MARINE SAFETY INVESTIGATION

REPORT 183

Independent investigation into the fire in the hold of

the Netherlands flag general cargo vessel

Marion Green

off the coast of Western Australia

on 28 July 2002

AU S T R A L I A N G OV E R N M E N T – D E PA R T M E N T O F T R A N S P O R T A N D R E G I O N A L S E RV I C E S
 Department of Transport and Regional Services
Australian Transport Safety Bureau

Navigation Act 1912

Navigation (Marine Casualty) Regulations
investigation into the fire in the hold of the Netherlands flag general cargo vessel
Marion Green
off the coast of Western Australia on
28 July 2002

Report No 183
November 2003
ISSN 1447-087X
ISBN 1 877071 42 0

Readers are advised that the Australian Transport Safety Bureau investigates for the sole purpose of
enhancing transport safety. Consequently, Bureau reports are confined to matters of safety significance
and may be misleading if used for other purposes.

Investigations commenced on or before 30 June 2003, including the publication of reports as a result
of those investigations, are authorised by the Executive Director of the Bureau in accordance with the
Navigation (Marine Casualty) Regulations 1990, made pursuant to subsections 425(1)(ea) and 425
(1AAA) of the Navigation Act 1912.

Investigations commenced on or after 1 July 2003, including the publication of reports as a result of
those investigations, are authorised by the Executive Director of the Bureau in accordance with the
Transport Safety Investigation Act 2003 (TSI Act). Reports released under the TSI Act are not
admissible as evidence in any civil or criminal proceedings.

It is ATSB policy to publish and widely distribute in full such reports as an educational tool to
increase awareness of the causes of marine accidents so as to improve safety at sea and enhance the
protection of the marine environment. Reports on serious marine casualties are also provided to the
IMO.

Australian Transport Safety Bureau

PO Box 967
Civic Square ACT 2608
AUSTRALIA

Phone: 02 6274 6478

1800 621 372
Fax: 02 6274 6699
E-mail: [email protected]

ii
CONTENTS

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1

Sources of information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2

Marion Green . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3
Fixed firefighting installation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4

Narrative . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5
Cargo loading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5
Vessel’s call at Fremantle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6
The fire . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7
Port of refuge – Albany . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9
Cargo discharge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11

Comment and analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15

The investigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15
Cocoa beans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15
Shipping cocoa beans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15
Ventilation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16
Fumigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17
Information provided by the shipper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17
The fire . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .18
Self-heating or spontaneous combustion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19
Phosphine as a possible source of ignition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .20
Cigarettes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .21
The cargo light . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22
The ship's staff response to the fire . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22
The port authority response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .24
Albany Port Authority emergency procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .25
Authority in fighting ship fires in port . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .26

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27

Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .29

Submissions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .31

Marion Green . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .33

iii
Figures

1. Marion Green at Albany, 30 July 2002 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .v

2. CO2 room, showing banks of gas bottles for CO2 flooding of cargo holds . . . . . . . . . . . . . . . .4

3. Cargo in No. 1 hold, covered with Kraft paper and showing timber dunnage and fumigation
sachets of aluminium phosphide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6

4. Smoke issuing from No. 2 hold access hatch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7

5. Position of Marion Green when the fire alarm activated at 1130 on 28 July 2002 . . . . . . . . . .8

6. No. 2 hold aft hatch panels partially opened . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10

7. Smoke from No. 2 hold after opening hatches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10

8. Albany fire brigade tackling fire in No. 2 hold . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12

9. Smouldering cargo at aft end of No. 2 hold showing burned cavities and channels . . . . . . . .12

10. Discharging damaged cargo onto the wharf . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13

11. Narrow ventilation channel between stacks of cargo in No. 1 hold. Note that some layers are
almost touching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17

12. Cigarette ends discarded around No. 1 hold access . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .21

13. Electric cable for cargo light hanging into No. 2 hold from hold access . . . . . . . . . . . . . . . . .23

14. Burned end of cargo light cable in No. 2 hold . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .23

iv
FIGURE 1:
Marion Green at Albany, 30 July 2002

v
vi
 The report concludes that the investigation was
Summary unable to determine, exactly, the cause of the
fire, but four distinct possibilities were
examined. These were:
On 28 July 2002, the Netherlands flag general- • Self-heating of the cargo due to fungal
purpose cargo vessel Marion Green, of growth
11 894 gross tonnes, carrying 6 000 tonnes of
cocoa beans, was off the coast of Western • Ignition caused by the flammable character-
Australia on passage from Fremantle to istics of the phosphine used for fumigating
Adelaide. the cargo
• Cigarette ends discarded in the hold during
At 1130 that morning a fire alarm was activated
loading of the cargo in Makassar
by the smoke detection system in No. 2 cargo
hold. After a brief inspection through the open • A cargo light that had been left in the hold on
hold access by the mate and second mate, sailing from Makassar
during which they saw some flames on top of
It also concludes that:
the cargo, all openings were closed and the
discharge of CO2 from the fixed firefighting • the vessel’s ‘no smoking’ policy was not
installation was started. By 1715 that afternoon, properly enforced during cargo loading
86 bottles of CO2 had been released into the
• insufficient CO2 was released into the hold in
hold.
the early stages of the fire
At 0700 on the following day, a slight increase • inadequate information on the hazards of
in hatch cover temperature was recorded. The shipping cocoa beans was provided to the
master was advised by the vessel’s managers to ship’s staff and
discharge the remaining 11 bottles of hold CO2
• the response to the fire, once the vessel was
and to divert to Albany as a port of refuge. alongside, lacked co-ordination and a clear
understanding of who had the responsibility
Marion Green berthed in Albany at 1615 on
and authority for dealing with it.
29 July. Additional bulk CO2 was delivered to
the ship from Perth and this, too, was The report recommends that:
discharged into the hold over the next few days.
• the shipowners enforce a strict ‘no smoking’
On the morning of 31 July, the after panels of
policy in the vicinity of cargo operations
No. 2 hatch were opened for an inspection.
Flames were seen on the top layers of cargo and • the shipowners ensure that ship’s masters are
these were doused by the fire brigade after provided with all relevant information on the
which the hatches were again closed and more hazards of carrying organic cargoes and their
CO2 discharged into the hold. At 1400 the fumigation
following day, 1 August, the hatch covers were
• shippers, stevedores and ship’s officers
once more opened and, after further flare-ups
ensure that adequate ventilation channels are
had been doused by the fire brigade, stevedores provided when stowing such cargoes and
began discharging the cargo into sand bungs on
the wharf. By 10 August all the cocoa bean • deck watchkeeping officers log the isolation
cargo had been discharged and Marion Green and stowage of all electrical equipment from
sailed for Adelaide. the holds on completion of cargo operations.

1
Sources of
information
Master and Chief Officer of Marion Green

The Harbourmaster, Albany Port

Fire & Emergency Services Authority of

Western Australia

Australian Quarantine and Inspection Service

(AQIS)

United States Department of Agriculture

Transport Information Service, German

Insurance Association

IPCS Inchem – Chemical Safety Information

from Intergovernmental Organisations

UK P&I Club

References:
Thomas’ Stowage

‘Fire Aboard’ by Frank Rushbrook CBE, 3rd

Edition 1998, Brown, Son & Ferguson

Certain reproductions of chart sections in this

publication are reproduced by permission of The
Australian Hydrographic Service.

© Commonwealth of Australia 13 October 2000. All

rights reserved.

Other than for the purposes of copying this

publication for public use, the chart information from
the chart sections may not be extracted, translated, or
reduced to any electronic medium or machine
readable form for incorporation into a derived
product, in whole or part, without the prior written
consent of the Australian Hydrographic Service.

