'TAIL-END OF A WILLY-WILLY.'

The West Australian, 4 April, 1912

THE STORM AT PORT HEDLAND.

CONDITIONS OH MARCH 20.

CAPTAIN ALLEN'S FEARS.

Port Hedland, April 3.

As if to palliate the tedium of the scorching, 

enervating heat Port Hedland was visited 

on March 20 with the tail-end of a

willy-willy which coming with equinoctial

tides, did a fair amount of damage. It was

fairly certain that had it been a full willy-

willy it would have spelt disaster to the low

lands of Hedland. 

Confirming that Port Hedland was only subjected to the outer fringe of the cyclone - centre 50+ miles away in the vicinity of Balla Balla. Note that the excessive heat was a precursor to cyclonic activity. 

It started on Tuesday night with a strong 

easterly wind, and the waves thundered on 

the beach with indications of heavier and 

deadlier surges out at sea. 

It simply cannot be argued that those at Port Hedland were unaware of a cyclonic system in the region prior to 20 March (Wednesday).

Wednesday morning saw about forty

luggers running to shelter off Port Hedland,

where they were soon safely anchored. These

luggers had a rough time on Tuesday night

near Turtle Island, and the cutting away of

masts in some cases was seriously 

contemplated. The wind increased its velocity, 

and it continued for three days and nights. The

waves surged up the harbour's entrance, and

the breakers crashed into the sandy Esplanade, 

and washed thousands of tons of it away.

Confirmation of physical warnings.

Captain Allen, of the Koombana, seemed

disinclined to go out, and when he decided

to do so he said, "I am going straight out to

sea, and will be lucky if I get to Broome on

Saturday." She left the port at half-past

10 o'clock on March 20 with propeller 

beating out of the water. When passing 

the entrance to the harbour she rolled 

deeply. 

Consistent with the ground swell out at sea and very light condition.

Captain Allen stated that he would fill the 

ballast tanks when he crossed the outer bank.

The ship headed due north, and was followed 

about an hour afterwards by the Bullarra

going south.

It took more than 2 hours to 'fill' the ballast tanks. In this report Bullarra followed an hour later (incorrect).

The damage done to the railway line on

the Causeway will cost a good deal to repair. 

The Esplanade suffered most, being washed 

away to the allotment boundaries in several 

places; in fact, the Esplanade has disappeared, 

and in its place we now have a sloping sandy 

beach from the rear of the Resident Magistrate's 

residence right round to the jetty. The light tower 

and oil house were undermined, while the roadway 

connected with the town bridge was rendered

unsafe for vehicular traffic. One lugger in port 

anchored too far in the stream, and drifted onto 

rocks. She was only slightly damaged. The 

schooner Alto was also damaged at the jetty.

Damage sustained at Port Hedland related to the storm surge and rain rather than hurricane-force winds.

It was a left-handed circular cyclone, and

from the directions of the wind it is certain

that the Koombana could not possibly have

avoided getting right into the centre. 

The centre of the cyclone was to the southwest of Port Hedland, whereas Koombana was headed northeast. This is not saying much about her 3,000 ihp!

The Koombana would be abeam Bedout at 4

p.m. on Wednesday, but out of course to the

south. 

Speculation: meaning Koombana would have been averaging 11.75 knots. Captain Allen had said that it was unlikely he would make Broome by the spring tide the following day, implying that he was anticipating less than 10.5 knots (strong headwind) and delays filling tanks. Out of course to the south of Bedout Island is strange - if anything, with a prevailing NE gale, out of course to the west makes more sense.

The wind was then blowing from

the east, but at 10 p.m. it changed from 

east to east north-east, and back in quick 

succession. At 10 a.m. on Thursday it was 

still blowing from the east, at noon east-north-east, 

and at 1 p.m. in the same direction. 

Confirmation that the cyclone stalled when it hit land (Balla Balla), but did not lose intensity.

It shifted at 2 p.m. to north-east, and 

remained in that direction till 6 p.m.. At 

10 p.m. it was blowing north-north-east, 

and at midnight it was north. 

Finally the system moved southeastward inland towards Marble Bar.

When Hedland got the worst of the hurricane 

on the 22nd, the wind was from the north-north-west 

from 4 a.m. till noon.

