Report

IN THE HIGH COURT OF JUSTICE (ADMIRALTY COURT)

REPORT OF THE RE-OPENED FORMAL INVESTIGATION INTO THE LOSS OF

THE MV DERBYSHIRE

THE HONOURABLE MR JUSTICE COLMAN

XXXX

REPORT OF THE RE-OPENED FORMAL INVESTIGATION INTO THE LOSS OF THE M/V DERBYSHIRE

i

Contents

Executive Summary .........................................................................................1

1. Introduction ......................................................................................................25

2. The Vessel ........................................................................................................35

3. Trading Activities and Last Voyage.................................................................43

4. Navigation ........................................................................................................51

5. Sea Conditions and Vessel Movements on 9th September 1980 ......................95

6. The MARIN Model Tests ..............................................................................101

7. The Relevant Evidence from the Wreckage...................................................109

8. Frame 65 Defects ...........................................................................................127

9. Inferences as to the Cause of the Loss ...........................................................137

10. Development of International Ship Design Regulations and

Classification Society Rules...........................................................................145

11. Towards Improved Ship Safety: Hatch Cover Strength and

permissible Freeboard ....................................................................................163

12. Towards Improved Ship Safety: Navigation..................................................185

13. Towards Improved Ship Safety: Other Matters .............................................189

14. Summary of Recommendations .....................................................................199

15. Availibility of Underwater Survey Capability ...............................................203

16. Answers to Questions.....................................................................................209

Appendices

XXXX

XXXX


ii

Parties to the Re-Opened Formal Investigation

The Attorney-General

Mr Michael Thomas CMG QC

Mr Nigel Meeson

Mr Guy Blackwood

Instructed by the Treasury Solicitor

Solicitor: Laurance O’Dea

Derbyshire Families Association

Mr Andrew Moran QC

Ms Karen Troy-Davies

Instructed by Messrs Boote Edgar Esterkin

Solicitor: Stephen Cantor

Swan Hunter SHSEGL Real Ltd

Mr RF Stone QC

Mr Simon Bryan

Instructed by Messrs Mills & Co

Solicitors: Geoff Mills


iii

Bibby Lines Ltd

Mr Michael Howard QC

Mr Robin Hay

Instructed by Messrs Holman Fenwick & Willan

Solicitor: Alan Walls

Lloyd’s Register of Shipping

Mr Bernard Eder QC

Ms Claire Blanchard

Instructed by Messrs Elborne Mitchell

Solicitors: Tim Akeroyd, Paul Green

The Department of the Environment, Transport and the Regions

Mr Charles MacDonald QC

Mr Nigel Cooper

Instructed by The Treasury Solicitor

Solicitor: Jacqueline Duff


iv

Inquiry Personnel

Secretariat

Laurance O’Dea (Inquiry Solicitor)

Jan Gladysz

Keith Barnes

Paul Mowbray

Adrian Drury

Maxine Ross

Jai-J Moss

Anita Bhatia

Ashley Dayman

Sue Doherty

Court Staff

Alison Tighe (Judge’s Clerk)

Peter Coates (Usher)

Peter Wilson (Usher)

Dave Reynolds (Court Service)

Technical Staff

Will Handley

Wil Shave

Megan Dale

Daniel Woolstone

Transcript Operators

Amanda MacPherson

Lisa Minsky

Jaqueline Gleghorn


Executive Summary


1

Executive Summary

Introduction

1 The tragedy and mystery of the DERBYSHIRE began in September 1980

when that enormous vessel disappeared without trace in the Pacific about 350

miles south east of Japan. She was an oil/bulk/ore carrier (O.B.O.), in length

nearly three times the length of the playing area at Wembley Stadium and

considerably larger in width than the Titanic or the QE2. All those on board

perished – 42 members of crew and two of their wives. She was the largest

British ship ever lost at sea.

2 Yet the DERBYSHIRE was only four years old, apparently well maintained

and manned by a competent and very experienced master and crew. She had

been built by highly reputable shipbuilders, namely, Swan Hunter at their

Haverton Hill Yard on Teeside. She was fully classed with Lloyd’s Register

of Shipping. All her surveys and certificates were up to date.

3 She was on a voyage from Quebec to Japan laden with iron ore concentrates.

She was within a few days of arrival at Kawasaki when she encountered a

tropical revolving storm known as Typhoon Orchid moving from the east.

Her northerly route was crossed by the worst sea conditions associated with

the typhoon on 9th September 1980. By 0300z on that day she reported that

she was hove to in a violent storm with a force 11 wind and a wave height of

30 feet. The vessel’s last radio message was sent at 1019z that day. This

Report finds that by 1700z on that day the conditions had deteriorated: the

average height of the highest one third of the waves (the significant wave

height) was 10.86 metres and the wind speed was over 56 knots.

EXECUTIVE SUMMARY

2

4 Such conditions are severe but not exceptional amongst north west Pacific

typhoons. Vessels of the size and design of the DERBYSHIRE were at the

time assumed to be quite capable of withstanding such conditions, even if they

had to reduce speed or be hove to. Nevertheless, as a matter of good

seamanship, masters would always do their best to avoid coming within 50-75

miles of the storm centre and would generally try to maintain a distance of at

least 200 miles.

5 Not surprisingly, the families of those who perished in this tragedy were

extremely anxious to know what had caused such a catastrophic loss of life.

They wanted a formal investigation. The UK Government took the view that,

because of the total absence of material evidence, a formal investigation could

not be expected to establish clearly the cause of the loss.

6 The DERBYSHIRE was the last of six bulk carrier sister ships built by Swan

Hunter at Haverton Hill Yard. The first – the FURNESS BRIDGE – was

completed in September 1971.

7 Following casualties which were sustained by the second of the series of sister

ships – the TYNE BRIDGE – in 1972 and by the third of the series – the

KOWLOON BRIDGE – in 1986, both involving fractures in way of Frame 65

located at the after end of the No.9 hold, just forward of the superstructure and

engine room, the finger of suspicion as to the cause of the loss of the

DERBYSHIRE was strongly pointed at a possible design defect in that part of

the vessel. Consequently, in December 1986 a Formal Investigation was

appointed under Mr Gerald Darling QC as Wreck Commissioner.

8 The hearing of that Formal Investigation took place during 46 days between

October 1987 and March 1988 and the Report was issued on 18th January

1989. The formal decision was:

“For the reasons stated in this Report the Court finds that the

DERBYSHIRE was probably overwhelmed by the forces of nature in


3

Typhoon ORCHID, possibly after getting beam on to wind and sea, off

Okinawa in darkness on the night of 9th/10th September 1980 with the

loss of 44 lives. The evidence available does not support any firmer

conclusion.”

9 Amongst the reasons for this conclusion were:

(i) Structural failure due to excessive loading at amidships was

highly improbable even under the severe wave loading

conditions and since at Fame 65 the vessel was relatively

stronger than at amidships and that bending moments were at a

maximum close to amidships, it was also highly improbable

that there had been structural failure at Frame 65.

(ii) Fatigue cracking was unlikely to have caused the total loss of

the ship by failure of the hull girder, either at amidships or in

way of Frame 65.

(iii) Separation of the hull at Frame 65 due to brittle fracture must

be regarded as very unlikely though some element of

uncertainty remained.

(iv) Although the flooding of the bosun’s store and chain locker

followed by loss of freeboard by the bow and then sequential

failure of the No.1, 2 and 3 hatch covers due to hydrostatic

loading was a possible cause of the loss, it was not thought

probable.

10 The members of the Derbyshire Families Association were greatly

disappointed that the Report was so inconclusive. They pressed for an

underwater survey but the exact location of the wreckage was not known and

until the 1990’s there was no adequate technology for obtaining sufficiently

clear photographic evidence at a probable depth of some 2½ miles.

EXECUTIVE SUMMARY

4

11 Eventually in 1994 the ITF funded an underwater search to try to locate the

wreckage. Against all the odds this succeeded. The depth was 4200 metres,

some 2½ miles.

12 The Department of Transport then appointed Lord Donaldson of Lymington to

carry out an assessment as to what further steps could be taken to obtain

evidence as to the cause of the loss, the cost of taking such steps and what

benefit to ship safety could be secured if the cause of the loss were

established.

13 In his Assessment, Lord Donaldson concluded that the cost of an appropriate

underwater survey would be about £2 million and that such a survey ought to

be mounted in the interests of international ship safety. His recommendation

was accepted by the Minister of Transport and the UK/EC Survey was duly

conducted, funded partly by the UK and partly by the EU.

14 That survey was conducted in two phases in 1997 and 1998 respectively by

the Woods Hole Oceanographic Institution by means of the United States

Research Vessel, THOMAS G THOMPSON. Three Assessors were

appointed to oversee the survey and report upon the results, namely Mr Robin

Williams, Dr Remo Torchio of Genoa, both naval architects, and Professor

Douglas Faulkner, Professor of Marine Architecture and Ocean Engineering at

the University of Glasgow.

15 The technical achievement of that survey went far beyond anything in scope

and detail that had previously been attempted on any underwater wreck at a

comparable depth.

16 In the event, 135,774 individual electronic stills were obtained from a carpet

plot of over 98 per cent of the entire wreckage field, which measured 1500

metres x 1000 metres. The result was to identify 2500 separate items of

wreckage. By means of the joining up of the separate stills into larger pictures


5

(mosaicking) it was possible to obtain and usually identify very clear black

and white pictures of continuous expanses of wreckage.

17 High definition video filming was also carried out. Some 200 hours of video

were made.

18 Two of the three Assessors (Mr Williams and Dr Torchio, Professor Faulkner

having resigned) concluded on the basis of the photographic evidence that the

vessel’s loss had been caused by seawater entering the bow section which

caused the vessel to develop a forward trim, thereby exposing its No.1 hatch

covers to wave heights great enough to impose loading in excess of their

collapse strength. Water then poured into the large empty space above the

cargo in the No.1 hold. That put the vessel further down by the bow until

No.2 and No.3 hatch covers suffered the same fate sequentially. The vessel

would then sink.

19 However, the disturbing aspect of this Report was that the main reason for

entry of seawater into the bosun’s store in the first place was found to be the

failure of the crew to secure the lid to the hatch on the foredeck.

20 This conclusion clearly involved the imputation of serious negligence against

the officers and crew. It was deeply upsetting to the families of those on

board. However, the conclusion also acquitted the design and construction of

the vessel in way of Frame 65 of any causal contribution to the loss.

21 Following that report, on 17th December 1998, the Deputy Prime Minister,

The Rt Hon John Prescott, MP announced that there was to be a full re-

opening of the Formal Investigation and that this was to be held in the High

Court under section 269(1) of the Merchant Shipping Act 1995. I was

subsequently appointed to conduct that Investigation. This is the first time

that such a reference to the High Court has been made.

EXECUTIVE SUMMARY

6

22 The hearing commenced on 5th April 2000 and lasted for 54 days, concluding

on 26th July 2000. Twenty scientific and technical experts, six master mariners

and five other witnesses gave oral evidence. Robes had nothing to contribute

to the hearing and were therefore dispensed with. The parties to the

Investigation were:

The Derbyshire Families Association

Bibby Tankers Ltd, the shipowners

SHSEGL Realisations Ltd, the successors to Swan Hunter, the ship

builders

Lloyd’s Register of Shipping, (LRS) the classification society

The Department of the Environment, Transport and the Regions

(DETR)

The Attorney General and his counsel acted as in effect counsel to the tribunal.

23 Assessors to sit with the court were not appointed. Instead, Dr P S J Crofton

of Imperial College and Professor John van Griethuysen of University

College, London, were appointed as its technical and scientific advisers. They

took no part in the making of the ultimate decisions expressed in this Report.

24 It is to be recorded that without a high degree of commitment and co-operation

between all parties and between the many expert witnesses, it would have

been impossible to produce this Report as early as it has been.


7

The Cause of the Loss

25 The DERBYSHIRE would certainly not have been lost had it not encountered

extremely severe seas close to the centre of Typhoon Orchid. The first

question to be answered is therefore why the vessel ever found itself in these

conditions.

26 This Report concludes that, whereas the vessel would have avoided the

typhoon if, on 7th September or up to about 1000z on 8th September, the

master had altered course to the west south west or south west, instead of

proceeding on a northerly course towards Japan, his failure to do so almost

certainly arose from his belief that the sea conditions along his route on 8th and

9th September would enable him to maintain sufficient speed to keep well

ahead of the typhoon which was then advancing on an east-west track.

27 The vessel was receiving weather forecasts from the United States Navy/Air

Force Joint Typhoon Warning Centre broadcast by Guam radio and from the

Japanese national weather forecasting service over Tokyo radio. The vessel

was also receiving occasional weather reports from Oceanroutes, a routeing

agency appointed by the charterers. One such report was sent at 2146z on 5th

September, before Typhoon Orchid had developed, and the next at 0113z on

8th September.

28 At no time did Oceanroutes advise the master to alter course to avoid the

typhoon.

29 The forecast track of the typhoon given by Guam radio on 7th and 8th

September differed substantially from that given by Tokyo radio: on 7th

September Guam put the track curving north west on 8th and 9th September,

whereas Tokyo radio put the track much further to the west. Further, the

Guam forecasts between 1200z on 7th September and 0600z on 8th September

very substantially under-estimated the windspeeds ahead of the centre of the

typhoon. One of several unusual characteristics of Typhoon Orchid was that

EXECUTIVE SUMMARY

8

the wind field and area of consequential high swell running in front of it were

abnormally extensive. Another unusual feature was its extremely long

duration.

30 The continuing under-prediction by the Guam forecasts of the extent of the

wind field running before the typhoon and their failure to anticipate that the

vessel would encounter 50 knot winds over 200 nautical miles ahead of the

storm centre and 30 knot winds over 400 nautical miles ahead of it would have

led the master to believe not only that the DERBYSHIRE would be able to

maintain sufficient speed to keep ahead of Orchid but that, if, on 8th

September, it became necessary to change course to the south west or west

south west, that option would not be lost due to the danger of turning beam on

in severe sea conditions.

31 Further, the fact that Oceanroutes at no stage advised the master to alter course

would have been understood by him to be a tacit endorsement of his keeping

to his present course.

32 Since the master could not reasonably have anticipated from the local wind,

sea and isometric conditions in the period from 1200z on 7th to 1000z on 8th

September that the vessel would not be able to maintain sufficient speed to

keep ahead of the typhoon, then closing from the east south east as much as

250 nautical miles from the vessel, no criticism can be attached to his decision

to hold to his rhumb line course or subsequently at some stage after 0300z on

8th September to take a slightly more northerly course probably in order to put

the sea 4 points off the starboard bow.

33 The DFA has strongly criticised Oceanroutes for failing properly to perform

its duty as a routeing agency in as much as it failed on 7th and early on 8th

September to provide the vessel with adequate information as to the track of

the typhoon and to advise the master to alter course to the west south west or

south west.


9

34 This Report accepts that the message sent to the vessel by Oceanroutes at

0113z on 8th September was seriously deficient in information. However, this

Report does not accept that it has been established that Oceanroutes were

negligent or otherwise at fault in failing to advise the master to alter course. In

as much as it has not been shown to be negligent for a typhoon forecasting

service to have failed to anticipate the unusual forward projection of the wind

field of Typhoon Orchid, it was not unreasonable for a routeing agency to

continue to endorse the vessel’s present course up to 1000z on 8th September.

Failure to anticipate the extent of the wind field and the unusually distant

spread of heavy swell running before the storm centre is not shown to have

been professionally negligent. Even if the message sent at 0113z on 8th

September had contained adequate information then available, that

information would not have caused the master to change course. It can

therefore be concluded that the vessel’s encounter with the most severe sea

conditions of the typhoon arose without the fault either of those on board or of

Oceanroutes.

35 However, the brochure issued to masters by Oceanroutes in which it set out

what routeing services it would provide in respect of typhoons was obscurely

worded. If Oceanroutes intended to suspend positive routeing advice in such

conditions, that should have been much more clearly expressed.

36 The vessel encountered such severe conditions by about 1800z on 8th

September that she was hove to, that is to say the master’s speed would be the

minimum necessary to maintain steerage, so as to keep the seas on the vessel’s

starboard quarter. From that time onwards, as the storm centre closed on her

course from the south east, she was inevitably going to be subjected to very

severe sea conditions.

37 These conditions had deteriorated in the course of 8th and 9th September. The

hindcast evidence is to the effect that for about 12 hours from 0500z on 9th

September the route of the DERBYSHIRE experienced significant wave

EXECUTIVE SUMMARY

10

heights above 9.5 metres, rising to a peak of 10.85 metres by 1700z. The

vessel’s average forward speed in these conditions would be in the range 1.50

to 2 knots over the ground depending on wave drift forces. That would have

taken the vessel to the wreckage site by about 1700z to 2000z on that day.

38 When a vessel is near to head on to the waves it will be subjected mainly to

pitching effects. The extent of these effects is dependent on the relationship

between the length of the vessel and the wave length. It will also be

dependent on the damping characteristics of the vessel’s configuration,

notably the relationship of draught to beam. When the wave frequency is

relatively low and the wavelength relatively large, the vessel will tend to

follow the slope of the waves with the effect that there will be little relative

vertical motion. By contrast, where there is a high wave frequency and a short

wave-length relative to the vessel’s length, the vessel will not respond to the

waves in terms of pitch and heave, thereby creating relative vertical motion

broadly equal to wave elevation. However, between those positions, as the

wavelength shortens, the slope of waves of the same wave height increases.

That causes the vessel’s pitch response to increase. There eventually comes a

point when the wavelength is broadly equal to the length of the vessel. It is at

this point that the vessel experiences maximum relative vertical motion. A

pattern of movement is thereby created which involves the vessel pitching

head-down into the trough of a wave and then, before its bow has lifted back

to horizontal, meeting the next wave crest.

39 Where the waves are very high, they tend to have very steep wave fronts

which allow little time for a ship to rise above them. The effect of the vessel’s

pitching in the manner described is to increase the relative vertical motion of

the vessel quite substantially. Where relative vertical motion exceeds the local

freeboard, green water will begin to load the vessel’s bow. The more the

relative vertical motion, the greater will be the head of green water on the

bow.


11

40 The length of the DERBYSHIRE (at 294.1 m overall) was slightly less than

the wavelength (about 300 m) that would typically have prevailed during the

period 1700z to 2000z on 9th September. The conditions therefore gave rise to

accentuated pitching and considerable amplification of relative vertical motion

to wave amplitude, so exposing the bow and forward hatch covers to heavy

and repeated green water loading; a response by the vessel strongly evidenced

by the model tests.

41 These model tests were conducted at the MARIN Research Institute,

Wageningen, Netherlands, and were designed to measure the magnitude of

green water impacts on the No.1 hatch cover in different conditions of trim

and at different significant wave heights. The various conditions of trim were

designed to ascertain what difference it made to green water loading on the

No.1 hatch cover that water had entered the bosun’s store and machinery

spaces, the forepeak ballast tank and the forepeak fuel tank. They were also

directed to ascertaining the rate of seawater entry to those spaces if ventilators

and air pipes or their covers had been destroyed.

42 The results of those tests indicated that at speeds of zero or 2 knots, without

water ingress, at the peak significant wave height of 10.85 metres, and even at

12.5 metres, the maximum loading on No.1 hatch would be well below the

collapse strength. However, in the damaged condition, with the bow flooded,

the maximum measured impacts on No.1 hatch at a significant wave height of

10.85 metres exceeded its collapse strength, even at zero speed.

43 In order to ascertain the probability of the No.1 hatch being subjected to green

water loading in excess of its collapse strength in sea conditions during

periods of time comparable to those experienced by the vessel and in different

states of bow flooding, it was necessary to extrapolate from the new model test

data by statistical methods. This exercise was conducted by the court-

appointed expert Professor Jonathan Tawn of the University of Lancaster. His

extrapolations indicate that at the most probable speeds of 70% maximum and

EXECUTIVE SUMMARY

12

2 knots, there was a very substantial risk of a hatch-breaking wave at an un-

enhanced significant wave height when both the stores and the ballast tank had

been flooded and a considerable risk when the ballast tank alone had been

flooded, whereas there was no risk if the stores alone had been flooded.

Obviously, the magnitude of the risk would be commensurately reduced if

there were only partial flooding of the ballast tank together with full flooding

of the stores.

44 If, however, there were a 10% increase in the significant wave height, not only

was it certain that there would be a hatch-breaking wave when the stores and

ballast tank were flooded, but there was a 1% to 10% risk that such a wave

would eventuate if, at 70% maximum speed, the stores alone were flooded

(Dam Stores single max) and a 4% to 42% risk of such a wave in that

condition if a speed of 2 knots were assumed. The expert evidence as to the

hindcast wave height was subject to a 10 per cent probability range.

45 The evidence derived from the underwater survey of the wreckage and the

inferences to be drawn from it can be summarised as follows:

(i) The wreckage of the bow section was found approximately 590

metres from that of the stern section with most of the wreckage of

the hatch covers within a corridor between the two sections about

230 metres wide. That pattern suggests that no substantial section

of the hull and none of the hatch covers separated from the main

part of the vessel on the surface.

(ii) All hatch cover panels located in the wreckage exhibited a similar

initial mode of failure and some exhibited a similar secondary

mode of failure. In particular, their initial mode of failure

consisted of a symmetrical pattern of bending inwards along an

axis transverse to the hull located at or forward of the centre girder.

The secondary mode of failure consisted of an asymmetrical


13

pattern of inward bending along a longitudinal axis very close to

the centre line of each panel. The transverse inwards bending

occurred before the longitudinal inwards bending. This condition

is consistent with the initial mode of failure being caused by green

water loading from the forward end and inconsistent with the

vessel having developed any significant list by the time sinking

began. Sea water would have entered the holds once the transverse

bending had occurred.

(iii) The hatch covers were in good condition with seals not excessively

worn. It was unnecessary for the centre line catches to be secured,

provided that the quick acting cleats were secured, which was the

case.

(iv) Ventilators and air pipes located on the foredeck and leading into

the bosun’s store, machinery space, ballast tank and possibly in one

case, the fuel tank, were in damaged condition and had sustained

that damage prior to the commencement of sinking.

(v) The seals on the spurling pipes were destroyed. This probably

occurred before sinking. Water would thereby enter and flood the

chain lockers.

(vi) Seawater would therefore have entered the bosun’s store, the

machinery space, the ballast tank and, possibly, to a lesser extent

the forward fuel tank

(vii) Seawater could not have entered the fuel tank prior to sinking

either by reason of the collapse of the floor of the bosun’s store or

through the manholes leading from the store. The covers to those

manholes would not have been left open. Nor would water have

entered the ballast tank through the access manhole from the

EXECUTIVE SUMMARY

14

bosun’s store, for it is distinctly improbable that it would have been

opened much less left open in the course of the voyage.

(viii) The lid to the foredeck bosun’s store hatch was missing from the

wreckage of the bow section. The after side of the hatch coaming

was severely dented and split. The toggle wing nuts on that hatch

were found in various conditions and at differing positions on their

threaded shanks, with some missing altogether.

(ix) The condition of the store hatch as found does not suggest that the

lid was left unsecured by the crew or that the lid could not properly

be closed because a rope was protruding from the hatch. On the

whole of the evidence the lid was adequately secured both by

properly tightened toggles and by a complicated roping device

designed to prevent the lid coming loose because the wing nuts had

ridden up the toggle shanks with the motion of the ship. The rope

seen in the wreckage to emerge from the hatch was a mooring rope

one end of which was originally attached to the inside of the hatch

lid. This Report rejects the Assessors’ conclusion that the crew had

left the hatch lid inadequately secured prior to the DERBYSHIRE

entering the typhoon.

(x) The starboard windlass is shown to have become detached from the

foredeck. Its position on the seabed suggests that severance

occurred immediately before or soon after the commencement of

the sinking process. This was probably caused by its being

subjected to a succession of powerful waves the impact of which

initiated low cycle fatigue cracks in the welding where the windlass

was joined to its bed. These would eventually open when

subjected to subsequent intense wave loading. This probably

occurred after the commencement of water entry to the bow


15

through damaged ventilators and air pipes and may well only have

occurred after the failure of the No.1 hatch cover.

(xi) The loss of the lid to the bosun’s store hatch and the loss of the

rope leading from it were probably caused by the impact of the

whole or part of the starboard windlass after the latter had become

detached from its seating and after seawater had already begun to

flood the store, machinery space and probably the ballast tank.

(xii) The condition of bulkhead 339 as found in the wreckage does not

support the proposition advanced by the UK/EC Assessors that the

bow spaces and particularly the fuel oil tank ullage space were

substantially full of seawater when the vessel began to sink. Rather

its condition is consistent with the deep fuel tank and perhaps part

of the ballast tank having been substantially empty of water when

sinking commenced, but having admitted additional water during

sinking.

(xiii) The wreckage evidence is inconsistent with cracking of the deck or

hull in way of Frame 65 having made any causal contribution to the

loss. However, there is shown to have been a substantial

misalignment of longitudinal bulkheads at bulkhead 65. This error

in construction when added to the design discontinuity at this

position created a very low additional safety risk on account of the

danger that it would cause fatigue cracks in the after bulkhead of

No.9 hold which could cause oil leakage into the pumproom

followed by explosion.

46 On the basis of the condition of the wreckage and of the data derived from the

model tests conducted at MARIN, it can be concluded with reasonable

confidence that the initiating cause of the loss was the destruction of some or

all of the ventilators and air pipes located on the foredeck by sustained green

EXECUTIVE SUMMARY

16

water loading over many hours in the course of 8th and more probably 9th

September 1980. Water was thereby able to enter the bosun’s store,

machinery spaces and probably the ballast tank in substantial quantities and

possibly to a minor extent the fuel tank. The DERBYSHIRE then developed a

trim by the bow which, although imperceptible from the bridge, had the effect,

as the bow dropped lower and lower, of accentuating green water loading on

the No.1 hatch cover as the sea conditions became more severe in the course

of that day. By about 1700z those conditions had deteriorated so greatly that

there was likely to have been green water loading in excess of the collapse

strength of No.1 hatch cover. Once that hatch cover gave way, water would

enter the No.1 hatch, very rapidly filling the large ullage space above the

cargo and thereby causing the vessel to go still further down by the bow by

another 3.7 metres. It is estimated that the filling of No.1 hold might take as

little as 5 minutes or as much as 16.5 minutes.

47 This flooding in turn caused the green water loading on No.2 hatch cover

progressively and rapidly to increase until it exceeded the collapse strength of

that hatch cover and water then entered No.2 hold.

48 No.3 hatch suffered the same fate. At that point the vessel was irretrievably

lost.

49 Although it is possible that the No.1 hatch cover collapsed under excessive

green water loading before there had been flooding of any of the bow spaces,

this is very unlikely indeed.

50 There is no reasonable likelihood that any of the other possible causes

considered by Lord Donaldson or by the UK/EC Assessors’ Report in any way

contributed to the loss of the vessel. If, which is extremely improbable, the

vessel ever went beam on to the wind and sea, as suggested as a possibility by

the Report of the Formal Investigation, that circumstance did not cause the

loss of the vessel.


17

International Ship Design Regulation and Design Standards:

Recommendations

51 The DERBYSHIRE was designed in compliance as to freeboard and hatch

cover strength with the regulations made by the United Kingdom Government

in 1968 – the Load Line Rules – which gave effect to most of the provisions of

the International Load Line Convention 1966, (ILLC 66), in particular

Regulation 16(2).

52 Bulk carriers were permitted to maintain a minimum freeboard of a dry cargo

vessel reduced by 60 per cent of the difference between the lower permissible

tanker freeboard and the higher permissible dry cargo vessel freeboard, subject

to having adequate internal flooding protection, so that the vessel could

survive one compartment flooding. Mild steel forward hatch covers were

required to have a minimum strength capable of loads of not less than 1.75

tonnes per square metre. This loading would give a collapse strength of 42

kPa. Bulk carriers designed by reference to these requirements were known as

B-60 carriers.

53 Although, in the course of the international conference which gave rise to

ILLC 66, the United Kingdom delegation put forward proposals for increased

minimum forward hatch cover strength of about 2.1 tonnes per square metre,

these proposals were dropped in view of the opposition of a majority of the 52

delegations. Having regard to the limited state of contemporary knowledge as

to the effect of green water loading, the UK Government cannot be criticised

for failing to secure agreement to its proposals.

54 The contemporary state of knowledge as to wave loading on forward hatch

covers which existed when the design for the DERBYSHIRE was developed

in 1968 and up to 1974 when the keel was laid did not suggest that the

minimum hatch cover strength requirements prescribed by ILLC 66 were

deficient or that a vessel designed to do no more than comply with those

EXECUTIVE SUMMARY

18

requirements might be unsafe. In that state of knowledge competent architects

could in 1974-76 properly rely on the requirements of ILLC 66 as a design

benchmark.

55 In the course of the 1980s the rising incidence of loss of bulk carriers caused

classification societies, in particular LRS, to investigate the adequacy of hatch

cover strength in conjunction with other design features of bulk carriers. The

International Association of Classification Societies (IACS) set up its Working

Party on Strength in 1992. By 1997 IACS had agreed a series of ten unified

requirements directed to single side skin bulk carriers of over 150 metres in

length. Amongst these UR S21 laid down minimum load and strength

requirements for hatch covers. There underlay the adoption of these new

classification requirements the belief, at least on the part of some members of

IACS, that the ILLC 66 requirements for hatch cover strength were not

adequate.

56 The effect of UR S21 was, in substance, to increase the minimum permissible

forward hatch cover strength from 42 kPa under ILLC 66 to 83 kPa,

equivalent to a corroded value collapse load of 8.3 metres, assuming good

design and application of the specified safety factor on first yield. The

DERBYSHIRE’s collapse strength was 4.2 metres.

57 This requirement was to take effect for bulk carriers contracted for

construction on or after 1st July 1998, but it was not to apply to existing

vessels or vessels under construction or already contracted for construction.

58 UR S21 expressed the minimum strength requirements for hatch covers by

means of a formula which quantified the minimum strength by reference to the

height of the freeboard.

59 Had the DERBYSHIRE been laden down to her marks, thus operating on her

minimum permissible freeboard, there would have been an unacceptably high

risk of her No.1 hatch being subjected to green water loading in excess of its


19

collapse strength, even at zero speed or at a speed as low as two knots and

even without flooding of the bow spaces, when encountering wave heights of

a similar order of magnitude to those met in Typhoon Orchid.

60 It can be concluded that the requirements of ILLC 66 are set at a minimum

level in relation to which there is a substantial risk of exceedance at the

forward hatch covers if a vessel such as the DERBYSHIRE is caught in a

typhoon similar to Orchid or in similar conditions and making any speed over

the ground and some risk if it is not. Consequently, these requirements could

only be regarded as “adequate” if that risk were very substantially discounted

because of the low incidence of loss experience since ILLC 66 came into force

and/or because of the additional cost that would be involved in increasing the

level of protection for new buildings and for existing vessels.

61 Taking fully into account the lack of hard evidence of loss experience, the

likely additional cost of fitting stronger hatch covers and the technical and

mechanical problems to which that might give rise, this Report is unable to

accept in the light of the loss of the Derbyshire that modern concepts of

appropriate safety standards for vessels and those on board can accommodate

as acceptable the level of risk of exposure to potentially catastrophic

consequences presented by ILLC 66.

62 This Report concludes that the minimum hatch cover strength requirements

laid down for forward hatches in ILLC 66 in conjunction with the prescribed

minimum permissible freeboard for B-60 bulk carriers of similar size to the

DERBYSHIRE are seriously deficient in the context of present day concepts

of acceptable safety levels. The precise extent of that deficiency should be

ascertained by means of a programme of further model tests to be conducted at

MARIN with reference to realistic extreme sea conditions.

63 Such a test programme has been agreed between the DETR and IACS and is

currently in progress. With a proper regard to the United Kingdom’s

EXECUTIVE SUMMARY

20

leadership in matters of international maritime safety this programme of tests

is being funded by the DETR.

64 Although the results of those tests are not yet known the strong probability is

that the ILLC 66 hatch strength requirements will be found to be very

seriously deficient. In these circumstances, it is important that the United

Kingdom should press strongly and urgently for the Convention to be

amended by the introduction of a new formula for the calculation of minimum

hatch cover strength in relation to minimum permissible freeboard. This

amendment should be applicable not only to new buildings but to existing bulk

carriers. The present world population of such vessels is about 476 of which

68 are of similar length to the DERBYSHIRE.

65 It is unsatisfactory that there should be the present dual level regulatory

regime in respect of minimum hatch cover strength – that under ILLC 66 and

that operated by those classification societies which are members of IACS,

particularly as 96 per cent of the world’s bulk carrier fleet are entered with

such members. The evidence before this Investigation has at least raised

serious doubts as to whether even the UR S21 formula provides for an

adequate minimum strength requirement. Those doubts can only be resolved

by the further model test programme now in progress. There is strong

evidence that the UR S21 formula includes certain conceptual inadequacies in

addition to any question of deficient strength requirements, which suggests

that it ought to be re-formulated in any event. What is clear beyond doubt,

however, is that the inapplicability of this formula for enhanced hatch cover

strength to vessels built or contracted to be built before July 1998 is a very

serious short-coming. The exposure to risk of those bulk carriers of the size of

the DERBYSHIRE whose hatch covers do no more than comply with ILLC 66

poses an unacceptable risk to the safety of those vessels and their crews.

66 Accordingly, this Report recommends that as soon as possible after the results

of the present model test programme have been obtained and analysed, IACS


21

should consider the adequacy of UR S21 and should replace it with a formula

which provides for adequate minimum strength requirements for the hatches

of bulk carriers. That formula should apply both to new buildings and to

existing vessels. The flexibility intrinsic to a formula which establishes a

relationship between required hatch cover strength and minimum freeboard

should make its retrospective introduction more acceptable to the shipping

industry. Evidence as to the likely additional costs that would be involved in

strengthening the hatches of existing vessels does not suggest that this

requirement will subject the operators of such vessels to unacceptably large

expenditure. Since any amendment of ILLC 66 to introduce such a formula,

particularly with retrospective effect, is likely to take 4 to 5 years to come into

effect, it is strongly recommended that IACS should in the meantime urgently

bring into operation its own revised formula.

67 The DETR, in conjunction with IACS, should use its best endeavours to

persuade all parties to ILLC 66 to agree to amend the Convention to introduce

a new formula providing for adequate minimum hatch cover strength

requirements for bulk carriers. That formula should apply retrospectively.

68 This Report does not recommend that the UK Government should act

unilaterally in introducing statutory requirements for enhanced minimum

hatch cover strength. Nor does it recommend that LRS should unilaterally

take this course.

69 These are the most far-reaching recommendations included in this Report.

EXECUTIVE SUMMARY

22

Other Recommendations as to Improved Ship Safety

70 Many other recommendations for improved ship safety are to be found in

Sections 12 and 13: Towards Improved Ship Safety: Navigation and Towards

Improved Ship Safety: Other Matters. They relate to the following matters:

(1) Compulsory daily reporting of position by all vessels.

(2) Amendment of the Mariners’ Handbook advice on navigation

in tropical storms.

(3) Advice to masters of bulk carriers on the dangers of bow

flooding.

(4) Wider participation in the World Meteological Organisation’s

Voluntary Observing Ships scheme.

(5) Explicit advice to masters by weather routeing agencies as to

the scope of their functions.

(6) A research programme to investigate minimum strength

requirements for ventilators and air pipes.

(7) Electronic indication of open or damaged ventilators and air

pipes.

(8) A research programme to investigate improved securing

devices for stores hatch lids.

(9) Electronic indication of displacement of foredeck hatch lids and

of their being unsecured.

(10) Improvements to hatch cover operating manuals.

(11) Electronic indication that hatch securing devices are in place.


23

(12) Automated hatch opening, closing and securing.

(13) Improvements in the sealing of spurling pipes.

(14) Access to chain lockers.

(15) Research into and development of minimum strength

requirements for securing fittings to the foredeck.

(16) Installation of lighting and video cameras on the foredeck of

capesize bulk carriers.

(17) Electronic indication on the bridge of bilge levels in forward

spaces.

(18) Research into independent pumping systems for dealing with

forward space flooding.

(19) Introduction by classification societies of improved design

approval and survey procedures for new buildings.

(20) Maintenance ashore by shipowners of accurate construction

plans showing the vessel as built and as subsequently altered.

(21) Establishment of a marine accident data base covering storm

damage.

(22) The fitting of a VDR (black box) system on all existing cargo

vessels and new buildings.

71 This Re-opened Formal Investigation could not have taken place but for the

perseverance of the Derbyshire Families Association which led to the location

of the wreckage by the ITF survey in June 1994 and the immensely impressive

scientific and technical achievement by the Woods Hole Oceanographic

Institution of the United States in the conduct of the underwater survey. This

EXECUTIVE SUMMARY

24

report concludes that until 1994 the Department of Transport was entirely

justified in taking the view that there did not exist photographic and filming

equipment capable of obtaining evidence of sufficient clarity and accuracy to

provide evidence of the cause of the loss. The long delay after the Formal

Investigation in organising an underwater survey cannot be the basis of any

criticism of the UK Government.

72 The present policy of the DETR in deciding whether to conduct underwater

surveys as described in Section 15 is entirely appropriate. The fact that the

Department is now co-operating under a research agreement with Woods Hole

on further research into the development of a new underwater remotely

operated vehicle is a course which this Report strongly endorses in the

interests of ship safety.


Main Report


25

Section 1

Introduction: a Mystery to be Solved

1.1 In September 1980 the MV DERBYSHIRE disappeared in the North West

Pacific about 350 miles south south east of the southern tip of Japan. She was

the largest British Ship ever lost at sea. In fact she was one of the largest

cargo ships then afloat, in length slightly more that the QE2 and the Titanic,

but several times wider and of a massively larger carrying capacity. The

immense size of the vessel is shown by the fact that her length was almost

three times that of the playing area at Wembley Stadium. Stood on end, she

would have been 60 metres taller than Canary Wharf. Yet she vanished

without trace and without a distress signal. All those on board perished: 42

crew members and two of their wives. Their names are to be found at

Appendix 1. Apart from an empty life boat found long afterwards and some

oil on the surface, nothing was found.

1.2 The DERBYSHIRE was only four years old at the time of the loss. She had

been built by the highly respectable yard of Swan Hunter Shipbuilders Ltd at

Haverton Hill, Teeside, and she was owned by Bibby Tankers Ltd and

managed by Bibby Bros and Co (Management) Ltd, an organization noted for

the care with which it maintained its vessels. She was classed +100A1 with

Lloyds Register of Shipping (“LRS”) and her survey record was up to date,

her last annual survey having been conducted at Sasebo, Japan, in April 1980.