2
 The vessel, which has a total container capacity
Marion Green of 962 TEUs1 (215 on deck and 446 underdeck),
is fitted with three deck cranes each of 60 tonne
lift at 16 m radius.
Marion Green, is a Netherlands flag, multi-
purpose, tweendeck, general cargo vessel, Marion Green is powered by a single 8-
registered in Rotterdam, Holland. At the time of cylinder, four-stroke, Wärtsilä 8L46B diesel
the incident it was owned and operated by engine developing 7 800 kW (10 460 shp) and
Genchart BV of Rotterdam, (since changed to driving a single shaft with a variable-pitch
Beluga Genchart BV) owners of the ‘Green propeller, giving the ship a maximum speed of
Fleet’. The ‘Green Fleet’ consists of some eight 16 knots. The ship is also fitted with a 750 kW
(and, at the time of the incident, two more bow-thruster unit.
building) identical ‘box-shaped’ vessels which
At the time of the incident, the ship had a
are employed in worldwide tramping.
complement of 13. This consisted of the master,
Marion Green, was built in 1999 at the shipyard mate, second and third mates, chief engineer,
of Schelde Scheepsnieuwbouw BV, at second engineer, bosun, cook, four able seamen
Vlissingen in Holland. The vessel has since then and a wiper. The master, mate, second mate and
been maintained in class with Lloyd’s Register chief engineer were Dutch nationals, the third
as
100 A1,
LMC, UMS, NAV 1, Ice Class mate was Belgian, and the second engineer was
1A. British. All the ratings were Filipinos. All of the
ship’s complement were appropriately qualified
It is a vessel of 142.81 m length overall in accordance with the requirements of the
(132.0 m between perpendiculars), with a STCW95 (Standards of Training, Certification
moulded depth of 13.3 m and a maximum beam and Watchkeeping, 1995) convention.
of 21.5 m. It has a gross tonnage of
11 894 tonnes and a summer deadweight of The master had first gone to sea in 1966 after
17 050 tonnes at a summer draught of 9.689 m. completing his training at a pre-sea training
college. He had served on Shell tankers,
All accommodation and machinery spaces are Nedlloyd container ships and general cargo
aft of the forward engine room bulkhead which vessels from the 1970s to the 1990s before
is located at frame 39. Forward of the engine joining Green Fleet in 1999. For most of the
room bulkhead are two fully box-shaped cargo time since joining Green Fleet he had sailed as
holds, the forward of which, No. 1 hold, master.
(between frames 107 and 151) has a cubic
The mate had been at sea for 12 years since
capacity of 7561 m3 and the aft hold, No. 2,
starting as an apprentice. He had served in the
(between frames 39 and 103) a capacity of offshore industry, on reefers, small coasters,
12 292 m3. The vessel can be configured with a general cargo vessels and heavy lift ships before
fully flush tweendeck, consisting of loose steel, joining the Green Fleet in 1999. It was his sixth
‘pontoons’ each of 17.75 x 6.52 x 0.90 m, voyage on this type of vessel. Neither the
which can be positioned individually at three master nor the mate had any previous
different heights in the holds. These pontoons experience in the carriage of cocoa, or other
can also be used to form vertical bulkheads at bean, cargoes.
intervals of six metres. The ship can be readily
converted from either a tween or even At the time of the incident, all ship’s certificates
tripledecker, to a full container vessel, or fitted required under international shipping
with any combination of tweendecks and conventions were valid. In addition, the vessel
bulkheads, thus making it extremely versatile. had been issued with a ‘Document of

1
Twenty-foot Equivalent Units.

3
Compliance with special requirements for ships Of these, 57 bottles are arranged for release into
carrying dangerous goods’ by the Netherlands the engine room, as a complete bank, in less
Shipping Inspectorate in June 1999. From July than two minutes, by pneumatically operated
2002 Marion Green, as a general cargo vessel, valves. The other 97 are arranged for manual
was required to hold an International Safety release and are dedicated to the holds. Release
Management Certificate. Such a certificate was of those CO2 bottles for the holds is done
issued to the vessel on 12 February 2002 by the manually after selecting the appropriate number
Netherlands Shipping Inspectorate. of bottles for the hold to be flooded. It is
possible, by use of a changeover valve, to also
Fixed firefighting installation release the engine-room bank into the holds.
Marion Green is fitted with a Unitor fixed CO2
The CO2 is discharged into the holds through
firefighting installation for gas flooding either,
nozzles arranged on a ring-main around the top
or both, Nos. 1 & 2 holds and for the engine
of each hold. In No. 2 hold, there are eight such
room. The major part of the installation is
nozzles, four on each side. These same nozzles,
contained in a CO2 compartment situated
under normal circumstances, are also used to
beneath the main deck on the port side aft of withdraw samples of air on a continuous basis
No. 2 hold (between frames 26 and 39) from for passing through the Autronica smoke
where the release of the CO2 can be controlled. detection system. This system, driven by small
An additional remote release cabinet for the fans, constantly samples air from both holds
bank of engine room bottles, only, is situated on drawing the samples through a photo-electric
the starboard side of the main deck, just off the cell which, if it detects obscuration due to
main cross-alleyway within the accommodation. smoke, will sound an alarm on the fire detection
There are a total of 154 CO2 bottles, each of panel situated on the after bulkhead in the
67.5 litre capacity and containing 45 kg of gas. wheelhouse.

FIGURE 2:
CO2 room, showing banks of gas bottles for CO2 flooding of cargo holds

4
 hatch. The vessel’s own cranes were used for
Narrative loading the bags, slung in cargo nets, into the
holds where the stevedores arranged their
stowage. Each sling contained five tiers of five
bags each.
Cargo loading
Cargo loading continued, with either three or
On 17 July 2002, Marion Green arrived at the four gangs working from 0800 on 18 July, until
Indonesian port of Makassar at the beginning of 2200 on 21 July. The weather throughout was
voyage 6 of 2002, to load a cargo of fine, with sunny and cloudy periods, but with
6 096 tonnes of Sulawesi cocoa beans in burlap high ambient temperatures, usually above 30°C,
bags (96 156 bags of approximately 63 kg each) and high humidity.
for the Brazilian port of Ilheus. Upon the ship’s
arrival alongside in Makassar, moveable steel On a number of occasions during this period,
pontoons were arranged as a vertical bulkhead the mates on deck watch had cause to draw the
in No. 2 hold at frame 79 (the aft end of No. 9 attention of the stevedores to the ‘NO
pontoon position – 19.56 m from the forward SMOKING’ signs painted on the ship’s structure
end of the hold) thus partitioning the hold. The around the holds, as a some men had been
forward section was for two items of machinery observed smoking in the vicinity of the holds.
to be loaded later in Australia, while the after,
major, part was for the loading of cocoa beans. On 19 July, the vessel’s liferafts underwent their
annual checks. During this process, the drinking
Prior to commencing loading the bagged cocoa water in the liferafts was replaced and the
beans, the vessel received 42 m3 of wooden removed sachets were taken by the stevedores
planks and pallets for use as dunnage, 800 working in the holds for some refreshment in
bamboo mats for placement over the dunnage their hot, humid working conditions.
on the tanktops and 2 200 kg of Kraft paper.
The paper was to be used to line the holds for Loading of the 6 000 tonnes of cargo was
absorption of any condensation which might completed at 2145 on 21 July, whereupon the
form during the voyage, due to changes in top layers of cocoa bags were covered by
ambient temperature, on the sides of the holds absorbent paper and the hatches closed.
and the undersides of the hatches Following checks to ensure that all openings to
both holds were able to be thoroughly sealed,
On the 16th July, the day before the ship’s both holds were then finally fumigated by the
arrival, all the dunnage, bamboo and Kraft distribution of aluminium phosphide, in sachets,
paper had been fumigated with methyl bromide. around the holds. The aluminium phosphide,
The cargo was similarly fumigated for a period when exposed to atmospheric moisture,
of seven days leading up to the loading and, in generates phosphine (hydrogen phosphide) gas,
addition, on 17 July, the ship’s empty holds a powerful insecticide. The quantity of
(Nos.1 and 2) were sprayed throughout with a aluminium phosphide required to provide for
5 per cent solution of pyrethrins. five days of ‘in transit’ fumigation of the cargo
had been calculated by the fumigation
Shortly after 0100 on 18 July, the holds were contractors before it was placed in the holds.
again opened. At 0230, after preparing the This process was completed at 0030 on 22 July
dunnage and other stowage materials, two gangs and both hatches were then closed. Shortly
of stevedores started loading the cargo of afterwards, the pilot boarded and Marion Green
bagged cocoa beans. 48 156 bags were to be sailed for the Western Australian port of
loaded into No. 1 hatch and 48 000 into No. 2 Fremantle.