Three days later, and when Koombana was long gone to a watery grave.

The general congratulations possessing

people's minds on having escaped the full

force of the willy-willy in Hedland on top

of the equinoctial tides gave place to feelings

of horror when the tidings came that the

Koombana and all aboard were lost.

There must surely have been those who wondered why Koombana foundered steaming steadily away from the cyclone rather than into it, and, with this in mind, why the aging Bullarra escaped intact after steaming directly into the heart of the cyclone?

Courtesy Trove.

AIRLIE ISLAND BOTTLE MESSAGE

The West Australian, 26 July, 1913.

Captain Winzar, acting chief harbourmaster 

at Fremantle, has in his possession a piece 

of paper supposed to contain a message 

from the sea. This was found in a bottle on 

the beach at Airlie Island by Captain Williamson, 

of the Government steamer Una. during his

recent trip to the Nor'-West, for the purpose 

of establishing some new lights along

the coast. While anchored off the islands

a party went ashore, and in the course of

their wanderings along the seashore came

across a pickle bottle, which on being 

opened, was found to contain a water-sodden

piece of blotting paper. On the latter could be 

seen some writing, and with the aid of a 

microscope Captain Williamson made out 

the following message :-Vessel hove to. . . 

but can't live through it. I think it is no good. 

Good-bye. J. Shannon. As Airlie Island is 

only about 200 miles from Bedout Island, off 

which the Koombana was wrecked last year, 

it was thought at first that it might have been

a message from one of her ill-fated company, 

but no name answering to the one on the message 

could be found in her passenger or crew list. 

Shipping people at the Port are inclined to believe 

that it is only one of the many hoaxes that some 

people are so fond of playing on the public.

Cruel hoax.

courtesy Google Earth

the bottle message would have to have negotiated these islands to arrive at Airlie Island - unlikely
courtesy Google Earth

4000 TONS OF CARGO.

"On a draught of 19 ft. she

can carry 4000 tons of cargo."

This figure of 4000 tons was widely quoted in newspapers. But it is incorrect. Koombana's realistic cargo carrying capacity was in the region of 1800 tons - 1000 tons general cargo and 800 tons refrigerated cargo, stowed 100 cubic feet to the ton. With ballast tanks accounting for 900 tons of her overall volume 4000 tons was a stretch of someone's imagination. What is interesting, however, is the draught of 19 ft., which is the figure (aft) the Inquiry thought they could get away with for Koombana departing Port Hedland with a paltry 260 tons of cargo and empty ballast tanks!

To confirm the correction we must return to a comparison with the SS Yongala:

                        Koombana (1908)                   Yongala (1903)

gross tons            3668                                             3664

net tons                2182                                             1825

length                   340 ft.                                           350 ft.

beam                    48 ft. 2 in.                                     45.2 ft.

depth (of hold)      25 ft. 8 in.                                     27.2 ft.

hull depth              29.5 ft.                                          30.5 ft.

draught                 20 ft. 8 in.                                      24 ft.

cargo                    1800 tons                                     1800 tons

What becomes apparent are the marked similarities between the two steamers, particularly with regard to specifications and prominent top hampers. However, Yongala's narrower beam would likely have improved stability marginally compared with Koombana's slightly broader beam and flat bottom. It is patently clear that Koombana operated with a significantly reduced draught (related to Northwest coastal ports access); a differential of 3.25 ft.. This would have had a significantly negative impact on GM stability and although both steamers were visually top heavy, Koombana took the prize.

SS Koombana (courtesy wikipedia)

SS Yongala (courtesy Michael McFadyen's scuba diving website)

courtesy Trove.

LUGGERS, 20 MARCH

"Wednesday, March 20th.—Big sea with N.E. wind, strong; 11 a.m., hove up and made for the Creek under double-reefed mainsail and jib; just before heaving up big sea caught Voladora taking away dinghy, dress, starboard rail and bulwarks;... went out to Mt. Blaze at low water ..."

Boyd, Annie. Koombana Days (p. 244). Fremantle Press. Kindle Edition.