1.3 There was no doubt that at the time of her last message the DERBYSHIRE

had been in close proximity with a typhoon – known as Typhoon Orchid. In

the course of her voyage from Sept-Iles in Quebec to Kawasaki in Japan laden

with iron ore concentrates she had traversed the Lombok Channel in Indonesia

and was within five days of arriving in Japan when she encountered the

SECTION 1: INTRODUCTION

26

typhoon. It was approaching from the east during 7th and 8th September and

therefore converging with the vessel’s route towards Kawasaki. By 0300z on

9th September the vessel was reporting that she was hove to in a violent storm

with a force 11 wind and a wave height of 30 feet. Her last message was

transmitted to the vessel’s Japanese agents at 1019z on the same morning. It

confirmed receipt of an earlier message by the vessel but contained nothing to

indicate that the vessel’s safety was imperilled.

1.4 Not surprisingly the families of those who perished in this tragedy were

extremely anxious to know what had caused such a catastrophic loss of life.

They wanted a formal investigation. The UK Government took the view that,

in view of the total absence of material evidence, a formal investigation could

not be expected to establish clearly the cause of the loss [Statement by the

Minister for Trade (Lord Trefgarne) May 1981].

1.5 The DERBYSHIRE was the last of six bulk carrier sister ships built by Swan

Hunter at Haverton Hill Yard. The first – the FURNESS BRIDGE – was

completed in September 1971.

1.6 In March 1982 the second of the series – the TYNE BRIDGE - completed in

1972, while on a ballast voyage across the North Sea in severe conditions,

developed a crack in her deck in way of the after end of No.9 hold at frame 65.

She had to be towed into Hamburg and surveyed and it was discovered that

she had developed cracks of 2.8 metres length in the deck on the port side and

4.7 metres long on the starboard side. She was repaired and a tougher grade of

plating was inserted into the deck. Similar modifications to the deck were

made to two other sister ships (SIR JOHN HUNTER and ALEXANDER

GLEN).

1.7 In July 1985 a departmental draft report on the loss of the DERBYSHIRE

concluded that the most likely cause of the loss was major cracking in way of

Frame 65, that is to say in the region of the after end of No.9 hold which had


27

led to a major structural failure of the hull girder, thereby causing severance of

the after end of the vessel. However, after this draft report had been the

subject of consultations with a number of consultees, including Swan Hunter

and LRS, it was substantially altered and the final version dated March 1986

reached the conclusion that there were five possible causes of the loss –

explosion, shift of cargo, failure of hatch covers, external hull damage and

structural failure (that is in way of Frame 65) and that, although none of them

could be eliminated in the absence of further evidence, the circumstantial

evidence tended to suggest structural failure as the most likely cause.

1.8 That report concluded with the words:

“In the last analysis the cause of the loss of the DERBYSHIRE is,

and will almost certainly remain, a matter of speculation”.

1.9 Then, in November 1986, the third of the sister ships to be completed – The

KOWLOON BRIDGE, built in March 1973 as the ENGLISH BRIDGE -

while on a laden voyage from Canada to the UK developed a crack between

No.9 hatch and the pump room, that is to say in the region of Frame 65. The

crack was expanding and the vessel took refuge in Bantry Bay, South West

Ireland, where temporary repairs were carried out. However, when the anchor

chain broke, the vessel had to make for the open sea and, in so doing, collided

with a rock which incapacitated her steering gear. She had to be abandoned,

whereupon she drifted onto the Stag Rocks and sank. She was later found to

have broken her back close to Frame 65.

1.10 So strongly was the finger of suspicion now pointed in the direction of a

design defect associated with Frame 65 that in December 1986 a Formal

Investigation was appointed to enquire into the loss of the DERBYSHIRE.

This was conducted by Mr Gerald Darling QC as Wreck Commissioner

together with three assessors. The hearing lasted 46 days between October


28

1987 and March 1988 and the Report was eventually completed on 18th

January 1989.

1.11 The formal decision was:

“For the reasons stated in this Report the Court finds that the

DERBYSHIRE was probably overwhelmed by the forces of nature

in Typhoon ORCHID, possibly after getting beam on to wind and

sea, off Okinawa in darkness on the night of 9th/10th September

1980 with the loss of 44 lives. The evidence available does not

support any firmer conclusion.”

1.12 Amongst the reasons for this overall conclusion were the following:

(i) Structural failure due to excessive loading at amidships was

highly improbable even under the severe wave loading

conditions and since at Fame 65 the vessel was relatively

stronger than at amidships and that bending moments were at a

maximum close to amidships, it was also highly improbable

that there had been structural failure at Frame 65.

(ii) Fatigue cracking was unlikely to have caused the total loss of

the ship by failure of the hull girder either at amidships or in

way of Frame 65.

(iii) Separation of the hull at Frame 65 due to brittle fracture must

be regarded as very unlikely though some element of

uncertainty remained.

(iv) Although the flooding of the bosun’s store chain locker

followed by loss of freeboard by the bow and then sequential

failure of the No.1, 2 and 3 hatch covers due to hydrostatic

loading was a possible cause of the loss, it was not thought

probable.


29

1.13 A number of other specific causes were considered but dismissed as of

extremely low probability or possibilities which because of the lack of

evidence could not be wholly dismissed.

1.14 The families of those who perished were extremely disappointed at this

inconclusive result. Further suspicion was cast on the DERBYSHIRE design

features in way of Frame 65 by a number of learned journal articles and

lectures in the course of 1989 to 1994. On television and in Parliament

questions continued to be raised as to the safety of the vessel’s design. The

DERBYSHIRE Families Association pursued the matter with great

determination. Its members were not unnaturally determined that the cause of

this tragedy should be conclusively established. This Report in Section 15

considers the circumstances in which over a long period the Department of

Transport declined to finance an underwater survey and the technical problems

which confronted such an exercise, given that the wreckage was likely to be

some 4000 metres (2.5 miles) below the surface.

1.15 Eventually in March 1994 the ITF agreed to fund an eight-day exploratory

survey to find the wreckage, using Oceaneering Technologies, an American

firm. There was little evidence to go on other than oil slicks sighted a week

after the date of the last communication from the vessel. Against all the odds,

this operation located the bow section on the seabed at 4200 metres depth.

1.16 In view of the developing availability of equipment which could possibly be

used to conduct a full survey capable of providing sufficient evidence to

determine the cause of the loss, in March 1995 the UK Government appointed

Lord Donaldson of Lymington to carry out an assessment as to what further

steps could be taken to obtain evidence as to the cause of the loss, the cost of

taking such steps and what benefit to ship safety could be secured if the cause

of the loss were established and whether this would justify the likely costs

involved.


30

1.17 In his Assessment, Lord Donaldson concluded that the cost of an appropriate

underwater survey would be about £2 million and that such a survey ought to

be mounted in the interests of international ship safety. His recommendation

was accepted by the Minister of Transport and the UK/EC Survey was duly

conducted, funded partly by the UK and partly by the EU.

1.18 That survey was conducted in two phases in 1997 and 1998 respectively by

the Woods Hole Oceanographic Institution by means of the United States

Research Vessel, THOMAS G THOMPSON. Three Assessors were

appointed to oversee the survey and report upon the results, namely Mr Robin

Williams, Dr Remo Torchio of Genoa, both naval architects, and Professor

Douglas Faulkner, Professor of Marine Architecture and Ocean Engineering at

the University of Glasgow.

1.19 The technical achievement of that survey went far beyond anything in scope

and detail that had previously been attempted on any underwater wreck at a

comparable depth. The DSL 120 deep towed side sea sonar was deployed to

delimit the wreckage field: See diagram at Appendix 5. The Sonar Survey

map is at Appendix 6. Argo II was the system used to obtain the photographic

images which made up the mosaic pictures. For this purpose, highly accurate

positioning of the towing vessel and the suspended camera unit was vital in

order to stick to the precisely parallel photographic tracks with just enough

overlapping. This positioning was achieved by means of an acoustic

navigation system referred to four transponders (beacons). In the event,

135,774 individual electronic stills were obtained from a carpet plot of over 98

per cent of the entire wreckage field which measured 1500 metres x 1000

metres. The result was to identify 2500 separate items of wreckage. By

means of the joining-up of the separate stills into larger pictures (mosaicking)

it was possible to obtain and usually identify very clear black and white

pictures of continuous expanses of wreckage as is to be seen in the hatch cover

images at Appendix 29.


31

1.20 The effect of mosaicking is also to create three-dimensional images thereby

assisting identification of the items of wreckage.

1.21 A drawing of the Argo is shown at Appendix 7.

1.22 High definition video filming was carried out by the Jason system. A drawing

of the camera is to be found at Appendix 8. Unlike Argo, this camera could be

manoeuvred by communication from the towing vessel sent down a fibre optic

steel umbilical. The images within the camera field were transmitted back to a

screen for the controllers (in this case the Assessors) on board the towing

vessel to direct the camera to particular pieces of wreckage and, if necessary,

to within a few inches of the material to be filmed. The video stills of the

bosun’s store hatch at Appendix 33 well illustrate the products of this

equipment. Some 200 hours of video was made.

1.23 This equipment was made available to the UK Government by the extremely

supportive response both of Woods Hole and of the United States

Government. Without the co-operation and support of both this Re-opened

Formal Investigation could never have taken place.

1.24 The Assessors were unable to agree on certain matters directly material to the

Report and Professor Faulkner resigned before it came to be written. Mr

Williams and Dr Torchio delivered their Report on 19th January 1998. It

concluded that the wreckage evidence supported the view that the primary

cause of the loss of the vessel was the flooding of the bow section which, in

turn, caused the vessel to develop a bow trim, which in turn, gave rise to

exposure of the No.1 hatch cover to green water loading which exceeded its

collapse strength, followed by flooding of the No.1 hold and sequential failure

of the No.2 hatch cover and flooding of the No.2 hold. However, the

disturbing aspect of this Report was that the main reason for entry of seawater

into the bosun’s store was found to be the failure of the crew to secure the lid

to the hatch on the foredeck.


32

1.25 This conclusion clearly involved the imputation of serious negligence against

the officers and crew. It was deeply upsetting to the families of those on

board. However, the conclusion also acquitted the design and construction of

the vessel in way of Frame 65 of any causal contribution to the loss.

1.26 Against this background – the enormous quantity of additional photographic

and video evidence now available, the reported irrelevance of Frame 65 and

the reported primary cause of the flooding of the bow section – on 17th

December 1998 the Deputy Prime Minister, The Rt Hon John Prescott, MP

announced that there was to be a full re-opening of the Formal Investigation

into the loss of the DERBYSHIRE and that this was to be held in the High

Court under Section 269(1)(a) of the Merchant Shipping Act 1995. See

Appendix 2.

1.27 The hearing of this Investigation commenced on 5th April 2000. Since no

order had ever previously been made for a formal investigation to be re-

opened in the High Court and since no procedure is specifically laid down for

such proceedings in the CPR it was necessary to develop a fair, flexible but

economic system of defining the issues and adducing the evidence. Although

broadly similar to the procedure before a wreck commissioner conducting a

formal investigation, the procedure was fashioned to achieve as rapid a

conclusion of the evidence as possible.

1.28 In the event, the hearing lasted for a total of 54 days. Robes had nothing to

contribute to the hearing and were therefore dispensed with.

1.29 This was somewhat longer than anticipated for three main reasons. Firstly, it

became necessary at very short notice in the course of the hearing to conduct a

new series of tests on the model of the DERBYSHIRE in the MARIN test

tank, the earlier tests having been conducted on a wrongly dimensioned

model. Secondly, when issues arose as to the statistical extrapolations from

the test data obtained by MARIN it became necessary for the Court to instruct


33

its own expert on extreme value statistics – Professor Jonathan Tawn – and for

him to report to the Court on the MARIN results. Thirdly, it became

necessary unexpectedly at a late stage in the course of the hearing to adduce

additional evidence on meteorology and the functions of routeing

organisations. This evidence was given by Captain Mackie.

1.30 In the event, a total of twenty scientific and technical experts gave oral

evidence. In addition, evidence was given by six master mariners and five

other witnesses. A list of all witnesses is contained in Appendix 3.

1.31 A procedural feature of crucial importance to the efficient conduct of the

hearing was the system of experts’ meetings. This involved meetings of all

experts on particular disciplines, for the purpose of narrowing issues. In all

there were 14 such meetings, many of them, such as those relating to

wreckage identification, lasting several days. The reports produced at those

meetings proved to be of immense assistance, not only in the efficient conduct

of the hearing but in the preparation of this Report. A list of the topics

covered by the meetings is set out at Appendix 4.

1.32 Consistently with this approach to the issues, the hearing was divided into

separate periods, each dealing with a discrete technical issue. With one or two

unavoidable exceptions, all the witnesses giving evidence in respect of that

area of the Investigation were called during the relevant period. This greatly

enhanced the efficient conduct of the hearing.

1.33 Although it was open to the Court to appoint assessors to sit with the judge

and to participate in the ultimate preparation of this Report, that was not a

course that was adopted in this case. Instead, the Court was assisted by two

technical advisers who neither gave evidence nor participated in the ultimate

decision-making as to the conclusions of this Report. They were Dr P S J

Crofton of the Mechanical Engineering Department of the Imperial College of


34

Science Technology and Medicine and Professor John van Griethuysen of the

Department of Mechanical Engineering, University College London.

1.34 Their contribution to the work of the Court and as technical advisers to the

Attorney General’s team was very considerable.

1.35 The DETR was separately represented because its conduct in relation to the

International Load Line Convention 1966 and to statutory ship design

regulations as well as in relation to the investigation of the cause of the loss

was at least potentially open to criticism. The Attorney General and his team

acted as, in effect, “counsel to the tribunal”.

1.36 The following were made parties to the Investigation and were represented by

those counsel and solicitors indicated at page ii.

The Derbyshire Families Association (DFA).

Bibby Tankers Ltd, the shipowners.

SHSEGL Realisations Ltd, the successors to Swan Hunter, the ship

builders.

Lloyd’s Register of Shipping, (LRS) the classification society.

The Department of the Environment, Transport and the Regions (DETR).


35

Section 2

The Vessel

2.1 The DERBYSHIRE was a single screw ore/bulk/oil carrier, (an O.B.O.)

registered at the port of Liverpool, of 91,654 tons gross and 67,428 tons net,

powered by a diesel engine developing 30,400 bhp. She was 294.1 metres in

overall length and 44.28 metres in extreme breadth. She had 9 holds/tanks and

9 hatches and her bridge, accommodation and machinery were situated aft.

Her capacity plan is to be found at Appendix 9

2.2 The DERBYSHIRE was built by Swan Hunter Shipbuilders Limited at their

Haverton Hill yard on the River Tees. She was commissioned under the name

Liverpool Bridge, on handing over to Bibby Tankers Limited at Hamburg on

the 10th June 1976. Between June 1976 and the date of her loss the

DERBYSHIRE was owned by Bibby Tankers Limited and managed by Bibby

Bros. and Co. (Management) Limited. Her operational life was shortened

because she was laid up at Stavanger between 28th March 1978 and 11th April

1979.

2.3 On completion of her construction the DERBYSHIRE was classed by Lloyd’s

Register of (LRS) Shipping with the following notation:

‘+100 A1 strengthened for ore cargoes holds 2 and 6 may be empty or

oil cargoes.’

2.4 She was also approved to load ore cargoes in alternate holds, that is with

numbers 2, 4, 6 and 8 holds empty. She continued to maintain this

classification with LRS until the date of her loss.

SECTION 2: THE VESSEL

36

2.5 The DERBYSHIRE was the last of a series of six sister ships built by Swan

Hunter Shipbuilders Ltd., at their Haverton Hill Shipyard.

2.6 The six ships were:

Yard No. Completed Original Name Original Owners Comments

25 Sept. 1971 Furness Bridge Furness Withy Scrapped 1992

26 Sept. 1972 Tyne Bridge Huntings Scrapped 1987

27 March 1973 English Bridge

(renamed 1985 Kowloon Bridge)

Bibby Line Total loss 1986

31 Jan. 1974 Sir John Hunter Hilmar Resksten Scrapped 1997

55 April 1975 Sir Alexander Glen Hilmar Resksten Scrapped 1994

57 June 1976 Liverpool Bridge

(renamed 1978 Derbyshire)

Bibby Line Total Loss 1980

2.7 The Furness Shipbuilding Co. Ltd. at Haverton Hill, Teeside was an

independent shipyard which had closed in mid-1968 and was later taken over

by Swan Hunter Shipbuilders. The Haverton Hill Shipyard closed in March

1979.

2.8 Although the initial design work for the so-called Bridge Class, including

calculations and class submissions, was carried out in the design offices of

Swan Hunter Shipbuilders in 1968, the steel working drawings for the hull

including structural modifications were prepared by the Hull Steel Office of

the Teeside technical division.

Certificates and Survey Status

2.9 At the time of her loss the following certificates had been issued in respect of

the vessel:


37

(a) Cargo Ship Safety Equipment Certificate.

(Issued at Oslo, Norway by the Norwegian Directorate of

Shipping on the 6th April 1979. Valid until the 9th April 1981).

(b) Cargo Ship Safety Radio Certificate.

(Issued at Sasebo, Japan, by the Japanese Ministry of Transport

on the 16th April 1980. Valid until the 15th April 1981).

(c) Cargo Ship Safety Construction Certificate.

(Issued by LRS on the 25th October 1977. Valid until the 30th

June 1981).

(d) International Load Line Certificate.

(Issued by LRS 4th June 1976. Valid until 3rd June 1981. Last

annual survey carried out at Sasebo, Japan, on the 17th April

1980).

2.10 Between the 3rd and 17th April 1980 the DERBYSHIRE was at Sasebo,

Japan. During this period surveyors from LRS carried out annual and dry-

docking surveys for classification purposes. In addition, special classification

survey was commenced and a general examination was made for the purposes

of permitting a postponement of the completion of the special survey until

sometime prior to April, 1981. The vessel had been in service only for about

three years by the time of the last voyage.

2.11 While the vessel was in dry-dock repairs were carried out including:

• repairs to a fracture in welds on the transverse bulkhead at

Frame 65 along the boundary with the starboard topside tank

including re-welding carried out in accordance with

classification society rules and verified as sound by dye-testing;


38

• repairs to fractures in hatch end beams at attachment of knuckle

tank bottom plating of topside tanks, including welding of ring

shape reinforcement pads;

• other sundry minor repairs.

2.12 The repairs were carried out by Sasebo Heavy Industries Co. Ltd. and

surveyed by surveyors from LRS.

2.13 Other work performed during the period of drydocking included the renewal

of the lifeboat falls and the servicing of the 4 life rafts. In addition, the

hatches were subjected to a hose test. No leakages were observed. Shell

plating damage forward, including a fracture of her starboard bilge keel, was

also repaired.

2.14 On completion of the period of drydocking referred to above, the LRS

surveyors concerned issued a report dated 17th April 1980 recommending that

the vessel be credited with her annual and drydocking surveys and that the

completion of her special survey be postponed until sometime prior to April

1981.

Equipment

2.15 Navigational equipment included a gyro compass, a Satellite Navigator, 2

radars, an echo sounder and a speed log. The DERBYSHIRE was equipped

for radio transmission and reception of MF, HF, and VHF frequencies. Her

call sign was “GULK”. Her main aerial was of the cage type and her

emergency aerial was of the inverted “IL” type. The emergency transmission

system was powered by a 24 volt D.C. battery supply. The lifeboat radio was

an IMR Solas III A type powered by a hand driven generator. The

DERBYSHIRE was also equipped with a radio telex and Koden weatherfax

machine for the receipt of facsimile weather charts.


39

2.16 The DERBYSHIRE had 2 G.R.P. motor lifeboats each capable of carrying a

total of 61 persons and 2 twelveman inflatable life rafts, 2 sixman inflatable

life rafts and 52 life jackets.

2.17 The DERBYSHIRE also had a loadicator.

Design

2.18 The design was for an oil/bulk/ore carrier (“OBO”) of approximately 150,000

tonnes deadweight. The ships were to be capable of carrying:

Ore cargoes in alternate holds with appropriate heavier scantlings.

Oil cargoes in all holds with surfaces pressed up to hatch coaming

levels.

Bulk cargoes in all holds.

2.19 The OBO first emerged as a distinct design type in the mid 1960’s. Its

versatility made it very popular. It was in some respects similar to a bulk

carrier: it did not generally have longitudinal bulkheads and its holds could be

used for all types of cargo carried.

2.20 In the early 1960’s there emerged designs which provided for the loading of

heavy ore cargoes into alternate holds. This enabled very dense ore cargoes to

be loaded to the full dead weight in a small number of holds somewhat shorter

than the adjacent spaces, while less dense cargoes, such as coal or grain could

be loaded in all the cargo spaces.

2.21 The alternate hold loading arrangements for ore generates high shear forces in

the hull girder and for this reason special Rules and Class notations such as

“Strengthened for Heavy Cargoes-Specified Holds may be Empty” were

introduced by Lloyd’s Register of Shipping in 1964. This innovation of

loading in alternate holds had become established practice by the mid 1960’s.


40

As with the introduction of dual purpose holds in ore/oil carriers it was a

significant step in the development of the O.B.O. ship.

2.22 The first true O.B.O. ships, which appeared in 1965/66, embodied the

principal characteristics of an O.B.O. carrier and were designed to carry either

ore or oil or less dense cargoes whilst still obtaining their full deadweight.

These ships had essentially the same hold configuration as the modern O.B.O.

i.e. no longitudinal hold bulkheads, full width holds served by single

hatchways, a normal depth double bottom structure and upper and lower wing

tank spaces with sloping sides.

2.23 The basic design philosophy governing the design of O.B.O. ships during the

1960’s and 1970’s included:-

(i) The need to provide full facilities for both dry and liquid

cargoes resulted in increased cost and complexity of the O.B.O.

for example, both a bilge system and a cargo oil handling

system had to be installed in association with wide hatches for

grab handling of dry cargoes.

(ii) Hatch covers had to provide watertight integrity of the main

hull in addition to withstanding the internal sloshing loads from

oil cargoes.

(iii) The particular characteristic of the O.B.O., as with nearly all

bulk carriers when acting as an ore carrier, is the ability to carry

very dense cargoes. This results in only a small proportion of

the volumetric capacity of the ship being required for cargo

stowage in order to achieve full deadweight.

2.24 The design of the Bridge Class vessels incorporated some typical O.B.O.

features including:


41

(i) Provision of a clean interior cargo hold surface to permit free

flow of ore, bulk or oil cargoes and fast and easy cleaning of

holds between successive voyages.

(ii) Maximum cargo volume for the carriage of grain and light oils.

(iii) Nine cargo holds of similar length, having hatch openings of

equal breadth of 71 feet (21.64 metres).

(iv) Hydraulically operated, side rolling steel hatch covers divided

at the centre line. These covers had to be both oil tight and gas

tight.

(v) Cargo holds with hopper-shaped bottoms at the sides and ends

for ease of unloading.

(vi) Upper surfaces sloped at 30° so that dry cargoes, such as grain,

completely filled the holds without the need for trimming.

(vii) Ballast tanks formed by the upper and lower wing tanks.

Double bottom tanks and wing tanks between the upper and

lower hopper tanks.

(viii) Ballast spaces of sufficient capacity to ensure propeller

immersion.

2.25 These vessels were amongst the first O.B.O. carriers to be built in the UK and

quite the largest bulk carriers yet built in the UK.

2.26 The Swan Hunter design office produced design calculations and principal

drawings. LRS approved the drawings subject to certain amendments, by

reference to their rules for Tanker and dry cargo vessel design, but there was

no design standard specifically for O.B.O.’s. The Haverton Hill Yard’s

production drawing offices then produced from the design drawings, the

working drawings for production purposes and these were in turn approved by


42

LRS. These drawings were then used as the basis for work in the fabrication

sheds.


43

Section 3

The Vessel’s Trading Activities and Last Voyage

3.1 The DERBYSHIRE departed Sasebo, Japan, on the 18th April 1980 bound for

Hay Point, Australia. She was under the command of Captain P. Boyle. On

the 29th April 1980 the vessel arrived at Hay Point and began loading a cargo

of coal for carriage for Fos-Sur-Mer, France.

3.2 On the 1st May 1980 the DERBYSHIRE departed for Fos-Sur-Mer. During the

voyage to Fos-Sur-Mer she experienced a period of heavy weather. As a

result of the heavy weather she sustained damage to a pipe casing situated

under the foremast on the starboard side, to 3 ventilators situated on No.1

hatch and to 4 steel plates forming the walkway platform around the windlass.

The heating coil covers around no.1 and no.8 hatches were torn away, the

forward liferaft was dislodged from its stowage position and inflated and five

catwalk pieces were removed from hatches 2,4,6, and 7. In addition water

entered her forecastle space through a ventilator which had had its lid washed

off. This water was subsequently removed manually as the discharge pump

serving the forecastle space was inoperable. The damage referred to above

was not reported to the classification society and was repaired by the ship’s

crew before the vessel arrived at Fos-Sur-Mer. The vessel arrived at Fos-Sur-

Mer on about the 20th June 1980 whereupon discharge of her cargo of coal

commenced. During the vessel’s stay at Fos-Sur-Mer, Captain Boyle was

relieved by Captain G.V. Underhill.

3.3 On the 23rd June 1980 the DERBYSHIRE sailed from Fos-Sur-Mer bound for

New York for bunkers.

SECTION 3: THE VESSEL’S TRADING ACTIVITIES AND LAST VOYAGE

44

3.4 During the course of that voyage the shipowners’ safety consultants conducted

a safety training programme on board, including an emergency muster and

lifeboat stations exercise. They inspected the vessel’s life saving equipment.

Defects were listed and the master informed of them. Steps were then taken to

remedy these defects.

3.5 The DERBYSHIRE arrived at New York on the 5th July 1980. After having

bunkered the vessel sailed for Sept-Iles, on the 7th July 1980. She arrived at

Sept-Iles on the 10th July 1980. On the 11th July 1980 loading operations were

completed and the DERBYSHIRE sailed for Kawasaki, Japan, laden with a

cargo of 157,447 tonnes of iron ore concentrates.

3.6 The loading equipment at Sept-Iles enabled the cargo to be delivered in such a

way as to achieve a reasonably level stow in the hold under the hatchway as

distinct from a free fall cone. At completion of loading each hold the cargo

would slope away at the edges outside the hatch area.

3.7 Holds 2 and 6 were left empty.

3.8 In holds 1,3,5,7 and 9 the average depth of cargo under the hatchways would

have been between about 10 and 12 metres sloping away to the vicinity of the

top of the sloping sides of the lower wing ballast tanks. In hold 4 the average

depth of cargo under the hatch way would have been about 6 metres sloping

away to about 3 metres up the sloping sides of the lower wing ballast tanks. In

hold 8 the average depth of cargo under the hatchway would have been about

5 metres sloping away to about 2 metres up the sloping sides of the lower

wing ballast tanks.

3.9 Cargo in holds, 1,3,5,7 and 9 received passive support from the vertical sides

of the holds, especially in 1 and 9 due to the hull shaping in towards the bow

and stern.


45

3.10 Cargo in holds 4 and 8 only received the support of the sloping sides of the

lower wing ballast tanks and therefore had less resistance to movement should

it be subjected to sufficient inducement to move.

3.11 The average moisture content (3.59 per cent) was within the safe limits for

carriage by sea set out in the IMCO “Code of Safe Practice for Bulk Cargoes”,

1977. During at least the first part of the voyage, while crossing the Atlantic,

the hold bilges required frequent pumping as water drained off the cargo. That

would not necessarily be inconsistent with the measured water content of

samples of the second half of the cargo loaded, namely 4.5 and 4.54 per cent.

3.12 The voyage from Sept-Iles commenced on 11th July 1980. The vessel was to

pick up stores while passing Cape Town.

3.13 The vessel had been instructed by her charterers to be weather-routed from

Cape Town onwards by Ocean Routes Inc. of California, and was also

participating in “AMVER” (Automated Mutual-Assistance Vessel Rescue)

system. Whilst on passage to Cape Town Charterers instructed the Master to

reduce to a minimum speed consistent with safety, prudent seamanship and

engine operations. On 31st July the DERBYSHIRE reduced speed to about 69

revolutions per minute.

3.14 The DERBYSHIRE picked up stores by helicopter from Cape Town and then

proceeded across the Indian Ocean, through the Lombok Strait, then the

Macassar Strait and into the Celebes Sea.

3.15 The details of the vessel’s messages in September 1980 are set out in section 4

of this Report. They can be summarized as follows.

3.16 At 0300z on 3rd September the master reported that the vessel was south of

Mindanao in the Philippines and doing 10 knots. At 10.30z he reported

having increased speed to 12.5 knots in view of the forecast of a tropical storm

over 100 miles to the eastwards. On the following day, Oceanroutes informed


46

the master of the forecast track of the depression and recommended that the

vessel should take a northerly course so as to clear the track by at least 200

miles. At 2146z on 5th September Oceanroutes gave the master further

information about the depression TD16 – and confirmed that the route was

still valid.

3.17 At 0300z on 6th September the vessel informed Oceanroutes that it was

resuming its rhumb line course to Japan.

3.18 In the course of that day, a further tropical depression developed which by

early on 7th September had been officially identified as typhoon “Orchid”.

That was over 650 nautical miles east of the vessel by midnight on 6th/7th

September 1980.

3.19 It will be necessary to consider in some detail the vessel’s movements during

the period 7th to 9th September for it is necessary to explain how it came about

that an extremely experienced master and his high quality deck officers

navigated the DERBYSHIRE to positions where by 0300z on 8th September it

was necessary due to the sea conditions to reduce speed to 8 knots, where by

0825z on that day the vessel was stormbound and where by 0300z on 9th

September it was, as it reported, hove-to in a violent storm with force 11

winds and with 30 feet seas.

3.20 A later report form the MV ALRAI stated that at 0300z (noon local time) on

9th September it received from the DERBYSHIRE the latter’s speed and

position, then 80 miles away from the ALRAI. It was calculated that by 1400z

(2300 local time) the distance would have reduced to 60 miles. The ALRAI’s

report indicated that by that time the wind was force 12 and wave heights 60

to 100 feet. Later in the course of 9th September the DERBYSHIRE came

very close to the centre of typhoon Orchid.

3.21 The last message received from the DERBYSHIRE was transmitted at 1019z

on 9th September.


47

3.22 There may have been a number, or combination, or reasons why a distress

message may not have been effectively transmitted. None was received.

Assuming that the main transmitter was fully operational and that the vessel’s

full mains supply was available it is possible that:

(i) There was insufficient time for the Radio Officer to man the

radio station and to transmit any message.

(ii) The aerials had been damaged by high winds, lightning or

heavy seas. The MAS 20 Transmitting Mast Aerial and the

Reserve Twin Wire Aerial were both fitted on the same side of

the ship. (The MAS 20 Aerial on the “ENGLISH BRIDGE”

was carried away during a storm force 10-11).

(iii) The insulation of the aerials was so badly affected by spray

and/or heavy rain that no transmission could be fully effective.

3.23 Assuming that the vessel’s main power supply was not available:

(i) The range of transmission would have been reduced to that of

the relatively low power of the Reserve Transmitter.

(ii) The effective range of the Reserve Transmitter could have been

further reduced by spray and/or rain affecting the aerials.

3.24 There could have been damage to the Radio Room or some of its equipment

preventing transmission of a distress message. The vessel’s only Radio

Officer could have been injured or incapacitated and untrained personnel may

have been unable to tune the transmitter.

Search and Rescue

3.25 On 13th September 1980, due to the absence of any radio communication from

the vessel since the 9th September, Bibby Bros. And Co (Management)


48

Limited instructed their local agents in Japan to request the Japanese Maritime

Safety Agency to mount a search for the vessel. The Japanese Maritime

Safety Agency indicated, however, that a full-scale search could not be

mounted until the vessel’s arrival was at least 24 hours overdue. Various

Japanese radio stations and ships agents were nevertheless requested to

maintain a watch for the vessel. The DERBYSHIRE failed to arrive at

Kawasaki on the 14th September 1980.

3.26 On the 15th September 1980 the Japanese Maritime Safety Agency organized

an air and surface search for the DERBYSHIRE. The major part of the search

was conducted by the Japanese patrol vessels OSUMI and MOTOBU assisted

by two reconnaissance aircraft Nos. 791 and 811.

3.27 At 1503 hours (local time) on the 15th September 1980 one of the

reconnaissance aircraft spotted an oil slick in a position approximately 25° 50’

North 133° 30’ East. The patrol vessel OSUBMI was instructed to proceed to

this position. At 0545 hours (local time) on the 16th September 1980 OSUMI

discovered oil rising to the surface of the sea in a position approximately 25°

48’ North 133° 37’ East.

3.28 Samples of one of the oil slicks were analysed and found to have had ‘a degree

of similarity’ to that of the DERBYSHIRE, although positive identification

was not possible.

3.29 On the 17th September 1980 the search was suspended due to the presence of

tropical revolving storm SPERRY in the vicinity of the search area.

3.30 On the morning of the 18th September 1980 the search was resumed. At about

1810 hours on the 20th September 1980 the search was terminated, no sign of

the vessel, her crew, or of any further material which might have come from

the vessel having been found.


49

3.31 On the 24th October 1980 one of the DERBYSHIRE’s lifeboats was sighted by

the Japanese tanker DAIEI MARU in a position 21° 14’ North 122° 18’ East.

The lifeboat was empty. It was not recovered.


50


51

Section 4

Navigation

4.1 The area of the wreckage field corresponds to a surface position which

experienced extremely severe sea conditions in the course of 9th and 10th

September. By midnight on 8th/9th September, the storm centre of Orchid had

reached a position about 164 nautical miles south south west of the vessel’s

reported position at 0300z on 9th September, some three hours later. It is not

possible to identify with any precision the vessel’s position at midnight on

8th/9th September. This is because at some time between its reported position

at 0300z on 8th September and its reported position at 0300 on 9th September it

departed from its rhumb line course to Tokyo Bay. It is not known at what

time during this 24 hour period it changed course. Nor is it known whether

during that period its course was changed more than once. Based upon the

hindcast of wind direction and speed and wave direction and significant wave

height prepared by Dr. V.J. Cardone of Oceanweather Inc an attempt has been

made by Mr J Hook of Burness Corlett & Partners, who gave evidence on

behalf of Swan Hunter, to reconstruct the vessel’s course and speed during the

relevant 24 hour period having regard to the manner in which a master might

reasonably be expected to navigate the vessel in the sea and wind conditions

assumed to be experienced. That exercise places the vessel’s position at

midnight on 8th/9th September at about 7 nautical miles due south of its

reported position at 0300z on 9th September.

4.2 Assuming that Mr Hook’s reconstruction is correct, the vessel was at midnight

on 8th/9th September about 209 nautical miles north north west of the storm

centre. Although it was then just outside the 200 mile radius of the centre of a

revolving tropical storm then recommended by the Mariners’ Handbook to be

SECTION 4: NAVIGATION

52

avoided, the vessel was nevertheless already experiencing very heavy seas.

According to Dr Cardone’s hindcast, the wind speed had reached 44 knots

from the north east and the significant wave height, that is to say the average

height of the highest one third of the waves, was 8 metres. The vessel, which

reported at 0300z on 9th September that it was hove to in a violent storm, was

by that time making very little headway over the ground, having covered only

about 7 nautical miles in 3 hours. It lay in a position which was very close to

the track of the centre of the typhoon during 9th September as forecast at

18.00z on 8th September. Since the vessel was able to make such little

headway due to the severe sea conditions, and could not safely be turned on to

a course which would place its beam to the seas, it can be concluded that by

midnight on 8th/9th September it was already trapped in circumstances in

which there was little a master could do to alleviate its position except to

maintain its heading as far possible with the sea on the starboard bow in

accordance with the advice in the Mariners’ Handbook.

4.3 The crucial question is therefore how the DERBYSHIRE ever found itself in

this dangerous and inextricable position.

4.4 It is first necessary to identify the sources of information relevant to the

navigation of the vessel available to the master. These were:

(i) Weather forecasts broadcast by Guam and Tokyo radios;

(ii) Faxed synoptic weather charts available from Tokyo;

(iii) The actual wind speed and direction observed from the vessel

and how they were developing;

(iv) The actual condition of the sea observed from the vessel and

how that condition was developing;

(v) Changes in barometric pressure;


53

(vi) Information received from Oceanroutes, the vessel routeing

service based in California which had been employed by the

charterers of the DERBYSHIRE to provide routeing

information on the voyage.

4.5 The weather forecasts broadcast from Guam emanated from the United States

Navy/Air Force Joint Typhoon Warning Center (JTWC). That governmental

organisation had international responsibility through the World

Meteorological Organisation for providing tropical cyclone warnings to ships

at sea in the north west Pacific. Those forecasts were issued every 6 hours.

They were derived from analysis of reports of actual conditions of wind,

seastate and weather systems supplied by observer vessels in the area and by

patrolling United States aircraft which were constantly in flight over the north

west Pacific. They gave the current position and forecast route of the typhoon

over the next 72 hours.

4.6 The forecasts available from Japan radio were produced by the Japanese

national weather forecasting service. They were transmitted in the form of

bulletins every 6 hours and gave the present position of the typhoon and a

forecast of its route over the next 24 hours.

4.7 It can confidently be assumed that the master of the DERBYSHIRE would

have been receiving forecasts from both Guam and Tokyo.

4.8 In addition to this information the vessel would also probably have been

receiving synoptic weather charts faxed from Tokyo. These showed a great

deal of meteorological information, including weather fronts and isobars, as

well as wind speed and direction symbols and barometric pressure readings

derived from observer vessels. The copies of these charts before this

Investigation were extremely difficult to read due to the lack of clarity of the

transmitted documents. It may well be that any fax transmissions to the vessel

were just as difficult to read. The evidence suggested that a reasonably


54

competent and experienced master would, if he could read these charts, have

been able to understand the general trends of weather development which they

showed but would not usually be sufficiently knowledgeable of meteorology

to form any independent analysis of the information to give him an insight into

typhoon movement which might cause him to question the forecasts issued by

Guam or Tokyo. The charts showed a range of storm centre positions by the

end of the succeeding 24 hours depicted by a 30° triangle with its apex at the

current position.

4.9 The master could also be expected to take account of wind speed and

direction, the condition of the sea and barometric pressure changes alongside

the Guam and Tokyo forecasts. In summary, as the typhoon got closer to his

current position, he could expect a freshening, veering wind increasing in

force, together with an increasing swell and falling barometric pressure.

4.10 The charterers of the vessel had organised Oceanroutes to provide routeing

services. In the course of the period 01.52z on 4th September to 01.23z on 9th

September Oceanroutes sent four messages to the vessel giving weather

information, including information as to the track of an earlier depression

(TD16) and of Orchid.

4.11 Later in this Report it will be necessary to consider further and in some detail

the content and frequency of advice given by Oceanroutes for this has been the

subject of strong criticism in the course of this Investigation.

4.12 Against that background it is now necessary to examine the course taken by

the DERBYSHIRE against the meteorological and other information available

to the master from 3rd to 9th September. The Mariners’ Handbook NP100,

published by the Hydrographer of the Navy, would have been on board. It

contained the following advice as to navigation in the area of a possible

tropical revolving storm:-


55

“In whatever situation a ship may find herself the matter of vital importance is to avoid passing within 50 miles or so of the centre of the storm. It is preferable but not always possible to keep outside a distance of 200 miles. If a ship has at least 20 knots at her disposal and shapes a course that will take her most rapidly away from the storm before the wind has increased above the point at which her movement becomes restricted, it is seldom that she will come to any harm. Sometimes a tropical storm moves so slowly that a vessel, if ahead of it, can easily outpace it or, if astern of it, can overtake it.