5
Vessel’s call at Fremantle, WA The tweendeck at the forward end of No. 2 hold
was then opened and the hold prepared for
The voyage from Makassar to Fremantle was
loading two large heat exchangers, also destined
uneventful. The hatches and all other openings
for Brazil, which were to be stowed forward of
to both holds remained sealed throughout the
the temporary steel pontoon bulkhead across the
voyage, while the phosphine fumigation of the
hold at frame 79. The ship’s lifting gear was
cargo took effect. All the hold ventilation fans
prepared and, during the afternoon, stevedores,
remained off.
assisted by the ship’s crew loaded the heat
At 0300 on 27 July, the Fremantle pilot boarded exchangers into the hold. The task took approxi-
and Marion Green was secured alongside No. 2 mately 45 minutes and, on completion at
berth, North Quay, at 0435. After arrival, the approximately 1700, the heat exchangers were
second and third mates, wearing gas masks, set covered with tarpaulins and the tweendeck
about opening all the hold ventilation flaps. The pontoons replaced. When the hatch had been
deck was then cleared of personnel and the hold closed, a 20-ft container containing spare parts
ventilation fans were started. Ventilation of the for the heat exchangers was then loaded onto
holds continued for approximately 30 minutes the top of the hatch as deck cargo.
longer than the required six hours, until about
That evening the ship took bunkers. Later that
1300 that day, when the mate then entered the
night the pilot boarded and, at 0145 on 28 July,
holds testing for traces of phosphine. Testing for
Marion Green sailed for Adelaide. After sailing,
phosphine was carried out using ‘Dräeger
the hold ventilation was switched off because of
tubes’2 supplied by the shipper of the cargo. The
heavy rain showers.
tests showed that the level of phosphine had
reduced to a zero reading.

FIGURE 3:
Cargo in No. 1 hold, covered with Kraft paper and showing timber dunnage and fumigation sachets of aluminium
phosphide

2
Test equipment to which can be fitted the appropriate glass tubes containing reactant chemicals designed to show up the presence of a
specific gas.

6
The fire investigate. They noticed that the access hatch
was warm and there was smoke issuing from
At about 0930 that morning, the mate went
the upper hold access opening. This information
down to the holds to measure the humidity and
was passed to the master in the wheelhouse,
dew point. He also again tested for the presence
who re-activated the fire alarm. Two fire suits
of phosphine, with a negative result. While at and four breathing apparatus sets were brought
the access to No. 2 hold he noticed that the hold to the hold access from the fire locker on the
was ‘soaking wet’. Moisture was running down main deck and from the wheelhouse.
the bulkheads and dripping from the hatch
covers. He measured the relative humidity as The ship’s crew closed all ventilation openings
being 93 per cent. By 1015 he had taken the and ran out fire hoses on deck while the chief
required readings in No. 1 hold. The relative officer and second officer donned the fire suits
humidity, at 82 per cent was significantly lower and breathing apparatus. They opened the
than that in No. 2 hold. Having taken the starboard aft access hatch to No. 2 hold and saw
necessary readings, he restarted the hold flames on the top layer of bags of cocoa beans,
ventilation fans then made his way to the not far from the hold access opening. A lot of
wheelhouse. smoke was coming from the hold access hatch.
They played a fire hose onto the area where the
At 1130 that same morning the ship was at the flames were just visible before closing the
approximate position 33° 50'S, 114° 49'E and access hatch and informing the master of the
making 16 knots, when a fire alarm was situation.
activated by the smoke detection system in
No. 2 cargo hold. At about the same time, The master instructed them to make
smoke was observed on the main deck coming preparations to flood the hold with CO2 from the
from the No. 2 hold port side ventilation ducts. fixed fire-fighting installation. Crewmembers
Hold ventilation was stopped and the mate went started cooling the hatch covers with two fire
down to the deck with the second mate to hoses.
FIGURE 4:
Smoke issuing from No. 2 hold access hatch

7
FIGURE 5:
Position of Marion Green when the fire alarm activated at 11.30 on 28 July 2002

115ºE

Position of Marion Green

at 1130 on 28 July 2002

NT
NT

Qld
WA

SA

NSW
NSW
ACT
ACT
Location of
incident
Vic

Tas

8
At 1150, the mate started to manually release one of the drains on the starboard side of the
the CO2, one bottle at a time, into the hold. hatch coaming to again test the oxygen level.
Initially, 25 bottles were released, after which it However, they were unable to take a reading as
was decided to wait for the CO2 lines to depres- they were confronted by a continuous flow from
surise and to don breathing apparatus again, as the drain of a brown, mud-like, substance. At
there were several leaking connections in the 0835, having again contacted the vessel’s
CO2 room. The vessel’s managers in Rotterdam managers, the master was advised to discharge
were advised of the situation. the remaining 11 bottles of hold CO2 and to
divert to Albany, approximately 100 nautical
At about 1210, one vent flap on the forward end miles distant, as a port of refuge. Agents for the
of No. 2 hold was opened in order to measure vessel, who were appointed in Albany, informed
the oxygen content of the issuing air/gas the harbourmaster of the situation and, at about
mixture. It was found that the oxygen content 0900 that morning, the harbourmaster contacted
had decreased only slightly and, consequently, the ship by VHF. The ship’s master explained
another 25 bottles of CO2 were progressively that there was a fire in No. 2 hold, that CO2
discharged from about 1220. At 1245, the hold flooding had been used and that the fire seemed
access was opened again to check the to be under control. The temperatures, however,
effectiveness of the CO2 flooding. The flames were not going down.
appeared to have been extinguished. However a
proposed entry by the mate and second mate The vessel proceeded towards Albany and, at
was cancelled on account of the thick smoke 1450 that same day, the harbourmaster, as pilot,
and the fear of a ‘backdraft’. The entrance was boarded the ship.
closed again, as was the forward vent flap.
At 1615 on 29 July, Marion Green was
At 1300, another 14 bottles were discharged and alongside at Albany. Shortly after berthing a
a fire watch was set up to monitor the temper- meeting was held between the harbourmaster,
atures of the hatch covers and coamings every the master, a representative from Customs and
15 minutes. At 1500, a further situation report the ship’s agent. As the vessel’s stock of CO2
was made to the ship’s managers and the had been severely depleted, the agent arranged
Australian Search and Rescue co-ordination for a bulk tank of 14 tonnes of CO2 to be
centre (AusSAR) was also notified. delivered from Perth. At 1730, representatives
of the local fire brigade arrived on board to
By 1715 that day, a total of 86 bottles of CO2
evaluate the situation. The hatch covers
had been released into the hold and the temper-
remained closed and temperatures were
atures of the hatch covers dropped steadily.
monitored throughout the night on a two-hourly
During the night, the hold remained sealed and
basis. At 2200 that evening, more hoses were
temperatures of the hatch covers were
rigged and boundary cooling was again applied
monitored, and recorded, on a regular basis. As
the temperatures were dropping, the to the hatch covers.
temperature observations were changed from
This situation remained unchanged throughout
every 15 minutes to hourly. Checks were
the following day, Tuesday 30 July. A surveyor,
maintained in the engine room on the forward
appointed by the P&I Club requested, through
bulkhead, being the aft bulkhead of No. 2 hold.
the master, that permission be given to open the
These temperatures, however, were found to be
hatch, but this was refused, the harbourmaster
steady.
requiring a further 24 hours of monitoring. Later
that day the bulk CO2 arrived from Perth and
Port of Refuge – Albany was placed on the main deck at the port after
At 0700 on the following day, an increase in corner of No. 2 hold. It was connected to the
hatch cover temperature of about 2°C was ship’s CO2 flooding system through the adjacent
recorded. The mate and chief engineer opened access to the CO2 room.

9
FIGURE 6:
No. 2 hold aft hatch panels partially opened

FIGURE 7:
Smoke from No. 2 hold after opening the hatches

10
At 0545 on the morning of 31 July, the temper- the hatch covers cool. No attempt, however, was
atures of the hatch covers appeared to be rising made to open the hold again before 1400 that
again and boundary cooling was resumed. Later afternoon, when a meeting of involved parties
that morning, a decision was made to open the was convened by the harbourmaster to
hold and ascertain the situation determine the future course of action.