Mount Blaze is 65 n miles from Port Hedland on the OTHER side of where the cyclone was preparing its onslaught on Balla Balla, 20 March. Allegedly, in this case, the wind was NE strong, with big sea, which matches Captain Upjohn's claim of a NE gale, exiting Port Hedland, the morning of 20 March. This extract suggests that although Koombana was making steady progress out of the immediate sphere of influence of the cyclone, she would ultimately have had big seas and strong winds on her stern and being in very light condition at night would have been a terrible combination, with the very real threat of losing steering control and broaching in a trough; 

a horrifying possibility.  

courtesy Google Earth

SUMMARY.

R.M.S. KOOMBANA

courtesy Trove

courtesy State Library of Western Australia.

Koombana was a magnificent, bespoke steamer with opulently appointed, well-ventilated, elevated passenger decks, cattle deck, and dedicated cargo holds (incl. refrigeration). But she was flawed; inherently top heavy, with a reduced draught of 20 ft. 8 in. to access harbours such as Port Hedland.

The indication of inherent top heaviness is illustrated by comparison with the similarly sized steamers Yongala and Grantala (greater draughts: 24 ft.) of the same Adelaide Steamship Company. The latter steamers, with significant top hampers, operated with ballast tank capacity of 400 tons, whereas Koombana required 900 tons to steady her.

Yongala and Grantala       3664 tons
Koombana                        3668 tons

Koombana's track record, fully ballasted, gives us the most significant confirmatory clue regarding inherent top heaviness. 8 February, 1910, Koombana allegedly heeled over to 45 degrees in a squall and took a considerable time to recover.

20 March, 1912:

All the signs indicated that Koombana and Bullarra would be departing Port Hedland before noon into unpredictable and potentially dangerous conditions. A period advisory referred to an easterly wind and falling barometer being the warning signs for pearlers to seek shelter. 'At 6 a.m. the barometer had been at 29.50', down from 29.9. Between 4 a.m. and 10 a.m., in normal conditions the barometer readings rise, not fall. The wind, a gale by noon, came from the NE. (some sources claim a half gale from ENE).

'Two or three layers of swiftly racing clouds could frequently be seen through the prevailing mist. It blew with terrific force.'

40 luggers sought shelter in the creek, bringing tidings of a heavy ground swell from the northwest, which in turn, signaled the presence of a significant cyclonic event out at sea.

'Vessels engaged in the pearl-shell fisheries, on the northwestern coast of Australia, run on shore in any available creek, on indications of the approach of a cyclone.' 

Alleged:

"I don't like the glass," was Captain Allen's remark,

But Captain Allen was compelled to depart at 10.20 am, due to pressure of maintaining a tide-dependent schedule; mail contract; cargo delivery and passenger demands. 

Koombana departed Port Hedland carrying:

- 260 tons cargo

- 480 tons coal

- 60 tons stores

= 800 tons

(plus an undisclosed tonnage of residual fresh water in her tanks)

157 souls on board.

260 tons of cargo was only 14.4 % of maximum load - 1800 tons (not 4000 tons as is commonly reported). 85 tons for Broome were stowed in 'tween decks 1 and 2, higher up in hull, raising centre of gravity, reducing stability and 175 tons for Derby in hold 2, low down, improving stability. These units of cargo were loaded and secured at Fremantle for the voyage.

I do not believe there was a significant component of livestock on the cattle deck due to steaming up the coast. Livestock was destined for the markets at Fremantle and required fresh water and fodder at sea. Moreover, Bullarra had been reintroduced on the run that month to bring livestock back from Port Hedland, relieving Koombana of this function. 

Drawing 16 ft. aft and 11 ft. forward (rather than the claimed 19 ft. aft 12 ft. forward), all ballast tanks empty, Koombana's propeller  was not fully submerged (decreased thrust and manoeuvrability) and she displayed 'an ugly list to port' confirming her very light lading and ballast status. 

The depth over the bar was 19 ft. at the time which means Koombana would not have cleared without incident, given the Inquiry's quoted 19 ft. aft figure. Furthermore, a dead weight of 800 tons (cargo in forward holds; empty after-peak and no 6 tank) could not have accounted for a draught 1.75 ft. short of maximum loading, aft.

Why was Koombana in such light, top heavy condition? 

She needed to be!