If a storm is expected in the vicinity, the vessel, whilst observing her barometer, should continue on her course until the barometer has fallen 5mbs (corrected for diurnal variation) below normal, or the wind has increased to force 6 when the barometer has fallen at least 3mbs. Then she should act as recommended in the following paragraphs, until the barometer has risen above the limit just given and the wind has decreased below force 6. Should it be certain, however, that the vessel is behind the storm, or in the navigable semicircle, it will evidently be sufficient to alter course away from the centre.

In the N hemisphere (ship initially moving slowly)

(a) If the wind is veering the ship must be in the dangerous semicircle.

The ship should proceed with all available speed with the wind 10º to 45º, depending on speed, on the starboard bow. As the wind veers the ship should turn to starboard, thereby tracing a course relative to the storm as shown in the following diagram.


56

(b) If the wind remains steady in direction, or if it backs, so that the ship should bring the wind well on the starboard quarter and proceed with all available speed. As the wind backs the ship should turn to port as shown in the diagram.

If there is insufficient room to run, when in the navigable semicircle, and it is not practicable to seek shelter, the ship should heave-to with the wind on her starboard bow in the N hemisphere.”

4.13 The evidence is that at the relevant time a master would be working to a

danger zone mentioned in the first paragraph of 75 nautical miles, rather than

50 nautical miles.

4.14 On 3rd September, when the vessel was south of Mindanao, the master

informed Oceanroutes that the Japanese weather fax showed a tropical

depression with a course and speed which intercepted the vessel’s present

course. He asked for a prognosis, stating that he had increased his speed to

12.5 knots in order to pass ahead of the forecast track. The tropical depression

referred to was known as TD16 and was a predecessor of Orchid.

4.15 At 0152z on 4th September Oceanroutes replied giving the forecast track of

TD16 together with its speed of movement and maximum predicted wind

speed. They recommended that, if conditions permitted, the vessel should take

a northerly course, thereby clearing the track of the storm by at least 200

nautical miles, then resuming its direct course to the destination. They asked

the master to advise his position daily until he was clear of the storm.

4.16 At 0300z on 5th September the master duly reported his position and speed of

11.5 knots. At 2146z on the same day Oceanroutes sent a message informing

the master of the present location of TD16 and its forecast speed and track

together with the information that it would reach storm intensity by noon on

6th September, including the anticipated wind speeds. The message informed

the master that his route was still valid “maintaining 200 nautical miles from

center”, this reflecting the advice in the Mariners’ Handbook.


57

4.17 At 0300z on 6th September the vessel reported its position to Oceanroutes and

that it was resuming its rhumb line to Japan, but reverting to a speed of 10

knots. The decision to resume the rhumb line was clearly taken by the master

not on the basis of the advice from Oceanroutes, but on other information

available to him, probably including both the Guam and Tokyo forecasts for

the track of TD16. The Guam forecast at 0000z on 6th September had

indicated that TD16 had subsided but that observations of local conditions

suggested a fair risk of regeneration of TD16 or for development of a new

tropical cyclone in the same area.

4.18 In the course of 6th September a distinct tropical depression centre developed

close to the area of TD16. This new depression had by 7th September come to

be officially identified as a typhoon, designated “Orchid”.

4.19 The vessel’s rhumb line route was maintained from 0300z on 6th September at

least until 0300z on 8th September its next reported position. During that

period its speed averaged 10 knots, but by the latter time its speed had been

reduced to 8 knots.

4.20 Captain Roberts, a master mariner who was called to give evidence on behalf

of the DFA prepared a series of plots which showed on the INT Chart 507 the

route of the DERBYSHIRE during the period 0300z on 6th September to

0300z on 8th September, a conjectural route between the latter time and its last

reported position at 0300z on 9th September and a further conjectural route

from the last position to the wreckage site. Captain Roberts also plotted the

actual positions of the centre of the tropical depression which came to be

called typhoon Orchid and its forecast tracks given by Guam and Tokyo.

4.21 Captain Roberts prepared such charts showing the forecast position at 6 hourly

intervals during that period. They are reproduced at Appendix 10.

4.22 Assuming, as seems probable, that the master would be plotting that same

information on his own chart, he would be able to see where on his projected


58

route he could expect to be over the next 72 hours in relation to the forecast

track of the centre of Orchid.

4.23 Thus at midnight on 6th/7th September the vessel was over 650 nautical miles

from the current storm centre identified by Guam. However, the forecast

storm track was predicted to cross the vessel’s rhumb line at about 1700z on

9th September. If the vessel maintained her current speed of 10 knots, it would

be about 270 nautical miles to the north north east of the storm at that time. At

midnight on 9th/10th September it would be at a position about 200 nautical

miles slightly west of north of the storm centre.

4.24 By 0600z on 7th September the forecast track of Orchid over the following 72

hours had swung a few degrees northwards to the effect that the storm track

was now forecast to cross the vessel’s rhumb line at about midnight on 9th/10th

September. Further, by 1200z on 8th September, some 30 hours later, the

vessel would enter the 200 nautical miles radius from the storm centre and, if

the speed of 10 knots were maintained, it would remain within that radius until

0600z on 10th September. Indeed, by 0600z on 9th September the vessel would

be only about 100 nautical miles north north east of the storm centre.

4.25 This Investigation has heard evidence from Captain Willey a former master of

one of the sister ships of the DERBYSHIRE – the Sir Alexander Glen – that at

0600z on 7th September, had he been master of the DERBYSHIRE, he would

not have held to the rhumb line course, but would have turned to the west or

west south west. His reason for doing so, would have been that, although the

vessel was still 630 nautical miles from the centre of Orchid, his course would

pass too close to the storm centre and, even if he kept ahead of the storm, there

would be the danger that, consistently with a propensity of revolving storms in

this area, it might re-curve to the north east thereby catching the vessel in a

dangerous position relative to the centre. He had considered but rejected

increasing speed to 12 to 12.5 knots because that would still leave the vessel

exposed to the risk that the storm might re-curve. He would have decided to


59

alter course at 0600z on 7th September even assuming that the vessel appeared

to be able to maintain 10 knots up to 8th September.

4.26 At 1200z on 7th September there emerged a substantial disparity between the

Guam and Tokyo forecasts for the succeeding 24 hours. The Tokyo forecast

indicated that the storm centre would move within a 30 degree triangle ranging

between points 15 degrees either side of due west, whereas the Guam forecast

had the track moving at a different speed to the north west and reaching a

point over 200 nautical miles to the north east of the most northerly point in

the range given by Tokyo. If it is assumed that the vessel maintained a speed

of 10 knots for that period of 24 hours, by 1200z on 8th September, if the

master held course, the vessel would be about 200 nautical miles from the

storm centre, due north of the most northerly position forecast by Tokyo and

north west of the position forecast by Guam.

4.27 Captain Willey’s evidence was that because of the apparent tendency to re-

curve to the north east after 1200z on 9th September he would have been even

more determined to alter course to the west or west south west at noon on 7th

September if he had not done so earlier. If he did not alter course, he might

well end up in the dangerous semi-circle.

4.28 It is to be observed that Captain Willey, while being firmly convinced as to the

course he would have taken, had he been master, acknowledged that it might

be perfectly reasonable for a master to have decided to hold to his rhumb line

course at those times to see how conditions developed.

4.29 At 1800z on 7th September the Guam forecast predicted the track of Orchid

significantly more to the west and then north west than 6 hours earlier. The

Tokyo forecast for the following 24 hours was substantially similar to that at

1200z. If the vessel adhered to her course she would be comfortably outside

the 200 miles radius of the storm centre forecast for the next 48 hours

provided that she averaged 10 knots. If, however, the average speed fell by 50


60

per cent she would be well within the 200 miles radius of the centre forecast

by Guam and of the northern extremity of the range forecast by Tokyo and

would be in a particularly dangerous position less than 100 miles east north

east of the centre a further 24 hours later. It is unclear whether the vessel

would have received these radio messages coming as they did at 3 am local

time. Captain Willey did not consider it likely that the master would have

required the radio officer to stay on duty at that time to receive such messages,

even having regard to the previously forecast typhoon.

4.30 At midnight on 7th/8th September the Guam forecast track of Orchid showed

its route to be tending yet more to the west and then north west. On the other

hand, Tokyo’s forecast track range was considerably to the north of where it

had been predicted 6 hours earlier. There was thus a very close proximity

between the position forecast by Guam and the northern-most part of the range

given by Tokyo. Both positions would be about 200 nautical miles south to

south south west of the vessel by midnight on 8th/9th September. It has been

calculated by Captain Roberts that in order to reach the vessel’s reported

position at 0300z on 8th September, by which time the speed had been reduced

from 10 knots to 8 knots, the speed reduction must have been made at about

midnight. It can therefore be assumed that at that time the master’s decision-

taking would be based on the assumption that he would be able to maintain

about 8 knots for 24 hours. If that assumption proved to be wrong, the vessel

might be caught well within the dangerous semi-circle, within 200 nautical

miles from the storm centre.

4.31 At 0113z on 8th September Oceanroutes broadcast to the vessel a forecast of

the position of Orchid at midnight on 8th/9th September, moving north at 10

knots for 48 hours and then recurving north eastwards and accelerating with a

radius of force 7 winds of 250 nautical miles. The forecast midnight position

was about the same as that given by Guam. The message contained no

routeing advice to the master.


61

4.32 To this message the DERBYSHIRE replied at 0610z, stating its position at

0300z. However, there was no request for routeing advice or for any further

information.

4.33 The hindcast carried out by Dr Cardone of the sea and wind conditions on 8th

September indicates that there was an hour-by-hour deterioration in those

conditions in the area of the vessel’s probable route from midnight on 6th/7th

September and that by midnight on 7th/8th the significant wave height had

reached 3.69 meters and by 0400z on 8th it was 4.11 meters with a wind speed

of 13.8 m/s (27 knots). At 0825z on 9th the vessel reported that it was

“stormbound”. According to the hindcast, that deterioration continued

throughout 8th September to the effect that by midnight the significant wave

height is calculated as 8 meters and the wind speed was 22.4 m/s (44 knots).

By 0300z on 9th there had been further deterioration to a significant wave

height of 4.86 meters and a wind speed of 23.7 m/s (46 knots). The vessel

then reported in its last message that it was hove-to in a violent storm with 30

feet seas. That estimate is consistent with the highest waves likely with the

hindcast significant wave height. Throughout this period the wave drift forces

acting on the vessel would also have been increased on a progression related

to the significant wave height.

4.34 The precise route followed by the master after 0300z on 8th September to his

position 24 hours later will never be known, but the prediction work done by

Mr Hook, to which reference has already been made, probably reflects roughly

the course taken. This is shown in the diagram in Appendix 11. This work

was based on a mathematical manoeuvring computer program called

MATHMAN which had been created by Mr Hook. Into that program were fed

certain basic assumptions, in particular the hindcast calculated by Dr Cardone

and the hydrodynamic characteristics of the DERBYSHIRE as derived from

the MARIN model tests. However, the only certainties are that the vessel did

change course at least once before 0300z on 9th and that her average speed was

of the order of 4.4 knots.


62

4.35 Captain Roberts stated in evidence that he found it hard to understand a

decision of the master to depart from the rhumb line as early as 0400z on 8th

September as postulated by Mr Hood’s prediction, having regard to the sea

conditions then prevailing. The problem here is that, if it be assumed that the

master held his course, for several hours more at 8 knots, he could only reach

the 0300z position reported on 9th by navigating a “dog-leg” from the east.

Such a manoeuvre would probably put the sea approximately on the vessel’s

beam. This would be undesirable as a matter of good seamanship in relation

to a bulk carrier loaded with such a heavy cargo and it is therefore very

improbable that the master would have taken this course. Moreover, if that is

what the vessel had done, it could only have reached its 0300z position on 9th

if it had substantially reduced its speed, yet the prudent course would be to

maintain all possible speed. On the other hand, a change of course soon after

0300z on 8th would be consistent with the Mariners’ Handbook advice on

navigation in the dangerous semi-circle, namely to put the sea to 10º to 45º off

the starboard bow. If the master believed himself to be so positioned, it is not

improbable that he would have taken this course.

4.36 Again, it cannot be established with certainty when, before 0300z on 9th, the

vessel was hove to. However, the master mariners’ evidence was that the

master would have hove to when the significant wave height was about 6 or 8

meters. If, as is appropriate, the hindcast by Dr Cardone is taken as the

definitive analysis of sea conditions, a significant wave height of 6 meters

would prevail at 1500z on 8th and 7 metres at 2000z on 8th. Mr Hook’s

prediction would put the time at 1800z on 8th and that would therefore appear

to be a realistic conclusion.

4.37 Once hove to, the master would employ no more than enough engine power to

maintain steerage so as to keep the sea 10º to 45º off the starboard bow. The

course of the vessel plotted by Mr Hook assumes 22.5º, which accords with

that evidence.


63

4.38 It will be necessary later in this Report to examine further the course and

speed of movement of the vessel between the position reported at 0300z on 9th

and the wreck site (see section 5 of this Report). For present purposes it can

be concluded that the master kept the seas about 22.5º off the starboard bow in

accordance with the approved practice for navigation in the dangerous semi-

circle.

4.39 Finally, one further broadcast to the vessel was made by Oceanroutes at 0123z

on 9th September. It set out the forecast positions of Orchid from 0600z to

1800z on 11th, giving the maximum wind speeds as 80 knots.

4.40 It is to be inferred that the master was navigating the vessel in consultation

with the other deck officers and it is established that all were extremely

competent seamen. The master was a careful and somewhat cautious master.

He would be concerned to strike a balance between taking the speediest course

to the port of discharge in Tokyo Bay and the reasonable safety of the ship.

With regard to the latter he would clearly have had well in mind the passage in

the Mariners’ Handbook quoted above. In that connection, however, it is to be

observed that, whereas vessels were advised to keep outside a radius from the

storm centre of 200 miles, it was stated to be “of vital importance” only to

avoid passing within 50 miles or so of the storm centre.

4.41 It must also be inferred as more probable than not that the vessel received

most if not all of the Guam and Tokyo forecasts, including the Tokyo weather

maps.

4.42 This being so, there can be no doubt that the vessel was kept on its rhumb line

course until a few hours after 0300z on 8th and thereafter for several more

hours on a more or less northerly course because such assumptions as the

master made about its future maintainable speed of 10 knots and subsequently

8 knots in the course of 8th September turned out to be mistaken. Thus, if at

1200z on 7th September it had appeared to the master that there was a serious


64

risk that within 30 hours he would have been hove to in the position plotted by

Mr Hook with a significant wave height of 6.5 meters and a wind speed of

18.7 m/s with the typhoon taking its forecast track, it is inconceivable that he

would not there and then have changed course to the south west or west south

west. The same must be true at 1800z on 7th and at midnight on 7th/8th, as well

as on the morning of 8th. The perception that there was no serious risk to the

vessel by continuing either on the rhumb line or on the course taken after

0300z on 8th must also have involved a perception that there was no serious

risk that it might not be possible to maintain sufficient speed to keep a safe

distance ahead of Orchid.

4.43 It is necessary to ask whether in all the circumstances and on the information

assumed to be available to the master, this perception was unreasonable.

4.44 None of the very experienced master mariners who gave evidence before this

Investigation considered that Captain Underhill had taken an unreasonable

course at any stage during 7th or 8th September. They were of the view that he

was at each stage up to 1200z on 8th entitled to assume that he could retain his

rhumb line course and keep the vessel a safe distance from the storm centre.

In particular Captain Roberts, Captain Boyle and Captain de Coverly did not

consider that there ought to be have been a change of course to the south west

or south south west at 0600z on 7th September. Captain Roberts stated that,

the typhoon centre then being 270 miles from the vessel, its course might well

change before the present rhumb line route took the vessel to within 200 miles.

The master had the objective of getting the vessel and her cargo to the

discharging port and he would try to avoid changing course unnecessarily

early. He still had 30 hours in hand and would be able to test the proximity

and direction of the approach of the storm centre by reference to the direction

and freshening of the wind. He would have kept a close check on the progress

of the storm. Such storms have a tendency to re-curve to the north east. He

would still be able to extract the vessel from the track of the storm if that were

maintained and by having regard to the Tokyo forecast due at noon. Captain


65

de Coverly agreed: the future movement of Orchid was too uncertain to justify

a change in course or an increase or decrease in speed.

4.45 At noon on 7th September the master was confronted with the disparate Guam

and Tokyo forecasts for the next 24 hours to which reference has already been

made. While accepting that the maintenance of the assumed speed of 10 knots

was important in order to keep out of the 200 miles radius, Captain Roberts

was confident that the vessel could have kept up that speed: the increasing

swell running ahead of Orchid would not be anticipated to be so great as to

impede the vessel. It could always increase speed to get ahead of the storm if

the track shifted. Captain de Coverly also considered that he would have held

course at this time, there being a reasonably good chance that 10 knots could

be maintained, although the weather might reduce the speed at some stage

during the 24 hours to noon on 8th September. He would not have taken a

westerly or south westerly course at that stage, particularly in view of the

Japanese forecast. If the typhoon followed that westerly track, the vessel

might be in trouble. The vessel might run out of sea space in view of the

position of land to the west. The master could keep a careful eye on the

weather which at noon on 7th was still quite reasonable, the wind speed being

only about 20 knots and the significant wave height 2.27 metres, and using his

“seaman’s instinct” he would probably press on with his present course.

Captain Boyle also considered that the master could still be confident of

maintaining 10 knots for the succeeding 24 hours.

4.46 Captain Roberts considered that at 1800z on 7th September although the Tokyo

forecast range of Orchid’s track had moved closer to the track forecast by

Guam, the master was entitled to believe that the wind speed and significant

wave height, which had by then increased on Dr Cardone’s hindcast to 23

knots and 3.1 m respectively would not increase to such an extent that the

vessel was unable to maintain 10 knots. It still had reserves of speed which

would, if the sea state permitted, enable it to get ahead of the storm if the next

(midnight) forecast looked different. Captain de Coverly also would have held


66

course if he had been the master, in spite of the freshening wind and

increasing swell. The barometric pressure had not dropped drastically. He

recognised that the sea conditions might impose a small speed reduction to 8

or 9 knots. He would certainly not alter course until after day break or until

after the midnight forecast. It was still quite likely that the track of Orchid

would move west and pass south of the vessel, a view shared by Captain

Roberts. Captain Boyle in general expressed similar views.

4.47 As to the position at midnight on 7th/8th September, by which time the vessel

was assumed by the master mariners to be proceeding at 8 knots, Captain

Roberts considered that it was difficult to say if the assumption that this speed

could be maintained was justified. Certainly the master must by then have

anticipated that within the next 24 hours he would probably experience high

seas although he was far enough away from the storm centre not to experience

“huge” seas. Captain de Coverly described the master as in a dilemma.

However, he considered that the master would proceed to hold course, aiming

to keep the wind on his starboard bow if and when he entered the dangerous

semi-circle. Given the distance the vessel then was from the storm centre and

the vessel’s reserves of speed, the master would hope to be able to steer away

from the track of the on-coming storm. He agreed with Captain Robert’s view

that the master might well have been influenced in deciding to continue on

course by the risk of the typhoon moving to the west. Nor would he wish to

put the sea anywhere near his beam with a significant wave height of 3.7

metres, which was quite a big sea. This was not an enviable position. A

further consideration was that typhoons occurred so frequently in the north

west Pacific that, if he turned to the west or south west, the vessel might be

further endangered due to its proximity to the nearest land.

4.48 Captain Boyle was of the view that, given that the Guam and Tokyo forecasts

were by midnight on 7th/8th both indicating a more westerly track for the

typhoon, the master might well have been deterred from taking a south-


67

westerly course. He would still be justified in assuming that he could maintain

a speed of 10 knots on 8th September.

4.49 Captain Roberts further expressed the view that, in the light of the 0113z

broadcast message from Oceanroutes on 8th September, the master was

justified in assuming that he could maintain at least 6-8 knots on 8th which

would justify him holding to his course. Captain Roberts considered that the

deterioration in the sea conditions on the morning of 8th would have prevented

the vessel changing course to the south west after about 1000z on that day.

That would have involved bringing the increasingly heavy seas abeam. Dr

Cardone’s hindcast showed winds of over 30 knots (28º) at that time and a

significant wave height of 4.9 metres. Thereafter there was little the master

could do but maintain his course with the sea on the starboard bow.

4.50 The navigational decisions which left the vessel in that position and

confronted by the seas which it encountered after 1000z on 8th and

subsequently on 9th September were the result of the master’s having to

balance the desirability of taking as direct a route as possible to the discharge

port and the need not unreasonably to endanger the vessel. It has to be

appreciated that in trying to achieve that balance the master was confronted by

disparate forecasts of track of the typhoon from Guam and Tokyo, either of

which could have proved to be correct, and further by significant differences

between the actual reported track of the typhoon and the previous forecasts.

4.51 It was, however, the inability of the vessel to keep ahead of the storm on 8th

and 9th September which proved to be fatal to it. The sea conditions

encountered after 0300z on 8th must have been considerably more severe and

must have had a much more detrimental effect on the vessel’s speed than was

ever envisaged by the master when he held to his course at 1800z on 7th and

from midnight on 7th/8th. Yet at midnight the storm centre was still about 344

nautical miles away to the south east and even at noon on 8th it was probably

not much less than 250 nautical miles distant. In substance what was


68

happening was that the condition of the sea ahead of the advancing storm by

significantly more than 200 miles was far worse than might reasonably have

been expected in the north west sector.

4.52 Although in terms of wind speed and significant wave height Orchid was not

exceptionally severe amongst Pacific typhoons, it demonstrated a number of

relatively unusual characteristics.

4.53 Firstly, for a tropical cyclone it was of unusually long duration – some 53

hours. Thus out of 25 tropical revolving storms examined by Dr Cardone, 11

had significant wave heights of more than 10 m, but the range of the periods of

duration was from 6 to 36 hours.

4.54 Coinciding with this long overall duration was an unusually long period of

peak intensity. Thus the significant wave height along the track of the storm

centre and at the wreck site remained within 10 per cent of the peak over a

period of 24 hours. That, according to Dr Cardone’s evidence was most

unusual. Based on his experience of analysing and hindcasting tropical

revolving storms during the period from 1945 to 1998 Dr Cardone expressed

the view that as a very broad estimate the probability of occurrence of storm

conditions of such duration was less than 10 per cent and might even be as low

as 1 per cent. One consequence was that the vessel was exposed to very high

significant wave heights for an unusually long period of time.

4.55 Secondly, it was casting ahead of its centre an unusually widespread field of

relatively high winds. According to Dr Cardone’s evidence, the radius of the

maximum wind speed in Orchid was at times nearly 100 nautical miles, some

three to four times the average radius of maximum wind speed for a North

Pacific typhoon.

4.56 Thirdly, associated with that wind field, there was an unusually widespread

field of relatively high swell. Thus by 0300z on 8th September, when the

vessel was still over 300 miles to the north west of the storm centre, the wind


69

speed had probably reached 26 knots and the significant wave height had

already reached 4 metres, which meant that the vessel might well be

encountering occasional waves of 7 to 8 metres. During the following 7 hours

the wind had probably freshened to about 30 knots, but the swell had also

probably worsened to a significant wave height of 4.9 metres which would

probably expose the vessel to occasional waves of over 9 metres. The

unchallenged evidence of Dr Cardone was that the peak significant wave

height was about 40 per cent higher than the average to be expected from the

maximum wind speed of Orchid, which would be about 9 metres.

4.57 Not only were these conditions unusual, but they were not forecast by Guam.

Thus, if one compares the radii in the north east semi-circle from the storm

centre of the 30 knot and 50 knot sustained wind fields forecast by Guam 24

hours in advance on the one hand with the actual radii reported by Guam 24

hours later, there are seen to be very substantial disparities. Analyses of these

forecasts carried out by Dr Cardone have shown that from 1200z on 7th

September up to and including 1200z on 8th September Guam consistently

under-predicted the 30 knot and 50 knot wind fields. Dr Cardone has also

hindcast the sustained wind speeds and significant wave height at the vessel’s

conjectural positions from midnight on 7th/8th September to 1200z on 9th

September. The following table indicates the extent of the under-prediction in

relation to the vessel’s positions and the hindcast sustained wind speed at each

such position.

4.58 Thus, for example, whereas the Guam forecast at 1800z on 7th September

predicted that at 1800z on 8th September the 30 knot wind radius would extend

for 225 nautical miles from the storm centre, in the event it extended 420

nautical miles or 460 on Dr Cardone’s hindcast and at 210 nautical miles from

the centre the vessel experienced sustained winds of 45 knots.

SECTIO

N 4: N

AV

IGA

TION

70

Vessel’s anticipated distance from the storm centre forecast 24 hours earlier if she maintained speed and course

194 (at 10 knots)

253 (at 10 knots)

229 (at 8 knots)

c. 250 (at 8 knots)

Hove to

Vessel’s estimated actual distance from storm centre (nautical miles)

260

210

165

130

160

Dr Cardone’s hindcast of sustained wind speed in position of vessel (knots)

38.9

45.0

53.8

62.3

55.8

30 knots

420

420

400

480

480

Guam reported actual radii

50 knots

50

50

60

150

150

30 knots

450

460

480

450

450

Dr Cardone’s hindcast of wind field radii

50 knots

200

210

220

220

260

30 knots

225

225

225

360

360

Guam forecast 24 hours earlier of sustained wind field radii (nautical miles

50 knots

60

60

-

60

60

The extent of the under-prediction in relation to the vessel’s positions and the hindcast sustained wind speed at each such position.

Time/Date of actual conditions

1200z 8th September

1800z 8th September

0000z 9th September

0600z 9th September

1200z 9th September


71

4.59 It can be inferred that the master’s decision – taking as to his route would have

been strongly influenced by these Guam forecasts. He would have been led to

believe that, if the Guam forecast track of Orchid, for 8th/9th September, as

distinct from that forecast by Tokyo, proved to be correct, the wind field and

therefore the increased wave heights at distances of 190 to 250 nautical miles

north west and north of the storm centre would be likely to be significantly

less than they turned out to be and that therefore he would be able to sustain a

speed of 10 knots and, subsequently, 8 knots.

4.60 The master’s decision to depart from his rhumb line course and to take a more

northerly course, probably early on 8th September, can be assumed to have

been taken in the continued belief that the vessel’s speed could be maintained

at a high enough level to keep sufficiently ahead of the typhoon to avoid

encountering seas of such severity that the vessel would have to be hove to in

a position which could place it dangerously close to the path of the on-coming

typhoon as forecast by Guam and Tokyo at midnight on 7th/8th September.

Further, the master’s adherence to that altered course past the time when the

seas were too high to alter course to the south west, which was probably

between 1000Z and 1200z on 8th September, strongly indicates that he

continued to believe that the advantages of holding course, including

considerations of the vessel’s safety, still outweighed the disadvantages. That

the master was reasonably confident in his decision can be inferred from the

fact that even when the vessel was encountering the conditions that she was,

he did not consult Oceanroutes as to an alteration of course, as he had done on

3rd September (see paragraph 4.14 above). It is highly improbable that,

without asking for advice, he would have held to his course until it was too

late to go to the south west or south south west if he had envisaged as a serious

likelihood that he would be obliged to hove to in what could be the path of the

on-coming typhoon or that his speed would have to be so much reduced that

the vessel would be within the 50-75 miles danger zone of the centre. The


72

very considerable worsening of the wind and sea conditions over the next 8

hours must therefore have been outside his actual anticipation.

4.61 According to a hindcast of the barometric pressure at the vessel’s position on

7th/8th, September, that pressure had only varied by 1 to 2 Mb between

midnight on 5th/6th and midnight on 7th/8th, a variance which would not

indicate close proximity with the typhoon. However, a sharp fall in pressure

occurred during the next 6 hours to 0600z on 8th. This would have indicated a

sudden and obvious closing between the vessel and the typhoon.

Nevertheless, in view of the still considerable distance of the vessel from the

storm centre, according to the Guam and Tokyo radio forecasts, the master

would probably not have concluded that the sea conditions on the intended

route would deteriorate on 8th/9th with the rapidity with which they did. By

1800z on 8th September, according to Dr Cardone’s hindcast, the vessel would

probably have been encountering a sustained wind speed of 45 knots and a

significant wave height of 6.5 metres with occasional waves of the order of 12

metres. Conditions of this severity with the storm centre still about 200 miles

away to the south east would not reasonably be contemplated by an

experienced master with recourse to the Mariner’s Handbook, for, although

that would place him just in the dangerous semi-circle, the vessel would still

not be in the danger zone within “50 miles or so” of the storm centre, to be

avoided as a matter of “vital importance” (see paragraph 4.12 above).

4.62 For these reasons, it is to be concluded that the master must have

underestimated by a very substantial extent the adverse state of the sea running

ahead of the advancing typhoon. The actual conditions proved to be far worse

at that distance from the centre than he would have been led to believe by

anything in the Mariners’ Handbook, by the Guam forecasts or, probably, in

his own experience.

4.63 It is further to be observed that the DERBYSHIRE would be seen by the

master as a well-maintained and extremely robust vessel, well able to


73

withstand the very heavy seas that might be encountered in the dangerous

semi-circle. There was no general experience of properly-navigated Capesize

bulk carriers being destroyed by typhoons. It can therefore be taken that he

would be extremely unlikely to see the risks involved in navigating in the

dangerous semi-circle outside a radius of about 50-75 nautical miles from the

storm centre as being life-threatening, although he probably would have

envisaged that there might well be superficial heavy weather damage.

4.64 Following the vessel’s arrival at its reported position at 0300z on 9th

September when it was hove to, it is to be inferred that the master

concentrated on keeping the seas on the starboard bow at an angle of 25° with

the purpose of riding out the storm. By that time the wind had probably

reached a sustained speed of over 54 knots and the significant wave height 8.6

metres. If it be assumed, as calculated by Mr Hook, whose evidence was not

challenged, that the vessel reached the wreckage site by about 1700z to 2000z

on 9th September, the wind and sea conditions continued to deteriorate, and

would have reached a speed of 56 knots and a significant wave height of 10.86

metres respectively by that time. In the course of that day there was therefore

little that the master could sensibly do other than to maintain the vessel’s

heading with the sea on the starboard bow. Turning away to the south west or

south south west was by then no longer open to him without exposing the

vessel to massive seas abeam and abaft which would be considered to be very

dangerous and contrary to good navigational practice.

4.65 This report therefore concludes that the master acted reasonably in relation to

the navigation of the vessel throughout the period from 3rd September until the

moment of her sinking at about 1700 to 2000z on 9th September. In taking the

hour-by-hour navigational decisions which he did, it cannot be said that he

was in any way at fault. That is not to say that other master mariners might

not have taken different decisions in identical circumstances, in particular

from 1200z on 7th September to 1000z on 8th September. Navigation was not

in 1980 and is not now a high-precision science. It is essentially the


74

application of experience and instinct to the whole body of information –

forecasts as well as actually encountered wind and sea conditions and

barometric readings – available to a master at any one time.

4.66 Captain Underhill’s navigation of the DERBYSHIRE, to the extent to which it

can now be re-constructed, cannot be open to blame. The vessel encountered

a typhoon with an unusually wide wind field and associated severe sea

conditions unusually far from the storm centre. The master’s navigational

decisions were probably heavily influenced by the consistent under-estimation

by Guam of the wind field radius and by the continuing disparity between the

track of the typhoon on 8th as forecast by Guam and by Tokyo in the course of

7th September.


75

Criticism of Oceanroutes

4.67 The DFA formulated its criticism of the conduct of Oceanroutes (“OR”) in the

following manner:

“Oceanroutes Inc. is criticised for the poor quality and frequency of its

weather routeing information and advice provided to the vessel,

including its failure to advise during 7th September 1980 that the vessel

steer an alternative course to the west or southwest.”

4.68 In the course of the hearing these criticisms were refined in a manner which

can be summarised as follows.

(i) OR held itself out to the masters of the vessels in respect of

which it was engaged to provide weather routeing services as

an organisation which could be relied upon to warn masters if

the maintenance of their current route might expose their vessel

to dangerously adverse sea conditions.

(ii) Specifically, the exchange of messages on and between the 3rd

and 6th September to which reference has already been made

(see paragraphs 4.14 – 4.17 above) established an advisory

course following which the failure of OR to give weather

routeing advice, as distinct from information as to weather

conditions, on 7th and 8th September would have led the master

into a sense of false security in respect of his route.

(iii) The information available to OR as to the track of Orchid

forecast by Guam and Tokyo which would be more complete

than that available to the master in as much as it would have

included that shown on the Guam and Tokyo weather charts, as

well as the bulletins, together with meteorological statistics and

its professional expertise, would have led a weather routeing

service acting with ordinary professional skill and care to have


76

monitored the vessel’s actual route and to have advised the

master to alter course to the south-west on 7th or early on 8th

September.

(iv) Had that advice been given to the master of the

DERBYSHIRE, he would have been influenced by it in his

choice of route and, although he might have enquired of OR as

to its reasoned prognosis, he would probably have taken that

advice and altered course to the south west at some stage on 7th

or 8th September before the conditions had deteriorated so

much that the vessel could not be turned to port without being

placed with heavy seas abeam and abaft (see paragraphs 4.49 &

4.60 above).

(v) If the master had altered course early enough in accordance

with such advice the vessel would not have been exposed to the

dangerous sea conditions which it experienced on 8th and 9th

September.

4.69 These criticisms give rise to the following questions.

(i) Should adverse criticism under a Formal Investigation such as

this be confined to situations in which there can be identified

breach of some legal duty to some other party or to some

person not a party to the Investigation?

(ii) If adverse criticism must be confined to breach of duty

situations, was OR in breach of any such duty to any relevant

person by reason of its failure to provide weather routeing

advice on 7th or 8th September?

(iii) If adverse criticism is not necessarily to be confined to breach

of duty situations, is the conduct of OR to be criticised


77

adversely because it was in a general and wider sense

blameworthy?

(iv) If the answer to the questions (ii) or (iii) is Yes, what affect

would the giving of proper advice at that time have had on the

master’s navigation of the vessel?

4.70 Is breach of a legal duty required? To this question there can be no doubt

whatever that the answer must be No. This conclusion follows from the nature

of the purpose of such an Investigation. That is essentially twofold. Firstly its

purpose is to ascertain what was the cause of the loss of life or of the loss of

the vessel, or of the vessel being placed in imminent danger, including not

only the immediate physical cause but also whether that immediate physical

cause was in turn caused or contributed to by the conduct of any person or

organisation, whether blameworthy or not.

4.71 Secondly, its purpose is to make recommendations for the safety of shipping

generally with regard to such matters, for example, as ship design, methods of

navigation and governmental and classification society regulations generally.

The making of such recommendations is in most cases likely to arise from

some perceived deficiency in the conduct of some person or body or to the

content of governmental or classification society regulations or rules

respectively which caused or contributed to the loss of life or the loss of the

vessel or which placed the vessel in a position of imminent danger by which

such loss might have been caused. If the making of adverse criticism of the

conduct of any person or organisation were to be constrained by a pre-

condition that such person or organisation had been in breach of some legal

duty, the whole recommendatory purpose of such an Investigation would be

seriously curtailed. Major deficiencies in conduct amounting to

blameworthiness might be identified on the part of individuals or corporations

or governmental departments which might well, largely because they were


78

blameworthy, give rise to recommendations aimed at avoiding future

repetition of such conduct.

4.72 It is not difficult to envisage such blameworthy conduct on the part of persons

or organisations which owed no legal duty to anyone to avoid such conduct.

An obvious example is a government department in respect of the content of

an international Convention or of legislation relating to ship-design or a

classification in respect of the content of its rules.

Was OR in Breach of any legal Duty?

4.73 The only person with whom OR was in contractual relations in respect of this

vessel was the charterer. OR, although made the subject of criticism by the

DFA and although provided with copies of those criticisms and of the

evidence relied upon by the DFA, namely the statements of Mr Johnson and

Captain Mackie, did not appear and was not represented at the hearing.

However, in the public interest it was given the opportunity of making written

representations and of putting in such written evidence as it saw fit,

consistently with the uninterrupted progress of the hearing. None of that

evidence has been tested in cross-examination and there has been no

comprehensive disclosure by OR of relevant documents. In determining what

were the terms upon which OR contracted it is therefore necessary to proceed

on somewhat limited evidence. However, the following conclusions can be

formulated with reasonable confidence.

4.74 The contract between OR and the charterers would not be made on any

standard contractual form. It would probably be governed by the law of

California where OR is located and carries on business. It is not known

whether this differs in any material respect from English Law. It is

appropriate to assume that it does not. The basis of payment would be a fee for

each voyage. OR would initially provide a route from port to port. This

recommendation would have regard to wind and sea conditions likely to be


79

encountered with a view to the most efficient utilization of the vessel. The

services which OR stated that it would provide to ships’ masters are set out in

its brochure produced in about 1973 and probably still current in 1980. A

copy of this document is to be found in Appendix 12. Although OR has been

given ample opportunity for introducing evidence of a brochure in different

terms, which was current in 1980, it has produced no such document. There

has however been produced in evidence an OR document entitled “Masters

Technical Paper”. This document is not dated, but internal evidence makes

clear that it was issued after 8th December 1980. Its contents are significantly

different from those of the 1973 brochure. It is to be inferred that the 1973

brochure was that current at the time of the DERBYSHIRE’s last voyage.

4.75 Although there is no direct evidence that the master of the DERBYSHIRE had

on board a copy of any part of OR’s brochure, having regard in particular to

the fact that OR were employed by the charterers and not by the ship owners,

it is more probable than not that he would have had a copy at least of that part

of the brochure entitled “Ship Weather Routeing Procedures for Masters”. It

is quite likely that he had the full Brochure.

4.76 The Substance of OR’s services to masters indicated by that brochure is to the

following effect.

(1) OR will monitor the vessel’s progress several times a day.

(2) If the weather pattern develops differently from that initially

predicted OR may recommend a diversion.

(3) Special weather messages will be sent warning the master of

anticipated adverse conditions.

(4) OR develops its forecasts from regular ships’ observations sent

to it four times each day, backed up by national meteorological


80

offices’ forecasts and from its own data accumulated every six

hours each day since 1955.

4.77 The substance of that part of the brochure entitled Ship Weather Routeing

Procedures for Masters is as follows.

(1) The master should send to OR his noon position and current

heading every second day to enable OR to check its dead

reckoning of the vessel’s position.

(2) OR will check the weather to be anticipated for several days in

advance along the vessel’s current route and maybe “for the

purpose of avoiding an area of unfavourable weather” will send

the master a recommended change of route.

(3) If there is no recommendation to change route, that is because

OR has no suggestion to make. It will however, continue to

monitor the ship’s position against anticipated weather several

times a day.