At 0930, the access hatch was opened, followed Cargo discharge

by the two aft panels of the cargo hatch cover
which were lifted into a partially open position. In spite of various evident conflicts of interest
Much smoke issued from the hold and some (see p25), it was eventually agreed that there
flames were observed through the thick smoke was no real option but for the cargo to be
on top of the cargo. The fire brigade started discharged and, at 1600 that afternoon, on
hosing down the top layers of bags of cocoa 1 August, the hatches were partially opened.
beans and the flames were eventually Several spot fires were again visible through the
extinguished. After a short period with no more smoke. Both the fire brigade and ship’s crew, all
flames appearing, the hatch covers were fully wearing BA sets, attacked the fires with several
opened. hoses.

As no decision had been reached at that stage Sand dams, into which the cargo would be
on what to do with the damaged cargo, the placed for final dousing, were constructed on
hatches were again closed and more CO2 was the wharf adjacent to the ship and arrangements
injected from the bulk tank on the main deck. were made, under the auspices of AQIS, for the
damaged cargo to be trucked to the Albany tip
At 1500 that afternoon the vessel was required where it would be buried as landfill.
to shift ship to the next berth. On completion of
the shift, steam was noticed coming from the At 1835 on 1 August, in consultation with the
port side coaming of No. 2 hatch. A small area fire brigade, the discharge of the cargo
of paint was scorching on the hatch cover and commenced. The port remained closed.
there was some noticeable deformation of the
hatch panel. It was immediately evident that the On the following day, 2 August, the spot fires
intensity of the fire was rapidly increasing and appeared to be under control. At 1120, the
the fire brigade was again called. At 1600, two tanker Scottish Bard was allowed to berth at
appliances attended on the wharf and more CO2 No. 2 wharf to discharge its cargo of petrol. The
was injected into the hold. Substantial boundary discharge of burning cargo from Marion Green
cooling was applied to the hatch covers and was suspended while Scottish Bard discharged
coamings, by both the fire brigade and the its cargo.
ship’s crew, throughout that night.
Over the next few days all the bags of cocoa
At approximately 1730 that evening, a tanker, beans in No. 2 hold were discharged by
Scottish Bard, arrived at Albany and was due to stevedores using the ship’s cranes and grabs.
berth close to Marion Green. The port, however, Any spot fires or areas of smouldering cargo
was closed due to the fire. were doused during the discharge. On 6 August
the ship’s agent was informed by AQIS that it
On the following morning, Thursday 1 August, had reviewed its decision regarding burial of the
the temperatures of the hatch covers, being cargo at the Albany tip and it could be
monitored by the ship’s staff, appeared reclaimed where appropriate. Approximately
relatively stable. The weather had deteriorated 44 per cent of the cargo from No. 2 hold,
overnight and the day became windy with rain relatively undamaged, was salvaged and moved
showers and drizzle. This assisted in keeping to a cargo shed on the wharf. Flare-ups in the

11
FIGURE 8:
Albany fire brigade tackling fire in no.2 hold

FIGURE 9:
Smouldering cargo at aft end of no.2 hold showing burned cavities and channels

12
FIGURE 10:
Discharging damaged cargo onto the wharf

cargo occurred frequently during the discharge, After completion of the discharge of cocoa
in the hold, on the wharf and in the cargo which beans and the cleaning out of residue in No. 2
had been moved to the shed. The fire brigade hold, Marion Green sailed for Adelaide at 1550
remained in attendance at the vessel, dealing on 10 August.
with these fires, until they were finally stood
down at 2020 on 7 August.

13
14
 each pod) consists of the seed coat which
Comment and encloses the cocoa kernel. The cocoa kernel is
the principal component for the production of
analysis cocoa products.

The history of cocoa and chocolate dates back

to the time of the Incas, Mayas and Aztecs of
Central and South America. It was at one time
The investigation the currency used by those peoples. It was
A marine investigator from the ATSB arrived on introduced to Europe in 1527. During the 16th
board Marion Green on the morning of 30 July and 17th centuries, the Spaniards established
and, while conducting the investigation over the cacao tree plantations in their colonies in
next five days, was also able to observe the Central and South America. The Dutch later
subsequent events and the measures taken to took trees to other countries including their
deal with the fire during that period. The master colony of Indonesia. Indonesia is not one of the
and chief officer were interviewed, copies of the largest of cocoa producing countries, unlike
ship’s logs and all other relevant documentation Brazil, the destination for the cargo aboard
were obtained and samples of cocoa beans were Marion Green. At the time of the incident, a
taken from both Nos. 1 and 2 holds for later worldwide shortage of cocoa existed and
expert examination. Evidence concerning the companies in Brazil were having to import
response of the port authority and the local fire cocoa beans in order to be able to fulfil their
brigade was obtained. Details of the cargo, its contracts.
loading in Makassar and its fumigation were
also obtained, as was information which had Cacao seeds are the source of commercial
been provided to the master by the shipper of chocolate, cocoa, and cocoa butter. Cocoa is
the cargo. made by removing most of the fat from the
beans, then roasting and grinding them. The fat
Owing to the nature of the incident, it was that is removed is called cocoa butter. Cocoa
inevitable that much of the physical evidence butter is used in confections and in the
relating to the source of ignition was destroyed manufacture of pharmaceuticals, soap, and
as the cargo, a large proportion of it burned, was cosmetics. Cocoa butter has been described as
soaked by water from fire hoses on several the world's most expensive fat and cacao beans
occasions before its discharge. In addition, the contain a great deal of this fat, some 40–50 per
action of the grabs during the discharge cent, in addition to small amounts of mildly
damaged many unburnt bags and caused the stimulating alkaloids, including caffeine.
cargo to finish up as an homogenous heap on Consequently, cocoa beans are a very high value
the wharf. The pattern, however, in which the cargo. The value of the cargo aboard Marion
fire spread through the hold could be observed Green on 28 July 2002 was approximately
through the smoke on those occasions when the US$10.5 million.
hatch covers were opened for fire-fighting
operations. Shipping cocoa beans
Cocoa beans are usually shipped in burlap
Cocoa beans (made from jute or sisal) bags, each bag
Cocoa beans are the seeds, contained in a weighing from 60–65 kg. Owing to the value of
cucumber-like fruit, of the cacao tree, the cargo, new, high quality bags are usually
Theobroma Cocoa Linné. (Theobroma, from the used. Cocoa beans require particular
Greek, literally means ‘food of the gods’) It is a temperature, humidity and ventilation
member of the Sterculiaceae family. The cocoa conditions. Due to its high fat content, the cargo
bean (there are usually about 30–40 seeds in is very apt to self-heat and there may be a risk

15
of cargo fire on contact with flammable the beans continually release water vapour
substances.3 The cargo requires to be stowed in during the voyage, this vapour must be removed
a dry place, normally below decks. The cargo to reduce the risk of condensation forming in
loses value, however, if allowed to become too the event of cooler ambient conditions, and the
dry and brittle; around 7 per cent moisture risk of mould growth due to high relative
content being acceptable. humidity in the hold. No. 2 hold of Marion
Green is ventilated by means of two supply
The temperature of the cargo is frequently hotter
fans, port and starboard, ducted through the
than the ambient temperature in areas through
after bulkhead and discharging into the hold on
which the vessel may pass. Too little ventilation
four levels. At the forward end of the hold the
results in hold temperatures rising above the
dew point with the consequence that conden- hold atmosphere is extracted by two extraction
sation forms on the ship’s shell plating, tank fans drawing from the four levels. The capacity
tops and hatches. Excess moisture content will of the fans is sufficient to provide 20 air
cause mould growth and, if excessive, will rot changes per hour when the holds are empty;
the cargo. The beans themselves release large considerably more than sufficient when the
amounts of water vapour during extended holds are loaded.
voyages after which the moisture content of the
beans may be observed to have decreased by 1 To facilitate the movement of air around the
to 3 per cent. The vapour given off may contain cargo, the holds were loaded with ventilation
acetic acid which can cause corrosion.4 In channels provided both in the fore-and-aft
addition to the moisture content, the beans direction and athwartships. The channels were
contain certain enzymes that bring about post- approximately 300 mm wide and extended
fermentation which is prevented only by good down the full depth of the cargo. At various
air circulation around the cargo. levels, individual bags were arranged as ‘ties’
across the ventilation channels to prevent the
To keep the bags dry should condensation form, layers of bags collapsing together and blocking
the cargo aboard Marion Green was loaded on the channels.
top of dunnage which had been made into
wooden frames, not unlike conventional pallets. The cargo in No. 2 hold was arranged with a
The wooden frames were then overlaid with single ventilation channel along the centre-line
bamboo matting which was in turn covered with of the hold and two channels in the athwartships
a layer of Kraft5 paper. Other wooden frames direction, such as to divide the length of the
were placed up the ship’s sides as the cargo was hold into three equal sections.
loaded and, again, covered with a layer of Kraft
During the investigation, the stowage of the
paper. Finally, once the loading of each hold had
cargo in No. 1 hold was assessed. It was
been completed, the cargo was completely
observed that the width of the ventilation
covered with a layer of Kraft paper to protect
channels in many areas was considerably less
the top bags from any condensation which could
than 300 mm and, indeed, in some spots had
drip from the hatch covers and coamings.
closed up completely, due either to improper
stowage on loading, or to subsequent movement
Ventilation
of the cargo. It was also noted that the Kraft
Good ventilation is essential, shippers requiring paper, covering the top layer of cargo, also
a minimum of 20 air changes per hour. Since covered the ventilation channels, hence

3
Transport Information Service, German Insurance Association, Berlin.
4
Thomas’ Stowage.
5
A brown paper used extensively in the packaging and manufacturing industries. Available in numerous grades and for numerous
purposes. In this instance the paper used was of a relatively heavy, absorbent grade.

