Deck 6. The bodies found on Decks 4 and 5 were recovered from the following areas:

All the deceased died as a direct result of the fire. For 125, the inhalation of carbon monoxide was probably the main cause of death. Heat injuries accounted for another 10 deaths, and hydrogen cyanide poisoning was responsible for many of the rest. The fire caused extensive damage to the after parts of Decks 3, 4 and 5 and to the whole of Decks 6 and 7.

4. Investigations into the Reasons for the Tragedy

Following the tragedy, Sweden, Denmark and Norway agreed to set up a committee to investigate the reasons for the tragedy. As well as taking evidence from survivors, the ship was also inspected fully. In addition, calculations and fire tests were later carried out at the National Institute for Testing and Verification (Dantest) in Denmark and at SINTEF NBL, the Norwegian Fire Research Laboratory. Much of the rest of this paper describes the findings and recommendations of the Committee of Inquiry.

Both the earlier fire that was extinguished and the one that led to the tragedy were almost certainly started deliberately by the application of a naked flame to bedclothes in the first instance and probably paper and bedding that had been placed at the site on the second occasion. No one has been charged with starting the fires.

1. The fire was ignited shortly after 02.00 hours. Between 2 - 8 minutes later, the heat release from the fire reached 200 kW which was enough to start the corridor wall burning rapidly.
2. Fire spread rapidly to staircase 2S and then upwards
3. Smoke reached Deck 4 about 1 minute after ignition and was drawn into the corridors fore and aft of the staircase. The fire door forward of the staircase remained open.
4. Smoke reached Deck 5 after 2-3 minutes and began seeping into adjoining corridors.
5. The fire spread to the port side along the transverse corridor on Deck 5.
6. On the port side the fire was drawn down through the 2P stairway.
7. Smoke was drawn into the port corridors on Deck 4 and to a lesser extent on Deck 5
8. The fire spread down to deck 3 where the fire door onto the car deck remained open.
9. The fire spread into the restaurant on Deck 6 through an open fire door at the top of the 2S staircase.

Thus within 8 to 12 minutes of the fire starting, most of the corridors where the people died were filled with smoke. While the ventilation was running, this maintained a positive pressure in the cabins keeping the smoke out, but when the ventilation was switched off possibly as late as 02.30 hours, smoke seeped into the cabins. It is considered that all 158 people who lost their lives in the tragedy had probably died by 02.45 hours.

a). SOLAS Requirements
The Committee found that in principle the Scandinavian Star complied with those requirements in SOLAS 1960 and SOLAS 1974 that a ship built in 1971 was supposed to comply with. However they did find the following deficiencies in the ship and its equipment:

• workshops and stores had been set up on the car deck
• some of the sprinkler heads on the car deck were blocked with rust
• pressure bottles were stored incorrectly
• there was a defective fire door on the port side of the car deck
• the motorised lifeboats were generally in poor repair
• a fire door was missing from the aft starboard part of Deck 6 and the door opening had been fitted with a glazed door
• three alarm bells were missing from the fire alarm system

As a result of the missing alarm bells it was found that only in about 37% of the cabins was the sound level of the alarm over 68 dB, which was considered to be "probably sufficient". In addition, as buttons had to be held down on the Bridge to maintain the sounding of the alarm, the alarm was not sounded for long enough periods

The carpets and cabin furniture were not considered to have been particularly significant in the development of the fire. However the laminated plastic coating on the walls and ceilings of corridors, although only about 1.5mm thick, was significant. Subsequent tests showed that the coating had a calorific value of 48 MJ/m2. SOLAS 1960 did not specify a maximum calorific value for such coating and the material was therefore acceptable. Indeed it is only 3 MJ over the maximum acceptable amount under SOLAS 1974. Nevertheless the material provided an uninterrupted surface in corridors and stairways that greatly assisted the spread of fire. In addition it was also found that the material, when it burned, produced large quantities of carbon monoxide and hydrogen cyanide, both of which were found to have been responsible for causing many of the deaths.

d). Fire doors
The fire doors in general were fitted with magnetic catches and could be closed either locally or from the Bridge. Although most of the doors were eventually closed some in the areas affected by fire, remained open. In particular, as no alarm was ever given from the zone on Deck 3 where the fire started, because no one was there to press the alarm button, the fire door from the zone to stairway 2S was never closed. This allowed the fire to spread to the staircase and hence to other decks. Other doors were also left open. The fact that some doors remained open while others were closed also created draughts which assisted the rapid spread of fire.

e). The ventilation system
The ventilation system aboard the Scandinavian Star may not have been stopped until 02.30 hours. While it was operating it did prevent the spread of smoke into cabins. However during the initial stages of the fire it also played a part in determining the route by which the fire spread, although, as the heat output from the fire increased, the buoyance of the hot combustion gases became the more dominant factor.

f). The escape routes
Many of the escape routes soon filled with smoke and this affected the evacuation of the accommodation. In addition, the routes involved changes of direction, corridors with dead ends and staircases that were not continuous. An example of the problem that this caused was the aft escape from the starboard corridor on Deck 5. The escape was not at the end of this corridor but about three metres forward set in the outboard bulkhead. In fact there was a door at the end of the corridor but this led only to a small storage cupboard. Some 13 bodies were found at the end of the corridor.

The layout of some of the escape routes meant that passengers unfamiliar with the ship needed the assistance of crew and signposts to find their way quickly. Following the change of ownership, the ship had been put into service without posting emergency notices in a Scandinavian language even though the ship was operating between two Scandinavian ports. In addition as passengers were not issued with boarding cards, they were unable to follow the colour coding system used to direct them to their allocated muster station. This led to an uneven distribution between the different muster stations. The assistance offered by the crew is considered in the next section.

g). The manning of the ship and the action of the crew
The ship was not under-manned and the officers possessed the necessary qualifications and certificates but the Committee found that the navigation officers should have had better training in safety matters. It also found that there was a language problem in that many of the Portuguese had little or no knowledge of English. However the most serious criticism made of the crew is that they never acted as an organised unit and that no real attempt was made to fight the fire. Furthermore it was found that the alarm was only sounded for a short period of time and that there was no organised waking of sleepers.

5. Overall Conclusion and Committee Recommendations

As a result of the previous factors the overall conclusion of the committee was that the ship was not ready to sail with passengers when it was brought into service and that it had been brought into service far too rapidly. The committee also made the following recommendations that all ships in passenger traffic to Scandinavian ports:

• should be fitted with sprinkler systems
• should be fitted with smoke detectors in corridors, stairways, saloons and cabins. The smoke detectors should be connected to indicators on the Bridge and be installed in sufficient numbers and arranged in such a way as to detect smoke as soon as possible and provide adequate indication of the spread of smoke.
• should be manned with a crew which has attended courses in safety procedures approved by the maritime administrations.
• should be inspected before coming in to service and then they should be subjected to further periodic scheduled and unscheduled inspections

The Committee also recommended that regulations were laid down governing the duty of ship owners to establish systems for the safe operation of ships.

6. Conclusion

The tragedy of the Scandinavian Star again illustrates just how important it is to detect a fire quickly, to start fighting it immediately and implement properly organised evacuation procedures supervised by properly trained people.