(4) The following passage is as it appears in the text: “In the

vicinity of typhoons or hurricanes the Master has frequent

weather reports and local data for planning evasion tactics. All

Masters seem to agree that this is their responsibility, and this

often becomes more a matter of seamanship than weather

forecasting. Many Masters cable their intentions concerning

typhoon avoidance for our analysis and suggestions”.

4.78 In its submissions to this Investigation OR has stated that in 1980 specific

routeing advice for vessels in the vicinity of tropical cyclones was not

provided unless specifically requested. Whereas this procedure would be

consistent with the general terms of the 1973 brochure, the passage quoted

above does not suggest that OR will automatically suspend its advisory service


81

every time a typhoon is forecast to coincide several days later with the vessel’s

current route. To give the passage that meaning would be inconsistent with

the sense of OR’s underlying on-voyage weather routeing advisory function

defined by the main part of the brochure. The meaning of the quoted passage

appears to be that, once a vessel has come close to a typhoon or hurricane, a

master’s responsibility is to work out evasion tactics according to good

seamanship. No doubt this would be understood as involving the kind of

navigational operations recommended in the Mariner’s Handbook (see

paragraph 4.12 above). In such circumstances the master must primarily rely

on “frequent weather reports” and “local data” but can always ask OR for its

suggestions for typhoon, avoidance. There is however, no specific statement

that OR will thereupon cease to provide routeing advice. If OR wished to be

entitled wholly to suspend its diversionary advice service in the event of the

approach of a TRS or typhoon however distant, it ought to have used much

clearer words than those quoted.

4.79 At the time of the loss of the DERBYSHIRE there were two other substantial

weather routeing services in operation. These were a department of the British

Meteorological Office providing a ship routeing service on a commercial basis

and KNMI, the Royal Netherlands Meteorological Institute, also operating on

a commercial basis. It is convenient to refer to the British Met Office service

by the name which it now bears, namely MetROUTE.

4.80 Captain Mackie, who was called on behalf of the DFA, had been one of the

founders of MetROUTE and was in charge of it in 1980. The routeing service

which it provided included advice as to the avoidance of typhoons. It would

provide information and advice to vessels closing with typhoons on a 6 hourly

or even a 3 hourly basis, whether or not such advice was specifically requested

by the master. The MetROUTE forecasters would study satellite data, in

particular, and, as soon as they identified a feature that might develop into a

TRS, MetROUTE would advise routed vessels, informing them that the

development would be monitored and vessels kept informed. If a tropical


82

depression developed, MetROUTE would start to track both the depression

and that of its routed vessels that might be affected and would then issue

advice to vessels on their course and speed to ensure that their nearest

approach to the storm centre was not less than 250-300 nautical miles. The

particular distance from the storm centre advised would depend upon the

vessel’s course relative to the meteorological system, the character and

projected movement of that system and, in particular the radius from the

centre of winds of storm force and above. The radius of 250-300 miles, which

was known to exceed the then conventional 200 miles in the Mariner’s

Handbook, was adopted because it allowed a better margin for manoeuvre and

enhanced ship safety. There was nothing in its brochure which suggested that

routeing advice would be withdrawn when the vessel was in the vicinity of a

typhoon unless specifically requested by a master.

4.81 The other main provider of routeing services, KNMI, operated in 1980 under

“Routeing Codes” which included the following passage:

“Hurricanes”

As soon as the master decides to change from shore based weather

routeing to ship based hurricane navigation, he should inform the

routeing officer of his decision. Only wave forecasts are issued during

this period. In a case where a ship’s master has not resorted to ship

based hurricane navigation, and KNMI consider the value of

“strategic” weather routeing is falling, the routeing officer will send a

telegram, the opening word being “HURNA”.

Masters can always ask KNMI for their views on the route to avoid a

hurricane, however, the advice given must be considered in

conjunction with the master’s views based on his “on the spot”

evaluation.


83

Caution

“Due to the fact that the uncertainty of the path and speed of a

hurricane is much greater than that of a normal depression, the advice

of the routeing officer must be considered with a fair amount of

reservation and care. In the vicinity of hurricanes, navigation is often

more a matter of good seamanship.”

4.82 These words make it absolutely clear that unless and until either the master or

KNMI expressly withdraw from shore-based weather routeing, such advice

will be provided.

4.83 Whereas it was unquestionably open to any routeing service to define the

service which it offered within whatever limits it chose, in the event that those

limits were not clearly defined, the understanding which those in the shipping

industry would have of the service offered would be likely to be influenced to

some extent by the kind of service then offered by the other main routeing

organisations. For that further reason the words of the OR 1973 brochure

would not be understood as having the meaning identified in paragraph 4.82

above.

4.84 Since OR had not contracted with the charterers so as to divest itself of an

obligation to provide on-voyage weather routeing advice, it would be in

breach of contract if it failed to provide advice to change course when the

exercise of due skill and care would require the giving of such advice. It

would further be in breach of a duty of care owed to those on board and indeed

to the owners of the vessel if it gave negligent routeing advice to the master.

The words of the Ship Weather Routeing Procedures for Masters included the

following:

“If advisable, we will send the Master a recommended change in route.

This may be for the purpose of avoiding an area of unfavourable


84

weather or to shorten the route and save time, if weather is better than

expected.

If the master does not hear from us again after receiving the original

route recommendation, it is because we have nothing further to

suggest, but we will be checking the position of the ship against

anticipated weather several times each day.”

4.85 Consequently a failure to recommend a change of route would be understood

by a master as an affirmative approval of the current route. It would thus be

capable of amounting to an implied representation to the master that it was

unnecessary to alter course. If that advice were negligent there would exist the

basis of a breach of duty of care by OR.

4.86 There was a strong body of evidence called by the DFA that by midnight on

7th/8th September at the latest, a weather routeing service, such as OR, ought,

in the exercise of proper professional skill and care, to have advised a change

of course to the south-west or west-south-west.

4.87 Thus, Captain Mackie was very firmly of the view that by 0600z on 7th

September MetROUTE would have advised the Master of the latest position,

central pressure, increased speed and projected direction of movement of the

typhoon and would have requested that the master should provide details of

his course, speed and local prevailing conditions. He would have been

requested to provide six-hourly reporting and although no alteration of speed

or course would have been suggested at that stage, the Master would have

been advised to maintain radio watch. At 1200z on 7th September MetROUTE

would and OR should have advised the Master to divert to port on a west-

south-west to south-west course at not less than 10 knots. It was more likely

having regard to the Guam and Tokyo weather maps and his meteorological

experience that the typhoon would move on a north westerly course, then re-

curving on to a northerly or north-easterly course, than that it would move


85

west, as had been suggested by the Tokyo forecast. The available data showed

that wind forces and barometric pressure falls were greater to the north-east.

This confirmed that the typhoon would follow a north-westerly or northerly

course rather than a westerly course.

4.88 For these reasons, by midnight on 7th/8th September, according to Captain

Mackie’s evidence, advice ought to have been given to divert to port on a

westerly or south-westerly course without delay. To continue on a present

rhumb line course would be hazardous and accordingly, MetROUTE would

have worded the advisory message to the Master so as to make it plain to him

that drastic action was then required. The assumption that a speed of 8 knots

could be maintained throughout the following day was not justifiable having

regard to the sea conditions to be anticipated. The message sent by OR at

0113z on 8th September, which did not indicate the present position of the

storm centre or its present speed or track and which was the first advice sent

by OR since 2146z on 5th September, was entirely inadequate in the

circumstances.

4.89 Mr Johnson of MetROUTE also gave evidence as to the advice which his

organisation would have given to the DERBYSHIRE on the basis of the

available meteorological information. He was not employed by MetROUTE

at that time and his evidence is therefore of very limited weight. However, he

was clearly of the view that as early as 0600z on 7th the master would have

been advised to consider altering course to the west or south-west. Similar

advice would have been given at 1200z and 1800z on that day. By midnight

on 7th/8th September the vessel was, as he put it, almost “in a no option

situation” in which MetROUTE would have strongly recommended altering

course. It would not have waited until the master asked for advice.

4.90 Although Mr Johnson’s evidence was not consistent with that of Captain

Mackie in all respects, its substance was that a routeing service ought properly

to have maintained a dialogue with the vessel from the dissipation of TD16 on


86

5th September onwards and on 7th and early on 8th September ought to have

advised the master to divert to the west or south-west.

4.91 OR has stated that its practice in 1980 was to estimate a vessel’s position by

hand calculations based on the previous day’s estimated steering distance

along the assumed track and to issue to any vessel predicted to be within the

area of influence of a storm centre within 48 hours advisory notices

identifying the position and forecast track of the tropical cyclone during the

organisation’s regular servicing shift for vessels in the western Pacific (1500-

2300UTC). Vessels predicted to be outside the “area of influence” were

earmarked for review during the next servicing shift. As subsequent storm

warnings were received, supplemental advisory notices would be issued to the

vessel if the forecast storm track changed significantly at the 48 hour forecast

position. It is not entirely clear what is meant by “area of influence”, but this

is probably intended to refer to the area within which the vessel would

experience wind and swell conditions that might adversely affect its

navigation in request of speed or route.

4.92 In the case of the DERBYSHIRE, according to OR, following receipt of the

initial warning of Orchid at 0200z on 7th September, no advisory message was

sent to the vessel on that day, probably because, based on its current course

and speed, the vessel was estimated to be outside the “area of influence” of the

storm at 48 hours. OR states that it relied exclusively on the JTWC forecasts

from Guam. It relies on the fact that these very substantially underestimated

the wind radii to be anticipated on 8th and 9th September and that, also up to 8th

September, these forecasts placed the track of the typhoon increasingly further

to the west. OR’s position appears to be that in reliance exclusively on these

forecasts it issued no advisory message to the vessel until that at 0113z on 8th

September. That message is said to have been prepared late on 7th September.

It must therefore have been based on a forecast issued by Guam earlier than

midnight on 7th/8th.


87

4.93 OR further state that they did not issue to the vessel six-hourly typhoon

warnings re-iterating those broadcast by Guam because of the amount of time

it would ordinarily have taken for their message to have been received by the

vessel via KPH Radio in San Francisco. This was likely to be as much as 3 to

6 hours. This might produce a situation in which the vessel received OR’s

message at a time when it had already been out-dated by subsequent six-

hourly Guam forecasts which the vessel was likely already to have received

direct. Having regard to OR’s reliance exclusively on the Guam forecasts, it

claims that it was in no better position than the vessel to decide when a change

in course was available.

4.94 If, as this Report concludes, OR was under a contractual duty to give to the

vessel such timely advice as to weather routeing, including evasion of tropical

revolving storms, as proper professional care required, the fact that a period of

delay even of 3 to 6 hours might elapse between the issue of a typhoon

warning and a forecast from Guam and receipt by the master of OR’s advice to

change course would clearly not justify OR in failing to give such advice. In

any event the insertion of a validity time would have avoided any confusion.

4.95 Further, the evidence which has been placed before this Investigation very

strongly suggests that, given that the vessel was fitted with the Selcal

communication system and that KPH Radio San Francisco had that system

available 24 hours a day, the maximum delay in the vessel receiving messages

from KPH was likely to be no more than 2 hours. The Selcal system involves

the vessel having a unique communication number and the agreement of a pre-

designated broadcasting frequency upon which the sender station will transmit

to that number and to which the receiver station will be tuned. The ship’s

equipment would print out the transmitted messages, so that the radio officer

would not have to be present at the moment of transmission and receipt. It can

therefore, be concluded, contrary to OR’s assertion, that, had advice been

given to the vessel to alter course by means of a message broadcast by KPH


88

on the Selcal system, it would have been received by the vessel at the most

within 2 hours of its having been sent by OR.

4.96 OR’s statement that it relied for its advice to vessels in the western Pacific

exclusively on the Guam forecasts is somewhat surprising. Both Tokyo and

Hong Kong were also broadcasting independent meteorological reports. The

DERBYSHIRE was receiving Tokyo bulletins and weather map faxes as well

as Guam bulletins. Given that OR was in business to provide weather routeing

services, it is hard to understand why, otherwise than for reasons of economy,

it confined its sources of meteorological data to Guam forecasts. Readers of

its brochure would assume that it relied on all reasonably available

meteorological information.

4.97 Given the meteorological information available to it, was the failure of OR to

advise the vessel to alter course to the west-south-west on 7th or early on 8th

September professionally negligent?

4.98 As is clear from the evidence of Captain Mackie and Mr Johnson, whether any

such advice should be given would depend substantially on the routeing

organisation’s view as to the vessel’s ability to keep a sufficiently safe

distance ahead of the advancing storm centre if it is continued on its present

course.

4.99 As is to be inferred from its message sent to the vessel at 2146z on 5th

September, OR was working on the assumption that it was desirable to

maintain a 200 nautical miles safety margin from the storm centre. Although,

according to Captain Mackie, MetROUTE worked on a 250-300 nautical

miles margin, there can be no criticism of OR for taking precisely the same

radius as that in the current edition of the Mariners’ Handbook.

4.100 Starting therefore from the assumption that the vessel ought to be given advice

to change course if its anticipated route would bring it within the dangerous


89

semi-circle at a distance substantially within the 200 nautical miles radius,

whether such advice was given would depend upon:

i. the routeing services’ perception of the track of the storm

centre over the succeeding 48 hours;

and

ii. the assumed speed of the vessel during that period if it

continued on the existing route or on any diversion route.

4.101 The latter would clearly depend on the wind and sea conditions to be

anticipated on the relevant route.

4.102 As to the anticipated track of the storm centre, there was a substantial

divergence between the forecast given by Tokyo and that given by Guam. As

appears from the positions plotted by Captain Roberts in Appendix 10 the

former consistently placed the track substantially to the west or west-north-

west, rather than to the north-west or north, which was preferred by Guam. As

already noted, Captain Mackie stated in evidence that it ought to have been

appreciated by OR by 1200z on 7th September that the typhoon was less likely

to follow the Tokyo track than the Guam track by reason mainly of the

current, wind and barometric pressure evidence in the area north-east and

north of the current centre.

4.103 This Report concludes that it would be quite unrealistic to assume that a

weather routeing service such as OR could be expected as a matter of

professional duty to improve on the track forecasting of national

meteorological organisations, such as Tokyo or Guam, which could be

expected to have far more comprehensive forecasting resources available to

them, unless it were established that OR had information on the basis of which

any such routeing agency ought to have appreciated that either forecast was

wrong.


90

4.104 In the absence of such evidence OR could not be said to be at fault by reason

of having failed to advise a change of route on the basis of preference for the

track forecast by Guam when the Japanese forecasters, who must have had

access to data similar to that available to Guam, concluded that a much more

westerly track was probable. The fact that OR, as it claims, simply relied on

Guam’s forecast cannot render its abstention from diversionary advice

blameworthy if a reasonably careful routeing service, having taken into

account the Tokyo forecast as well as the Guam forecast, could have

concluded that the track forecast by Tokyo was the more likely one or was at

least as likely as that forecast by Guam or one which was a realistic

possibility. In spite of Captain Mackie’s evidence it has not been made out on

the evidence that any reasonably competent routeing service should have

appreciated that the Tokyo forecast was untenable and should be ignored, as

distinct from being less probable than the Guam forecast.

4.105 Further, even the Guam forecasts from 1200z on 7th September, although

closer to the typhoon’s ultimate track, did not place the probable route of the

vessel within the dangerous distance of 75 nautical miles of the storm centre

until midnight on 8th/9th September, long after the latest time (noon on 8th

September) when it would have been possible for the vessel safely to alter

course to the south-west. Nor did the Guam forecasts at 1200z and 1800z on

7th or at midnight on 7th/8th suggest that by 1000z to 1200z on 8th the wind and

sea conditions almost 250 nautical miles west-north-west of the storm centre

would have deteriorated to such an extent that the vessel would no longer be

able safely to change course to the south-west. Indeed, as this Report has

already concluded, those forecasts consistently and substantially

underestimated the wind radius in the north east semi-circle.

4.106 This unpredicted wind radius in the north-east semi-circle was accompanied

by the development of a swell to the north-west of the storm centre which

substantially exceeded that which would ordinarily be encountered at 200 to

350 nautical miles from the storm centre. Just as the Guam forecasts, which


91

were based on high quality meteorological expertise and substantial local data

under-estimated the wind radius in the north easterly quadrant, it is to be

inferred that the omission of a routeing service such as OR to appreciate the

extent of the likely wind radius could occur without any lack of proper care on

its part. By parity of reasoning, so also could it be inferred that it was not

professionally negligent for a routeing service, having failed to anticipate the

extent of the wind radius and the consequent extent of the increase in swell, to

advise that the vessel should therefore change course to the south west.

4.107 It was suggested by Captain Mackie in a written report issued subsequently to

his oral evidence that OR ought to have spotted the consistent under-

estimation of the wind field radius by Guam and should have warned the

vessel at least from 1200z on 8th September that Guam had so under-

estimated. Although there is some force in this point, it is of limited weight.

Firstly Guam itself had already, at 0600z on 8th September, forecast a

substantially increased 30 knot wind field radius (360 nautical miles increased

from 225 nautical miles given 6 hours earlier). Secondly, that information

would not have assisted the DERBYSHIRE, for, by the time OR’s advice was

received by the vessel at some stage after 1200z on 8th September it would

already be too dangerous to attempt to alter course to the south west and

advice to that effect should not have been given by OR. If given, it could not

have been acted on safely by the master.

4.108 Accordingly, this Investigation concludes that throughout the period 1200z on

7th to the morning of 8th September a routeing agency in the position of OR

could, without lack of proper professional care, have formed the view that:

(i) it was just as likely that during 8th/9th September the storm

centre would follow a westerly track as forecast by Tokyo

distinct from a northerly/north-easterly re-curving track, as

forecast by Guam, or at least that there was a serious risk that it

would follow such a track;


92

(ii) that in either event there was no reasonable probability of sea

conditions on the current course of the vessel which were so

severe that it would lose speed to such an extent that it might

have to be hove to in the course of 8th September and thereby

caught in the path of the typhoon if it took a north-westerly or a

northerly track as forecast by Guam;

(iii) it was therefore both undesirable and unnecessary to advise the

master at any time prior to 1000z on 8th to alter course to the

south-west.

4.109 For these reasons this Report concludes that it is not established that OR was

at fault in failing to advise the master to alter course prior to noon on 8th

September.

4.110 The very strong expressions of dissatisfaction with the conduct of OR

expressed by Captain Mackie and Mr Johnson were based upon a failure to

appreciate that OR was entitled to rely on the forecasts provided by Guam and

that both the Guam and Tokyo meteorologists were finding it very difficult

accurately to predict, even 24 hours in advance, either the track or the wind

radius of the typhoon. The consequent under-estimation of the swell likely to

be experienced in the north-westerly sector and the failure to appreciate that it

would prevent the vessel from maintaining sufficient speed to maintain her

course is not, on the whole of the evidence, a matter for which a routeing

service such as OR could be held to blame.

4.111 The expert master mariners, Captain Roberts, Captain Boyle and, although he

had no personal experience of weather routeing, Captain de Coverly all said in

evidence that they would have expected a weather routeing service such as OR

to have advised on 7th and 8th whether to hold course or to deviate to the south-

west and specifically to turn to port. This Report cannot accept that evidence

as a basis for concluding that OR was at fault. The very striking feature of


93

these tragic events is that at no time after Orchid was identified as a tropical

revolving storm did Captain Underhill ask the advice of OR as to whether to

hold or change course. That hardly suggests that at the time of the voyage

experienced masters generally would have expected to receive unsolicited

routeing advice regardless of whether a change in route was recommended.

He was entitled to assume in these circumstances that OR did not have any

positive recommendations as to a change of route. As already indicated, that

was a view that OR was entitled to take.

4.112 It remains to add that, although the centre of Typhoon Orchid somewhat

unusually described three loops during the period of 24 hours between 1200z

on 9th and 1200z on 10th September, this characteristic had no material effect

on the ability of the vessel to make sufficient speed to evade the worst sea

conditions associated with Orchid. The vessel was already hove to in very

heavy seas many hours before the development of the first loop. Further, Dr

Cardone’s hind-cast of the track of the highest significant wave heights

attributable to the typhoon suggests that the incidence of the three loops did

not cause heavier seas to be nearer to the likely position of the vessel at any

time on 9th September than, would otherwise have been the case.

4.113 It is right to add that if on 7th or the morning of 8th September OR had advised

Captain Underhill to change route to the south-west, that advice would

probably have been taken. The master was, on the evidence, highly competent

and not the sort of man who would take undue risks with his vessel.

4.114 This Investigation has further considered whether OR’s conduct in this matter,

although not amounting to breach any legal duty, could be adversely criticised

because it was in a general sense blameworthy. The failure of OR to provide

the master with information between 2146z on 5th September and 0113z on 8th

was not, in the view of this Investigation, blameworthy. However, in view of

the development of Orchid by late on 7th, the content of OR’s 0113z message

was distinctly deficient. In particular, it failed to identify the current position


94

of the storm centre or its track to the 48 hours forecast position. Nor did it

give the wind speed during the succeeding 24 hours. However, in view of the

fact that the vessel was receiving the Guam and Tokyo bulletins and the

Tokyo weather map faxes, it is highly improbable that the master was

materially influenced by this message in deciding to hold to his course.

Although he might well have assumed that the absence of advice to change

course at that time was tacit approval by OR of the current course, such advice

would not, as this Investigation has concluded, have been professionally

negligent on the part of OR.

4.115 Finally, the contents of the brochures upon the basis of which OR conducted

its services in 1980 were far less clear than they ought to have been. This

Report has concluded that OR undertook a continuing duty to provide positive

routeing advice even in the event that a typhoon was identified and had not

divested itself of that duty by the specific references to the master having the

primary responsibility for routeing in the vicinity of a typhoon. The wording

used was less than clear and if it was indeed the purpose of OR to exclude its

obligation to provide such advice it should have said so in very clear words so

that a master knew without doubt what he could expect. The brochure issued

by OR in about 1981 was more explicitly worded. That currently issued by

KNMI and set out in Appendix 13 was clearer still.

4.116 For reasons which will be apparent from this report it is important that weather

routeing services, such as OR, should make very clear in their documentation

exactly the scope of the services which they are undertaking to provide.

Misunderstandings of that scope could have dangerous consequences for

vessels using those services.


95

Section 5

Sea Conditions and Vessel Movements on 9th September

5.1 This Report has already concluded that the DERBYSHIRE was probably hove

to from about 1800z on 8th September and that, given the sea conditions

prevailing after that time, the master would employ no more than enough

engine power to maintain steerage so as to keep the sea 100 to 450 off the

starboard bow. The vessel reported its position at 0300z on 9th September.

Mr Hook plotted the track of the vessel from that position to the position of

the wreckage site, making certain assumptions, including the significant wave

heights hindcast by Dr Cardone and the hydrodynamic characteristics of the

vessel derived from the MARIN tests. By employing the MATHMAN

computer programme Mr Hook was able to plot a conjectural course to the

wreckage site see Appendix 14 and to estimate the time range during which

the vessel must have arrived there. There can however, be no certainty about

this. The highest it can be put is that the justification for Mr Hook’s course

and arrival time is based on a sound logical approach which, according to the

evidence before this Investigation, appears to take into account of all known

and estimatable relevant factors.

5.2 The evidence of Mr Hook, which has not been effectively challenged by any

party, can be summarised as follows.

5.3 In order to proceed to the wreck site from the 0300z reported position on 9th

September the vessel would be most unlikely to be hove to for any sustained

period with the seas head on to the bow or very fine off the bow. This would

be because the speed would then have to be so low, about ½ knot, that the

vessel could not have maintained her heading in the prevailing sea conditions.

She must therefore clearly have been making some headway for most of this

SECTION 5: SEA CONDITIONS AND VESSEL MOVEMENTS ON 9TH SEPTEMBER

96

time because otherwise she could not both have held her course and reached

the wreck site. At best, therefore, it is possible only to assume that the vessel

would have been making some headway continuously and that, depending on

her heading in relation to the seas, she would have a certain average forward

speed over the ground. Mr Hook calculated that mean speed as in the range

1.5 to 2.5 knots. His evidence was that there could be a small variation about

those means due to wave drift forces, but that such variation would not amount

to more than +/- ½ Knot. The finer the angle of the seas off the bow, the

lower the speed over the ground would be.

5.4 Within that average range there might be a good deal of variation in speed

from hour to hour, depending on the immediate wave conditions. So, if the

engines were run at a constant engine speed, as would be expected, the

vessel’s speed over the ground would vary according to the fluctuating wave

height. Thus, upon encountering a lower wave in a series, the vessel would

speed up, maintaining that enhanced speed until meeting a higher wave. In the

result the velocity of impact would increase with increasing wave height. The

speed variance would however not be likely to exceed ½ knot from the mean.

5.5 Starting therefore from Dr Cardone’s hindcast, which is reproduced at

Appendix 15, it will be seen that for 12 hours the route of the vessel

experienced significant wave heights (that is the average height of the one

third highest waves) above 9.5 metres, and rising to a peak of 10.85 metres at

1700z. It is calculated that the master would probably hold the sea at about

250 off the starboard bow. An engine speed of 80-90 RPM would probably

have to be maintained to hold this heading. If that were adopted, the average

forward speed would be about 1.75 to 2 knots, depending on wave drift forces.

On those assumptions, the vessel would have arrived at the wreck site between

1700z and 2000z on 9th September. Since, by hindcast, the worst conditions

by reference to significant wave height on her route were experienced at

1700z, it must be somewhat more likely that the sinking commenced shortly

after 1700z rather than later in that period. However, that probability has to be


97

discounted to a limited extent by the possibility referred to by Dr Cardone that

the individual highest wave could occur outside the precise measurement

period of highest significant wave height.

5.6 The submission was advanced on behalf of Lloyd’s Register that, based on a

mean speed of 1.7 knots, at the period of encounter with the most severe

conditions the DERBYSHIRE would have been making zero speed over the

ground. This is not based on any firm evidence. Nor does it effectively

displace on any logical or mathematical grounds the speed calculations put

forward by Mr Hook. Above all, it does not take into account variations in

velocity attributable to the differing heights of the individual consecutive

waves in a series. Nor can the evidence of the master mariners, Captain

Davies and Captain Willey, of their handling of the DERBYSHIRE’s sister

ships be said to support it, for, on the voyages of which they spoke, those ships

were not carrying comparable cargo in comparable sea conditions.

5.7 This Report concludes that the most probable mean speed of the vessel, during

the period of one hour prior to the commencement of sinking, was in the range

1.5 to 2.00 knots. Although it is impossible to estimate precisely the highest

impact velocity with the highest waves, it is likely to have exceeded 2 knots

and might have reached 2.5 knots.

5.8 Dr Cardone’s hindcast analysis of the conditions on the vessel’s assumed route

suggests that from 0400z on 9th the vessel was encountering a significant wave

height above 9 metres and reaching 10.85 metres 13 hours later (1700z), after

which the significant wave height did not reduce below 10 metres until about

0600z on 10th September, except for a short 2 hours period between 2100z and

midnight on 9th/10th. During this 13 hours period the peak spectral period

varied from 13.02 seconds to 13.96 seconds (at 2000z on 9th). A wave period

of 14 seconds would ordinarily give rise to a wave-length of about 300 metres.


98

5.9 When a vessel is near to head on to the waves it will be subjected mainly to

pitching effects. The extent of these effects is dependent on the relationship

between the length of the vessel and the wave length. It will also be

dependent on the damping characteristics of the vessel’s configuration,

notably the relationship of draught to beam. When the wave frequency is

relatively low and the wavelength relatively large, the vessel will tend to

follow the slope of the waves with the effect that there will be little relative

vertical motion. By contrast, where there is a high wave frequency and a short

wave-length relative to the vessel’s length, the vessel will not respond to the

waves in terms of pitch and heave, thereby creating relative vertical motion

broadly equal to wave elevation. However, between those positions, as the

wavelength shortens, the slope of waves of the same wave height increases.

That causes the vessel’s pitch response to increase. There eventually comes a

point when the wavelength is broadly equal to the length of the vessel. It is at

this point that the vessel experiences maximum relative vertical motion. A

pattern of movement is thereby created which involves the vessel pitching

head-down into the trough of a wave and then, before its bow has lifted back

to horizontal, meeting the next wave crest.

5.10 Where the waves are very high, they tend to have very steep wave fronts

which allow little time for a ship to rise above them. The effect of the vessel’s

pitching in the manner described is to increase the relative vertical motion of

the vessel quite substantially. Where relative vertical motion exceeds the local

freeboard, green water will begin to load the vessel’s bow. The more the

relative vertical motion, the greater will be the head of green water on the

bow.

5.11 The length of the DERBYSHIRE (at 294.1 m overall) was slightly less than

the wavelength (about 300 m) that would typically have prevailed during the

period 1700z to 2000z on 9th September. The conditions therefore gave rise to

accentuated pitching and considerable amplification of relative vertical motion

to wave amplitude so exposing the bow and forward hatch covers to heavy and


99

repeated green water loading; a response by the vessel strongly evidenced by

the model tests conducted at MARIN which must now be considered.

• = Please see the accompanying CD-ROM for an animation of the actions

described in paragraphs 5.9 – 5.11


100


101

Section 6

The Marin Model Tests

6.1 The first model tests relating to the DERBYSHIRE were carried out by the

Ship Stability and Research Centre, University of Strathclyde in 1998.

6.2 These were based on a model of the vessel of a scale 1:65. They were also

directed to investigate the response of the vessel in the seastate which had

been hindcast by Dr Cardone for the Original Formal Investigation and in

unidirectional seas. Following completion of these tests, Dr Cardone revised

his hindcast exercise for the purposes of the present Investigation using the

latest technology. This was based on multidirectional seas and led to results

significantly different from those originally arrived at. The results only

became known in October 1999. It was then appreciated that the Strathclyde

model tests might be of little assistance. Further, the effect of multidirectional

waves might be of considerable relevance to the thesis advanced by Professor

Faulkner that in a tropical typhoon pyramidal waves of extreme height could

have been encountered.

6.3 Since the facility at Strathclyde was not capable of producing multidirectional

waves, arrangements for further tests on the same model of the

DERBYSHIRE were made with MARIN Research Institute, Wageningen,

Netherlands. The immediate purpose was to measure the magnitude of green

water impacts on No.1 hatch cover in different conditions of trim, namely that

which she would have had if no seawater had first entered the bow spaces (the

“intact” condition) and second when she had already taken water into the bow

spaces (the “damaged” condition). Tests were conducted in November 1999.

Further tests were conducted in January 2000 directed to ascertaining the rate

of water ingress attributable to shipping green water in the seastates replicated.

SECTION 6: THE MARIN MODEL TESTS

102

Measurements of green water loading on No.1 hatch at speeds of zero and 4

knots were also taken.

6.4 Finally, a further series of tests was commissioned to be conducted by

MARIN some weeks after the commencement of the hearing of this

Investigation. These additional tests, which were much more comprehensive

than either previous series, became necessary because the validity of the water

ingress rates found in the January series was called in question by Professor

Faulkner on the grounds that the foredeck furniture and freeing ports on the

DERBYSHIRE had not been properly replicated in the model. Secondly, it

was thought important to position ingress holes to correspond with each of the

11 damaged ventilators and air pipes on the foredeck. Thirdly, it was thought

more helpful to conduct load tests on a 12 panel hatch cover than on one

continuous one as in the previous tests. Finally it was found that certain

important dimensions of the model used for the two previous test series had

been inaccurate. Finally, as to speed, MARIN was asked to conduct load tests

at 2 knots as well as at zero and 4 knots and also to calculate the maximum

achievable speed which the vessel could have attained in the prevailing

conditions as hindcast by Dr Cardone.

6.5 The MARIN speed tests gave results broadly consistent with Mr Hook’s

calculations (see paragraph 5.3 above). Thus, for a 10.85 metres significant

wave height MARIN gave a mean 2.135 knots and Mr Hook calculated 2.65

knots. MARIN did not record a speed for a significant wave height of 12.5

metres, whereas Mr Hook calculated 1.7 knots. This report accepts that during

conditions of the severity encountered at about 1700z on 9th September, it is

unlikely that the master would have kept up his speed to full maximum

attainable speed unless he needed to do so to maintain the vessel’s heading.

The more likely speed was about 70 per cent of maximum attainable speed,

which would be on average 2.1 knots over the period from 12 noon on 9th to

1700z on 9th. It is impossible to be very precise in spite of the extent of the

tests and calculations and it is for this reason that this Report concludes that


103

the actual mean speed over the ground is unlikely to have been much more

than 2 knots, and might have been as low as about 1.5 knots.

6.6 The MARIN results are shown in the table at Appendix 16.

6.7 It will be observed that in the intact condition the model tests for 1700z on 9th

September at speeds of 0 and 2 knots, with a significant wave height of 10.85

metres, gave maximum impact pressure on the No.1 hatch cover which did not

exceed 24 kPa. At a significant wave height of 12.5 metres (10.85 + 15%) the

maximum impact pressure at 0 and 2 knots was 29.9 kPa. These are well

below DERBYSHIRE’s No.1 hatch cover collapse load of 42 kPa. It was only

when the speed was increased to 4 knots in the 12.5 metre significant wave

height that a maximum loading slightly in excess of 42 kPa was recorded.

6.8 When it came to the damaged condition, however, the results gave a very

different picture.

6.9 As can be seen from Appendix 16, the maximum measured impacts were all in

excess of 42 kPa, even at zero speed, for significant wave height of 10.85m.

For a significant wave height of 10.85m increased by 10% the maximum was

51.6 kPa and it was 55.5 kPa for 10.85m increased by 15% to 12.5 m.

• = Please see the accompanying CD-ROM for a video sequence showing

the model tests at 10.85 m to 12.5 m in a bow damaged condition.

6.10 The MARIN tests had their limitations. Thus, questions have been raised

about the results obtained as to the rate of bow flooding, in particular whether

the rate of ingress recorded through the apertures in the model may have been

distorted by comparison with the actual rate of ingress on the vessel from a

similar head of water. Secondly, MARIN’s damaged condition tests all

assumed that all the available bow spaces were fully flooded: there were no

tests with only parts of the bow, such as the bosun’s store, fully flooded.

Further, the tests were all conducted over periods of 90 or 120 minutes. Yet it


104

is clear that the DERBYSHIRE was exposed to a period of significant wave

heights of about 10 m or more for a much longer period, viz from 0800z on 9th

September to 0500z on 10th September.

6.11 Therefore the first area of vital information needed to be derived from these

tests, so far as this Investigation was concerned, was to ascertain to what

extent it was possible that over a period of time much longer than the actual

periods of the tests and closer to that of the vessel’s true exposure to severe

conditions she might have encountered wave loading of sufficient intensity to

cause her No.1 hatch cover to collapse, even if such loading had not been

recorded during the tests.

6.12 The second vital area of information needed to be derived from the MARIN

tests was to ascertain the effect on the experience of wave loading recorded in

the tests of partial as distinct from total flooding of the bow spaces.

6.13 In order to investigate and, if possible, conduct the complex exercises in

extrapolation which the need for this further information appeared to require,

this Court appointed Professor Jonathan Tawn of Lancaster University as sole

expert in extreme value statistical extrapolation., He was appointed late in the

course of the hearing and with the assistance of Dr Heffernan, he produced

with remarkable speed detailed reports upon which he then gave oral evidence

of an extremely explicit and helpful nature. It is no exaggeration to report that

the combination of his evidence and the results of the MARIN tests has proved

to be of absolutely fundamental importance to the outcome of this

Investigation.

6.14 Before coming to Professor Tawn’s extrapolations, it is important to stress, as

he emphasised in the course of his evidence, that the extrapolated results

depend for their accuracy in replicating realistic probabilities of the occurrence

of events, such as the incidence of a wave of a height sufficient to destroy a

hatch cover, on the accuracy with which the MARIN test data replicate true


105

conditions. Here again there is an intrinsic area of uncertainty. Not only are

there questions surrounding MARIN’s measurements of water ingress rates,

but the hindcast of seastates by Dr Cardone to which the MARIN tests were

directed was, according to his evidence, subject to a 10% accuracy range, that

percentage itself being subject to a 10% range. Thus, a 10.85 m significant

wave height hindcast might represent a true significant wave height of +/- 10%

(Hs=97.65m to 11.94m) or even yet 1% lower or higher. If one also factors

into the MARIN data a small range to accommodate casual inaccuracy, the

variance between reality and the results of a combination of hindcast and

model data and statistical extrapolation could be significant. However, given

the generally self-consistent nature of the MARIN results at different speeds

and freeboards, it can be reasonably safely concluded that such variance as

there may be is not so great as to invalidate the conclusions to be drawn. The

most recent tests included significant wave heights increased by 10% and 15%

for certain freeboard conditions.

6.15 The final and most comprehensive report from Professor Tawn and Dr

Heffernan (see Appendix 17) set out by extrapolation from the MARIN data to

provide answers to the following question:

What was the risk of a wave impact on DERBYSHIRE’s No.1 hatch

cover which exceeded 42 kPa during given time periods in the course

of Typhoon Orchid and with given speeds over the ground and in

given states of freeboard, including intact and various states of partial

bow flooding? The underlying significant wave heights were those

relied upon by MARIN and provided by Dr Cardone for the relevant

times.

6.16 Five different time regimes were used in this exercise for each of the speeds 0

knots, 70% maximum attainable speed (assumed to be 1.5 knots), 2 knots and

4 knots. The time regimes were:


106

i. Initial Risk (“Init”): the risk of an impact exceeding 42 kPa by

1200z on 9th September, a time less than 2 hours after the time

(1019z) at which there is evidence that the vessel received a

message from its agents and was therefore still afloat;

ii. Intermediate Risk (“Inter”): the risk of an impact exceeding

42kPa by 2300z on 9th September;

iii. Cumulative Risk (“Cum”): the risk of an impact exceeding

42kPa by 0000z on 10/11 September (taken as the end of the

typhoon conditions);

iv. Conditional (Intermediate) Risk (“Con 2300”): the risk of an

impact exceeding 42kPa by 2300z on 9th September conditional

on there having been no such impact by 1200z on 9th

September;

v. Conditional (Cumulative) Risk (“Con Cum.”): the risk of an

impact exceeding 42 kPa by the end of the typhoon (0000z on

10/11 September) conditional on there having no such impact

by 1200z on 9th September.

6.17 Eight conditions of bow flooding were investigated. These were:

a. Intact;

b. Ingress of water into the stores through one smaller and one

larger orifice: (“Dam Stores, single min”) resulting in a

reduction of 0.4m, in freeboard;

c. Ingress of water into the stores as in b followed by subsequent

flooding of the fuel tank ullage space (“Dam Stores, single

max”) resulting in a reduction of 1.6m in freeboard;


107

d. Ingress of water into the fuel tank alone through one orifice,

(“Dam Deep Tank”);

e. Single damage to one orifice (3) which then admitted water into

the ballast tank (“Dam Ball Tank, single min”);

f. Single damage to one larger orifice (8) which then admitted

water into the ballast tank (“Dam Ball Tank, single max”);

g. Single damage to one orifice (3) into the ballast tank, but

assuming only 50% of the ingress rate under (e) (“Dam Ball

Tank, reduced single min”);

h. Single orifice damage giving ingress both to the stores and the

ballast tank (“Dam Stores and Ball Tank”).