16
restricting any vertical movement of air through Phosphine is generated by allowing atmospheric
these channels. moisture in the surrounding air to react with
aluminium phosphide in the form of either
Fumigation tablets, pellets or sachets. Aluminium phosphide
is also referred to as the ‘dry gas’. Once spent,
To both protect the cocoa bean cargo from
the aluminium phosphide degenerates into a
damage by insect pests and to prevent the
grey-white powder composed of aluminium
transport of such pests from one country to
hydroxide and some inert ingredients.
another, cargoes of cocoa beans are fumigated,
both before loading and for the first few days in On completion of loading of the cocoa bean
transit. During the fumigation period, in this cargo at Makassar, sufficient aluminium
case for five days, the holds must remain tightly phosphide, in the form of sachets, to provide for
sealed and all ventilation shut down. Only when five days ‘in transit’ fumigation was scattered
the fumigant has had sufficient time to penetrate randomly across the top layer of the cargo in
all of the cargo is ventilation restored and each hold. In No. 1 hold, some of these sachets
maintained for the remainder of the voyage. could be observed to have fallen down into the
ventilation channels provided between the
One of the most common fumigants used is
stacks of bags.
hydrogen phosphide, usually referred to as
phosphine gas (PH3). Phosphine is a colourless
gas which is toxic to insects, humans and other Information provided by shipper
forms of animal life. It is very mobile with a Neither the master nor the mate of Marion
high vapour pressure which enables it to Green had any previous experience of the
penetrate to all parts of the cargo. This carriage of cocoa bean cargoes and both had to
combination, together with high molecular rely on information from the sub-charterers
activity and toxicity at low doses, accounts for provided to the ship, in a telex from Genchart, a
its wide acceptance as a fumigant. month before the incident.

FIGURE 11:
Narrow ventilation channel between stacks of cargo in No. 1 hold. Note that some layers are almost touching

17
The information related to the way the bags of The fire
cocoa beans were to be slung to avoid damage,
The quantity of wooden dunnage used in the
the arrangement for dunnage on the tank-tops
holds was 42 m2, the number of bamboo mats
and at the ship’s sides to avoid contact with any
was 800, and the quantity of Kraft paper used
condensation and the requirements for
was 2 200 kgs. The total amounted to a
ventilation, temperature and dew-point
significant quantity of dry, combustible fuel for
monitoring.
a fire in the hold, quite apart from the burlap
The information went on to discuss the bags and the cargo itself with its relatively high
problems for the cargo caused by either fat content. There is no question as to the fuel
excessive or too little ventilation, and the effects load available for the fire, as these materials
of passing through zones of ambient were observed to be burning on each occasion
temperature below, or above, that of the cargo. that the hold was opened. By the time the cargo
Examples used were generic in nature, had been discharged from No. 2 hold, the
discussing the problem of ‘ship sweat’ (conden- smouldering fire had spread throughout the bags
sation on ship’s sides and hatches) for bagged of beans with frequent flare-ups of the timber,
rice loaded in a tropical climate, and ‘cargo paper and burlap.
sweat’ (condensation on the cargo) for steel
The issue facing the investigation was the
coils loaded in a cold climate. No information
source, or origin, of the fire.
was provided to the ship specifically on the
carriage of cocoa beans, nor on the subject of During the course of the investigation, four
the possibility of self-heating of the cargo due to possible sources were identified:
the growth of fungi as a consequence of warm,
moist conditions. • Self heating of the cargo, promoted by fungal
growth
Information on fumigation and the use of
• Phosphine used for fumigation
phosphine as a fumigant was provided to the
ship by the Pan Asia Superintendence • Cigarette ends found amongst the discharged
Corporation at Makassar. The five-page cargo
document described in detail phosphine and its
• A cargo light
detection, the process of fumigation, the
requirements for ventilation, the toxic effect that In each case, the fire would have started at the
phosphine has on the human body, precautions time of, or shortly after, the loading of the cargo
to be taken and medical treatment should in Makassar. In Fremantle, the mate noticed that
poisoning occur. The company also provided the the humidity in the hold was very high. This
gas masks and gas detection equipment would indicate that heat was being generated in
(Dräeger tubes) to be used by the ship’s officers the lower part of the hold, driving moisture
during the voyage. However, nowhere in the from the cocoa beans. However, no sign of
information provided was the hazard of flamma- smoke or flame was evident at that time. It was
bility of phosphine mentioned. only after air had been admitted, through hold
ventilation and opening the hatch covers in
Further information on the carriage of Fremantle, that the fire gained a hold.
fumigated cargo is contained in the IMO’s
‘Recommendations on the safe use of pesticides Observations when the hold was first opened in
in ships’. This document, however, was not Albany indicated that the source of the fire was
carried on board Marion Green. in the lower tiers of cargo near the starboard aft

18
corner of the hold. The way that the fire had phosphine) are formed. The resultant abrupt
burned slowly up towards the top layers of increase in temperature gives rise to the
cargo, and then also spread along the bottom of smouldering cavities and channels within the
the hold to the forward end would be consistent bulk cargo.
with smouldering, or slow combustion, having
Observation of the progress of the fire during
taken place during the five day passage between
the period 29 July to 1 August showed clearly
Makassar and Fremantle. It continued in this
that the fire had spread from the starboard after
manner, albeit accelerated, after the vessel
corner of No. 2 hold through such channels,
arrived at Albany.
with cavities appearing at the top of the cargo in
various places. This, however, does not in itself
Self-heating or spontaneous provide sufficient evidence that this was the
combustion source of ignition for the fire as the fire could
Cocoa beans are not generally regarded as an have spread in this manner, whatever the source
hazardous cargo and are not mentioned as such of ignition.
in the IMDG Code. However, under suitable
ambient conditions (temperature > 25°C, high Samples of the cargo were taken from both No.
relative humidity and lack of oxygen) and due 1 hold (undamaged cargo) and from the
to their elevated fat content which, in damaged cargo in No. 2 hold. The samples were
conjunction with moisture, results in submitted to the Eastern Creek Plant Quarantine
hydrolytic/enzymatic fat cleavage, cocoa beans Laboratory of the Australian Quarantine &
have a tendency for postfermentation and self- Inspection Service (AQIS) near Sydney, NSW.
heating. Some species of fungus, such as There, the seeds were examined by a quarantine
Aspergillus Fumigatus, participate in the self- plant pathologist and placed in a humid
heating process. Spontaneous combustion of environment in petri dishes for the culture of
broken raw cocoa beans in bulk usually takes any fungi that may have been present.
the form of smouldering cavities or channels,
The results of the tests showed no Aspergillus
such a fire developing in four distinct phases:6
contamination could be isolated from the
• first phase: a general biological phase in undamaged sample of cargo from No. 1 hold.
which mesophilic micro-organisms multiply However, the tests revealed that a mould,
in a wet spot, raising the temperature to Aspergillus Niger, which is closely related to
37°C. Aspergillus Fumigatus, was present in the
• second phase: this involves the highest level sample of cargo from No. 2 hold. The tests were
of activity from thermophilic micro- repeated and, again, the same results were
organisms at temperatures of up to obtained. Aspergillus Niger was present in the
approximately 70°C. cargo in No. 2 hold and, in the opinion of the
AQIS pathologist, this could have had the same
• third phase: the thermophilic decomposition heating effect.
phase which is characterised by exothermic
chemical breakdown reactions, in particular If the condition of the cocoa beans in the hold
by oxidation reactions between unsaturated was conducive to mould development, then a
fatty acids and atmospheric oxygen (the number of different fungi, including Aspergillus
cocoa beans having a fat content of >50 per Niger, could take advantage of the favourable
cent). growing conditions of warmth and high
• fourth phase: the pyrophoric gas phase, in humidity to proliferate through the stored beans.
which pyrophoric carbon and gasses (eg Once the mould activity is encouraged to