Captain Allen:

"I might bump that outer bar going out on a sea like this."

"My passengers think they will get to Broome to-morrow (Thursday)," he remarked; "but they will be lucky if they get there by Saturday. I am going to put right out to sea, and as might bump the bar going out I will leave my ballast tanks until I get outside and fill them out there."

It was both risky and laborious to fill tanks at sea, particularly in a heavy ground swell and gale, and it would have taken more than 3 hours (in ideal conditions) to achieve this task and which delayed progress significantly. Two sets of eye witnesses referred to Koombana being in sight for only 2 hours. The tanks were unlikely to have been successfully pressed up during this time frame, which in turn would have created a free surface effect, enhancing the list; reducing what GM there was and impairing heeling recovery.

Why did Captain Allen not take the necessary time to fill all tanks? Quite simply, the Light at Bedout was not functioning and he had to clear the dangerous reef adjacent to the island before dark (6.38 p.m.). 

Captain Allen anticipated and was ultimately confronted by a stiff head wind (gale from NE; ENE) slowing progress. Distance to Broome = 265 n miles. An average of 10 knots (14 knot steamer) or less would miss spring tide access to Broome the following day. Marked pitching would have exposed the propeller periodically, reducing thrust.

Bullarra:

'When we came on deck for the mid-day meal, the Bullara and the Koombana were stern on to each other, and the distance apart being about five miles.'

Koombana had not made much progress by noon due to the attempt to fill tanks.

With a gale and a significant wind catchment factor in the form of towering superstructure and funnel enhancing the existing list, Koombana was presented with challenges from the start. I believe Captain Allen made his decision to pursue the standard course rounding the northern aspect of Bedout Island based on the fact that he could see evidence of the cyclone to the westward on the horizon; not north or northeastward in which directions he was headed; confirmed by a witness account aboard Bullarra heading southwest:

"I noticed away on the starboard bow and 

just above the horizon, a dense mass of 

inky black clouds and from descriptions 

I'd had from old pearlers, I was convinced 

we were in for it."

Therefore, I believe Koombana only ever skirted the outer margin of the cyclone which targeted Balla Balla from the north-northwest, having curved round. The official meteorology report presented at the Inquiry claimed a cyclone diameter of 90 miles, which could not have broached Koombana's intended steamer track. This is further substantiated by evidence of minimal storm impact at Port Hedland and Bedout Island (lighthouse glass not encrusted and no signs of a disturbance).

"Scarcely any damage

was caused by the 

hurricane at Port Hedland."

However, this said, Koombana in her unstable, top heavy condition would have been subjected to broad, deep sea swells; strong winds shifting through NE ---> N ----> NW, further unsettling stability and hindering heeling recovery. Listing heavily, she would have been vulnerable to water ingress via the cattle loading doors or the ventilation ports on the main deck, which would further have contributed to the free surface effect, worsening the list and further diminishing heeling recovery.

Cargo might also have shifted, contributing to the deteriorating situation.

In an attempt to correct the dangerous list, Captain Allen might have elected to pump out one of the longitudinal ballast tank sections on the side of the list, with the resultant catastrophic effect as illustrated by the S.S. Vestris disaster (November, 1928). Captain Carey pumped out tanks in the hope of decreasing a progressive list. Paradoxically, in part due to residual free water in tanks, the list increased dramatically.

Ultimately, when Koombana went over I believe she did so suddenly and violently, not allowing souls successful escape in lifeboats and if a few did, they did not survive to tell the tale.

No cargo or bodies were ever recovered suggesting that she turned turtle trapping both people and cargo within. Absence of bodies and cargo goes against the theory that Koombana was battered to pieces by the cyclone. Damaged pieces of wreckage, ultimately released from the wreck, could have been caused by superstructure striking the seabed or movement of the wreck itself along the seabed. It would also explain why wreckage was only discovered 12 to 14 days after the disaster - rising from the seabed - rather than due to immediate hurricane-force damage.

Insufficient time was allowed for wireless operator Harry Lyon to send out a distress signal via Marconi wireless, if indeed it was still operational. It is important to note that there was a backup wireless unit independent of the ship's electrical supply, but with only a 100 mile range. 