6.18 Special variants on the significant wave heights were also investigated for

Damaged Stores, single maximum (10% higher waves) Intact (10% higher

waves, 15% higher waves and 10% lower waves). A time variant as to the

Damaged Stores and Ballast Tank was also investigated, namely ingress

having begun at 0600z on 9th September. With, that time variant a wave

height variant of 10% was also investigated.

6.19 Professor Tawn, by application of a curve-fitting methodology, known as the

Generalised Pareto statistical model, which he found to be entirely

appropriate, arrived at the extrapolated risk values set out at Appendix 17. He

also estimated a 1 hour time period during which in each of the bow flooding

and wave height conditions investigated the No.1 hatch cover would have

sustained the worst possible impact. These results are set out at Appendix 17.

6.20 It will be seen that for the risk extrapolations at 70% maximum maintainable

speed and 2 knots respectively in the Intact condition, without an increased

significant wave height, there was no risk of a hatch-breaking wave. It is only

when the vessel was making 4 knots that there was any such risk. When,


108

however, the significant wave height was increased by 10% the risk of such a

wave was 1% at 70% of maximum attainable speed and 4% at 2 knots. When

the wave height was increased by 15% the risk increased to 2% (at 70%

maximum speed) and 26% to 34% (at 2 knots). These results demonstrate the

considerable sensitivity of the risk of a hatch breaking wave to the vessel’s

speed.

6.21 The results also demonstrate the high sensitivity of risk of encountering a

hatch breaking wave to loss of freeboard in the various states of bow flooding.

Thus, for example, at 70% maximum speed there would be a nil chance of a

hatch-breaking wave if the stores, but no other part of the bow were flooded

(loss of freeboard: 0.4 m), whereas if the ballast tank alone were flooded

through one larger orifice (see f above), thus reducing freeboard by 1.1 m, the

risk of a hatch-breaking wave would increase dramatically to 20% to 70%.

6.22 These results indicate that at the most probable speeds of 70% maximum and

2 knots, there was a very substantial risk of a hatch-breaking wave at an un-

enhanced significant wave height when both the stores and the ballast tank had

been flooded and a considerable risk when the ballast tank alone had been

flooded, whereas there was no risk if the stores alone had been flooded.

Obviously, the magnitude of the risk would be commensurately reduced if

there were partial flooding of the ballast tank but full flooding of the stores.

6.23 If, however, there were a 10% enhancement to the significant wave height, not

only was it certain that there would be a hatch-breaking wave when the stores

and ballast tank were flooded, but there was a 1% to 10% risk that such a

wave would eventuate if, at 70% maximum speed, the stores alone were

flooded (Dam Stores single max) and a 4% to 42% risk of such a wave in that

condition if a speed of 2 knots were assumed.


109

Section 7

The Relevant Evidence from the Wreckage

Distribution of the Wreckage Field

7.1 The bulk of the wreckage was found to be concentrated within a relatively

small area by comparison with the depth (4200 metres) at which it lay. The

angle of the distribution cone was only about 8 degrees, as illustrated in

Appendix 18. Histograms of the concentration of the distribution of

approximately 2500 separate items of wreckage were prepared on an east-west

axis (X) and a north-south axis (Y). These are at Appendix 19. From these

histograms it can be seen that on the east-west axis practically all the items lay

within 1033 m, with the heaviest concentration within 563 m, whereas on the

north-south axis the spread is 1337 m and 515 m respectively. As appears

from Appendix 20 a plot of hatch cover wreckage distribution, the bow

section (B) was located approximately 590 m from the stern section (S). It can

also be seen that most of the hatch wreckage is concentrated in an area

stretching along the bow/stern axis with a breadth of about 230 m, with most

of the wreckage in that part of the area between the bow and the stern.

7.2 This pattern of distribution very strongly suggests that no substantial section

of the hull and none of the hatch covers separated from the main part of the

vessel earlier than the commencement of the sinking process.

Hatch covers

7.3 The vessel had nine sets of side-rolling hatch covers designed and supplied by

MacGregor & Co (Naval Architects) Ltd of Whitley Bay, Tyne & Wear. Each

SECTION 7: THE RELEVANT EVIDENCE FROM THE WRECKAGE

110

set consisted of a pair of equal area panels which opened by rolling outboard

on tracks. The dimensions were as follows:

Overall Length

Overall Width

Hatch No.1 14.719m 21.946m

Hatch No. 2 12.66m 21.946m

Hatches No. 3 to 9 14.947m 21.946m

7.4 The panels were of welded steel, open web construction. Each had 10

longitudinal girders spaced 994 mm apart together with 3 deep transverse

girders. The top plate was 11 mm in thickness. It can be inferred that Grade

A steel was used. The design was approved by Lloyd’s Register of Shipping

which also surveyed the construction of the hatches.

7.5 The watertight and oil-tight integrity of the hatch covers was provided by a

rubber seal running continuously round the edge of the hatch cover panels

including along the cross-joint. The foam rubber neoprene coated seal was

penetrated by the compression bar. There would be compression of about

13mm + 3mm. This is illustrated in Appendix 21.

7.6 The hatch covers were secured and kept in position by a number of devices.

Dealing with the No.1 hatch, there were 102 quick acting cleats at

approximately 0.5 metre intervals around the hatch coaming edge. These

cleats, illustrated in Appendix 21 are not required primarily to create

compression of the rubber seal, but rather to keep the hatch cover in place

while maintaining sufficient flexibility for that cover to “ride” the relative

movement as the hatch coaming flexes in the course of a sea passage. The

weight of the hatch covers is sufficient to create the required compression into

the rubber seals for dry cargo purposes.


111

7.7 Along the longitudinal centre line between the hath cover panels at intervals of

approximately 0.61 metres there were positioned locating pegs which engaged

in matching holes. These ensured correct alignment of the cross-joint seal and

compression bar at the time of closure, as illustrated in Appendix 22.

7.8 These were the features which provided restraint of the hatch covers against

vertical forces.

7.9 Restraint in the longitudinal horizontal plane was provided by the centre stop

on the coaming at each end of the cross-joint. The cross-joint locating pegs

ensured transmission of differential horizontal forces across the cross-joint.

Restraint in the transverse plane was, provided by centre devices at each end

of the cross-joint in the form of a wedge which located into a socket as the

hatch panel was homed into its correct location. These are illustrated in

Appendix 23.

7.10 In addition to the securing and locating devices described which are normal

equipment for a bulk carrier, there were three heavy duty securing catches

along the centre line in line with the three main transverse girders in the hatch

covers as illustrated at Appendix 24. These centre line catches were manually

dropped into place and then located by a hand operated hydraulic jack to

tension the fitting so that the pin could slide into place.

• = Please see the accompanying CD-ROM for an animation of this

process.

7.11 It is to be noted that these cross-joint catches were not to be found on the

majority (perhaps two thirds) of bulk carriers then afloat, the other centring

devices already described providing the necessary transverse restraint. The

function of these cross-joint catches appears to be exclusively related to

internal loading and gas-tightness. Such internal loading would arise in the

case of liquid cargoes when the vessel was pitching and rolling or when there

was high localised loading due to sloshing of the cargo. Gas tightness was


112

required particularly to control inert gas. These cross joint catches were not,

however, a class requirement. The quick acting cleats were treated by

classification societies as providing all the required restraint on the hatch

covers.

7.12 The MacGregor Hatch Cover Manual for the DERBYSHIRE at Section II,

Operating Instructions, contained the following:

“Oil Cargo”

“It is essential that the heavy duty cross-joint catches and all periphery

QA cleats are secured at all times when carrying any form of liquid

cargo.”

7.13 The implication is that the cross-joint catches and some of the quick acting

cleats are not required to be tightened when dry cargo is carried. As regards

the quick-acting cleats, this could reflect the fact that the DERBYSHIRE had

twice as many such cleats as were required for dry cargoes by Lloyd’s

Register’s Rules.

7.14 However, at a later section – V Pipes and Fittings - there is a further reference

to these catches:

“IMPORTANT – TO BE FASTENED AT ALL TIMES

NB During carriage of oil, bulk or ore cargoes”. (Emphasis added)

7.15 These references would certainly tend to confuse a master or chief officer.

The evidence strongly suggests that the other securing and positioning devices

already referred to provided quite sufficient weathertightness and that the

cross-joint catches did not provide an improved level of weathertightness.

Had those catches been required to make any significant contribution to the

structural strength or weathertightness of the hatch covers, it is inconceivable

that Lloyd’s Register Rules would not have expressly so provided. A majority

of the expert witnesses were of the view that the hatch centring wedges would


113

have the effect of providing an as-designed 16mm compression of the cross-

joint seal. It is not established that, without the cross-joint catches being

secured, there would be insufficient compression to maintain weathertightness

or that the hatch cover strength would have been in any respect materially

reduced, provided that the rubbing faces on the centring devices and stops

were not excessively worn. Given that in the course of the Sasebo dry

docking for the Class Continuous Survey in April 1980 those tests of the hatch

covers were conducted with all cleats and centre-line catches unsecured and

proved effective weathertightness, it would be most improbable that there was

such excessive wear. This is born out by the appearance of these seals in the

wreckage imagery.

7.16 At the time of the DERBYSHIRE’s last voyage her hatch covers complied

with the minimum strength requirements of ILLC 66 and of the Lloyd’s

Register of Shipping Rules. Further, the most probable collapse head for the

No.1 hatch cover lay in the range 4.1 to 4.3 metres head. As can be seen from

the wreckage imagery these covers had every appearance of being well-

maintained.

7.17 In an extremely helpful report Mr David Byrne, a marine architect of

Transmarine Ltd., prepared, by reference to the wreckage imagery, a plan

showing a possible allocation of the hatch cover wreckage (see Appendix 25)

together with sketches of the wreckage of each hatch cover or parts of a hatch

cover. The sketches of the hatch cover identified as No.1 starboard and of the

parts of that identified as No.1 port are at Appendix 26. The sketches of the

hatch covers identified as No.2 starboard and No.2 port are at Appendix 27.

7.18 All the hatch cover panels appear to exhibit the same initial or primary mode

of failure. Some of the covers demonstrate a secondary mode of failure. The

No.2 hatch covers do not. The two modes are to be seen at Appendix 28. The

primary mode of failure is shown by the yellow line and consisted of a

bending inwards along an axis transverse to the hull. That axis was in all


114

cases at or forward of the centre girder. The secondary mode of failure

consisted of an inward bending along a longitudinal axis very close to the

centre line of each panel. The experts were agreed, and this Report accepts,

that the transverse inwards bending was caused prior to the longitudinal

inwards bending. They were further agreed, and this Report accepts, that the

position of the transverse bending axis indicated that the panels had first been

subjected to loading in their forward area which would indicate waves moving

from forward to aft and would be consistent with the vessel being down by the

bow. The effect would be that the weight of green water on the forward area

of each panel would exceed the weight further aft. The positioning of the

longitudinal bending axis on the individual panels did not show the same

symmetry as the transverse axis damage. These damage locations strongly

indicate that when the vessel began to sink it had no significant list. The

symmetry of the transverse bending axis across the whole hatch cover width is

shown by comparing the images of the no.2 starboard and port covers – the

only pair which can be positively identified by hatch number (see Appendix.

29).

7.19 Once the hatch covers had sustained the transverse bending deformation they

would have lost water-tight integrity and seawater entry into the holds would

take place as soon as there was green water above the hatch coaming.

Ventilators, Air Pipes and other Deck Openings

7.20 The main bow section spaces are shown in elevation and at tank top level in

plan at Appendix 30. The fittings on the foredeck, including the ventilators

and air pipes leading down into the various bow section spaces are shown on

Appendix 31. Each of the ventilators and air pipes which the experts agreed,

by reference to the relevant imagery, had sustained damage at some stage is

also listed in Appendix 31. The numbering in the damages list has been

transposed onto the foredeck arrangement plan.


115

7.21 This Report accepts the location and description of the ventilator and air pipe

damage agreed by the experts.

7.22 The damage found to have been sustained by these ventilators and air pipes

was at least in part sustained on the surface and before the commencement of

the sinking process.

7.23 The evidence suggested that mushroom ventilators are particularly vulnerable

to damage in the course of green water loading. There was previous

experience of instances of such damage on the DERBYSHIRE and evidence

of such damage on the sister ships SIR ALEXANDER GLEN and

KOWLOON BRIDGE and other unidentified vessels.

7.24 It is much less probable that the Winel Vents, which were self-closing

ventilators, were damaged by wave loading on the surface. They were located

in a relatively sheltered position behind the bulwark plating. Since that

appears to be intact at the relevant position, it is more likely that these were

damaged by mechanical impact. Such damage could have been caused by

other deck fittings, such as the plates from the windlass platform. Whether

that did happen and, if so, whether on the surface, must remain in doubt.

7.25 It is possible that one of the 7 inch air pipes to the fuel oil tank was so

damaged on the surface as to admit seawater before the start of the sinking

process, but the tests conducted at MARIN to ascertain ingress rates indicate

relatively small rates through such an opening.

7.26 The spurling pipes would ordinarily be sealed at the commencement of the

voyage by means of a mesh of chicken wire and cement. The wreckage

imagery shows this to have been dislodged. It is almost certain that, being

fairly vulnerable to wave impact damage, this sealing would have been

destroyed before the commencement of sinking. As the vessel shipped green

water over the bows, the chain lockers would have filled through these

openings. It is, however, most improbable that water could have travelled


116

beyond the chain lockers because the only access from those spaces was,

according to the shipowners’ evidence which this Report accepts, by means of

bolted covers, as distinct from the access doors shown on the builders’

drawings.

7.27 Damage to or destruction of the mushroom ventilators would cause the

admission of seawater to the bosun’s store and adjoining machinery space.

Although they were separated by gas-tight doors opening from the machinery

space into the store, the probability is that water could find its way past the

doors.

7.28 Water which had already entered the store or machinery space could not enter

the fuel tank unless the stores deck collapsed. But by finite element analysis it

has been established that, provided that the floor was properly welded, the

downward pressure on the floor required to cause collapse would be of the

order of 42 metres head above foredeck level, which would not have occurred

in the relevant sea conditions prior to the commencement of sinking. If the

floor welding were materially defective, it has been established by finite

element analysis that a downward pressure on the floor would be of the order

of 12 metres head above foredeck level. That certainly would be a

theoretically possible head of green water in extremely severe conditions, but

it is most unlikely that the welding would have been materially defective.

According to the written evidence of Mr M Turner, the Group Welding

Engineer of Swan Hunter at the time of construction, which was not

challenged by any party, and which this Report accepts, it is highly

improbable that any of the three types of welding defect found to give rise to

collapse of the floor with a 12 metre head of water above the foredeck would

have occurred. All the relevant fillet welds joining the structure to the

underside of the main deck at the stores flat deck would have been effected in

the builders’ fabrication shops and defects and omissions would almost

certainly have been spotted on inspection. Although defects to other fillet


117

welding which had to be carried out on the berth could in theory have gone

undetected, that was distinctly improbable.

7.29 Although it was suggested in the Assessors’ Report that seawater could have

entered the deep fuel tank if man hole covers in the stores had been removed

in the course of the voyage for the purpose of ventilation of the tank (see

Assessors’ Report paragraph 4.74), this suggestion, which Mr Robin Williams

no longer advanced, can be rejected. If, as must now be accepted, there were

about 2600 m.t. of fuel oil in that tank at the relevant time, there is no realistic

likelihood that a competent crew, such as that on board the DERBYSHIRE,

would have ventilated a partially-filled fuel tank.

7.30 It was also suggested in the Assessors’ Report that it was possible that

seawater could have entered the foredeck ballast tank if the access manhole

cover in the bosun’s store had been left open for any reason (see Assessors’

Report paragraph 4.72). It is distinctly improbable that it would have been

opened, much less left open, in the course of the voyage. There would be no

reason to enter this tank in the course of the voyage, unless there was a faulty

valve or gauging system or as part of the planned tank maintenance inspection

system. There is no evidence of the existence of any such fault or of any such

maintenance work. At the relevant time the tank only had a few centimetres

depth of ballast in it and it is unlikely that any remedial work had to be done.

Further, the tank would have been inspected at Sasebo in April 1980 so that it

is unlikely that its condition would again have to be investigated. Additionally

the lid to this tank, like that to the deep fuel tank, was so heavy that two men

with portable lifting gear were needed to remove and replace it. Such removal

and replacement had to be recorded in writing in accordance with the

shipowners’ strict control procedures. Taking all these considerations into

account, it is highly unlikely that the ballast tank lid was left off.


118

The Bosun’s Store Hatch

7.31 This was located on the foredeck slightly to port of the centre line between

frames 353 and 356 as shown on the plan at Appendix 31. It gave access to a

metal ladder down into the store. The lid opened on two hinges located at the

forward end of the coaming. There was a balance bar attached to the hinges

which extended for the whole width of the lid. The latter was secured to the

coaming by 8 hinged toggles each of which screwed down on to a cleat. There

were two such toggles on each of the 4 sides of the coaming. Photographs of a

model of the hatch are shown at Appendix 32.

7.32 Its condition as shown in the wreckage imagery was as at Appendix 33

7.33 In the wreckage imagery the lid is seem to be missing and has never been

located.

7.34 The after side of the hatch coaming was severely dented and split as shown at

Appendix 33.

7.35 The toggle wing nuts as found were in various conditions and at differing

positions on their threaded shanks; some were missing altogether. Much time

was spent in the course of the hearing investigating whether it could be

inferred from the apparent condition and position and absence of wing nuts as

found on the wreckage site that the lid had not been properly secured before

the vessel encountered the typhoon. The toggles assumed considerable

importance by reason of the conclusion arrived at by the Assessors in their

Report that the lid had not been properly secured. This was based primarily

on the assertions that it could not be securely closed. This was said firstly to

be because a length of rope had been attached to and wrapped round the aft

toggle on the portside longitudinal coaming and had been attached to and

wrapped round the port toggles on the aft transverse coaming, thereby

preventing the sufficient tightening of those two toggles. Secondly it was


119

asserted that proper closure of the lid had probably been prevented by a rope

seen in the imagery to emerge from the open hatch and to lead over the top of

the coaming. It was said that this rope could be seen running amid the

wreckage to the outboard drum of the port windlass.

7.36 Upon the evidence which has been adduced before this Investigation it is clear

beyond doubt that:

i. the rope attached to the two toggles was the remains of an

elaborate lashing system employed on board this vessel and

some of its sister ships to ensure that the toggles did not work

loose and fall out of their lugs with the working of the vessel in

heavy seas;

ii. the rope seen to be protruding over the hatch coaming was not

the same kind of tope as that which was attached to the port

windlass but was passed through a loop in another rope

amongst the wreckage and would have been attached to the

underside of the bosun’s store hatch cover.

7.37 As to the rope lashing system, the evidence of four master mariners was that

the toggles tended to loosen in as much as the wing nuts worked up the shanks

as the vessel worked. That could lead to the shank disengaging from the cleat.

Some masters solved this problem by using locking bars as an added means of

securing the lid (see Appendix 34), but others used ropes in the form of a cat’s

cradle. This was demonstrated in Court by Mr Malpass, a former Chief

Officer of the DERBYSHIRE, who rigged up this type of roping device on a

half-size model of the hatch which was then photographed (see Appendix 35).

The particularly pertinent evidence of Mr Malpass, which this Report accepts,

was that the carpenter on board the DERBYSHIRE on her last voyage, Mr

Tommy Blease, was a very competent crew member who was familiar with

that method of lashing the hatch. It can confidently be inferred that in


120

anticipation of the onset of bad weather the bosun’s hatch would have been

secured by him in that way.

7.38 As to the protruding rope, it is seen from the imagery to be a four-square

polypropylene lay, whereas the rope attached to the drum of the windlass is

seen to be of braided nylon. The evidence of Mr Malpass, which this Report

accepts, was that he personally had caused a length of steel bar to be welded to

the underside of the bosun’s hatch lid on the DERBYSHIRE so that the last

mooring rope might be attached to it to make it easier to haul the mooring

ropes up on deck before berthing. The reason why on the wreckage site the

mooring rope was protruding over the edge of the coaming was almost

certainly that when the hatch lid was broken off, it dragged the rope out with

it, leaving a length of it strewn over the foredeck.

7.39 In the course of his evidence the assessor, Mr Robin Williams, was at first

quite unconvinced that, in spite of these considerations, the bosun’s hatch was

properly secured. His eventual concern was in relation to 3 out of the 8

toggles which he felt were found to be in a condition which suggested that the

dogs had not been hardened down as they ought to have been.

7.40 There are many ways in which, when seen on the wreckage field, the toggles

might have had the appearance which they did, not least the effect on the wing

nuts of hydrodynamic forces in the course of sinking which might well have

caused the nuts to mount or descend the toggle shanks to positions quite

different from that which they occupied at the surface. Further, the effect of

the destruction of the cat’s cradle rope lashing on the position of the wing nuts

might well have been quite considerable. Additionally, the appearance of

grease on one of the toggle shanks is hardly indicative of failure to deploy the

wing nuts.

7.41 The evidence, which is accepted, that Mr Bayliss, the Chief Officer, and Mr

Blease, the carpenter, were very competent and the very strong inference that a


121

cat’s cradle lashing had been set up on that last voyage suggest that it is almost

inconceivable that any of the toggles would have been accidentally left

untightened. It is also improbable that any of them were defective and for that

reason could not be tightened. Replacing them was a routine job which would

ordinarily be carried out by the crew during routine maintenance and, on a

well-run vessel such as the DERBYSHIRE, without undue delay. Moreover,

the hatch would probably have been subjected to inspection during the load

line surveys conducted in April 1980 at Sasebo.

7.42 The bosun’s hatch was clearly struck with great force on its transverse after

coaming by a sharp metal object. The inwards-bending configuration of the

longitudinal coamings strongly suggests that the lid was in place at the

moment of impact. There can be little doubt that the impact would have

severed the rope lashing. It might also account for the damaged condition of

some of the toggles. It is possible that the lid itself was broken off by that

blow.

7.43 Taking into account all this evidence, it can be concluded with reasonable

confidence that the bosun’s store hatch was properly secured by the crew

before the DERBYSHIRE encountered the typhoon and that if the lid opened

or was lost so that the hatch admitted seawater before the commencement of

sinking, that was through no fault of those on board. To his credit Mr

Williams eventually appreciated the force of this evidence and accepted in the

course of his cross-examination that it was more probable than not that the

crew had fully secured the hatch before the DERBYSHIRE encountered the

typhoon.

The Starboard Windlass

7.44 The images of the bow section as found clearly show that the starboard

windlass had became detached from the deck at an earlier stage. See


122

Appendix 36. A labelled diagram of the windlass is to be found at Appendix

37. Part of that equipment was found on the seabed some 200 metres from the

position of the bow. This proximity is consistent with severance either

immediately before or soon after the commencement of the sinking process.

The position of the starboard windlass relative to the bosun’s store hatch

clearly suggests that it or a part of it must be a strong candidate for having

collided with the after end of the store hatch, thereby severing the lid from the

hatch coaming and destroying the rope lashing.

7.45 Close inspection of the foundations of the windlass shows that many of the

holding down bolts which were manufactured from high tensile steel, remain

in place and that most of the main structure of the windlass has separated from

its foundations at the welded connections with the bolting flanges. In other

places, the fillet welds to the web and tripping brackets had failed. The in

board wire drum pedestal had failed not at the welds but at the bearing housing

in as much as the top part was found to be broken off.

7.46 The remains of the three parts of the windlass on the seabed do not indicate

substantial impact damage except that part of the main gear wheel of the

middle gypsey section of the windlass was missing. Such a fracture must have

been due either to being struck by some loose object or, following severance

from the deck, to striking some other fixed object or objects. Further, as seen

on the imagery, part of the anchor chain to the starboard anchor has been

extracted on to the deck from the chain locker, but has then become separated

from the windlass.

7.47 It was suggested by a number of witnesses that the starboard windlass could

have been dislodged by the spare propeller which was carried on deck just

after No.1 hatch coaming and which has not been located in the wreckage

field. This item weighed 41 tonnes and was a very bulky object. To reach the

starboard windlass would have involved breaking loose from its attachments

to the deck and a comparatively lengthy journey forward along the deck beside


123

the No.1 hatch coaming, followed by a sharp turn to port. This chain of events

is improbable. The spare propeller, had it broken loose, would have been

much more likely to go over the side before it reached the foredeck, a journey

which, because of its mass, is very unlikely unless the angle of descent was

very steep. If that steep angle had been maintained, it is hard to see what

could have led to the propeller describing a sharp turn to the port after passing

the No.1 hatch coaming, as distinct from crashing through the starboard

bulwark on the foredeck, which is largely intact. Further, had it proceeded to

strike the starboard windlass, it is hard to see how that item could have

remained in such relatively good condition as discovered on the seabed.

7.48 This Report concludes that the most likely explanation for severance of the

starboard windlass is that it was subjected to a succession of powerful waves

striking the foredeck which imported sufficient energy to initiate low cycle

fatigue cracks in the welding of the windlass to its bed. These would

eventually open upon impact by subsequent wave loading. Following that

severance, the windlass or part of it would slew across the foredeck, striking

the bosun’s store hatch as it went, yet substantially missing the port windlass

which, as can be seen from Appendix 36 remained largely intact.

7.49 It is impossible to say at what point in the sinking process the windlass

became detached. However, it is most improbable that this occurred prior to

the commencement of water entry through the ventilators due to the

mushroom heads being damaged or destroyed. The forces necessary to

damage those ventilators would be less and less continuous than those

responsible for severance of the windlass welding. It may well be that, as was

suggested by Mr Andrew Squire in the course of his evidence, that it was only

when the freeboard had been very substantially diminished by the flooding of

the No.1 hold that the wave loading was sufficient to detach the windlass. The

portside windlass remained undisturbed. Although, at first sight that might

suggest that the starboard windlass had been hit otherwise than by waves, the

more likely explanation is that when the vessel was hove to the seas were on


124

the starboard bow and the starboard side was therefore more vulnerable to low

cycle fatigue cracking which had developed to a point where it could be more

readily fractured than the welding of the portside windlass at an earlier stage

relative to the commencement of sinking.

7.50 Accordingly, it is to be concluded that the bosun’s store hatch was probably

damaged and the rope lashing and lid destroyed by impact by the starboard

windlass or some part of it and that by that time a substantial volume of

seawater had already entered the bosun’s store and machinery space due to the

destruction of the mushroom ventilators and when water entry had already

begun into the forepeak ballast tank and quite possibly the No.1 hold.

Bulkhead 339

7.51 Bulkhead 339 and its associated cofferdam at 340 separated the No.1 hold

from the forepeak. They are indicated on the diagram at Appendix 31 and

photograph at Appendix 38.

7.52 As found in the seabed the bulkhead was attached to the bow section and

visible above ground from approximately the 28ft level upwards. Various

careful drawings were made of the somewhat complicated damage to the after

face of this bulkhead shown on the underwater imagery. That which is

particularly helpful was prepared by London Offshore Consultants and is at

Appendix 39. A model reconstruction incorporating all the salient features of

that damaged bulkhead was also prepared on behalf of the DFA and is shown

at Appendix 40. The imagery also shows a length of rope protruding from a

level inside the fuel tank.

7.53 The condition of this bulkhead assumed considerable importance in the

Assessors’ Report in as much as its condition in the wreckage was said to be

consistent with the bow section, including the fuel tank, having been

substantially full of seawater prior to the commencement of sinking. Thus, at


125

paragraph 4.57 it was stated that the pressure damage to that bulkhead took the

structure into the plastic region but there was no implosion; instead the only

damage was attributable to explosion. This was said to be evidence of a lower

inwards deflection and lower energy regime than that which prevailed for the

rest of the ship’s wreckage, except at the stern, and which was indicative of

those spaces further aft, in contrast to the bow spaces, having been

substantially empty of water until after sinking. It could therefore be inferred

that the various compartments of the bow were either full or mostly full at the

time when sinking commenced: see paragraph 4.58 of the Assessors’ Report.

This sequence played an important part in the Assessors’ preferred mechanism

of failure which involved the filling or substantial filling of the bow spaces,

including the fuel tank ullage space, with seawater which entered through the

stores hatch and ventilators and then caused the stores deck to give way which

caused water to be admitted to the fuel tank. The loss of freeboard caused by

that ingress was enough to expose the No.1 hatch to green water loading of

hatch-breaking pressure and thus start the domino affect which led to

progressive hatch cover failure.

7.54 In the course of the Investigation much time and attention was devoted to

exploring the wreckage condition of bulkhead 339. Mr Robin Williams

adhered to the view, consistent with the Assessor’s Report, that most of the

more salient features of the damage to the bulkhead had occurred on impact

with the seabed. Had the bow section not already been substantially full of

seawater when the vessel left the surface, by the time its deck sank to a depth

27-30 metres below the surface the bulkhead would have failed very

substantially by implosion/explosion giving rise to widespread collapse. To

support this thesis he relied on a finite element analysis by EATEC.

7.55 There are, however, serious difficulties in this approach, not least the enduring

support for the structure attributable to membrane stress following the

commencement of deflection once yield strength had been surpassed. The

stringers as brackets could still have residual strength. In consequence the


126

appearance of the damage suggests that the cofferdam has been pulled down

and ruptured near the centre line, causing a crack to run up to the deck plating

which has been drawn down at about the centre line. It was considered by Mr

Squire and Mr Corlett that this cofferdam fracture would have run into the

stores deck and also caused it to fracture. The associated fractures and

deformations would have caused water to be admitted to such parts of the bow

spaces as had not already been filled.

7.56 This Report is unable to accept that the condition of bulkhead 339 supports the

proposition that the bow spaces and particularly the fuel oil tank ullage space

were substantially full of seawater when the vessel began to sink. Had that

been so, it is inconceivable that the foredeck above the store would have been

found in a sagging condition in the wreckage field. On the contrary it would

almost certainly have been caused to bulge upwards and outwards by impact

of the bow section with the seabed. Mr Williams was asked about this

apparent inconsistency, but he was only able to suggest that it was due to a

column of entrained water descending with the bow. This is intrinsically

improbable. That circumstance could not have imparted sufficient forces to

the foredeck to create the depression now observed.

7.57 This Report concludes that the present condition of bulkhead 339 is consistent

with the deep fuel tank and perhaps part of the ballast tank having been empty

of water when the sinking process commenced, but having admitted additional

water during sinking. The collapse of the cofferdam and resultant bulkhead

fractures would have initiated water ingress into the bow which would have

been augmented when the saddle tanks collapsed due to implosion/explosion

in the course of sinking after dropping below the surface. On further analysis

of the whole of this evidence, therefore, the condition of the bulkhead lends no

support to the theory of a more or less completely filled bow prior to the

commencement of sinking.


127

Section 8

Frame 65 Defects

8.1 Before the Original Formal Investigation in 1988-9 there was a great deal of

expert evidence as to the design of the DERBYSHIRE immediately forward of

the superstructure and aft of No.9 hatch, in particular in way of frame 65. The

position of that member can be seen on the vessel’s Capacity Plan at Appendix

9. That Investigation had to consider a major issue identified in the Question

1(a) “Was the design of the DERBYSHIRE in way of frame 65 satisfactory?”.

There was a further question as to whether the repairs carried out in that

position in April 1980 at Sasebo were satisfactory (Question 1 (c)). To both

those questions the Wreck Commissioner answered Yes.

8.2 The underlying issue was whether, having regard to the design drawings and

evidence of construction, there was such discontinuity between the topside

tank vertical side which stopped at the forward side of the transverse bulkhead

at Frame 65 and the longitudinal engine room bulkhead which started on the

aft side of that bulkhead as to give rise to such an interruption in loading as to

create significant risk of areas of high stress concentration which might

endanger the integrity of the hull.

8.3 By the time of Lord Donaldson’s Assessment in September 1995 which

recommended a further more extensive underwater survey, there had been a

great deal of academic and more general media discussion as to whether the

DERBYSHIRE had been lost by reason of catastrophic structural failure due

to a design defect in way of frame 65. One of the main reasons for Lord

Donaldson’s recommendations was that he felt it necessary to explore this

possibility further in the interests of international ship safety.

SECTION 8: FRAME 65 DEFECTS

128

8.4 Further, on the basis of the evidence which emerged from the underwater

survey, the Assessors concluded that the vessel did not fracture on a vertical

plane, either at Frame 65 or between that frame and Frame 64. Their reasons

for this conclusion are set out in their Report and paragraphs 4.104 to 4.124.

In summary, the vessel could not have split open while still on the surface in

way of Frame 65 because, as appeared from the underwater imagery, the

pattern of plating failures and fractures, the integrity of the port and starboard

cofferdam structures and the damage to the slop tanks aft of Frame 64 and the

saddle tanks and hopper tanks forward of Frame 65, which showed that they

had imploded/exploded due to the effects of hydrostatic loading after the

commencement of sinking, showed that all the ascertained damage to that area

of the vessel occurred below the surface.

8.5 In the course of preparation of the evidence and definition of the issues to be

considered in this Investigation it emerged that no party contended that the

fracture of the vessel in way of Frame 65 was an initiating cause of the loss of

the DERBYSHIRE. However, the DFA contended that defects in design

and/or construction might have given rise to an unacceptable risk of loss or

injury to the crew and, therefore that question ought to be investigated and

reported upon and any necessary recommendations made. By a judgment

delivered on 6th August 1999, this Court ruled that it was in the public interest

that a limited investigation of this matter should be conducted in the course of

the Investigation, but subject to the constraints set out on pages 21-23 and

reproduced at Appendix 41. As explained earlier in the judgment, these

constraints were aimed at striking a reasonable balance, in a case where the

alleged hazardous state of the vessel had not caused the loss, between the

public interest in improving ship safety and the need to avoid undue delay and

public expense and to concentrate on the circumstances of the loss the subject

of the Investigation.

8.6 In February 2000 the DFA gave notice of amended criticisms which included

allegations directed at the builders and Lloyd’s Register in respect of the


129

design and construction of the vessel at Frame 65. This Court took the view

that the assignment of fault for design defects which might be established but

which had not caused or contributed to the loss of the vessel would cause a

risk of injustice, delay and additional expense. Two paragraphs from the

reasons of the Court are set out at Appendix 42.

8.7 Before this Investigation, Dr Remo Torchio, one of the Assessors, gave

evidence to the effect that the design features in way of Frame 65 could not

have caused or contributed to the loss of the DERBYSHIRE. He relied in

substance, on those features of the wreckage already referred to which had

been relied upon in the Assessor’s Report, in particular, on 4 basic points.

These were as follows:

i. There was evidence of implosion/explosion of the tanks

forward and aft of Frame 65.

ii. There was no encircling fracture round the cofferdam between

Frames 64 and 65 and the double bottom as far aft as Frame 55

was still attached to the hull at and forward of Frame 65.

iii. The cofferdam was not split in two.

iv. The wreckage of the bow was as little as 600 m. from the

wreckage of the stern which was consistent with fragmentation

after sinking beneath the surface and inconsistent with

fragmentation while still on the surface.

8.8 All the experts were agreed that there was misalignment of not less than

35mm (plate centre to plate centre) between the starboard outboard

longitudinal plating of the engine room longitudinal cofferdam and the saddle

tank vertical hatch side girder at their meeting point at Frame 65. This is

illustrated in the diagram of Appendix 43.


130

8.9 It was also agreed by the experts that the presence of this misalignment

increased the risk of the development of fatigue cracks in the transverse

bulkhead plating at Frame 65. This was due to undue stress concentration

associated with the discontinuity.

8.10 This Report accepts that view. It is very strongly supported by the experience

of the DERBYSHIRE itself. The vessel developed a crack in the aft starboard

upper corner of No. 9 hold where the hopper met the join of the aft and

starboard bulkhead. This was first reported in November 1997. That fracture

could permit oil leakage from No.9 hold thereby causing oil to enter the ballast

tank and the pump room. It was not repaired. The vessel carried oil on a

voyage commencing on 5th August 1979 and by 19th August Captain Underhill

was reporting to his owners that oil was leaking quite badly through the crack

into the pump room. The crack was reported to be in a position to which it

was very difficult to gain access for repairs. About 2 to 3 tons of oil had

leaked into the pump room, a point re-iterated by him on 17th October 1979.

Captain Boyle, who was in command from January to June 1980 observed an

oil stain on the bulkhead in the pump room but he was not very concerned

since he did not carry an oil cargo. He said it had been no more than “a very

very slow trickle”. The crack was repaired at Sasebo in April 1980. The

repair plan described it as 1 m. long on the No.9 hold side and 300 mm long

on the pump room side. The repair bill described it as 250 mm long.

8.11 The experts were further agreed that if fatigue cracks did develop on the

bulkhead, the risk of their then propagating un-noticed and un-repaired into

the deck plating would be low relative to the risk of their developing in the

bulkhead in the first place. The experts also agreed that the probability of

cracks in the bulkhead leading to catastrophic failure of the hull girder would

be extremely low. In view of the evidence of Dr T Baker before this

Investigation, this Report concludes that the risk of the misalignment causing

fatigue cracking of the bulkhead at frame 65 which in time gave rise to

cracking of the main deck and which in turn caused the vessel to split due to


131

propagation of such deck cracks round the hull can be regarded as no more

than miniscule. The fact that the DERBYSHIRE endured typhoon Orchid,

albeit after the recent repair of the Frame 65 bulkhead, and was thereby

subjected to any increased stress regime in that area shows very clearly that

risk of the above failure sequence must be regarded as minimal. This would

be consistent with there being a relatively low overall stress area at that

position towards the aft end of the vessel.

8.12 It was submitted on behalf of the DFA that there were other defects in addition

to the misalignment. These were the use of liner plates to make good gaps and

spaces, the use of Grade A rather than Grade D steel in this area and the

failure during construction to fit a bracket shown on the design plans as

attached between the underside of the hatch side girder and the bulkhead

plating of bulkhead 65.

8.13 There can be no doubt that in terms of the initiation of fatigue cracking of the

bulkhead it was the misalignment which was the major factor. Although all

experts were agreed that there was in any ship of that kind an underlying low-

level probability of fatigue cracking of the bulkhead even without

misalignment, there is little or no evidence that such cracking would ever have

occurred on the DERBYSHIRE without such misalignment. Accordingly

none of these three further factors could be regarded as independently

representing any substantial risk to the overall integrity of the hull girder.

8.14 In the course of the Investigation those representing the DFA emphasised the

risk to the vessel and crew, should bulkhead cracks occur, from the leakage of

oil through the cracks. It was also said that such leakage could cause pollution

by entering ballast which might then be pumped overboard. Or there might be

an explosion or crew injury due to the presence of gas entering a non gas-free

area, either in the pump room or elsewhere. The experts disagreed over the

magnitude of the risk involved. The majority held the view that these risks

were controllable because they were already catered for by the vessel’s


132

internal management and survey regime. In particular such controls ought to

ensure that, if there were such a crack, it would be detected before it reached a

dangerous length and, if it were not, that such oil leakage as might occur

would not cause pollution or explosion or injury to the crew. Accordingly the

risk of such occurrences was negligible.