6
Transport Information Service, German Insurance Association, Berlin.

19
develop and spread, so too will there be an driving off moisture, before the vessel arrived
increase in the heat generated by the fungal and the hatches were opened. During the
growth. passage from Makassar to Fremantle the hold
ventilation had been off, to maintain the
The conditions necessary to promote fungal integrity of the fumigation. In Fremantle,
growth are warmth and moisture. restoring the ventilation and opening the hatches
would have allowed the ingress of air which, in
If the moulds find favourable living conditions,
turn, would have allowed the fire to develop and
i.e. when the critical water content of 8.5 per
become evident after sailing on the morning of
cent is exceeded at an equilibrium moisture
28 July.
content of approximately 88 per cent, the
moulds rapidly develop within 3 - 4 days, at the Although it cannot be confirmed with any
end of which period thousands of spores have degree of certainty, there is sufficient evidence
formed on the surface of the cocoa beans7. to indicate that moisture, warmth and
During this process, considerable heat may be subsequent mould growth leading to self-
generated. heating and combustion should be considered as
one possible source of the cargo fire in No. 2
The cargo aboard Marion Green was loaded in
hold.
tropical weather conditions where the ambient
air temperature was consistently above 30°C
Phosphine as a possible source
with high relative humidity. In addition, during
the loading of the cargo the sachets of water in of ignition
the liferafts were replaced with fresh ones. Phosphine is a highly flammable gas which can
Those removed from the rafts were taken for auto-ignite at ambient temperatures. At concen-
refreshment by the stevedores working in the trations greater than 1.8 per cent, it forms
holds. During the discharge of the burned cargo explosive mixtures with air.8 To reduce the risk
at Albany, several of these sachets, having been that this poses, all commercial formulations
discarded in the hold, were found amongst the available contain ammonium carbamate which
remains of the burned cargo. Some were empty, releases ammonia gas and carbon dioxide to
but some still contained water. Such sachets, reduce the potential fire hazard posed by
discarded during the cargo loading in Makassar phosphine. The ammonia in the formulation also
may well have wetted small areas of some of serves as a warning agent, giving out a garlic-
the bags of beans and thus provided ideal levels like odour. Nevertheless, great care has to be
of moisture for the fungi to flourish. taken in how aluminium phosphide is applied
due to its inherent problem of flammability. The
During the later discharge of the, mostly ammonium carbamate reduces the potential
undamaged, cargo from No. 1 hold it was found danger, but it does not eliminate it.
that 14 bags of beans had localised areas of wet
staining. In February 2003, the UK P&I Club published a
Bulletin (No. 289), reproduced in Seaways, the
When the mate first entered No. 2 hold after the journal of the Nautical Institute, drawing the
vessel’s arrival in Fremantle, he noticed that No. attention of the maritime industry to the danger
2 hold was ‘soaking wet’, with moisture of phosphine explosions. The bulletin noted that
dripping from hatch covers and running down a number of such explosions were known to
bulkheads. No. 1 hold was not so wet. This have occurred. It also noted that the process of
would indicate that an area, or areas, of the aluminium phosphide reacting with moisture in
cargo in No. 2 hold were already hot, and the air sometimes produces small quantities of a
7
IBID.
8
IPCS Inchem – Chemical Safety Information from Intergovernmental Organisations.

20
gas known as diphosphine, which, unlike Such tablets could be envisaged as producing
phosphine, is spontaneously flammable reacting localised high concentrations of diphosphine
instantly with oxygen in the air. This is likely to leading to a very rapid reaction with oxygen and
to ignition.
occur when there is an imbalance between the
aluminium and phosphorus, with an excess of Aluminium phosphide tablets are routinely used
the latter. Such a situation may arise during in fumigation and a very large number of
production of the tablets if an excess of shipments are fumigated annually without any
problems. Incidences of explosions are therefore
phosphorus is inadvertently used during the
very rare and, as far as we have been advised,
preparation. fumigant explosions have only been encountered
when companies have used cheaper brands of
In describing a recent incident, the bulletin then aluminium phosphide tablets produced in
goes on to state: developing countries.
Although not definitively proven, we have been
The origin and chemical composition of the
advised that it is likely that potentially explosive
mixtures of air and phosphine are frequently aluminium phosphide used to fumigate the
encountered during the first 12 to 24 hours of cocoa bean cargo aboard Marion Green is not
phosphine fumigation when the phosphine known, however it cannot be eliminated as
concentration in the upper section of the hold having been a factor in the fire. This is partic-
reaches a peak concentration. The resulting high ularly so if sachets had fallen into confined
concentration of phosphine then disperses by
spots in the lower tiers of cargo, where there
diffusion, with the gas diffusing into the less
accessible lower sections of the cargo. In this was potential for the gas concentration to have
recent case, the explosion occurred some risen to extreme levels.
12 hours or more after the fumigation had been
started and the hatch covers had been closed. Cigarettes
Although no source of ignition was identified
conclusively, it is suspected that defective ‘NO SMOKING’ signs are prominently
aluminium phosphide tablets, containing displayed, painted in large letters, around the
localised excesses of phosphorus, were the cause. hatches of Marion Green. Nevertheless, during

FIGURE 12:
Cigarette ends discarded around No. 1 hold access

21
the loading of the bagged cocoa beans at compliance with the ship’s routine on
Makassar, the deck officers on cargo watch had completion of loading.
need, on several occasions, to reprimand certain
members of the loading gangs for smoking The cable hanging into the hold was the same
around the hatches and hold access spaces. type as that attached to the cargo lights and the
plug on the end had also been removed from its
During the investigation, empty Indonesian socket. However, the other end, in the hold, was
cigarette packets and numerous butt-ends of found to have been burned away. Evidence
cigarettes were found in the accesses to both given to the investigation indicated that it had
holds. In addition, it was later reported by the also been attached to a cargo light which, in
surveyor appointed by the P&I Club, that Makassar, was hanging on one of the timber
similar cigarette packets and cigarette ends had dunnage frames against the aft bulkhead
been found amongst the cargo from No. 2 hold adjacent to the hold access. This cargo light,
after it had been discharged ashore. however, could not be located during the
investigation, neither was it found amongst the
A cigarette end or match, discarded in the hold cargo either in the hold, or once discharged to
during cargo loading, without being properly the wharf.
extinguished, could well have initiated some
smouldering amongst the cargo and associated The whereabouts of this cargo light, or the
Kraft paper, dunnage and burlap bags. This reason for its cable having been left in No. 2
smouldering, in the absence of sufficient air, the hold after the loading of the cargo, could not be
ventilation being shut off, may have developed ascertained. Neither could the point in time be
into a significant slow combustion fire over the determined when its plug had been removed
next few days, developing more rapidly once from the power socket.
the ventilation had been restored and the hatch
covers opened while the vessel was in These 1500-watt cargo lights generate consid-
Fremantle. erable heat. If this light had been left, switched
on, in the hold at the completion of loading and
Smoking materials, discarded in No. 2 hold had come into contact with either the bagged
during the loading of the cargo in Makassar beans, the Kraft paper or other combustible
must be regarded as another possible cause of material, it may well have provided a source of
the fire. ignition.