I believe that the oil patch, discovered by Captain Upjohn of the Bullarra and represented on the image below, is Koombana's final resting place. An example of oil marking the spot was used to locate the Clan Ranald wreck, January 1909: 

"direct me to the spot where the wreck was sunk, 

which he said he knew exactly, having seen 

streams of oil rising from her."

This is further borne out by the distribution of the bulk of wreckage, drifting with the predominantly (south)westward trending current which predominates after alongshore cyclones in the Pilbara. This position is within the outer steamer track for Broome. 

Koombana was not off course.

Inquiry (courtesy Koombana Days online site):

"- It was in latitude 19.11 and 119.25 E.

What distance would that be off Bedout Island?

- About 27 or 28 miles - I cannot say which. 

Did you take any samples of this oily substance?

- Yes, two or three dozen bottles.

What depth was there at this place?

- 30 or 35 fathoms. (55 m - 64 m)

Did you see any trace of the vessel at that depth? There would be nothing to indicate that the Koombana or any other vessel would be there?

- It was getting dark and it looked like the outline of a vessel." 

After further analysis, 27.5 miles northeast of Bedout Island coincides with 19 15 51 S, 119 26 48 E (30 fathoms), a position 20 miles from additional wreckage to the west and which, in turn, lay 20 miles to the north of Bedout Island - as referenced by Captain Upjohn.

Although not represented on the graphic below, Captain Upjohn referred to:

They had taken the specimens (of

oily water) produced in Court at 

this spot. An awning spar and one 

of the planks exhibited in Court 

had also been discovered in close

proximity. Other articles were picked

up about 20 miles from that spot.

(marked 2 and 3 on graphic).

1. Stateroom door; painting stage; small pieces of board. SS Gorgon

2. Motor launch starboard bow plank (with insignia). SS Bullarra

3. Small wreckage; (life) boat tanks; lifebelts; panel from saloon / smoke room ceiling. SS Bullarra
(see: https://koombanarevisited.blogspot.com/2019/11/sufficient-warning.html)

4. Bottom board from (life) boat; white painted board. Lugger McLennan. 

5. (life) boat mast and small wreckage (rising from bottom). SS Una

6. Miscellaneous wreckage. SS Una   

7. Miscellaneous wreckage. SS Una. 

(see https://koombanarevisited.blogspot.com/2019/10/captain-rantzau.html)

8. Cabin paneling. Lugger Mina.

9. Smoking room cushion; cabin door. SS Minderoo.

10. Straw envelopes (Leech's fortune). SS Minderoo and SS Gorgon.
(see:  https://koombanarevisited.blogspot.com/2019/11/sufficient-warning.html)

11. Bottom boards (lifeboat); drawer; small teak panel. SS Minderoo.

Official cyclone, 90 miles in diameter, as per Meteorological Office. Note that the cyclone approached the coast from the direction of S.S. Moira's position, 8 a.m..
Courtesy Google Earth.

courtesy Trove

this map which appeared in a period newspaper report gives an idea of the route taken by Koombana, 'shaping a course' around the northern aspect of Bedout Island.
(the 'last seen here' annotation is unconfirmed, but could substantiate Koombana being on course this far)

presumed track Port Hedland to Broome

The residents of Hedland had a final say on the matter:

The ship's propeller was showing

when anchored at the jetty, and

raced out of the water as she sailed

over the rolling seas at the harbor's

entrance ; the boat also rolled heavily

when the wind struck her on the

starboard side—so much so that

several who were watching her exclaimed.

"She'll be over directly " !

And a final, cryptic comment:

Sunday Times, Perth, 31 March, 1912.

Therefore, as she has not been found along the 

Ninety-Mile Beach or in La Grange Bay, we are

led to the terrible alternative that she was engulfed 

somewhere north-east of Bedout Island, where 

she was last spoken.

courtesy ANU archives

In this image of Koombana apparently steaming, the main livestock loading door is open, one assumes due to the excessive heat and humidity along the Nor'West coast. Of concern would have been a potential portal of water ingress during heavy seas and rolling.

courtesy Trove extracts; Australia Pilot, 1920.

For those with an in depth interest in this subject should obtain Annie Boyd's outstanding Koombana Days.