8.15 Experts called on behalf of the DFA did not agree with that assessment. Mr

Squire and Mr Deegan considered the risk to be “moderate” because it might

not be controllable in practice.

8.16 The evidence of Mr Squire was that “all vessels will naturally carry out the

normal risk-reducing measures or the normal procedures that are required of

vessels to assist in the safety of their crew, but that they are not always

implemented in a satisfactory manner and that they do not always make up for

a failure of a primary system, such as the structure of the ship”.

8.17 Mr Deegan characterised the chances of such occurrences due to oil leakage

through fatigue cracks as very small but the “risk” of that happening, taking

into account not merely such small chances but also the potentially

catastrophic consequences (for example, an explosion), could be described as

moderate.

8.18 There is some support for the submission that all such safety regimes are to

some extent fallible. The bulkhead 65 crack on the DERBYSHIRE itself

apparently went undetected when the vessel had been laid up for a year and

was surveyed prior to returning to service in April 1979, no doubt due to the

extremely inaccessible area in which the crack was located. In consequence,

no repair had been carried out before the vessel loaded her next oil cargo in

August 1979 and nothing was done to effect repairs until Sasebo in April

1980, over two and a half years after Captain Underhill had first drawn

attention to the problem, albeit only one and a half trading years.


133

8.19 The pump room was to be treated as a hazardous space, because in the

ordinary way oil leaks could occur from the pumping equipment. It therefore

should not be entered unless it were first gas-freed and ventilated. The ballast

tank should similarly be gas-freed and ventilated before being entered by the

crew.

8.20 As to pollution, that would certainly be a possibility if it were not known

before de-ballasting that an oil leak existed. In the ordinary way the ballast

water ought to be sampled thereby providing an opportunity to detect the

presence of oil and to avoid pollution.

8.21 On analysis the issue as to whether the chances of crew injury, explosion or

pollution occurring by reason of leakage of oil through an un-repaired fatigue

crack in the Frame 65 Bulkhead are to be treated as “negligible” or

“moderate” is in substance somewhat unreal. What really matters is that

because of the combination of a defect in design (discontinuity) and in

construction (misalignment) the chances of fatigue cracking were increased.

Consequently the chances of oil leakage into the pump room and/or ballast

tanks were increased. Since there was in place a vessel management and

survey regime, there was a safety system designed to prevent dangerous

consequences of leakage. That system was not infallible. However, the

increased chance of leakage meant that there would be an increased chance

that the fallibilities of the safety system might eventuate. Having regard to the

nature of the safety regime, this Report concludes that the chances of its not

preventing injury or explosion were very slight and that the chances of its not

preventing pollution were at highest very little greater.

8.22 The submission of the DFA included a description of the risk of injury,

explosion or pollution from such a crack as “an accident waiting to happen”.

Mr Squire said in his evidence that these were not words that he would have

used. In as much as they suggest that the risk of any of those eventualities


134

occurring could be regarded as imminent and substantial they are a gross

exaggeration and quite unsupported by the evidence.

8.23 The question remains whether there is room for improvement in class rules or

otherwise to reduce still further this area of risk. This really goes mainly to

avoiding the primary causes of the fatigue cracking.

8.24 The root cause of the Frame 65 problem was the misalignment which was a

construction defect superadded to the break in longitudinal continuity which

was part of the design. The salient features of this problem were that the hatch

side girder forming the inboard side of the upper wing tanks was not scarphed

through bulkhead 65 and continued aft of that bulkhead, but was instead

butted on to Frame 65 and was supposed to line up with the much more

slender longitudinal cofferdam bulkhead, outboard of the pump room, that

abutted the hatch side girder aft of the bulkhead 65. The mere existence of the

discontinuity might have induced some bulkhead fatigue cracks and so that

design feature in itself was not an ideal structural arrangement. It was,

however, clearly born out of a desire to facilitate the block construction of the

stern section and the section from No.9 hold forward.

8.25 The evidence suggested that the process of accurate alignment of the hatch

side girder with the longitudinal cofferdam bulkhead proved to be difficult to

achieve. The builders and Lloyd’s Register inspected the jointing of the

blocks on the berth. The testing for alignment was done by means of hammer

tests and visual checking of marks on the metal members. The Lloyd’s

Register surveyor who was present at the yard during the later stages of

construction said in his evidence to the original Investigation that these

methods would have enabled surveyors to detect misalignment even more

accurately than within 10 mm. The 35 mm misalignment discovered from the

wreckage shows either that this was not correct or that the relevant tests were

not or not effectively carried out.


135

8.26 This Report concludes that Classification Societies should have in place rules

which would discourage designed discontinuities of the kind demonstrated by

the DERBYSHIRE and which would provide for their surveyors in attendance

at construction to satisfy themselves by more effective testing that any

unavoidable abutment of members of the nature of that which ought to have

been achieved on the DERBYSHIRE is accurately assembled and constructed

to within an acceptable tolerance. Whereas it appears from the evidence

provided by Lloyd’s Register that there may not be any vessels afloat with a

cruciform configuration precisely similar to that in way of Frame 65, it may

well be that broadly similar design features may be put forward in future or

may still be found in some of the 290 bulk carriers and 12 OBO’s of

deadweight in excess of 150,000 m.t which are still in service.

8.27 Lloyd’s Register, which at the present time is the Classification Society for

18.8 per cent of the world’s tonnage and 17.6 per cent of bulk carriers over

20,000 m.t, has since 1996 issued its Ship Right Fatigue Design Assessment

Procedure – a structural detail design guide, including bulk carrier design,

containing a vast spectrum of design details which Lloyd’s surveyors are

instructed to require. The experts were agreed that these details adequately

covered the aspects of design similar to the frame 65 features. Mr Gavin of

Lloyd’s Register described the fatigue data which it provides as part of its

Ship Right Fatigue Design Assessment software as “the most comprehensive

fatigue data available in the marine industry”. This may well be correct.

Certainly Lloyd’s has, most impressively, devoted massive resources to

research and development in this field. The evidence suggests that not all

members of IACS have such high design and construction standards.

8.28 For these reasons this Report accepts the experts’ recommendation that IACS

should now consider requiring all its members to introduce design approval

and survey procedures for new buildings and for vessels in service which

would implement standards similar to those in the Lloyd’s Register Ship Right

Fatigue Design Assessment procedure.


136

8.29 Amongst the submissions advanced on behalf of the DFA was the proposition

that Grade A steel should not have been used in the area of Frame 65 and

further that there should be a recommendation that Grade A steel should not

be used in primary (main hull) and secondary (contributing superstructures)

structures unless certified as having an acceptable minimum notch toughness.

This repeated the Assessors’ Report recommendation at paragraph 8.96. This

recommendation was directed to preventing potentially catastrophic brittle

fractures developing in consequence of the propagation of fatigue cracking.

8.30 This issue is well outside the scope of this Investigation. There was no brittle

fracture on the DERBYSHIRE. Nor was there any substantial risk that fatigue

cracks would propagate undetected to cause brittle fracture, catastrophic or

otherwise. No lessons about this field have been learned from the evidence

before this Investigation except that a great deal of further work and research

would have to be embarked upon before an informed view could be expressed.

That work and research was not available to the Assessors and it is not

available to this Investigation. All that can be said is that Grade A steel may

well be inappropriate material in areas likely to be subject to high tensile

stress. Whereas the 0.4 midships area is probably one such area, there may

also be other similarly vulnerable areas which have yet to be identified. It

would clearly be sensible for IACS to investigate this matter further.

8.31 Finally, as to the repairs to the cracks in the Frame 65 bulkhead carried out in

April 1980, there is nothing to suggest that these were not carried out in a

proper manner or that an inappropriate repair (by welding) was done.


137

Section 9

Inferences as to the Cause of the Loss

9.1 From the evidence of the condition of the wreckage photographed and video-

filmed on the seabed and from the data derived from the model tests

conducted by MARIN, very strong inferences can be drawn as to the cause of

the loss of the DERBYSHIRE.

9.2 The Imagery from the wreckage field described in section 7 of this Report

most strongly suggests that:

i. ventilators and airpipes located on the foredeck and leading

down to the bosun’s store, machinery space and ballast tank

were damaged before the commencement of sinking, in such a

way as to admit substantial volumes of seawater to those

spaces;

ii. the hatch covers were damaged by impact from forward before

they collapsed into the holds and before they were subjected to

bending on a longitudinal axis;

iii. the hull girder had not lost its longitudinal integrity at any time

until after sinking had already commenced;

iv. the bosun’s store hatch lid did not admit water to the bosun’s

store until it was destroyed by impact, probably by some part of

the starboard windlass;

v. the starboard windlass broke loose by reason of hydrodynamic

loading, probably after some or all of the ventilators and

airpipes to the bosun’s store and ballast tank and probably also

SECTION 9: INFERENCES AS TO THE CAUSE OF THE LOSS

138

the No.1 hatch covers had already been destroyed and had

started to admit seawater thereby reducing the vessel’s

freeboard and increasing the exposure of the windlass to weld-

cracking wave forces.

9.3 The MARIN model test results, as extrapolated by Professor Tawn, strongly

point to the conclusion that the flooding of both the stores and the ballast tank,

and possibly even of the stores alone, could cause sufficient loss of freeboard

in the seastate experienced by the vessel to have exposed the No.1 hatch cover

to at least one hatch-breaking wave during the period 1700z to 1800z on 9th

September, that is to say to a wave imparting more than 42 kPa on those hatch

covers.

9.4 It is further demonstrated that, if none of the bow spaces were flooded, it is

extremely improbable that a hatch-breaking wave would be encountered

unless the vessel encountered seas of significant wave height enhanced by

15% above Dr Cardone’s hindcast while travelling at a minimum of 2 knots

or, which seems to be highly unlikely in such severe conditions, 4 knots.

9.5 It follows that there can be little doubt that the initiating cause of the loss of

the vessel was the destruction of some or all of the ventilators and airpipes

located on the foredeck and leading down into the bosun’s store, machinery

spaces and ballast tank and the consequent loss of freeboard at the bow due to

the gradual flooding of those forward spaces.

9.6 Once the vessel began to lose freeboard forward the incidence and magnitude

of green water loading on the No.1 hatch would have progressively increased,

as the available freeboard was reduced and the significant wave height

increased after 1200z on 9th September. This gradual deterioration in

freeboard would on the evidence have been imperceptible from the bridge.

Eventually the combination of diminished freeboard and a wave of a height

towards the top end of the hindcast range encountered at about 1700z to 1800z


139

on 9th September brought about loading on the No.1 hatch cover in excess of

its collapse strength.

9.7 Once the No.1 hatch covers had collapsed upon impact by such a hatch-

breaking wave, there would have been put in motion an irreversibly severe

loss of freeboard. The filling of the ullage space in No.1 hold would have

been very rapid. No expert has been able to advance any very precise period

of time, but Professor Vassalos and Mr Squire were of the view that this might

have been as much as 15 to 16.5 minutes. The former gave a range starting as

low as 5 minutes. Upon completion of flooding of No.1 hold, the calm water

freeboard would have been reduced by about 3.7 m. This would have caused

the foredeck to be almost continuously awash in the prevailing seastate and

would have exposed the No.2 hatch covers to intense green water loading.

This sequence is well illustrated by comparing three diagrams prepared by Mr

Squire of London Offshore Consultants Appendix 44. That considerable loss

of freeboard would cause the waves to run up the main deck over the hatch

covers aft No.1 hatch, thereby exposing the No.2 hatch covers to the risk of

more frequent loading by wave forces in excess of collapse strength, as shown

in Appendix 44. This would lead to the rapid collapse of No.2 hatch covers

and the filling of the No.2 hold ullage space as illustrated in the diagram also

at Appendix 44. The loss of freeboard would then be so great that the No.3

hatch covers would be caused to collapse and the vessel would sink.

• = Please see the accompanying CD-ROM for an animation of this

process

9.8 The UK/EC Assessors’ Survey was planned by reference to 13 “loss

scenarios” derived from Lord Donaldson’s Assessment. These are analysed in

a diagram in Appendix 3 of the Assessors’ Report, reproduced at Appendix

45, in relation to the initiating event of each and the consequential events

leading to the loss of the ship. In view of the strength of the evidence of the

cause of the loss discussed in this section of this Report and in sections 7


140

relating to the wreckage and 8 relating to Frame 65, it is unnecessary to

consider at length other causes of the loss thought to be possible before the

survey had been carried out.

9.9 The experts agreed by reference to the appearance of the wreckage, and this

Report accepts, that each of the following could be eliminated or at the most

was highly unlikely as a significant factor contributing to the loss.

C1: The cracking of the deck at Frame 65;

C2: The cracking of the deck at mid-sections. This could not have

occurred because the implosion/explosion damage to the

wreckage of this part of the vessel showed that the hull girder

must have been intact at sinking.

C3: Torsional fatigue cracking of hatch covers giving rise to

leakage and consequent cargo liquefaction. This would have

caused the vessel to capsize but the wreckage shows that it did

not.

C5: Failure of hatch cover attachments giving rise to leakage

followed by cargo liquefaction and capsize of the vessel or

sequential hatch failure. All the hatch covers had been

collapsed by seawater loading from forward and there was no

evidence that any of them were deficiently attached at the time

of sinking.

C6: Foredeck corrosion and fracture. The foredeck was not shown

to have been corroded or to have separated from the main part

of the vessel before sinking.

C11: Explosion and/or fire in the engine room. There is no evidence

of fire or smoke damage in the engine room or any other parts

of the wreckage.


141

C12 and C13: Pooping from forward waves or from running with the

sea. The lack of stern air pipe damage shown by the wreckage

did not suggest such wave impact.

9.10 A majority of the experts agreed, and this Report accepts, that C9 Propulsion

loss and C10 Rudder loss and steering failure were highly unlikely factors.

Both would have led the vessel to capsize, which was not the case. If the

effect had been to cause the vessel to swing beam on to the seas, it would

probably not have demonstrated damage to its hatch covers of the kind found

in the wreckage field, namely by longitudinal as distinct from transverse –

impact, unless, improbably, it had subsequently swung back with its bow to

the seas while hove to.

9.11 As to C8 cargo shift or liquefaction, a majority of the experts agreed and this

Report accepts that this was a highly unlikely factor. The effect would either

be to cause the vessel to capsize, which does not appear from the wreckage, or

to develop a list which somehow exposed it to more severe green water

loading on the forward hatches from the bow. The wreckage shows that there

was no substantial list and the causal connection of such list with the loss

appears highly unlikely.

9.12 Finally, a fourteenth scenario – some unknown cause - was added but rejected

by the majority of experts. This Report accepts that conclusion. The evidence

in support of the initiating cause of the loss being ingress of seawater into the

bow spaces due to damage to ventilators and airpipes is so compelling that as a

matter of probability there is no room for any further unidentified factor.

9.13 Following the conclusion of the evidence and before the final submissions,

this Investigation received from Mr Shaun Kent a detailed submission on the

conclusions to be drawn from the photographic and video evidence and as to

the cause of the loss. Mr Kent had previously acted as one of the advisers to

the DERBYSHIRE Families Association. Some years ago he purchased part


142

of the submerged wreckage of the KOWLOON BRIDGE in South West

Ireland. He has a deep interest in underwater surveys and in explaining the

loss of the DERBYSHIRE. His thesis as to the cause of the loss was at

variance with that set out in this Report. He said, in substance, that the vessel

had encountered huge waves caused by the clash of the counter-rotating

typhoon and the “oceanic re-circulation zone” associated with the apex

between the north equatorial current, and the Kuroshio current the effect of

which was augmented by lunar influence. These huge waves caused massive

green water loading on the foredeck which gave rise to structural failure of the

fuel tank, bosun’s store, machinery space, the stores flat cofferdam, the No.1

hold and the No.1 port and starboard saddle tanks. There was thus structural

failure of the side shell plating and the deck in way of the forward spaces. The

watertight integrity of No.1 hold was destroyed by the consequent failure of

Bulkhead 339.

9.14 This submission raised a number of matters which had not hitherto been

specifically considered by all of the experts or by the Court. In view of the

overriding public interest that all material evidence should be investigated

unless it was manifestly immaterial, but within the confines of the procedural

framework prescribed for this Investigation, it was ordered that, after the

conclusion of the hearing, Mr Kent’s submissions and evidence should be

considered by certain of the experts and if those submissions and evidence

were found to have sufficient substance to merit fuller investigation, the

hearing would have to be re-opened to enable Mr Kent to present his evidence

and to give other parties the opportunity of testing that evidence and his

submissions.

9.15 It is to be observed that at the time when he presented his written submissions

Mr Kent was clearly well aware from the Internet that the hearing of this

Investigation was about to close. It must have been obvious to him that what

he was doing would inevitably be to some extent disruptive.


143

9.16 This investigation has been advised by Dr Cardone that Mr Kent’s theory of

massive wave generation is quite unsustainable and provides no scientific

basis for displacing the hindcast significant wave heights presented in Dr

Cardone’s evidence. Advice has also been given by Mr Robin Williams and

Mr Andrew Squire. In essence, both expressed the view that the central thesis

advanced by Mr Kent as to the detachment of the bow section and the setting

down of the cofferdam and deformation of the hull in way of bulkhead 339 by

the impact of a huge wave or waves could not be sustained by the appearance

of the wreckage, the relevant features of which could only have been caused

after sinking had already commenced. Moreover, the wreckage showed

clearly that the bow section had not become wholly detached from the main

part of the hull.

9.17 In the light of this advice it was considered that a further hearing would not be

justified.


144


145

Section 10

The Development of International Ship Design

Regulations and Classification Society Rules

10.1 The freeboard of a ship is the height of its main continuous weather tight deck

above the water line. Freeboard determines the vessel’s reserve buoyancy. It

determines how much flooding a vessel can accept before it sinks. It also

influences the amount of green water which can invade the deck during heavy

weather. It directly affects the depth of the hull for a given draft and that

depth affects the longitudinal strength of the hull girder.

10.2 The frequency and amount of green water which boards the vessel in heavy

weather depends on a combination of factors inter-acting with the freeboard.

These include the height and frequency of the waves, the length of the vessel’s

hull and its shape. These factors affect the way in which the vessel moves in

the sea, in particular the extent to which it sustains leave, pitch and roll and

whether its hull rises with the waves or the bow pitches into the waves. The

height of a wave crest above the deck at the bow is likely to be greater than

amidships. Vessels are navigated in very heavy seas with the sea on the

starboard bow rather than abeam. Hence, in such conditions, greater freeboard

is required forward than amidships. This additional forward freeboard may be

permanently provided by a forecastle or a deck not parallel to the water line.

The DERBYSHIRE had neither of these features. That was not unusual at the

time of its design and construction.

10.3 Tankers do not require as much freeboard as other vessels because they have

very limited deck openings and therefore present a much reduced opportunity

for seawater entry by comparison with dry cargo ships with their hatches.

SECTION 10: THE DEVELOPMENT OF INTERNATIONAL SHIP DESIGN REGULATIONS AND CLASSIFICATION SOCIETY RULES

146

Bulk carriers, such as the DERBYSHIRE, which can carry dry bulk cargoes as

well as oil, have substantial hatches. It is obviously essential that the hatch

covers can be fastened so as to remain closed in heavy seas and that they have

sufficient strength to withstand such loading by green water as may reasonably

be anticipated. Such hatch covers must also be sufficiently strong to

accommodate internal loading from liquid cargoes when the vessel heels or

lists.

10.4 International regulation of hatch cover strength in conjunction with freeboard

first appeared in the International Load Line Convention 1930. However,

metal hatch covers had not yet come into use and the Convention did not

provide for such hatch covers. Tankers were permitted to have reduced

freeboards by comparison with dry cargo ships, but other vessels such as bulk

carriers, were permitted reduced freeboard if they possessed similar features

relevant to water-tight integrity and subdivision of cargo spaces to tankers.

Control of this concession was left to the individual countries’ administrative

authorities. The ILLC 1930 also contained an express provision inviting

classification societies to confer for the purpose of achieving uniformity in

standards of strength on which freeboard was based. The resulting

conferences of classification societies gave rise to the creation of the

International Association of Classification Societies (IACS).

10.5 By the time that the classification societies met in Paris in 1955 there were

already several oil tankers which had been converted into ore carriers but

which retained the same freeboard as when they had been tankers.

Discussions took place as to safety implications of the larger hatch covers used

on such vessels by comparison with tanker deck openings. The appropriate

strength for steel hatch covers was specifically considered. LRS took the

position that the scantlings for steel hatch covers on ore carriers should be

greater than those proposed for ordinary dry cargo vessels. It emerged that the

American Bureau of Shipping (“ABS”) had a somewhat higher minimum

strength requirement than the other societies. It was agreed to recommend to


147

the respective governments that the 1930 Convention should be amended to

deal with steel hatch covers on ore carriers.

10.6 Following the 1955 conference, LRS proposed to the other classification

societies that, where bulk carriers were assigned tanker freeboards, the section

modulus and thickness of hatch covers should be increased by 15 per cent.

Although many of the societies agreed, ABS did not. However, the United

Kingdom Government accepted the need for stronger hatch covers for ore

carriers assigned tanker freeboards and unilaterally adopted a 15 per cent

increase by introducing that for new buildings into Rule 109 of the Load Line

Rules 1959.

10.7 It had originally been intended that there should be an International Load Line

Conference in 1961, but that was eventually postponed until 1966. In the

meantime, the classification societies met in London in 1959. There emerged

differences of view as to whether freeboard should, as LRS proposed, be

determined by reference only to external strength of the vessel (including that

of hatch covers) or should be dependent upon internal watertight divisions by

analogy with the Safety of Life at Sea Convention. ABS proposed that steel

hatch covers should have an ultimate tensile strength capable of withstanding

loading of 360 1bs per sq ft (just over 1.75 tonnes per sq metre) multiplied by

a so-called “safety factor” of 4.50, subject to a limit of deflection of 0.0028

times the span subjected to that load. This proposal was generally accepted,

but not formally introduced.

10.8 The United Kingdom Ministry of Transport set up a load line working party to

consider proposals for amendments to the Convention. Its membership

included representatives of the biggest classification societies, LRS, Bureau

Veritas, ABS and Det Norske Veritas (“DNV”). In 1964 LRS took the

position that the ABS proposals provided inadequate hatch cover strength.

They wanted the reference loading to be increased to 370 1bs per sq.ft, with a

“safety” factor of 4.75, which was equivalent to their current rules


148

requirement. The Ministry of Transport took the position that ore carriers

should not have tanker freeboards unless the hatch covers were of sufficient

strength. It subsequently proposed that bulk carriers with gasketed hatch

covers of adequate strength should be given freeboards equivalent to 1930

tanker freeboards extended up to 950 feet in length beyond which the

freeboard would be a constant 126 inches, but the hatch cover strength would

have to be raised by 15 per cent to the present UK level of 425 1bs per sq ft.

10.9 This proposal was considered at a further meeting of the classification

societies in New York in 1965. The discussions highlighted a number of

considerations, which have in later years assumed increasing importance.

Germanischer Lloyd argued as to the relationship between permissible

freeboard and ship length by reference to data from model tests, a vital source

of relevant information, as this Report demonstrates. An issue arose as to

whether the UK proposals did in truth involve any greater hatch cover strength

than the ABS proposal because of the relationship between the level of

allowable stress required by the UK proposal and the minimum strength of

steel in the ABS rules. In connection with this issue DNV’s observation that,

in the context of the safety of the hatch covers against collapse, minimum

strength ought to be defined by reference to yield strength rather than ultimate

tensile strength is particularly pertinent to the present need to reformulate the

minimum strength requirements as recommended by this Report.

10.10 The major shipping nations approached the International Load Line

Conference in March/April 1966 with divergent views.

10.11 The UK delegation took the position that loadline should be based on the

following considerations:

(1) prevention of water ingress from water shipped on deck;

(2) maintenance of sufficient reserve buoyancy;


149

(3) protection of crew moving about or working on deck;

(4) provision and maintenance of structural strength of the hull.

10.12 The UK’s material objectives were to achieve deeper loading (reduced

freeboard) for tankers and dry cargo ships with steel hatch covers and to avoid

the dependence of freeboard on internal subdivision aimed at flooding

survival.

10.13 It was argued in the UK briefing document that there was a case for deeper

loading than that permitted under the 1930 Convention. Advent of gasketed

metal hatch covers had given greater deck integrity and lessened the need for

constant attendance by the crew to maintain hatch covers. The paper stated

that “maximum loadlines should be defined as those which correspond to an

average of the smallest freeboards at which competent and courageous

captains are prepared to take the ordinary risks of familiar voyages on ships of

known characteristics, without undue fear of damage to crew, ship or cargo”.

Trial and error was thus the only sure criterion for determination of a proper

freeboard.

10.14 This paper has to be seen against the very limited scientific data available at

that time. For example, nobody had at that stage assembled sea state analyses

of the scope and detail, which have been put before this Investigation by Dr

Cardone. Nor had anybody conducted model tests based on similar data.

There were no extant green water loading values of the nature of those

provided to this Investigation by the Ship Stability Research Centre of the

University of Strathclyde (“SSRC”) and MARIN. It is improbable that there

was any developed understanding of the effect on green water loading of the

relationship between vessel length, significant wave height, wave period,

relative vertical motion, speed and freeboard. The only reliable information

material to the issues before the Conference were the experience of vessels in


150

heavy seas and the extent to which damage had been sustained due to green

water loading.

10.15 The hatch cover strength proposed by the UK for reduced freeboard was 425

1b per sq ft (about 2.1 tonnes per sq m) with a safety factor of 4.75 with

reference to mild steel having a mean ultimate tensile strength of 29 tons per

sq inch. This was not accepted. It represented a minimum strength about 18

per cent higher than that ultimately adopted for all mild steel hatch covers by

the Conference, which was expressed in Regulation 16(2) as follows:

“Where weathertight covers are of mild steel the strength shall be

calculated with assumed loads of not less than 1.75 metric tons per

square metre (358 pounds per square foot) on hatchways in Position 1,

and not less than 1.30 metric tons per square metre (266 pounds per

square foot) on hatchways in Position 2, and the product of the

maximum stress thus calculated and the factor of 4.25 shall not exceed

the minimum ultimate strength of the material. They are to be so

designed as to limit the deflection to not more than 0.0028 times the

span under these loads. Mild steel plating forming the tops of covers

shall be not less in thickness than one per cent of the spacing of

stiffeners or 6 millimetres (0.24 inches) if that be the greater”.

10.16 The reference to Position 1 is to the forward 25 per cent of the vessel’s length.

10.17 Nor did the Conference accept the UK’s argument that if bulk carriers were to

be accorded a reduced freeboard they ought to have greater hatch cover

strength than a tanker of similar freeboard. The Conference adopted the

qualifying requirement of adequate internal flooding protection. Thus larger

bulk carriers with steel hatch covers were to be permitted to load 10 per cent

more deeply than under the 1930 Convention but, if they had a one

compartment standard of sub-division, they could load to 60 per cent of the


151

difference between that draught and the permissible tanker draught. This

qualification was subject to the conditions that:

“(a) That there were adequate measures for protecting the crew

from water on deck.

(b) That exposed hatch covers ‘have adequate strength and

securing arrangements’.

(c) That the ship should be able to survive one compartment

flooding.”

10.18 This was the “B-60 class” of vessels, to be contrasted with the B-100 class,

which was permitted to load down to tanker draught conditional on a two-

compartment standard of sub-division. The DERBYSHIRE was a B-60

vessel.

10.19 Investigation of the reports of what passed in the Technical Committee of the

Conference between 21st and 27th March 1966 make it clear that the UK’s

proposals for increased hatch cover strength, as distinct from internal sub-

division against flooding, were rejected and that the UK accepted the latter

solution with no increase in the hatch cover strength requirements for B-60

vessels. There is no substance in the suggestion advanced before this court

that the omission of any such requirements was due to there being insufficient

time fully to discuss the point. The UK proposal was simply not acceptable to

the vast majority of national delegations.

10.20 It is to be concluded that the conference and the UK Government cannot be

criticized for reaching this solution. The Ministry of Transport and the UK

delegation did all that reasonably could be done to obtain agreement to

enhanced hatch cover strength. The failure to persuade a majority of the 52

national delegations to accept these proposals has to be seen against the

background of the limited amount of contemporary information about sea


152

conditions and hatch cover loading to which reference has already been made

(see paragraph 10.15) and of the following specific considerations.

10.21 The stronger hatch covers proposed by the UK were put forward in response

to application to bulk carriers of the tanker freeboard. The B-60 minimum

freeboard provided an increased safety margin by increase of the maximum

permissible draught. This has to be seen against the background that ore

carriers had operated successfully at permitted tanker draughts for 10 years

and the UK had experienced no failures of steel hatch covers over the previous

25 years. The UK Government had carried out green sea boarding tests on 20

tankers of 600 feet and over which had been temporarily assigned deeper load

lines as proposed by UK and had been traded in northern winter conditions.

Except in one case there was no appreciable increase in water shipped over the

deck. There was no record to suggest that hatch covers were ordinarily

susceptible to green water damage.

10.22 The definition in ILLC 66 of hatch cover strength by reference to ultimate

tensile strength, on the face of it not a directly relevant factor, is unsurprising.

At that time the ultimate tensile strength of mild steel was more reliably

known than the yield strength. It would have been appreciated at the time that

by reference to yield strength the equivalent safety factor was 2.5 as distinct

from 4.25.

10.23 Following ILLC 66, there was a period of two years before it came into force.

During that period the UK maintained in force Rule 109 of the Load Line

Rules with its 15 per cent additional hatch cover strength margin over and

above the requirements of the 1930 Convention. However, by August 1966 it

had become clear that, apart from NKK, the Dutch classification society, other

countries were not interpreting the ILLC 66 “adequate strength” provision as

referable to the structural strength of hatch covers but rather to the sealing and

securing arrangements. In this respect, UK and the Netherlands became

isolated. There were at that time many bulk carriers on the UK register. This


153

Report accepts the evidence that if the British Government had continued to

insist on more stringent hatch cover strength requirements, above what was

required by ILLC 66, than any other major shipping nation, shipowners could

simply have “flagged out”. Had they not retained UK registry hatch covers

need not have been strengthened or replaced. In the event, there would have

been little effect on international safety standards overall and serious damage

to the British national interest as a leading merchant marine flag state.

10.24 In the event, the UK enacted the ILLC 66 in 1968, thereby bringing into force

the new B-60 regime. It did not proceed with its proposal to replace Rule 109

with a requirement that B-60 vessels should have a 10 per cent strength

increase and B-100 vessels a 20 per cent hatch cover strength increase.

Indeed, in the new Load Line Rules, which were enacted to bring into force

the ILLC 66 regime the UK, did not even include the requirement that B-60

vessels’ hatch covers should have adequate strength set out in Regulation

27(7)(c). That omission has not been explained. However, once the policy

decision had been taken not to depart from the minimum strength

requirements for No.1 hatches of other major maritime states, it is highly

improbable that the inclusion of this provision would have had any practical

effect.

10.25 The DERBYSHIRE’s keel was laid in November 1974. Her design had been

substantially that of the first of her sister ships and had originated in 1968.

10.26 It was suggested by Mr Brian Corlett of Burness Corlett & Partners in the

course of his evidence before this Investigation that, prior to construction of

the DERBYSHIRE, there was “at the research level” enough information

available to lead to the conclusion that loading on foredecks and Position 1

hatch covers on such large ships could be much greater than the load catered

for in hatch cover design rules and regulations in force at the time. However,

on further investigation it emerged that none of the sources of information


154

referred to by Mr Corlett was of sufficient substance to raise any real doubt as

to the sufficiency of the loading assumptions underlying ILLC 66.

10.27 Thus, a colloquium on Naval Construction in the Shipbuilding Industry

organised at Hamburg in February 1968 by Germanischer Lloyd and recorded

in the trade periodical Schiff and Hafen referred to 10 examples of recent

experience of considerable heavy weather damage to the forecastle deck

construction and/or forward deck of larger size bulk carriers recorded over the

previous 12 months. Amongst the damage referred to was sagging of part of

the forecastle deck and damage to the forward transverse hatch covers.

Reference was made to critical loads of 4 to 6 t/m2 having been calculated.

These would have exceeded the hatch cover design load under ILLC 66.

However, it was stated that, as no observations had been made during the

periods of impact, it was uncertain what the level of static pressure was. A

dynamic pressure component was also likely to be present. The reference

concludes with an acknowledgement that more exact investigations were

needed on the behaviour of large and small ships of a similar mass distribution

in the same seaway. Until then, damages could be reduced by strengthening

forecastle deck construction.

10.28 This material does not explicitly raise the question of position 1 hatch cover

strength. It leaves open to question a considerable number of matters relevant

to the level of hatch cover loading and although it refers to the theoretically

calculated loadings on the elements of and attached to the forecastle, those

figures do not apply to the hatch covers. To indicate such calculated loadings

in the context of suggesting further research would not in the current state of

knowledge be expected to put on notice the designer of a vessel such as the

DERBYSHIRE that forward covers ought to be substantially strengthened

beyond the ILLC 66 minima.

10.29 Secondly, reference was made to an article in Germany by Dr Hans J. Hansen

of Germanischer Lloyd, published in Schiff und Hafen in May 1972 entitled


155

“Uber die Vorhersage von Decksbelastungen durch Grunes Wasser”

(Concerning the Forecast of Green Water Deck Loading). This article records

a series of calculations referable to a Panamax bulk carrier showing static

loading on the forecastle and No.1 hatch in a sea state of significant wave

height of 12.5m and a wave period of 10 seconds on the basis of three

different speeds, namely 10, 14 and 16 knots. It assumes that freeboard

exceedance at the deck edge is equivalent to a static head of water. The

resulting calculated loadings are very high: for example 17 tonnes per square

metre at 10 knots. These results, however, give little practical indication of

realistic conditions or responses by the vessel. Thus, a speed of 10 knots, let

alone 14 or 16 knots, in such a sea state is virtually inconceivable. The

calculation of static head also rests on highly improbable assumptions. Above

all, it does not present any serious doubt about the ILLC 66 loading

assumptions. It is so unreal that it would not alert one to question the

sufficiency of ILLC 66. When Dr Hansen returned to this type of theoretical

investigation in 1982 in his article “Belastungen durch Grunes Wasser auf

Wetterdecks und Lukendeckel” (Loads due to Green Water on Weather Decks

and Hatch Covers) he accepted that there were significant unknowns in the

pressure calculations and that “in spite of statistical uncertainty inherent in

model tests, such tests were necessary for the elaboration of rules for

dimensioning”. It is to be observed that even in 1982 two years after the loss

of the DERBYSHIRE, one finds no suggestion in this article that ILLC 66

may be seriously deficient.

10.30 Finally, Mr Corlett drew attention to the Report of Committee 8 at the 5th

International Ship Structures Congress 1973 on “Slamming and Impact”. This

Report referred to work done by the Ship Structures Committee of West

Japan. The experimental work to which this Report refers appears to have

involved very high speeds, including apparently, damage to a No.1 hatch

cover sustained at 16 knots. The Report makes it clear that much further work

is required to be done in this field, including model testing. It concludes with


156

the recommendation that efforts to study the frequency and severity of deck

wetness and bow flare impact should be continued in an attempt to provide

useful information for design “and for use in future modification of the sheer

formula given by the International Load Line Convention”.

10.31 This certainly does not suggest that those in the forefront of research had come

anywhere near reaching the conclusion that there was a serious likelihood that

ILLC 66 was providing a level of protection which did not cater for realistic

extreme sea conditions. The area was one which called for extensive further

research.

10.32 Thus, assuming that 1974 was the last year in which a change in ILLC 66

could have affected the design of the DERBYSHIRE, for which the final work

was carried out up to 1976, it can be concluded that the contemporary state of

knowledge as to wave loading on forward hatch covers did not call for any

action towards amendment of the Convention. It was finally accepted by Mr

Corlett that, as a careful marine architect, he would have been content in 1974

to design a bulk carrier to the standards of ILLC 66.

10.33 In 1976, when the DERBYSHIRE was built, none of the major classification

societies, other than DNV, imposed No.1 hatch cover strength requirements by

reference to sea loading in excess of the ILLC 66 level of 1.75 tonnes per

square metre. The fact that Germanischer Lloyd did so some indication that,

in spite of the work and opinions of Dr Hansen, that classification society did

not regard the ILLC 66 standard for providing for too low a level of

protection. The enhanced strength requirements of DNV (approximately 15

per cent above ILLC 66) was due to a record of some damage to DNV classed

ships before this time. There is, however, no evidence as to the incidence of

such damage or the vessels concerned or the precise circumstances in which

the hatch cover damage was sustained.


157

10.34 In 1976 there was an experimental study of green water impacts on a 200

metre ore carrier in the course of model tests. This was put before the 1976

International Ship Structures Congress. It investigated green water loading on

the foredeck at significant wave heights of 6.5 m, 9.8 m and 13.1 m at speeds

of 8, 12, 16 and 20 knots. Loads substantially in excess of the ILLC 66

reference level of 1.75 tons per square metre were recorded. It is unclear to

what extent these were hydrodynamic, as distinct from sustained hydrostatic

pressure heads and that makes it impossible to make any accurate comparison

with ILLC 66.

10.35 Mr Corlett referred to three further research papers published during the

period 1976-1978. These were (i) a 1976 paper by Kawakami Tanaka

published by the Japanese Society of Naval Architects referring to model tests

of foredeck loading in North Atlantic conditions at a speed of 10 knots and

showing a maximum pressure of 23 t/m2 in 24 hours; (ii) a conference paper

presented in 1977 by Mr RG Lockhart of Lloyd’s Register of Shipping, in

which he reported on a statistical analysis of heavy weather damage carried

out by LRS which indicated some heavy weather damage to the forward hatch

covers of bulk carriers, and as to which, in the course of discussion, doubts

were expressed as to the adequacy of the requirements of ILLC 66; (iii) a

paper for the Japanese Society of Naval Architects in 1978 concerned with

model tests on an ore carrier conducted at a speed of 5 knots and indicating

foredeck impact pressures above the levels catered for by the ILLC 66.

10.36 The year before DERBYSHIRE was lost, 1979, a summary of the data upon

which Mr Lockhart had based his 1977 paper was included in an internal

paper to Lloyd’s Register Technical Association. It showed that the vast

majority of incidence of hatch cover leakage in bulk carriers were due to

faulty gaskets, seals or drainage, as distinct from panel collapse, and that

incidence of damage to complete hatch covers occurred at a frequency of only

2.6 per thousand ship years.


158

10.37 These materials do not provide any substantial basis for the suggestion that the

design of the DERBYSHIRE would, at the time when it was built, have been

seen by competent naval architects generally as unsafe or presenting any real

risk of undue hazard by reason of the impact of green seas on the forward

hatch covers. These research papers go no further than pointing out the need

for further research into hatch cover loading. On what was known at that time,

the belief that ILLC 66 provided for an entirely safe design regime, would

have been a perfectly acceptable professional view in the field of bulk carrier

design. Just as Mr Corlett would in 1974 have designed a bulk carrier by

reference to those requirements, so also would other marine architects in 1976.