The cargo light Although possible, however, it would seem

unlikely that the holds would have been closed
When the after panels of the hatch covers were up with the light still switched on and without it
first opened up on 31 July, it could be seen that being noticed. More likely is that it was
an electrical cable was hanging from the upper unplugged at the time that the other lights were
access at the starboard after corner of No. 2 unplugged, but the fact that the cable still led
hold, down into the cargo of bagged beans. This into the hold, instead of being coiled up in the
was close to the area where the fire had been access space, went unnoticed.
first observed and to where it is believed to
have originated. See fig. 13. The ship’s staff response to the
In each hold access space there were a number fire
of halogen cargo lights, stowed with their The fire was first discovered at about 1130 on
cables, and the sockets into which these lights the morning of 28 July with a fire alarm being
would be plugged during cargo operations. All activated on the Autronica detection system, at
the plugs in the starboard aft access had been the same time as smoke was seen issuing from
removed from their sockets, this being in the No. 2 hold vents. The initial response by the

22
FIGURE 13: FIGURE 14:
Electric cable for cargo light hanging into no.2 hold Burned end of cargo light cable in no.2 hold
from hold access

ship’s staff was correct in that all ventilation in the hold was roughly 7 003 m3. The total
was stopped and all openings were sealed off. volume of No. 2 hold is 12 282 m3 hence, by
After a brief inspection by the mate and second volume, the hold was approximately 57 per cent
mate through the opened hold access and full.
playing a hose on the flames, the decision was
taken to flood with CO2 and the first of 25 With this volume of cargo loaded in No. 2, by
bottles was discharged at 1150. simple proportion, it would have been necessary
to release, as a minimum, 110 bottles of CO2 in
According to the instructions on CO2 release, order to raise the concentration of gas to
issued to the vessel by the manufacturers of the between 30 and 40 per cent; the level necessary
fixed firefighting installation, the number of to smoother the fire.
CO2 bottles required for flooding No. 2 hold
was 77 if the hold was full, and 116 if it was After discharging the first 25 bottles, releasing
half full. the gas was suspended for a while due to leaks
on several pipe connections in the CO2 room. At
No. 2 hold contained 3 048 tonnes of cocoa 1210, a forward ventilation flap on No. 2 hold
beans with a stowage factor of approximately was opened to check on the effect of the CO2
1.95 m3 per tonne. Hence the volume of cocoa flooding, by analysing the oxygen content of the
bean cargo was approximately 5 950 m3. It also issuing air. However, as insufficient gas had
contained the two machinery modules in the been released for even a full hold, and as CO2 is
forward part of the hold. Allowing an heavier than air, it is unlikely that any
approximate volume for these two crates as significant effect on the oxygen level would
220 m3 and taking the volume occupied by the have been recorded at the ventilation flap on the
steel pontoons, from the specifications, as main deck. A large proportion of the gas would
833 m3, the total volume of cargo and pontoons have been discharged into the virtually empty

23
part of the hold, forward of the pontoon should never be skimped. Technically, provided
bulkhead, and descended to the bottom of the that hold sealing is adequate, it would be possible
hold. In doing so, it would slowly displace the for a ship with cargo heating in all her holds, to
sail safely across the Pacific Ocean with her CO2
air upwards towards the top of the hold.
supply exhausted (assuming a sufficient reserve
for the engine room). However, such action
A little more than half an hour later, at 1245, the would only be recommended if sealing efficiency
hold access hatch was opened for an inspection. could be guaranteed. Under normal circum-
Acknowledging that the action to be taken is stances where there is obvious progressive
always a matter of judgment for the master or heating a ship would be recommended to a port
officer in charge, in the opinion of the Inspector, of refuge to obtain adequate CO2 supplies. This
this was too short a period to allow before any often involves fitting a bulk tank containing
several tonnes of CO2. If considered necessary,
hold penetrations should be opened, for further sealing should be performed whilst the
whatever reason. ship remains in port.

It was not until 1715 that afternoon that 86 There is often a tendency for personnel trained
bottles of CO2 had been discharged, with the in firefighting to exercise an ingrained urge to
remaining 11 bottles not being discharged until get to grips with the seat of a fire with the least
the following day. It is unlikely that, at any possible delay. In a well-respected reference
time, there was sufficient CO2 within the hold to book on fighting ship fires, ‘Fire Aboard’, the
effectively stifle the fire which, indeed, was still author states that there is ‘…much to be said for
active after the vessel arrived at Albany. playing the waiting game’.9 He goes on to say:
A safety bulletin published by the UK P&I Club If 95 per cent of a cargo can be saved by a
on 13 August 1996 describes at some length the certain line of attack and one which involves also
minimal risk of structural damage to the vessel
measures to be taken in the case of self-heating,
itself, what matters it that final extinction of the
or fire, in a cargo. The bulletin deals specifically fire may take 48 hours or so? …..He would urge,
with bagged fishmeal cargoes, but the principles rather that, wherever a fixed extinguishing
involved apply to combustion in any bagged, system is fitted in an endangered hold, it be
organic cargo. The bulletin emphasises the actuated at once and the hold in question be kept
importance of sealing, and keeping sealed, the tightly battened down until temperature, smoke
affected hold. It states: and other tests indicate that the fire has been
beaten down or suppressed.
It is basically undesirable to inject less CO2 than
is recommended in the manual (provided by the The port authority response
manufacturers of the CO2 system) even though
this means that only a few cargo spaces can be so The ‘waiting game’ was indeed played when
treated. Marion Green arrived alongside in Albany on
29 July, although there was considerable
It should be appreciated that any cargo heating
results from an oxidation process. This means surprise and consternation on the part of the
that the oxygen concentration in a hold is ship’s master and officers that the local fire
depleted and the concentration of nitrogen, an brigade was not in attendance on the wharf
inert gas, increases. Hence in a sealed hold, cargo when the vessel berthed. The brigade, having
heating tends to be self-quenching. It is therefore not been advised of the ship’s correct arrival
of paramount importance that the Master has all
time by the agent, arrived to assess the situation
necessary materials on board to allow very
efficient sealing of cargo spaces in order to approximately 45 minutes later. Shortly after the
minimise atmospheric interchange. Very efficient vessel’s arrival, a meeting was held between the
sealing may be a time-consuming operation but master, the harbourmaster, a representative from

9
‘Fire Aboard’ by Frank Rushbrook CBE, 3rd Edition 1998, Brown, Son & Ferguson Ltd. Glasgow.

24
Customs and the ship’s agent. The master’s determine the future course of action. No
initial consternation was compounded by the decision had been taken at that stage as to what
impression which he gained that the meeting could be done with the damaged cargo.
was focussed far more on administrative
procedures and paperwork than on the measures The meeting revealed several conflicting
which should be taken to deal with the fire. interests. The master and the vessel’s owners
wanted the fire extinguished as soon as possible
Albany is a small port and the unit of the Fire and the cargo discharged. The Australian
and Emergency Service of Western Australia, Quarantine and Inspection Service (AQIS) did
based in the town, is also numerically small and not want the cargo, which was not destined for
has little experience of ship fires and, indeed, Australia, to be landed in Australia. This
little familiarity with ships. From the time of reluctance was shared by the harbourmaster,
Marion Green’s arrival it was evident that there who did not want 3,000 tonnes of burned,
was a lack of co-ordination in the response to soaked, cocoa beans landed on the wharf. The
the incident. The master expected the fire harbourmaster also wanted Marion Green off
brigade to take charge of the firefighting, while the berth to enable the waiting tanker, carrying
the fire brigade expected the master to tell them fuel for the Albany area, to berth and, in
what measures he wanted taken. The fire addition, there were concerns from the
brigade suggested that the hold should be filled Environmental Protection Agency about any
with high-expansion foam but this was vetoed possible environmental impact which might
by the P&I Club surveyor who was concerned result from the incident.
about further damage to the high-value cargo.
Eventually, however, it was agreed that there
The vessel’s owners in the Netherlands
was no option but to discharge the cargo.
consulted with the Rotterdam port fire brigade
to obtain advice on how to deal with a fire in a
cargo of cocoa beans. It appeared, however, that
Albany Port Authority emergency
there had been little experience in fighting such procedures
fires and the advice forwarded to the master was The Albany Port Authority is a corporate body
to flood the hold with more CO2, to inspect the constituted under the Port Authorities Act 1999
hold later, then flood with more CO2, until the and is responsible to the Minister for Planning
fire had been overcome. & Infrastructure. Under the Act, the Port
Authority is required to have in place a compre-
On the morning of 30 July, the P&I Club
hensive Marine Safety Plan which is to include
surveyor requested that No. 2 hold be opened up
an Emergency Response Plan. At the time that
but this request was refused by the harbour-
the incident involving Marion Green occurred
master who, correctly, required a further 24
hours monitoring of the situation and who this plan was in draft state and had not been
wanted to wait for the bulk CO2 to arrive. signed off by the Minister. (However such an
Emergency Response Plan was in place prior to
On the following morning, 31 July, the hold the publication of this report.)
temperature appeared to be rising in spite of the
earlier use of the bulk CO2. There had been, Albany Port Authority did have in place, at the
however, some confusion as to whether the bulk time of the incident, emergency procedures for
CO2 was actually getting into the hold, since fire, serious injury, collapse, chemical spillage,
there was uncertainty regarding the correct explosion or bomb threat. The procedures,
alignment of various valves on the tank and the however, contained very little and were
accuracy of the contents gauge. essentially directed at the wharf labour force.
The only procedures relating to fire instructed
At 1400 that afternoon, a meeting of involved Supervisors to clear the area of labour and order
parties was convened by the harbourmaster, to them to the amenities block, to call the fire

25
brigade, to advise the harbourmaster, to hold a and the Northern Territory on their
roll call and liaise with the ship’s master. understanding of the answer to this question.
The responses varied enormously, some stating
It is the policy of the Albany Port Authority: that the fire brigade had complete authority and
…to achieve Best Practice in the management of others that the master retained complete
all risks that threaten to adversely impact the authority.
Authority, its customers, people, assets,
functions, objectives, operations or members of The response from Western Australia indicated
the public. Risk Management will form part of that, in the opinion of the Fire and Emergency
operational and line management responsibilities Services Authority of Western Australia, the
and be integrated into the Business Planning
master or chief engineer of the vessel remains
Process10.
responsible for the vessel at all times. The
It is to the credit of the Albany harbourmaster response added, however, that in regard to the
that he did accept Marion Green into the port as legal aspects, the Authority believed that ports
a port of refuge. The issue of ports of refuge, are beyond the Fire District and remain under
following the Erika and Prestige disasters in the control of the harbourmaster.
Europe, is one which is the cause of much
Consideration of this issue revealed a number of
concern and contention both among countries of
anomalies between the various laws of the
the European Union and at the International
different states and Commonwealth law. Several
Maritime Organisation. states have laws clearly detailing the authority
of the fire brigade, whereas the authority of the
In terms of safety, the operation was a success,
master, although traditional, is less clearly
the fire was extinguished without injury or loss
defined in terms of legislation. However,
of life and the ship was able to resume its
sections 6 and 278 of the Navigation Act 1912
voyage, undamaged, and with the cargo in No. 1
imply that ultimate responsibility does rest with
hold intact. the master.

Authority in fighting ship fires in To further pursue these complex legal anomalies
port is beyond the scope of this investigation. In the
opinion of the Inspector, however, what is clear
The question of who has ultimate authority (and
is that, whatever anomalies exist, where a
responsibility) in the case of a fire aboard ship
trained fire brigade is involved in fighting a ship
in port, the master, the harbourmaster or the
fire, the fire brigade should be regarded as the
senior fire brigade officer on the scene, is not a
authority on firefighting procedure, yet working
simple one. It is a question which has arisen a
in close co-operation with the ship’s staff and
number of times in the past, (eg ATSB Report
harbourmaster.
No. 142, Fire aboard the Australian flag tanker
Tasman), and is one about which there is great In the Port of Albany’s Emergency Response
uncertainty in the minds of many on those Plan, promulgated since this incident but before
involved. the publication of this report, the roles of the
harbourmaster, the ship’s master and the
In 1999, the Inspector of Marine Accidents
emergency services in such incidents have been
canvassed the senior officers of the fire and
more clearly defined.
emergency services of all the Australian states

10
Albany Port Authority, Internet www.albanyport.com.au/risk_management_policy

26
 considerable evidence of this in the form
Conclusions of empty cigarette packets, cigarette ends
and matches found in the hold access
spaces and amongst the cargo once it had
These conclusions identify the different factors been discharged onto the wharf at Albany.
which may have contributed to the cargo hold d) A cargo light left in the hold after the
fire aboard Marion Green and should not be completion of cargo loading.
read as apportioning blame or liability to any
particular individual or organisation. 2. The discovery of discarded smoking materials
in the discharged cargo indicates that the
1. It was not possible to positively identify the vessel’s ‘no smoking’ policy around the
source of the fire in No. 2 hold, although holds was not sufficiently policed by the
four possibilities existed. These were: deck watch.
a) Self-heating of the cargo of cocoa beans 3. With the volume of cargo in No. 2 hold, and
promoted by the growth of a fungus, the number of CO2 bottles discharged before
Aspergillus Niger. This may have been the vessel arrived at Albany, it is unlikely
initiated at a damp patch caused by that the concentration of gas in the hold was
spillage of the liferaft water, taken by the at any time sufficient to stifle the fire. The
loading gangs working in the holds, and number of bottles released was less than
exacerbated by poor ventilation in an area indicated in the instructions provided by the
where the ventilation channels did not manufacturer of the CO2 system.
allow sufficient air flow. 4. Information provided to the ship’s staff on the
b) Combustion initiated by the decompo- shipping and stowage of cocoa beans, their
sition of aluminium phosphide into fumigation and the associated hazards was
phosphine gas used for fumigation. inadequate.
Combustion is known to have occurred in 5. The response to the fire, once the vessel had
other instances, particularly when there is berthed in Albany, was lacking in co-
an imbalance in the phosphorus content ordination and there was no clear
during preparation of the chemical. understanding of who had the authority and
c) Smoking materials discarded in the hold responsibility for actions to be taken in
during loading of the cargo. There is dealing with the fire.

27
28
Recommendations

MR20030037
That Beluga Genchart BV take measures to
ensure that a strict ‘no smoking’ policy in the
vicinity of cargo operations is enforced on board
their vessels.

MR20030038
That Beluga Genchart BV ensure that ship’s
masters are provided with all relevant
information regarding the stowage of cocoa
beans, their fumigation and any associated
hazards. As a minimum, the IMO’s
‘Recommendations on the safe use of pesticides
in ships’ should be available on board any ship
loading a fumigated cargo.

MR20030039
That shippers, stevedores and ship’s officers
should all ensure that, when bagged cocoa bean
or other organic cargoes are being loaded,
adequate ventilation channels are provided and
that cargo is stacked with sufficient ‘cross ties’
to ensure that the channels remain clear during
the voyage.

MR20030040
That Beluga Genchart BV take measures to
ensure that, on completion of loading cargoes,
the deck watchkeeping officer makes an entry in
the deck logbook confirming that all electrical
equipment in the holds has been isolated and
stowed as applicable.

29
30
Submissions
Under sub-regulation 16(3) of the Navigation
(Marine Casualty) Regulations, if a report, or
part of a report, relates to a person’s affairs to a
material extent, the Inspector must, if it is
reasonable to do so, give that person a copy of
the report or the relevant part of the report. Sub-
regulation 16(4) provides that such a person
may provide written comments or information
relating to the report.

The final draft of the report was sent to the

following:

The Master, Marion Green

The Mate, Marion Green

Beluga Genchart BV

The Harbourmaster, Albany Port Authority

The Area Manager, Fire Services of WA

Australian Maritime Safety Authority

Submissions were received from Beluga

Genchart BV and from the Harbourmaster,
Albany Port Authority. The text of the report
has been amended where appropriate.

31
32
Marion Green
IMO Number 9164029

Flag Netherlands

Classification Society Lloyd’s Register

Ship Type Multi-purpose general cargo

Builder Schelde Scheepsnieuwbouw B.V. - Vlissingen

Year Built 1999

Owner C.V. Marion Green

Ship Manager Beluga Genchart B.V., Rotterdam

Gross Tonnage 11 894

Net Tonnage 5 920

Summer deadweight 17 050 tonnes

Summer draught 9.689 m

Length overall 142.81 m

Breadth 21.5 m

Moulded depth 13.3 m

Engine 8 cylinder Wärtsilä 8L46B four-stroke diesel

Power 7 800 kW

Crew 13 – mostly Dutch & Filipino nationals

33
34
 Marion Green. 11.03
Independent investigation into the fire in the hold of the Netherlands flag general cargo vessel Marion Green
off the coast of Western Australia, on 28 July 2002
ISSN 1447-087X
ISBN 1 877071 42 0
1800 621 372
www.atsb.gov.au