10.38 In the course of the 1980’s there was a rising incidence of losses of bulk

carriers. This caused LRS to conduct an investigation aimed at explaining

these losses. In the result it was established that the vast majority of such

losses occurred in older vessels due to failure of side shell frames on account

of corrosion. There were also more instances of failure of hatch covers due to

wave impact, but quite clearly this was not seen by Lloyd’s as the main

problem, which was loss of side shell and hull girder strength due to corrosion.

In consequence in 1991, LRS amended its Rules for side frames and brackets,

but it clearly did not at that time see the need to address the level of hatch

cover strength.

10.39 Arising out of the incidence of bulk carrier losses IACS by its Working Party

on Strength commenced work in 1992 on a review of Classification Society

Rules for bulk carriers. By June 1997 IACS had agreed on a total of ten

unified requirements (“URS”) directed to single side skin bulk carriers of over

150 metres in length. UR S21 related to Evaluation of Scantlings of hatch

covers of bulk carrier cargo holds and it came into force for such bulk carriers

contracted for construction on or after 1st July 1998. The Working Party’s

mandate had been to develop load and strength criteria for hatch covers

sufficient to preserve their integrity, including realistic extreme loads.


159

10.40 The deliberations of the Working Party on Strength of IACS which gave rise

to UR S21 do not have to be considered in any detail. For present purposes it

can be recorded that all 10 full members and 3 associate members of IACS

were invited in December 1995 to provide calculations indicating their

minimum strength requirements for two examples of bulk carriers in the

context of a proposal to substitute a more stringent wave load factor for that in

ILLC 66. The Working Party chairman, Mr D Cervetto of Registro Italiano

Navale, indicated that it would be very profitable to conduct a comparison of

the different class rules.

10.41 The response of ABS supported the adoption of more severe wave loads for

forward hatch covers compared to ILLC 66. It stated that a green water

loading “corresponding to severe storm conditions seems appropriate”. An

appendix set out the ABS “procedure for calculating boarding seas forward

corresponding to a 10 hour storm”.

10.42 The returns to the Working Party showed that ABS, DNV and others had in

place formulae for calculating hatch cover strength which referred to strengths

in excess of those provided for by ILLC 66, although that of ABS was

apparently not a compulsory class requirement. Over the period January 1996

to January 1997 the proposals for UR S21 were discussed at five separate

meetings of the Working Party. It was decided to adopt as a basis for

discussion the load model provided by ABS. This was the subject of detailed

requests for explanations by various of the IACS members and considerable

detailed correspondence. In the event the ABS model formulae were revised

upwards to take as reference load strengths the highest heads in both the ABS

and DNV formulae. It is, however, apparent that the raw data (some model

tests and North Atlantic sea state statistics) were not explored in the course of

these discussions. Nor were the effects of certain of the features of the

formula, such as the interposition of the square root in √H and the function of

a minimum design speed of 13 knots. The ABS formula is said to have been

based on a 20 years period for North Atlantic storms using Dr Ochi’s extreme


160

value theory resulting in 1/10-8 risk which gave approximately a 1 per cent

chance of the maximum loading in that period.

10.43 However, it can be inferred from the proceedings of the Working Party that

the members were generally of the view that their mandate “to develop load

and strength criteria for hatch covers sufficient to preserve their integrity,

including realistic extreme loads” required that more stringent strengths than

those prescribed by ILLC 66 were necessary. The Working Party’s report to

the Council of IACS prepared in February 1996, well before finalisation of

UR S21, stated that “The WP/s considered the requirements of ILLC 66 in

comparison with each Society’s criteria and concluded that the ILLC 66

requirements are not sufficient for the hatch covers located in the forward area

(precisely, forward of 0,25-L from the forward perpendicular”). It is clear

from the context that these words can only be taken as meaning what they say.

They are not expressed as meaning that the Working Party had found that

some Classification Societies had criteria above those of ILLC 66 and for that

reason only the other societies in IACS that did not have sufficient

requirements. This Report views that statement as an objective expression of

the intrinsic insufficiency of the ILLC 66 requirements.

10.44 It is to be observed that the new UR S21 formula was to apply only to new

buildings. There is no evidence in the records of the Working Party or the

IACS Council’s deliberations that consideration was given at any stage to

applying UR S21 retrospectively. This is not entirely surprising. Although

the view was taken that there ought to be a substantial increase in the

minimum load head strength for forward hatch covers (from 1.75 tonnes per

square metre to 3.5 tonnes per square metre), it had not at that stage been

established that hatch cover collapse had caused the loss of the

DERBYSHIRE, albeit that this was amongst the possible loss scenarios

advanced in Lord Donaldson’s Assessment. The previous experience of

catastrophic loss due to such causes was extremely rare. Further, all the

members did not have access to the results of model tests monitored and


161

extrapolated with great sophistication such as those at MARIN and to some

extent SSRC. Nor does it appear that any of the members of IACS had

investigated the economic implications of making such a Classification

Society rule retrospective in effect.

10.45 The effect of the application of UR S21 to Capesize bulk carriers, such as the

DERBYSHIRE, was to increase the ILLC 66 design pressure loading on the

No.1 hatch position from 1.71 metres to 5.5 metres and the collapse load from

4.2 metres (the DERBYSHIRE’s actual collapse load) to a corroded value

collapse load of 8.3 metres, assuming good design and application of the

specified safety factor on first yield.

10.46 In the course of this Investigation a major issue arose as to whether, on the

presently available evidence, the minimum level of hatch cover strength

provided for B-60 bulk carriers by ILLC 66 should now be regarded as

inadequate.

10.47 It was also suggested, principally by Mr J S Milne of the DETR, that there was

evidence at least to suggest that the minimum forward hatch cover strength

required by UR S21 might also be inadequate. The substance of this evidence

was that model tests conducted by SSRC in 1998 on a model of the

DERBYSHIRE and in 1999 on a model of a Capesize bulk carrier, together

with the MARIN model tests on the DERBYSHIRE, as extrapolated by

Professor Tawn, demonstrated (at least) that there was a materially high

probability that in realistic extreme sea conditions a B-60 bulk carrier whose

No.1 hatch covers did no more than comply with UR S21 could encounter

green water loading which imposed hatch-breaking pressure on the No.1

hatch. Professor Faulkner also considered that UR S21 was inadequate

because it did not cater for sea states of sufficient severity having regard to

Pacific typhoons. This Report must now consider these issues.


162


163

Section 11

Towards Improved Ship Safety: Hatch Cover Strength

and Permissible Freeboard

The Concept of Adequate Strength

11.1 The level at which hatch cover strength is set by international Convention or

Classification Society rules involves a degree of subjectivity. To conclude

that any given strength requirement is “adequate” begs the question adequate

by reference to what criteria? For example, it would in theory be possible to

require all vessels to have hatch covers of such a high strength that no wave

loading that had ever been known to occur could cause collapse. That would

be an expensive and unnecessary exercise. Or it might be enough to require a

level of strength sufficient to withstand wave conditions that had been

experienced but which might occur, very rarely indeed.

11.2 In identifying the precise strength level one might also have regard to the fact

that major hurricanes and tropical typhoons can often, if not usually, be

avoided by careful navigation, thereby reducing the risk of any given vessel

encountering sea conditions of relatively rare severity in the course of its

trading life. Navigability renders a vessel less vulnerable to such encounters

than, for example, a fixed oil and gas platform for which design loads may

often be based on a return period of a hundred years or, more recently, up to a

thousand years. There is therefore an argument that strength requirements

need not be tailored to cater for the most extreme wave severity imaginable,

but rather something less which reflects the extreme improbability of such an

encounter. On the other hand, the ultimate purpose of any such minimum

strength requirement is to protect the vessel and those on board from total

SECTION 11: TOWARDS IMPROVED SHIP SAFETY: HATCH COVER STRENGTH AND PERMISSIBLE FREEBOARD

164

catastrophe as in the case of the DERBYSHIRE. The consequences of such

extremely improbable encounter eventuating may be potentially catastrophic.

In terms of risk assessment the degree of protection demanded must take into

account that level of seriousness. However, the precise level at which

protection is set will obviously be a matter on which views can differ.

11.3 The starting point for identifying the relevant level of protection clearly has to

be the extent of the risk of sea states of particular severity occurring anywhere.

This in turn involves a review of statistical data on reported sea states. The

evidence suggests that a significant wave height of not less than 15 metres

might fairly reflect realistic probabilities, at least for the North Atlantic trade.

This was the level upon which UR S21 was apparently based and it reflects the

incidence of reported worst case experience over a 20 year period in the North

Atlantic.

11.4 The unchallenged evidence before this Investigation – that of Dr Cardone – is

that during the period 1971 to 1986 there were 77 tropical typhoons and that,

by reference to wind speed, Orchid ranked 34. Of the 24 most severe

typhoons investigated in the North West Pacific the peak significant wave

height ranged from 7.2 metres to 17.6 metres. Orchid ranked 8th with

reference to significant wave height, on the basis of Dr Cardone’s hindcast of

a peak of 12.6 metres. The recurrence interval of a 12.6 metre significant

wave height in that area of the Pacific was not more than about 5 years. In as

much as hindcasting may have an inaccuracy range of +/- 10 per cent and

given that Capesize bulk carriers very regularly trade to the North West

Pacific, it may reasonably be safely concluded that the UR S21 formula basis

of 15 metres is the minimum appropriate wave height factor when considering

adequate strength of the hatch covers.

11.5 Further, the risk of heavy seas boarding a bulk carrier depends not only on the

wave height but also on the vessel’s length relative to the wave length. The

Buckley Survivability Envelope suggests that a 15m significant wave height


165

combined with a peak wave period of about 14 seconds is an appropriate

reference sea condition for a 300m Capesize bulk carrier.

11.6 The concept of adequacy of strength will also necessarily take loss statistics

into account. The number of reported losses of bulk carriers due to forward

hatch cover collapse under green water loading is extremely small. In this

connection, Lloyd’s Register very helpfully investigated its database and its

documentary archive in order to ascertain the number of incidents of hatch

cover damage apparently due to heavy weather caused to vessels on Lloyd’s

Register during 1960 to 1988. During this period there were 29 reported

incidents of such damage relating to 25,346 ship years, an average of one

incident every 874 ship years. In addition, only one further loss of a ship not

on Lloyd’s Register was identified between 1978 and 1999 as probably

attributable to seawater damage to hatch covers. However, that vessel was

much smaller than the DERBYSHIRE, and very much older.

11.7 The weight to be attached to a very low incidence of losses in the

quantification of adequate hatch cover strength must be a matter of judgment.

However the approach which this Report would strongly advocate is that a

maximum wave loading for realistic extreme sea conditions should first be

established and that the hatch cover strength considered to be adequate to

withstand that loading should be identified. That must be the irreducible

starting level of protection for the vessel and for those on board: irreducible

because, as the fate of the DERBYSHIRE shows, improbabilities can

eventuate to produce catastrophe. That level of protection should then be

increased by a safety factor over and above the figure for extreme conditions,

that safety factor to be set at a level which reflects the incidence and nature of

losses. That level will be a matter for judgment.

11.8 In determining the design requirements for Capesize bulk carriers the cost of

added strength is a further factor that must be considered. The position here is

different depending on whether one is looking at pre-existing vessels or at new


166

buildings. For the latter the position is relatively simple. The additional cost

of No.1 and No.2 rolling hatch covers of sufficiently increased strength

somewhat in excess of UR S21 requirements is unlikely to exceed about

£100,000, which represents a very small proportion of the total cost of hatch

covers for a nine-hatch bulk carrier.

11.9 As to existing vessels, mechanical problems may arise in ensuring that the

underlying structures are strong enough to support heavier hatch covers and

some down time may be involved in carrying out the work if it cannot be

combined with special survey work, but the additional cost of reinforcement of

the No.1 and 2 hatch covers is unlikely to exceed £150,000. This is a

relatively small proportion of the annual hire revenue of a Capesize bulk

carrier.

11.10 Accordingly the additional cost of a requirement of substantially increased

strength in forward hatch covers, even in existing vessels, should carry very

little weight indeed in judging the level of the requirements. Like the

incidence of past losses, it is a consideration which should go only to the

magnitude of the safety factor and not to whether the requirement should be

introduced or made applicable to existing vessels.

11.11 Against that background, it is now necessary to consider the level of

protection provided by the ILLC 66 regime and by UR S21 by which it has

been supplemented for vessels contracted to be built since 1998 which are

classed with members of IACS.

ILLC 66: the Level of Protection

11.12 The design pressure head loading required by ILLC 66 for all the hatch covers

on a B-60 bulk carrier, such as the DERBYSHIRE, is 1.71 metres. The

application of the regulations in the case of the No. 1 hatch on the

DERBYSHIRE gave rise to a collapse strength of 4.2 m (42 kPa).


167

11.13 It is quite clear that at 70% of the DERBYSHIRE’s maximum sustainable

speed for sea states in excess of about 9 metres significant wave height, there

would have been a growing risk that the No.1 hatch cover would have been

subjected to loading in excess of the design load of 1.71 metres sea water

head. For example, from Professor Tawn’s calculations for the vessel in its

intact (un-flooded) condition in a sea state of 10% below the significant wave

height (Hs=9.76m) it can be calculated that in as short as a 6 hour period there

would be a 97% chance of the No.1 hatch cover being subjected to green

water loading in excess of the design load.

11.14 This is very much lower than the worst significant wave height sustained by

the DERBYSHIRE (10.85 metres and possibly up to 11.95 metres) and for a

very much shorter period.

11.15 If one considers the ultimate collapse load implied by the ILLC 66 regulations

for the No.1 hatch on the DERBYSHIRE, that is a pressure head of 4.2 metres

(42 k Pa), it can be seen from the MARIN tests as extrapolated by Professor

Tawn’s statistical calculations that, when laden to her marks, thereby utilizing

all available freeboard, the DERBYSHIRE would have been at some risk (in

the order of 1 per cent) that her No.1 hatch covers would have been subjected

to collapse loading with a significant wave height of 10.85 metres +10%, even

at no forward speed. This is to be derived from the extrapolations shown by

Professor Tawn in respect of “Damaged Stores, single max, 10% Higher

Waves” (Table 2). The incidence of hatch cover loading in an intact fully

laden condition does not differ significantly from that encountered in the

actual “partly laden condition” with the stores space fully flooded. This

comparative exercise is justified by the results of the MARIN January tests

which show that when comparing the vessel’s relative motion in the full bow-

flooded condition with that in the intact (un-flooded) condition, there is close

correspondence in the frequency response curves.


168

11.16 If the vessel is assumed to be proceeding at 2 knots with a significant wave

height of 10.85 metres + 10% the risk of boarding seas causing pressures in

excess of the required collapse strength of the number 1 hatch covers would

rise to a range of 31% – 42% according to the period of exposure in question.

11.17 The so-called safety factor specified by ILLC 66 in terms of a ratio of 4.25

ultimate tensile strength to maximum working stress is found on investigation

to be illusory. It is equivalent to a ratio of 2.5 between maximum design stress

and yield stress for mild steel. On the DERBYSHIRE there was in reality no

significant reserve strength beyond the pressure load to cause the first yield.

That is to say the collapse load was virtually the same as the nominal load to

first yield.

11.18 This method of hatch strength calculation is not only misleading but

intrinsically defective because it does not require the true collapse load of the

hatch covers to be calculated or guarantee any reserve strength beyond first

yield. In no sense therefore can the 4.25 factor be described as a “safety”

factor.

11.19 Any reformulation of the minimum hatch cover strength requirements should

clearly be based on limit state methods. That is to say, it should refer to the

ultimate collapse strength of the structure considered as a whole as distinct

from the yield strength of the hatch cover panels. The hatch cover might well

retain a substantial load margin over and above yield strength before it became

so deformed as to admit seawater.

11.20 So far, it is the application of ILLC 66 to the forward holds of a bulk carrier

that has been considered. The limited tests conducted at MARIN were

directed to investigation of the causes of the loss of the DERBYSHIRE. To

the extent that those tests were directed to hatches aft of 0.25L the results

suggest that ILLC 66 requirements are sufficient. Further tests to be


169

conducted should investigate loading on all hatch covers in beam seas with

large rolling.

11.21 In summary, it can be concluded that the requirements of ILLC 66 are set at a

minimum level in relation to which there is a substantial risk of exceedance at

the forward hatch covers if a vessel such as the DERBYSHIRE is making any

speed over the ground and some risk if it is not. Consequently, these

requirements could only be regarded as “adequate” if that risk were very

substantially discounted because of the low incidence of loss experience since

ILLC 66 came into force and/or because of the additional cost that would be

involved in increasing the level of protection for new buildings and for

existing vessels.

11.22 Taking fully into account the lack of hard evidence of loss experience, the

likely additional cost of fitting stronger hatch covers and the technical and

mechanical problems to which that might give rise, this Report is unable to

accept in the light of the loss of the DERBYSHIRE that modern concepts of

appropriate safety standards can accommodate as acceptable the level of risk

of exposure to potentially catastrophic consequences presented by ILLC 66.

11.23 It is surprising and regrettable that witnesses called from LRS felt unable to

agree that the ILLC 66 requirements for forward hatch covers should now be

regarded as inadequate. Having regard to the enormous amount of work

which that organisation has put into the formulation of ship safety standards

over the years as a leading participant in IACS, not least in relation to the

introduction of UR S21, the failure to accept that the level of safety offered by

ILLC 66 was inadequate, suggested undue caution in the face of what on the

evidence before this Investigation is an obvious need.

11.24 Although it is clearly appropriate that the forthcoming comprehensive test

programme should further calculate the extent to which the requirements of

ILLC 66 fail to protect against realistic extreme sea conditions, this Report


170

concludes that on the evidence now available the standards are so seriously

deficient for Capesize bulk carriers that the Convention must be amended as a

matter of urgency to introduce enhanced requirements and that such

requirements should be made applicable to existing Capesize bulk carriers.

The precise quantification of such enhanced strength requirements will have to

be settled on the basis of the further test programme now envisaged.

11.25 It is not clear on the evidence before this Investigation what is the size of the

existing population of B-60 Capesize bulk carriers which have forward hatch

covers which do no more than comply with ILLC 66. Many of the total world

population of about 476 vessels (68 of them of similar length to the

DERBYSHIRE) may already have been given hatch covers which are of

equivalent strength to that specified by UR S21 or which at least have hatch

covers substantially stronger than the minimum that would be required by

ILLC 66. It may be that the strength levels achieved are such that acceptable

safety can be achieved by an increase in the minimum permissible freeboard

of existing vessels. The adoption, as recommended by this Report, of a

composite formula which, like UR S21, provides for hatch cover strength to be

calculated by reference to different levels of freeboard ought to provide

shipowners with some reasonable flexibility in deciding how they should

satisfy new requirements for additional hatch cover strength, relative to

minimum permissible freeboard. For example, it may be economically more

advantageous for the owners of older B-60 vessels to accept a substantial

increase in minimum freeboard rather to incur the cost of strengthening hatch

covers.

11.26 A relatively easily implemented and speedy method of increasing the safety of

forward hatch covers on Capesize bulk carriers would be to withdraw the 60

per cent freeboard concession for such vessels, thereby assimilating their

freeboard to that of ordinary dry cargo vessels. However, this would achieve

only a relatively small accretion to safety.


171

11.27 The order of magnitude of this increase can be derived from the MARIN test

results as extrapolated by Prof. Tawn. Thus, if the DERBYSHIRE had

relinquished her B-60 freeboard concession of 60 per cent and was laden down

to her marks, she would have had a roughly equivalent freeboard at the bow to

that which she had in her actual intact laden condition on the voyage as

assumed for the purposes of the MARIN tests. Consequently, the

extrapolations of those test results for the vessel in its intact condition give a

reasonable indication of risk of the No.1 hatch covers being collapsed in the

applicable sea conditions. For a significant wave height of 10.85m + 10% at

70 per cent of maximum sustainable speed, which would be just over 1.5 knots

over the ground, there would be a 1 per cent probability of a hatchbreaking

wave eventuating during the assumed periods and for significant wave height

of 10.85m + 15% at the same speed a 2 per cent probability. However, if the

vessel were proceeding at a speed of 2 knots over the ground, for significant

wave heights of 10.85m + 10% and 10.85m + 15% the probability of a hatch

breaking wave rises to 4 per cent and a range of 26% to 34%, according to the

applicable time period, respectively.

11.28 Thus, even if the 60 per cent freeboard concession were removed, a vessel

similar in shape and size to the DERBYSHIRE would have remained exposed

to a material risk of waves of hatch-breaking loading at significant wave

heights as low as 11.95m and 12.48m. This would represent an improved, but

still unacceptably low, level of protection. Accordingly, although this

expedient could be introduced with a minimum of commercial disruption, it

would not go far enough and could not be regarded as more than a very

temporary expedient.

UR S21: the Level of Protection

11.29 One of the more remarkable matters to emerge in the course of this

Investigation has been that IACS adopted this formula on the basis of


172

extremely slender information as to its underlying assumptions with regard to

environmental loading conditions. It is impossible to infer from the contents

of the regulation what it assumed in respect of the sea state, storm duration or

the statistical basis for the calculation of loading and in particular, whether the

UR S21 values are based on low probability extreme values. There seems to

be no logical explanation for the use of the square root in the formula p = 19.6

√H over the whole material range of H. No witness has been able to justify it,

although, according to evidence from ABS, this is a calibrated formula. There

are grounds for believing that in the case of Capesize vessels it can lead to

distortion for values of freeboard exceedance approaching the reference sea

state of 15 metres significant wave height. Nor has there been any satisfactory

explanation for the inclusion or applicability of the reference to a speed of 13

knots. It would be irrelevant to Capesize vessels in extreme sea conditions.

11.30 Doubts as to the accuracy of the loads predicted by UR S21 for Capesize bulk

carriers emerge from the model tests conducted by SSRC at Strathclyde and

from one of the model tests conducted on the DERBYSHIRE model at

MARIN. Thus, SSRC Test Case 3 which was on the basis of nil speed and a

significant wave height of 12.78m recorded a load of 81.56 kPa which

compares with a nominal design pressure load of 55 kPa under UR S21 and a

corroded value collapse load of 84 kPa. This appears to show an extremely

slender margin between the UR S21 minimum collapse load and the wave

loading at a significant wave height (12.78m) materially lower than the 15m

reference height for which UR S21 is supposed to cater.

11.31 This test result has been challenged as questionable by reference to its

apparent inconsistency with other tests on the DERBYSHIRE model, namely

a result obtained by SSRC in the course of the 1998 series and one obtained by

MARIN. The SSRC 1998 DERBYSHIRE model test gave a maximum hatch

cover loading of only 4m where the freeboard was only 5.56m, whereas with a

freeboard of 6.51m the SSRC Capesize model test gave a loading of 8.156m.

The MARIN test gave a maximum load of 6.83m on a freeboard of 6.47m but


173

with a significant wave height of 14.74m. However, that maximum load

exceeded the load for which UR S21 is supposed to cater, namely 5.5 metres.

The SSRC result has also been challenged on the grounds that it was

conducted on a model having a freeboard which was materially less than that

of a B-60 bulk carrier. There is conflicting evidence as to whether it is

possible to extrapolate from that result back to what would have been the

loading if the model had been on a B-60 freeboard.

11.32 Professor Vassalos of SSRC gave evidence about the model tests at

Strathclyde, and in particular the test which gave rise to the high loading of

81.56m (81.56 kPa) by comparison with the results obtained by SSRC on the

DERBYSHIRE model. It may be that, as he suggested, the apparent

inconsistencies do not really exist. They may arise from differences in the

configuration of the bow of the DERBYSHIRE and the bow of the Capesize

model used by SSRC. It is unnecessary for present purposes to attempt to

resolve this problem. However, there remains a large area of doubt as to the

reliability of the SSRC results as evidence of the true levels of protection

given by UR S21. This doubt can only be resolved by a further

comprehensive test programme specifically designed to explore the forward

hatch loading in sea conditions up to and beyond 15m significant wave height

and a wave period of about 14 seconds. This condition ought to be an

appropriate benchmark for Capesize bulk carriers.

11.33 Although it cannot be said with confidence, without further model tests,

whether UR S21 provides for sufficient hatch cover strength to withstand

realistic extreme sea state loading, it appears likely that at worst it may not fall

very far short of what might reasonably be required. Thus, given that for the

DERBYSHIRE the design load would have been 55 kPa and the minimum

required first yield strength would have been about 70 kPa on the basis of

which 84 kPa would have been achievable with good design, it appears that

had UR S21 been applied to the design of that vessel, the No.1 hatch cover

would not have failed in the course of Orchid even had the bow spaces been


174

fully flooded. There would probably have had to be a significant wave height

of approximately 14.7m, assuming a speed of 2 knots, and of about 16m

assuming nil speed for there to have been a 1% chance of the vessel

encountering a hatch-breaking wave in her actual loaded condition.

11.34 Further, if one assumes that the DERBYSHIRE had been loaded down to her

B-60 marks with her forepeak intact and that her No.1 hatch covers had been

designed to a collapse strength of 84 kPa on the basis of UR S21, the critical

significant wave height before there was a 1% risk of a hatch-breaking wave

would be 14m-14.5m if her speed were 2 knots and just over 15m if her speed

were zero.

11.35 UR S21, like ILLC66, directs itself to yield strength and not to collapse

strength. A safety factor of 1.25 on first yield is introduced. This bears no

necessary relationship to the reserve of strength before collapse strength.

Therefore there may be little or no reserve of strength beyond the safety factor.

It would be essential for any re-formulation of hatch strength requirements to

be based on limit state design values. This would give rise to more efficient

design and would give confidence that a defined collapse strength/failure

criterion had been achieved.

Resolving the Doubts

11.36 In the course of the hearing it became apparent that, although witnesses from

Lloyd’s Register, such as Mr Gavin, Mr Zheng and Mr Sole, firmly

maintained that the inadequacy of ILLC 66 minimum hatch cover strength

requirements and of the UR S21 requirements had not been demonstrated by

the SSRC and MARIN test results, and other witnesses including Professor

Faulkner, Professor Vassalos of SSRC and Mr Milne of the DETR were

strongly of the view that such inadequacy had been demonstrated, all were

agreed as to the need for a further comprehensive test programme to be carried


175

out at an early stage, if only to demonstrate the correctness of their own

position. In the course of Day 49 of the hearing the DETR made it known that

it would, in principle, be prepared to finance the conduct of such a programme

if that course were recommended by the Report. This was a highly impressive

demonstration of the Department’s sense of public responsibility for

international ship safety.

11.37 At the same time Lloyd’s Register, with commendable speed, canvassed the

membership of IACS as to whether there could be developed a jointly agreed

specification for such further tests so that a contract could be placed with an

appropriate model testing organisation without undue delay. In the event this

Court indicated that, on the evidence which had been placed before it, there

would be a recommendation that such an agreed test programme ought to be

carried out without delay. The reasons for the need for further tests have

already been set out in this Report.

11.38 The result of this indication was that the members of IACS had by 18th August

2000 agreed on a comprehensive specification for such tests and a contract

was duly placed with MARIN. The specification was considered by and

approved in principle by this Court as likely to provide data necessary for

reconsideration of the adequacy of the protection given by the requirements of

ILLC 66 and UR S21. The test programme was fully carried out in the course

of August and September 2000, even before this Report had been completed.

Towards a Formula for future Ship Safety

11.39 While it is accepted that until the results of the new test programme have been

analysed and evaluated it will not be possible to develop with any precision

the structure of a formula which provides for hatch cover strength relative to

minimum freeboard, it is appropriate for the purposes of this Report to

indicate, without seeking to impose an inflexible methodology, certain general


176

considerations to which on the evidence before this Investigation those

concerned with forming new international regulations or classification Society

rules should have regard.

(1) The overriding objective of any such formula must be to

provide for survivability of the vessel in question in any

realistic extreme condition that it may encounter during its

service.

(2) It is essential that there should be developed a self-contained

composite formula which, subject to a minimum freeboard,

provides for variable hatch cover strength relative to levels of

freeboard additional to the minimum level.

(3) For the purpose of developing this formula the logical starting

point is to identify the relevant realistic extreme conditions in

terms of significant wave height, modal period and expected

type and from this to calculate the significant wave amplitude.

(4) It is then necessary to calculate the significant relative vertical

motion of the vessel, that is to say the overall degree of

magnification of relative motion with reference to wave

elevation, at the relevant position on the vessel. The data

necessary for this exercise is ideally to be derived from model

testing.

(5) The next step is to calculate the extreme relative vertical

motion at the relevant positions on board the vessel for a

defined risk level. That risk level would sensibly be set at a

low value, perhaps as low as 1 per cent for a storm duration of

3 to 24 hours. There must be a factor to account for negative

non-linearity.


177

(6) It is next necessary to calculate extreme freeboard exceedance

from the still waterline to the relevant places on the vessel, such

as the top of the hatch cover for the defined risk level.

(7) Finally it is necessary to calculate the extreme impact pressure

head on the hatch covers for the defined risk level. Here again

further model tests are necessary to establish the relationship

between varying values of freeboard exceedance at any given

location on the vessel and the impact pressurehead.

(8) The broad essential issues which have to be resolved in the

course of developing an effective safety formula are identified

sufficiently for present purposes in a document most of which

is reproduced at Appendix 46. It may be that, on the basis of

data to be derived from the further test programme,

modifications will have to be made to this general approach.

11.40 Having regard to the fact that, with a highly commendable demonstration of

its sense of responsibility for ship safety standards, the DETR has undertaken

to fund the further comprehensive model testing programme on the basis of a

specification mutually acceptable to it and to IACS, it is clearly to be

recommended that before finalisation of that specification:

(a) DETR, LRS and IACS should as soon as possible agree on the

relevant realistic extreme sea conditions referable to each type

of vessel, in particular Capesize and Panamax bulk carriers; and

(b) DETR, LRS and IACS should as soon as possible agree on the

relevant level of risk in the selected extreme sea conditions.

11.41 This recommendation had been complied with before this Report had been

completed and the further programme of model tests recommended to be


178

carried out at MARIN and be completed at as early a date as practicable had

also already been completed.

11.42 There should be agreed between LRS, IACS and the DETR a specification for

the Statistical extrapolation of values of extreme seas loading from the data to

be derived from the further model tests. This extrapolation ought to be

conducted by Professor Tawn in view of his familiarity with and great

contribution to this Investigation. It is further to be recommended that, once

the further test programme has been completed, IACS should, as a matter of

priority, proceed to review the data and the statistical extrapolation from it in

conjunction with the DETR and they should endeavour to agree:

(i) whether the level of protection provided by UR S21 accords

with the objectives of such a minimum strength formula

identified in paragraphs ( 1-7 ) of this Report;

(ii) the details and values of a formula to replace UR S21 so as to

provide a clear and logically structured calculation of the

appropriate level of protection, (based on limit state design),

whether the same as or different from that provided by UR S21.

11.43 It is of great importance that, as soon as possible after such a minimum

strength formula has been agreed upon by IACS, it should be introduced by

the members as part of their classification rules.

11.44 The question arises at this point whether the new minimum strength formula

should be made applicable only to new buildings or should be extended to

existing vessels.

11.45 Since there can be no doubt whatever that ILLC 66 provides for a level of

protection substantially too low by reference to modern safety standards, this

Report strongly recommends that it is essential that the new formula should be

made applicable at least to all existing bulk carriers which do no more than


179

comply with the ILLC 66 regulations for minimum hatch cover strength

referable to the relevant minimum freeboard. It should be left to ship owners

to decide how to apply the new formula to their vessels, in particular whether

to opt for additional hatch cover strength or for greater minimum freeboard or

for a combination of the two.

11.46 A question arises as to those vessels which already have an enhanced level of

protection. Several classification societies had introduced minimum hatch

cover strength requirements more stringent than those of ILLC 66 but before

the introduction of UR S21. There may therefore be quite a number of vessels

afloat whose hatch over collapse strength exceeds what it would have been

under ILLC 66 but which is below that requirement by UR S21 or by such

minimum strength formula as may now be developed in response to this

Report.

11.47 Should the new requirement be applicable to such vessels whatever the extent

of the deficiencies in strength or should there be a designated maximum

permissible deficiency, say 10 per cent?

11.48 Having regard to the permissible flexibility in compliance implicit in a

formula which tailors hatch cover strength to variable minimum freeboard as

envisaged by this Report, there can be little justification in principle, for

permitting any reduced strength level not matched by an additional minimum

freeboard level, given the all-important considerations of crew and vessel

safety which underlie the recommended minimum strength criteria.

Accordingly, it is recommended that all existing bulk carriers whose hatch

covers are shown to offer levels of protection below those that would be

required by the new composite formula should be required either to increase

hatch cover strength or to accept sufficient increased freeboard or both.

11.49 The period of time allowed to shipowners to comply with the new minimum

strength requirements should not exceed twelve months.


180

11.50 The response of IACS to the further information that will be derived from the

proposed test programme is of great importance to the development of

amendments to ILLC 66. The IMO relies heavily on the classification

societies and on IACS in particular to supply technical advice. There is

currently scheduled a meeting of the Maritime Safety Committee of IMO

(“MSC”) in May 2001. That would be the earliest time at which that body

could consider recommendations from IACS with regard to amendments to

ILLC 66. In order to decide whether any amendment should be made

retrospective, the MSC would require to see the results of a formal safety

assessment, currently in progress, so that it could take into account statistical

information about the incidence of losses, risk magnitude, the number of

existing vessels involved and the cost implications. Since amendments to

ILLC 66 might extend beyond hatch cover strength, this formal safety

assessment might have to take account of wider considerations than those

relating only to minimum freeboard and hatch cover strength. The

deliberations of IMO would depend heavily on advice from IACS. However,

even if IACS were to provide a fully-presented case for amendments to the

Convention by the May 2001 meeting of the MSC, it would be very unlikely

that any amendment to the Convention could take effect until well into 2004,

or even 2005, at the earliest.

11.51 Because there can be no reasonable doubt that ILLC 66 provides for a

minimum level of protection which is far too low for present-day safety

standards, this Report concludes that, as soon as the results of the new model

test programme have been analysed and the new minimum strength formula

has been developed by IACS in conjunction with DETR, the Department

should, as soon as practicable, use its best endeavours to persuade member

states of the IMO to agree to amend the ILLC 66 to make provision for higher

safety standards as expressed in the formula. Such amendments should take

effect both for new buildings and for existing vessels.


181

11.52 The present regulatory regime relating to hatch cover strength is extremely

unsatisfactory. There are two distinct sources of regulation. There is the

ILLC 66 regime which is given the force of law in domestic maritime codes

throughout the world. In addition to that, there is the UR S21 regime which

can only be enforced by those classification societies which have introduced it

withdrawing class from vessels entered with them. Since members of IACS

account for about 96 per cent of registrations of bulk carriers, there are

unlikely to be many new buildings contracted after July 1998 and in future

which do not comply with UR S21.

11.53 It is, however, in relation to the large number of bulk carriers built before UR

S21 was introduced and took effect that concern arises. That is by far the

greater part of the world’s bulk carrier population. It would be highly

unsatisfactory if these vessels were to remain unstrengthened and unregulated

beyond the requirements of ILLC 66 for another 4 or 5 years pending

amendment of the Convention. For these reasons, unsatisfactory though it

may be to have, as at present, two distinct regulatory regimes – the

Convention and the classification societies, this Report recommends that as

soon as IACS has finalised its new minimum hatch cover strength formula, its

members should introduce that formula into their class rules to come into

operation with retrospective effect in respect of existing bulk carriers, as soon

as possible and without waiting for IMO to agree on amendments to the

Convention. The evidence suggests that IACS can reach decisions

significantly more speedily than IMO and the need for additional protection

for those on board bulk carriers calls for speedy improvement.

11.54 Reference has already been made to the Maritime Safety Committee of the

IMO requiring to see the results of a formal safety assessment before it could

be expected to take action on the amendment of the 1966 Convention,

particularly so as to operate retrospectively so as to impose new strength

requirements on existing bulk carriers. There is obviously a pertinent question

as to whether IACS also should similarly refrain from introducing


182

retrospectively a new formula setting requirements as to increases in hatch

cover strength in relation to permissible freeboard until that formal safety

assessment has been completed. Any such assessment would involve a major

investigation of the economic consequences of requiring owners of existing

vessels to comply with substantially increased hatch cover strength

requirements for the forward hatches or to comply with a significantly reduced

permissible freeboard. One way, they would incur expense in hatch

reconstruction. The other way they would lose cargo capacity and revenue for

the remainder of the vessel’s trading life. As already indicated (see paragraph

11.9 above) the evidence, in so far as it goes, suggests that, unless the

remaining trading life were extremely short, reconstruction of the hatches

would be likely to be the commercially preferable solution.

11.55 This report concludes that the need to replace the 1966 Load Line Convention

minimum permissible load lines and hatch cover strength requirements, at

least in the case of existing Capesize bulk carriers, is so strongly established in

the interests of ship safety that once IACS has satisfied itself as to the essential

suitability of a new formula, it should make that formula applicable to existing

vessels without any further delay related to the completion of the formal safety

assessment needed by the IMO’s MSC.

11.56 There could obviously be room for flexibility in the retrospective application

of the new formula depending on the extent to which an existing vessel’s

hatch cover strength fell short of the permissible minimum and/or the

remaining trading life of the vessel. However, if the formula, offers the

alternative facility of an increased minimum permissible freeboard for vessels

with below strength hatch covers, as this Report recommends, the need for

special dispensation from compliance with the new hatch cover strength

requirements should be minimal regardless of the cost/risk factor.

11.57 This Investigation has carefully considered whether it would be appropriate

for the UK Government to lead the progress towards greater ship safety


183

recommended by this Report by unilaterally adopting a more stringent hatch

cover strength regime by amendment of the Regulations under the Merchant

Shipping Act in advance of any amendment of the Convention. The evidence

very strongly suggests that any such move would be counterproductive. There

are currently no Capesize bulk carriers on the British Register, although there

are significant numbers on the registers of governments, such as the Cayman

Islands, whose legislation is based on that of the Merchant Shipping Acts.

The adoption of more stringent requirements in advance of international

acceptance would cause shipowners to avoid the British Register and that of

administrations with similar requirements. They would be likely to “flag out”,

that is change flag to a less stringent jurisdiction.

11.58 Similar considerations apply to a classification society, such as LRS,

unilaterally changing its rules to require more stringent minimum hatch cover

strength. The reality is that shipowners would be most unlikely to continue to

class their vessels with a classification society that proceeded unilaterally in

advance of the adoption by all the members of IACS of a universal

requirement. Members of IACS are the classification societies for about 96

per cent of the world population of bulk carriers. Provided that all the

members of IACS agree to adopt a more stringent safety formula, ship owners

are likely to accept it, rather than evade it by changing class, even if it operates

retrospectively so as to affect existing vessels.

11.59 In view of the results of the continuing programme of model tests being

conducted at MARIN not being known at the time of completion of this

Report, it may be necessary, depending on the results, at a later date to issue

an Addendum to this Section.


184


185

Section 12

Towards Improved Ship Safety: Navigational Matters

Positional Reporting each Day

12.1 There was no standing instruction to the master of the DERBYSHIRE that he

should report his position to owners each day and he did not do so. In the

event this did not make it more difficult to locate the wreckage site or

materially prejudice the safety of the vessel because the vessel sent messages

at 0300z and 1019z on 9th September, giving her position. However, she did

not report her position on 7th September.

12.2 It is important that vessels should make daily reports of their position. Failure

to do so may result in a vessel being lost for more than 24 hours without any

indication that this might have occurred.

12.3 Accordingly, it is recommended that IMO should require the compulsory daily

reporting of position by all vessels. Some consideration needs to be given as

to the appropriate recipient of the report, particularly in the case of vessels

under charter. It is to be noted that, subsequently to the loss of the

DERBYSHIRE, Bibby introduced such a reporting practice for their vessels.

Mariners’ Handbook Advice on Navigation in Tropical Storms

12.4 The Original Formal Investigation Report recommended at page 45 as follows:

“Thirdly, we consider that the “rules for avoidance when in the

dangerous semicircle” ought at least to be subject to review. We have

no doubt that they are broadly sound but suggest that when a ship is in

SECTION 12: TOWARDS IMPROVED SHIP SAFETY: NAVIGATIONAL MATTERS

186

the dangerous quadrant but not far from the forecast track, it may well

– if she is still far enough from the worst weather to retain her freedom

of manoeuvre – be better to run with the wind on the quarter, as if she

was in the navigable semicircle, at least in fairly low latitudes when the

storm is likely to be slow moving.

We appreciate that there is some hazard in this proposal and do not

suggest that it should be followed up purely on the strength of this one

casualty, but feel that there is cause for the authorities responsible for

the books quoted, both of which are regarded as standard works, to

consider whether some revision is necessary.”

12.5 The 1989 Edition of the Handbook did not adopt this recommendation. It

retained the advice to bring the wind well on the starboard quarter in such

circumstances. It does not suggest that there may be circumstances where it

would be better to run with the wind on the port quarter, in effect turning well

across to port, as if the navigable semi-circle. Captain de Coverly and Captain

Roberts both held the view that the Handbook ought to be amended to take

account of this as a possible option provided that the immediate sea conditions

were such that the vessel retained her freedom of manoeuvre and in cases

where the storm was moving slowly.

12.6 This Report concludes that some amendment to the Handbook advice is

probably desirable. The problem with the 1989 text is that it does not

contemplate that running with the wind on the port quarter in the dangerous

sector may in some special circumstances be a more desirable course than

bringing the wind on the to starboard quarter and proceeding with all available

speed.

12.7 It is recommended that the Handbook ought to be amended so as not to close

off the option of entirely running with the wind on the port quarter. However,

the wording should be carefully qualified so as to convey the need for caution


187

as to the need for the storm to be advancing slowly and for the sea conditions

to admit of continuing manoeuvrability.

Advice as to the Danger of Bow Flooding

12.8 The experience of the DERBYSHIRE has brought to light the very

considerably increased risk of severe damage to the foredeck and to forward

hatches resulting from relatively small reductions in freeboard and small

variations in speed. If there is significant water entry into the forward spaces

when the bulk carrier is deeply laden, green water loading on forward hatches

may exceed collapse strength with potentially disastrous consequences.

12.9 Accordingly, it is important that masters of bulk carriers – particularly of

Capesize carriers – are made fully aware of this danger. This should be done

by means of IMO circular letters and British M notices.

Meteorological Data

12.10 There is still a real scarcity of meteorological information in some areas of the

seas through which relatively few vessels may navigate. The position would

be greatly improved if more vessels participated in the World Meteorological

Organisation’s Voluntary Observing Ships’ scheme. This scheme involves the

provision by WMO of tested meteorological instruments to each participating

vessel free of charge together with instruction on how to use the

instrumentation and to transmit data.

12.11 In December 1994 the IMO recommended member governments to bring the

scheme to the attention of the shipping industry and to encourage volunteers.

12.12 It is highly desirable that there should be a wider participation in this scheme

that may help to avoid the unfortunate inconsistencies of forecasts of the path

SECTION 12: TOWARDS IMPROVED SHIP SAFETY: NAVIGATIONAL MATTERS

188

and associated conditions of Typhoon Orchid evidenced in the case of the

DERBYSHIRE.

12.13 It is therefore recommended that a British M Notice should encourage wider

participation in the scheme and that the IMO should issue further circular

letters to that effect.

Weather Routeing

12.14 This Report has already concluded that routeing agencies should make very

clear to masters the precise circumstances in which they undertake to give

positive routeing advice in the course of the voyage (see paragraph 4.78

above).

Training of Deck Officers

12.15 It was strongly propounded by Captain Mackie, supported by Captain Roberts,

that the training of deck officers to prepare them for navigation in sea

conditions of extreme severity was inadequate in as much as it was too

theoretical and involved insufficient practical training on board ship and ought

to involve greater use of simulators. High quality training referable to

dangerous sea conditions is obviously an important part of the preparation of a

deck officer for his very responsible duties. However, it is quite clear that

those navigating the DERBYSHIRE were both highly trained and experienced

mariners. It is not suggested that their seamanship was at fault.

12.16 While acknowledging the importance of practical training in heavy weather

seamanship, it is not appropriate that this Report should make any specific

recommendations in respect of a field (of officer training) which has not been

specifically in question or under investigation.


189

Section 13

Towards Improved Ship Safety: Additional Matters

Ventilators and Air Pipes on the Foredeck

13.1 As appears from this Report, the initiating cause of the loss of freeboard in the

course of the typhoon was the entry of seawater into the bow spaces by reason

of damage to air pipes and ventilators located on the foredeck. The occurrence

of such damage in the course of the typhoon would not be detectable from the

bridge. The evidence suggests, that the occurrence of such damage would

have been treated as a routine incident, giving rise to an ordinary repair item,

not necessarily reportable to the vessel’s classification society unless it were

relevant to an annual class inspection.

13.2 Furthermore, little or no research appears to have been done on sustainable

green water loading for ventilators of a standard type, such as the mushroom

design. It is not the practice for classification societies to incorporate

minimum strength requirements into their rules. Obviously, the smaller the

risk of damage to air pipes and ventilators, and therefore of seawater ingress,

the smaller the risk of loss of freeboard and of the resultant increased green

water loading on forward hatches.

13.3 There is clearly a need in the interests of increased ship safety to reduce these

risks. However, research directed to minimum strength requirements for such

fittings involves difficulties not encountered in the field of research into hatch

cover loading. In particular, these items of deck furniture are so small relative

to the overall size of a bulk carrier that they do not lend themselves to model

testing in a facility such as that used by MARIN for testing the loading of

hatch covers. Further, they have peculiarities of configuration and disparities

SECTION 13: TOWARDS IMPROVED SHIP SAFETY: ADDITIONAL MATTERS

190

of location in relation to other features of the foredeck which could give rise to

widely varied exposures to green water loading. Consequently, research

aimed at such objectives as minimum ventilator strength and optimum location

and protection, is likely to be complex and time consuming. That is not to say

that it is not worth doing. On the contrary, this Report recommends that IACS

should set up a research programme aimed at ascertaining from vessels in

service the nature and extent of such damage, with a view to establishing

minimum strength requirements and, if necessary, location and protection

requirements for such deck fittings. This programme should be conducted

independently of the re-assessment of and introduction of the requirements for

a minimum strength for hatch covers. It is highly undesirable that the latter

should be delayed to accommodate what will inevitably be relatively time-

consuming research.

13.4 In order to improve the availability of material information to those on the

bridge, it is desirable that the fact that a ventilator or air pipe is open either

deliberately or by reason of damage or destruction, should be electronically

indicated on the bridge.

Foredeck Hatches

13.5 A great deal of evidence before this Investigation related to the bosun’s store

hatch. On board vessels, such as the DERBYSHIRE, which have no

forecastle in which the door to the forepeak store can be set aft, an access

hatch has to be located on the foredeck in a relatively exposed position. It is

thus especially vulnerable to green water impact and, if not watertight, to the

risk of seawater entry into the stores space. It was therefore surprising that

there was strong evidence that such hatches on the DERBYSHIRE and some

of her sister ships suffered from a design defect in as much as the toggles

tended to loosen and ride down their shanks with the working of the vessel in

the seaway. That this was recognised to present a risk of water entry is


191

demonstrated by the development of back-up securing devices such as locking

bars and the rope lashings used on the DERBYSHIRE.

13.6 A properly designed hatch should not need such makeshift expedients. The

more safety measures required to be taken the greater the risk that they will be

omitted or inadequately carried out. IACS should therefore set up a research

programme to develop a new kind of locking device for such hatches which

continues to provide sufficient pressure to the lid to maintain watertightness in

heavy seas with a view to introducing rules requiring such a device to be

installed.

13.7 Further, in the interests of providing improved information to those on the

bridge, displacement of the foredeck hatch lid should be electronically

indicated.

Hatch Cover Fittings and Cleats

13.8 This Report does not conclude that defects in the cargo hatch covers (as

distinct from their strength) or omission to secure them in any way caused or

contributed to the loss of the DERBYSHIRE. Nevertheless, it is apparent

from the evidence and the contents of the MacGregor Hatch Cover Manual,

Operating Instructions, that there was room for confusion over the need to

secure the cross-joint catches. It is essential that instructions to the crew as to

securing hatch covers are clearly and unambiguously expressed both as to the

method and extent of securing.

13.9 In order to assist the monitoring of the proper securing of hatch covers it is

desirable that there should be an electronic indication to the bridge that all

cleats and other securing devices are effectively in place.


192

13.10 There is also much to be said for the introduction of automated closing devices

for cargo hatches. These would ensure speedy securing of hatch covers

following completion of cargo operations.

13.11 The DETR should encourage IMO to develop and introduce guidelines to this

effect.

Spurling Pipes.

13.12 As concluded by this Report, the cement and chicken wire “seal” applied to

the spurling pipes was destroyed, probably prior to sinking by seawater

impact. There would therefore be water entry to chain lockers, but, by reason

of the bolted manhole as distinct from a door connecting the chain lockers

with the stores, there would be no further water ingress beyond that.

Consequently the contribution to the loss of freeboard would have been very

small having regard to the volume of the chain lockers.

13.13 It is recommended that the DETR should encourage IMO and IACS to

investigate alternative means of sealing up spurling pipes to give a less

vulnerable packing. Further, IACS should require that chain locker access

should be through bolted manholes.

Deck Fittings

13.14 There was much investigation in the course of the hearing as to whether

damage caused by the breaking loose of deck fittings, such as the starboard

windlass, or the windlass floor plates, might have caused damage to the

ventilators or the bosun’s hatch, thereby causing the entry of seawater into the

stores and ballast tank. This Report has rejected as improbable that sequence

of events, although concluding that at a later stage, after initial flooding of the

stores and possibly the ballast tank and No.1 hold, the starboard windlass


193

broke loose and probably caused the destruction of the bosun’s hatch lid and

damaged the hatch coaming.

13.15 It is, however important that the ability of such deck fittings to withstand

heavy green water loading, including the attachment of such fittings to the

foredeck, should be reviewed by IACS with a view to introducing

requirements for minimum securing strength.

Illumination of the Foredeck

13.16 The evidence strongly suggests that, even in daylight, the distance from the

bridge to the foredeck or a Capesize bulk carrier would be so great that in

heavy seas and poor visibility it would be extremely difficult for duty officers

to see if damage was being caused in the area of the foredeck and forward

hatches. At night the problem is obviously worse. The introduction of

electronic information monitors on the bridge connected to air pipes,

ventilators and hatch covers as already recommended would obviously

improve matters. However, the installation of a powerful lighting system

based on suitably strong fittings and also of industrial video cameras should be

considered by IMO and IACS with a view to recommending that such

facilities be installed on all Capesize bulk carriers.

Monitoring of Water Ingress into forward Spaces

13.17 Because of the immense length of Capsize bulk carriers it is exceptionally

difficult for those on the bridge to detect changes in trim and ship’s motion in

circumstances, for example, where water has entered the forward spaces. This

is particularly so in heavy seas when the bow is never seen to be in a settled

position horizontal to the calm water level. The evidence was to the effect that

in such conditions it might be difficult to detect a trim by the bow of as much


194

as 2 to 3 metres. The tests conducted by MARIN have demonstrated the

considerable sensitivity of the risk of encountering green water loading

sufficient to destroy a forward hatch cover of a loss of freeboard to that limited

extent.

13.18 It is therefore important that there be installed into forward spaces a

monitoring system which communicates bilge levels electronically to the

bridge. There should also be an audio alarm to attract the attention of the duty

officer to a flooding situation. The DETR should take steps to persuade the

IMO to make this mandatory.

Emergency Pumping of forward Spaces

13.19 One of the problems arising from the exposed position of the foredeck and its

immense distance from the bridge on a Capesize bulk carrier is that of

pumping out the spaces in response to the water ingress. Clearly it is

insufficient to depend on a pumping system which can only be effectively

activated from the foredeck area. On the DERBYSHIRE, for example, there

was a facility for 2-valve separation of the forepeak ballast tank only if a

manual valve operated from the bosun’s store were left open. It would then be

possible to pump out through the ballast line the contents of the ballast tank by

operation from the cargo control room. The bosun’s store could not be

pumped out by that system. If the manual valve were left closed, the forepeak

could not be pumped out. Although it was normal to leave the manual valve

open in the dry cargo condition, this would not be a desirable course with

regard to effective maintenance of insulation of the forepeak.

13.20 In conjunction with the forward spaces flooding monitoring system

recommended as desirable it is equally important that those on the bridge

should be able to take emergency pumping measures without having to go on

deck to the foredeck in what could be very dangerous conditions.


195

In conjunction with its actions in relation to facilities for the monitoring from

the bridge of forward space flooding the DETR should encourage the IMO to

investigate a pumping system to be recommended for dealing with forward

space flooding and which is sufficiently independent of the vessel’s main

pumping facilities and can be operated from the bridge. Such pumps ought to

be specially designed to accommodate a substantial solids component in the

water to be abstracted because of the possibility of there being floating stores

materials. The pumps should also be able to run dry without damage.

Design Features and Accuracy of Construction

13.21 In paragraphs 8.26 – 8.28 above this Report has already recommended that

IACS should now consider requiring all its members to introduce design

approval and survey procedures for new buildings and for vessels in service

which would implement standards similar to those set out in the Lloyd’s

Register Ship Register Fatigue Design Assessment procedure. This was with

reference to the designed discontinuity involved in the cruciform configuration

in way of Frame 65 on the DERBYSHIRE and to the constructional

misalignment involved in that discontinuity.

Availability of Plans

13.22 In the course of the evidence doubts rose as to whether certain details shown

on the vessel’s builders’ plans had been incorporated by the yard or had

subsequently been altered. Particularly where a vessel is totally lost, it may be

of considerable importance to have exact knowledge of the as-built and any

altered details of construction. For this reason it is important that accurate as-

built plans and plans showing subsequent alterations should be available. It is

clearly sensible that there should be a set of such plans both on board and


196

ashore. IACS should recommend that its members require the owners of

vessels entered with them to maintain such plans.

Accident Database

13.23 A particular feature of the evidence in this Investigation has been the difficulty

of ascertaining information as to comparable losses or incidents involving

similar vessels. If there were an international database on which were

required to be recorded all serious incidents affecting the safety of ships, for

example destruction of forward hatch covers, that might be of very

considerable assistance in monitoring loss and damage risks.

13.24 The DETR should seek to persuade IMO to examine how such a database

might be set up and administered and how and from what sources data might

be provided.

13.25 The sources of such data might include the existing Marine Accident

Reporting System ( MARS). This is essentially directed to collisions and

near-misses, but it could perhaps be extended to include storm damage

incidents. IMO should investigate whether this is feasible.

Recording Evidence of the Cause of Losses

13.26 For many years aircrafts have been required to carry black boxes which can

record information relevant to the cause of crashes and which can normally

survive the severe impacts involved in a crash. In 1988 the IMO adopted

amendments to the International Convention for the Safety of Life at Sea 1974

which required that all merchant vessels of over 300 tonnes should carry float

free Emergency Position Indicating Radio Beacons (“EPIRBS”) from 1st

August 1993 and that all passenger and cargo vessels should have an

automated global positioning facility, the Global Maritime Distress and Safety


197

System, which allows an automated distress message to be sent from ship to

shore by simply pushing a button. The insulation of the latter was to be

phased in during 1992-93. The EPIRB, which was made mandatory on UK

registered fishing vessels of over 12 metres in 1988, is designed to float free

when a vessel sinks. It transmits the signal which can be picked up orbiting

satellites, search vessels and aircrafts. These devices are, however, subject to

movement through wind and wave action. Neither of these systems existed in

1980. Had either been fitted on the DERBYSHIRE, location of the wreck, but

not rescue on board would have been much easier.

13.27 Apart from these facilities, ships do not carry any loss monitoring device

equivalent of an aircraft’s black box. The IMO, however, requires that black

boxes, known as voyage data recorders (“VDR”), should be fitted to passenger

ships and to new building cargo vessels to be phased in during 2002 to 2004.

Such devices are the direct result of a research programme conducted in 1981-

1987 by Lloyd’s Register in association with the Department of Transport. A

fully working VDR had been built by Lloyd’s Register by 1988 but it took 11

years for IMO to adopt the facility and it is not yet scheduled to consider

whether a VDR should be made mandatory for existing cargo vessels.

13.28 Such a device will record conversations on and from the bridge for the last 2

hours before the vessel’s electrical position shuts off. It would record the

vessel’s position, its trim, draughts, speeds, heading and rudder angles during

the period of 12 hours before the loss of power. All alarms would also be

recorded. Further, the vessel’s port and starboard hull girder bending stress

including the effect of waves passing the vessel would be recorded. This

would be done by means of port and starboard strain gauges. Also and

importantly, accelerometers in the forepeak area would measure the velocity

or acceleration due to pitching and heave accelerometers amidships would

measure roll motion. All this data would be recorded by the VDR.


198

13.29 Quite clearly had such a device been fitted on the DERBYSHIRE a very great

deal of data of vital relevance to the cause of the loss would have been

preserved provided that the device had been located. It is quite probable that,

had that data been available, there would have been no need to conduct the

underwater survey.

13.30 This report concludes that the availability of data of the kind recorded by a

VDR (black box) makes an extremely valuable contribution to future ship

safety. This is no less important for the well-being of the crews of cargo

vessels than for that of passengers on passenger vessels. It is therefore

recommended that the DETR should press IMO to require the fitting of these

devices on all existing cargo vessels and new buildings.


199

Section 14

Summary of Recommendations towards Improved

Ship Safety

Hatch Cover Strength and Permissible Freeboard

14.1 IACS, LRS and DETR should agree a specification for further model tests to

be conducted at MARIN designed to ascertain (i) to what extent the

International Load Line Convention needs to be amended with regard to the

minimum permissible hatch cover strength and freeboard of Panamax and

Capesize bulk carriers and (ii) whether the classification societies’ rule UR

S21 provides for adequate minimum hatch cover strength of such vessels.

14.2 Such model tests should be carried out and the data derived from them should

be analysed as soon as practicable.

14.3 Following the evaluation of the model tests data and analyses IACS should

introduce a new formula governing the relationship between minimum hatch

cover strength and minimum permissible freeboard amending UR S21 as far

as may be necessary.

14.4 As soon as practicable and independently of the amendment of the

International Load Line Convention IACS should make that new hatch cover

strength rule applicable both to new buildings and with retrospective effect to

existing vessels whose hatches permitted freeboard do not already comply

with its requirements.

SECTION 14: SUMMARY OF RECOMMENDATIONS TOWARDS IMPROVED SHIP SAFETY

200

Navigational Matters

14.5 The IMO should require the compulsory daily reporting of the position of all

vessels.

14.6 Mariners’ Handbook NP100 should be amended as regards navigation in the

dangerous semi-circle of a tropical revolving storm to include the possibility

of running with the wind on the port quarter in certain cases.

14.7 Masters of bulk carriers should be made fully aware of the possible dangerous

consequences of water entry into forward spaces and consequent reduction of

freeboard.

14.8 Increased participation by vessels in the World Meteorological Organisation’s

Voluntary Observing Ships Scheme should be encouraged by a British M

Notice and an IMO circular.

14.9 Weather routeing agencies should make clear to masters the precise

circumstances in which positive routeing advice will be given to the vessel in

the course of the voyage.

Additional Matters

14.10 IACS should set up a research programme with a view to establishing

minimum strength requirements for and, if necessary, location and protection

requirements for ventilators and air pipe fittings on deck.

14.11 There should be electronic indication on the bridge that a ventilator or air pipe

is open or damaged in such a way as to be open.

14.12 IACS should set up a research programme to develop a new kind of locking

device for stores hatch lids to maintain watertightness in heavy seas with a

view to introducing rules requiring installation of such devices.


201

14.13 Displacement of the foredeck hatch lid should be electronically indicated on

the bridge.

14.14 The DETR should encourage the IMO to develop and introduce guidelines to

the effect that (i) hatch cover operating manuals carried on board should

contain clear and unambiguous instructions as to how and to what extent hatch

covers should be secured; (ii) there should be electronic indication on the

bridge that all cleats and other devices for securing hatch covers are

effectively in place; (iii) otherwise automated closing and securing devices for

hatch covers should be installed.

14.15 IACS and IMO should investigate alternative means of sealing up spurling

pipes to prevent water entry.

14.16 IACS should introduce a rule requiring the use of such means and that chain

locker access should be by bolted manholes and not doors.

14.17 IACS should review the methods of securing deck fittings such as the

windlass, to the foredeck with a view to introducing by rule requirements for

minimum securing strength.

14.18 IMO and IACS should consider the installation of powerful lighting and

industrial video cameras on the foredeck of all Capesize bulk carriers.

14.19 The DETR should take steps to persuade the IMO to make mandatory the

installation on all Capesize bulk carriers a monitoring system for

communicating to bridge electronically the level of bilges in the forward

spaces, including an audio alarm.

14.20 The DETR should encourage the IMO to investigate a pumping system to be

recommended for dealing with forward space flooding which is sufficiently

independent of the vessel’s main pumping facilities and can be operated from

the bridge, such pumps to be capable of accommodating a substantial solids

component and of running dry without damage.

SECTION 14: SUMMARY OF RECOMMENDATIONS TOWARDS IMPROVED SHIP SAFETY

202

14.21 IACS should consider requiring all its member classification societies to

introduce design approval and survey procedures for new buildings and for

vessels in service which would implement standards similar to those set out in

the Lloyd’s Register Ship Right Fatigue Design Assessment procedure,

particularly with regard to discouraging designed discontinuities and

constructional misalignment.

14.22 IACS should recommend that its members should require shipowners to

maintain on board and ashore accurate and up to date as built construction

plans and plans showing subsequent structural alterations.

14.23 The DETR should seek to persuade the IMO to examine how a marine

accident database could be established sufficiently comprehensive to record

storm damage incidents.

14.24 The DETR should press IMO to require the fitting of a VDR (voyage data

recorder) system (black box) on all existing cargo vessels and new buildings.


203

Section 15

Availability of Underwater Survey Capability

15.1 The whole basis of this Re-opened Formal Investigation has been the

underwater survey conducted in March/April 1997 by the Woods Hole

Oceanographic Institution Deep Submergence Laboratory of the United States

by means of the research vessel Thomas G Thompson operated by the

University of Washington. That was organised following the preliminary

underwater investigation which in July 1996 had located the stern section in

the wreckage field. These surveys were conducted 16 years after the loss and

8/9 years after the Original Formal Investigation – a costly and largely

inconclusive exercise which involved a hearing lasting 46 days. This present

Report would be failing in its public purpose if it did not provide an answer to

the question why it took so long to put in place an underwater survey

sufficiently informative to reach a firm conclusion as to the cause of loss.

15.2 At the time of the loss of the DERBYSHIRE in 1980 the Department of

Transport had no developed policy as to whether there should be underwater

surveys of wrecks. Occasionally, however, the Royal Navy searched for new

wrecks, such as that of the mfv GAUL, using hull-mounted sonar on an

opportunity basis, and, when a wreck had been located, the Department might,

if appropriate, undertake a diving survey. Such surveys could only be

conducted in very shallow waters. Although remotely operated vehicles

(“ROVs”) existed at that time, most of them had a depth limit of 500 metres,

as well as limited endurance, power, lighting and camera capabilities. Some

submersibles operated by Soviet and Canadian organisations could work as

deep as 2200 metres and the United States Navy had developed a manned

submersible with a capability of 4500 metres. This was primarily for naval

SECTION 15: AVAILABILITY OF UNDERWATER SURVEY CAPABILITY

204

and intelligence operations and would not have been available for commercial

work. Seabed surveys using manned submersibles and bathyscaphes had been

jointly developed by the United States and France in the 1970’s but had

serious limitations relating to cameras and image transmission to the surface.

15.3 It can thus be concluded that at the time of the loss the technology did not

exist to conduct an underwater survey of the wreck area.

15.4 In 1982 Woods Hole established the Deep Submergence Laboratory and the

JASON programme began. This was directed to the development of remotely

operated equipment capable of undertaking multi-scale search and survey of

the ocean floor using sonar and photographic methods accompanied by a high

level of navigational accuracy. Thus the ARGO suspended camera sledge

operating above the ocean floor was first operated in 1984 and was used to

locate the Titanic in 1985 at a depth approximately the same as that of the

DERBYSHIRE. The prototype of the JASON ROV was used on the Titanic

in 1986. The quality of photographic coverage was very limited since

electronic still imaging and mosaicking had not yet been developed.

15.5 Some 3 months before the commencement of the Original Formal

Investigation a request for an underwater search was made by Ingledew

Brown Bennison and Garrett, solicitors acting for the DERBYSHIRE families,

suggesting the use of the Royal Navy’s vessel “Challenger” which was

capable of operating at 5280 feet. This request was put forward under the

mistaken belief that the wreck lay at 3300 feet. Nevertheless, the Department

took advice from the Chief Surveyor and from Captain de Coverly. It was

decided that such a survey could not be carried out without the development of

new suitable equipment and because of the impracticability of such an

operation, given the uncertainty as to the position of the wreck as well as the

costs involved and the imminence of the hearing in October that year. It must

be appreciated that the means of conducting an accurately detailed survey had

not yet been developed, in particular the accurate positioning of the survey


205

vessel and the ROV and the transmission of images of high quality by fibre

optic cable from the ocean floor with great clarity could not yet be achieved.

In fact, in August 1988 the deepest ship – fibre optic cable vehicle systems test

yet conducted had been at a depth of 3000 m.

15.6 Following the publication of the Original Formal Investigation Report there

were further requests on behalf of the DERBYSHIRE families for an

underwater survey but the Department concluded that the difficulties of

locating the position of the wreck, of carrying out accurate and sufficiently

clear photography and the probable inconclusiveness as to the cause of the loss

of any images obtained, as well as the cost, made it not worthwhile to attempt

such a survey. This view was entirely justifiable.

15.7 Meanwhile, in 1989 the DSL 120 deep towed side scan sonar, which was to

play such an important part in the 1997-8 DERBYSHIRE survey, was tested

by Woods Hole for the first time. The development of mosaicking of

electronic still camera images also began that year with the first operational

use of an Electronic Still Camera. The DSL 120 was first used operationally

in 1991. The image collation and retrievable system known as VISUAL

began development in 1991.

15.8 By 1993-4 the Department continued to entertain the view that, although it

would be technically possible to film the wreck, if it could be located, any

information that might be derived from an underwater survey of the wreck of

the DERBYSHIRE would be so limited that it would not shed further light on

the cause of the loss. This was not surprising. Apart from the major problem

of initially locating the wreckage, the effect of the impact of the hull on the

seabed was thought likely to be such as to make it impossible to tell from the

nature and quality of film that might be obtained whether, if it had fractured at

frame 65 (which was the major matter of interest) that had occurred prior to

sinking. It was for those reasons that the Department declined requests in


206

1993 and 1994 for funding for a proposed RMT/ITF underwater search for the

wreck.

15.9 The ITF survey which on 8th June 1994 located the bow section of the

DERBYSHIRE was conducted by the United States firm Oceaneering using a

deep towed side scan sonar similar in capability to the DSL 120 and a ROV.

As already described, that in turn led to Lord Donaldson’s independent

Assessment in 1995 and, in accordance with his recommendations, to the two

phases of the UK/EC Assessors’ Survey in 1997 and 1998. This has been the

most comprehensive and sustained deep seabed survey of a merchant ship

wreck ever conducted.

15.10 It is to be concluded that, although it would have been technically possible to

carry out an underwater survey as comprehensive as that eventually conducted

in 1998 some year or two before that time, the development of the equipment

necessary to obtain photographs and film of adequate detail and clarity would

not have been sufficiently advanced to do so at any earlier stage. In the event

the recommendations made in Lord Donaldson’s Assessment have been amply

born out. The most impressive photographic and video material provided by

the Woods Hole equipment has, together with the data from the MARIN

model tests, made it possible at last to resolve with reasonable confidence the

crucial question of how and why the DERBYSHIRE was lost.

15.11 Following the successful completion of the DERBYSHIRE survey, the United

States National Science Foundation and the DETR are co-operating in the

development by Woods Hole of a new generation deep sea ROV. This would

have a 6000 metre depth and a 1500 metre continental shelf capability.

15.12 The availability of the underwater survey equipment so far developed has

brought about a change in the Department’s approach to providing underwater

surveys of wrecks. This is reflected in the conduct in 1998 of the underwater

survey of the trawler Gaul at a depth 300 metres in the Barents Sea, which has


207

led to the re-opening of the Formal Investigation into the loss of that vessel. It

is also reflected in the Department’s responses to the recommendations of Mr

Roger Clarke in his recent report on the Gaul entitled “Why was no search

made for the Wreck?” Amongst these recommendations and responses were

the following:

15.13 In respect of the Marine Accident Investigation Branch’s access to the

equipment needed for examining a wreck, the recommendation was:

“R15 DETR should evaluate alternative approaches for giving MAIB

access to the appropriate equipment with a view to deciding as soon as

possible which of them, or which combination of them, will provide

MAIB with the most cost-effective means of reacting to a requirement

for an underwater survey.”

15.14 The Department’s response was:

“Accepted. The Department is following up on the excellent co-

operation with the UNSAF [and the] Woods Hole Oceanographic

Institution (WHOI) in the underwater survey of the MV

DERBYSHIRE. The Government has a research agreement to help

WHOI develop a new under water remotely operated vehicle (ROV)

that will have an enhanced wreck survey capability. The Government

intends to use these developments to enhance the UK’s Wreck survey

capability within the next few years.”

15.15 As to the involvement of the relatives of those lost at sea in any decision

whether to conduct an underwater survey, the recommendation was:

“R17 In future, before any decision is taken not to search for a

wreck or to investigate it when found, officials and ministers should

explicitly take into consideration the wishes and interests of bereaved


208

families, and any danger that their inaction will lead to their grief being

aggravated by rumour and prolonged uncertainty.”

15.16 The Department’s response was:

“The MAIB now regularly uses ROVs to examine wrecks underwater.

The Department is committed to providing MAIB with the equipment

both to locate wrecks and then to examine them. Decisions on the

conduct of marine accident investigations are for the Chief Inspector of

Marine Accidents reporting on such matters directly to the Secretary of

State. MAIB seeks to brief the relatives of those lost in shipping

accidents on the progress on marine accident investigations.”

15.17 This Report entirely endorses these responses. It is to be emphasised that the

Department’s approach to the first of these recommendations depends, as does

its approach to the investigation, by means of the continuing tests at MARIN,

of the adequacy of presently prescribed minimum hatch cover strengths, on the

continuing availability of sufficient public funding. This Report concludes

that in these two respects the Department is performing an international

service of immense value to the enhancement of ship safety and strongly

endorses the continuance of public funding sufficient for these purposes.


209

Section 16

Answers to Questions

16.1 What were the most probable cause(s) of the loss of the

“DERBYSHIRE”?

- Following weather forecasts which proved to have under-estimated

the severity of the sea conditions to be anticipated on the vessel’s

projected route it came into very close proximity with the worst of

Typhoon Orchid. This involved shipping on the foredeck for a

prolonged period on the 9th September 1980 green water of

sufficient head to destroy or damage ventilators and air pipes

leading to the bosun’s store, machinery space and probably the

ballast tank and to destroy the cement seals to the spurling pipes.

In consequence substantial quantities of seawater entered the

bosun’s store and machinery space, the chain lockers and probably

the ballast tank in the forepeak. This entry of seawater had by

about 1700z caused the vessel to develop a trim by the bow with

the consequence that, at or shortly after that time, the vessel

encountered a wave or waves of such height that the head of water

imposed on the No.1 hatch cover panels a loading in excess of their

collapse strength. Seawater then entered and began to fill the very

substantial ullage space above the cargo in No.1 hold, thereby

increasing the vessel’s trim by the bow and exposing the No.2

hatch-covers to a much increased risk of green water loading in

excess of their collapse strength. Those hatch covers, having been

subjected to such loading, then collapsed, admitting seawater to

and filling No.2 hold and so yet further increasing the trim by the

SECTION 16: ANSWERS TO QUESTIONS

210

bow. The same process was then repeated in relation to the No.3

hold and the vessel could then no longer remain afloat.

- It is possible, but considerably less likely, that excessive green

water loading destroyed the No.1 hatch covers before substantial

quantities of seawater entered the spaces under the foredeck. The

successive failure of hatch covers would then follow as described.

16.2 What possible causes (previously considered by the Formal Investigation

or Lord Donaldson’s Assessment [see a list of his 14 scenarios at

Appendix 47] can be eliminated in the light of the new evidence of the

wreckage of the “DERBYSHIRE”?

- All other possible causes

16.3 What other possible causes of the loss of the “DERBYSHIRE” remain

open?

- None.

16.4 In so far as material to the loss of the “DERBYSHIRE”, was the design of

the “DERBYSHIRE” in way of her fore-end (from frame 339 forward –

including her hull, bow height, deck, deck openings and fittings) in

accordance with the standards applicable at the time she was built?

- Yes

16.5 Is that design satisfactory in the light of what is now known?

- No. The Structural strength of air pipes and particularly mushroom

ventilators was insufficient given the low bow height, to withstand

green water loading to the extent and for the period experienced on

9th September in severe sea conditions, but not exceptionally

severe for a North West Pacific typhoon.


211

16.6 In so far as material to the loss of the “DERBYSHIRE”, were any repairs

carried out to the DERBYSHIRE in way of her fore-end? If so;

(a) what repairs were carried out?

(b) was her condition satisfactory after such repairs had been carried

out?

- No: (a) and (b) are inapplicable.

16.7 In so far as material to the loss of the “DERBYSHIRE” was the design of

the hatch covers of the “DERBYSHIRE” in accordance with the

standards applicable at the time she was built?

- Yes.

16.8 Is that design satisfactory in the light of what is now known?

- No: seriously deficient. The design strength was inadequate to

withstand the green water loads likely to be sustained during

extreme sea conditions which a vessel might encounter during her

service life.

16.9 At the time when the vessel was:

(a) designed; and

(b) built;

were the regulations and classification society rules for:

(i) assignment of freeboard;

(ii) design of her fore-end (from frame 339 forward – including her

hull, bow height, deck, deck openings, and fittings); and

(iii) design of her hatch covers


212

inadequate in any respect material to the loss in the light of the then

current state of knowledge and what ought reasonably then to have been

known or anticipated?

- No

16.10 When the “DERBYSHIRE” sailed on her last voyage from Sept-Isle was

she in all respects seaworthy for her contemplated voyage to Japan in so

far as material to her loss?

- In the sense that she was not fit to withstand an encounter with a

North West Pacific Typhoon of not unusual intensity because of

the insufficient strength of the air pipes and ventilators on the

foredeck and of the forward hatch covers, she was not objectively

seaworthy. She was, however, reasonably believed to be

seaworthy by her owners, classification society and officers and

crew, given the current state of knowledge in the shipping industry

in relation to the possible incidence and effects of green water

loading on air pipes and ventilators and forward hatch covers.

16.11 Did she cease to be seaworthy in any respect material to her loss at any

time prior to her loss, and, if so, in what respects, where and when?

- No, not prior to the time when she began to sustain damage to her

air pipes and ventilators after encountering the typhoon.

16.12

(a) Is there any evidence of defective design, construction or repair of the

structure in way of frame 65 in the wreckage of the “DERBYSHIRE”

which would materially impair the safety of the vessel or those on board

her?

- Yes.


213

- As to design, there was a discontinuity at the cruciform

configuration where the hatch side girder butted on to Frame 65

and was lined up with the longitudinal cofferdam bulkhead aft of

the bulkhead 65.

- As to construction, there was a misalignment of not less than

35mm between the starboard saddle tank vertical hatch side girder

and the engine room outboard longitudinal in way of Frame 65.

(b) If so, did any such defect materially impair the safety of the vessel or

those on board her?

- Yes, but only to a very small extent.

(c) Are the present –day classification society rules and instructions to

surveyors adequate as regards the quality of design, construction and

repairs.

- Yes, as regards Lloyd’s Register. Not known as regards other

classification societies.

16.13

(a) Was the information and advice provided to the “DERBYSHIRE” by

Oceanroutes Inc. adequate and appropriate in the circumstances?

- No. Oceanroutes failed to include in its message timed at 0113z on

8th September 1980 the current position of the storm centre or its

forecast track to the 48 hours forecast position or the forecast wind

speed during the succeeding 24 hours. The information and advice

given by Oceanroutes was not inadequate or inappropriate by

reason of not having advised the vessel to alter course to the west

or south west or not having given more information than that

contained in the Guam and Tokyo weather reports.


214

(b) If not, then did the inadequacy or inappropriateness of such advice and

information cause or contribute to the “DERBYSHIRE’s” close

encounter with typhoon “Orchid”.

- Probably not.

(c) Did the master navigate appropriately in the light of the weather

information available to him?.

- Yes.

16.14

(a) What steps should be taken to avoid a similar loss in future?

- This Report contains 24 recommendations which are summarised

in Section 15.

(b) Should the current regulations or current classification society rules or

instructions to surveyors be amended to avoid a similar loss in the future?

- Yes. The International Load Line Convention ought to be

amended to provide for increased strength for forward hatch covers

and for deck fittings located on the foredeck and in way of the

forward hatches. Current classification society rules as to forward

hatch cover strength ought to be amended to apply to new building

and existing bulk carriers in a manner to be ascertained from the

programme of model testing at MARIN which has just been

completed. As regards the structural strength of deck fittings a

programme of further research should be set up with a view to

introducing into classification society rules minimum strength

requirements for such fittings as soon as possible.


Appendices 1-47

Report

Documents

reopened formal investigation

mv derbyshire parties

treasury solicitor solicitor

mr rf stone qc mr simon

mv derbyshire bibby

improved ship safety

messrs mills

summary of recommendations