The Climate and Weather of Australia/Chapter 11
XI.(a).—WEST AUSTRALIAN HURRICANES.
Undoubtedly the most striking feature of the climate of Western Australia is the visitation of the north-west coast by hurricanes, locally known as "willy-willies," in the summer months.
These occur in connexion with, or are very violent forms of, cyclonic disturbances which originate over tropical seas, and are practically limited to the summer months—December to April inclusive.
"These storms sometimes give the first indication of their approach in the extreme north-eastern corner of the State, and occasionally it is believed at Port Darwin. They travel at first in a south-westerly direction, the centre keeping well out to sea; and their pressure and movement are shown by an easterly wind gradually veering north and west, accompanied by heavy rainfall. When they reach latitude 20° or thereabouts, their course alters and they recurve and commence to travel in a southerly or south-easterly direction, striking the coast generally between Condon and Fortescue, and frequently bring a "willy-willy" to wreck whatever happens to be in the way. They now travel inland, passing, as a rule, either over or to the east of the gold-fields, and bringing more or less rain according to their intensity. Thence they travel to the Southern Ocean."[1]
The path of a typical willy-willy disturbance is shown in Figs. 91-92.
Figs. 91-92.
Hurricanes in Western Australia.
(Tropical Cyclones or "Willy-Willies.")
| Year. | Month. | Time of Maximum. |
Chief Locality. | Direction of Movement. | Rain. | Remarks, Pressure, Minima, &c. |
| 1872 | No particulars available. | |||||
| 1877 | Feb. | 15 to 18th | Lacepede Island N. of Broome) |
From N. E. | Not tabulated | Barometer, 29.49 Winds veered from E.S.E., E., N., N.E. |
| 1878 | Dec. | 24th | Lacepede Island | Not tabulated„ | Barometer, 29.52. Wind at maximum from N.W. Runnymede lost | |
| 1879 | Jan. | 23rd | Broome | |||
| 1879„ | Jan.„ | 25th, 5 a.m. | Roeburne | Towards S. W. | Not tabulated„ | Barometer at Roeburne, 28.92. Sulina and Manfred lost |
| 1880 | Jan.„ | 9th | Yarnmadarra Creek |
Not tabulated„ | Adalia lost. Tidal wave 25 feet over H.W.M. | |
| 1881 | Jan.„ | 7th | Cossack | Started Cape Carney; moved along coast to Cape Lockyer |
Not Tabulated„ | Barometer, 26.00, on Alpha (doubtful); four pearling boats reported 27.00. Many luggers lost. Wind veered from N.E. to N.W. |
| 1882 | Mar. | 7th, at 9 p.m. | Roeburne | Extended 25 miles E. and 50 miles W. but no distance inland |
6½″ at Cossack | More severe than 1872. Barometer at Cossack, 28.1. Wind from E.N.E. at maximum |
| 1887 | Feb. | 12th | Cossack, 8.6″ | Barometer, 29.176, at Cossack | ||
| 1887„ | Apl. | 22nd | S.W. of Wallal 90- mile Beach |
Over 200 lives lost | ||
| 1888 | Jan. | 7th, 8th | Derby | Wyndham, 4″; Derby, 17″ | Barometer at Derby, 29.134. Wind S.E., E., and later N.W. | |
| 1888„ | Apl. | 22nd | 90-mile Beach | Many pearling boats lost, and 114 men drowned | ||
| 1889 | Mar. | 1st | Roeburne | Travelled from near Broome to near Onslow |
Roeburne, 5.9″; Onslow, 3.4″; Fortescue, 5″ |
A "circular hurricane," i.e., wind veered from S. through N.E. to W. Three boats wrecked near Cossack |
| 1890 | Jan. | 27th | Kimberley | Wyndham, 11.6″; Derby, 5.4″ |
Barometer, 29.357, at Wyndham | |
| 1897 | Dec. | 26th, 10.30 a.m. |
Onslow | Travelled from Onslow to Eucla, 30th |
Port Headland, 5.9″; Condon, 5.3″; Onslow, 2″; To- wara, 6.5″ |
Minimum barometer, 28.57; wrecked Onslow. Terminated a hot spell in Victoria on 31st |
| 1898 | Apl. | 2nd, 5 p.m. | Cossack | From Wyndham 29th; through Derby and Broome, 30th; to Cossack |
Bamboo Creek, 16″; Whim Creek, 36″; Cossack, 15″ |
Wind at Cossack (at maximum of gale) as a hurricane from east, evidently towards deep barometric depressions at sea; s.s. Albany, 27.8 All boats in creek at Cossack, except police boat, wrecked |
| 1899 | Jan. | 12th | Off Wyndham (see map) |
Along coast to Condon, then recurved and traced to Tasmania |
Marble Bar, 7″; Nullagine, 9″; Mt. Margaret, 3″, on 16th |
s.s. Tangier encountered "willy-willy" on 12th off the N.W. coast |
| 1900 | Mar. | 5th, at 3.15 p.m. |
Cossack | Along coast from N.E., passed inland south of Cossack, and then crossed the gold- fields to the Bight |
Wyndham, 3″; Derby, 5″, 3rd and 4th; Marble Bar, 6″, 6th; Cossack, 8″, 6th; Cue, 1½″, 9th; Laverton, 2″, 10th; Whim Creek, 13″, on 5th and 6th |
Minimum barometer, 29.34. Hurricane from S.W., not so disastrous as previous cases. 13″ of rain fell at Whim Creek, near Cossack, on the 5th and 6th. The value of these violent disturbances in bringing rain to the interior is obvious |
| 1900„ | Mar.„ | 6th, at 8 a.m. | Onslow (see map) | |||
| 1901 | Jan. | 30th | s.s. Australind, south of Java |
Storm travelling W.S.W. | Ship hove to. Barometer 29.36. A very interesting hurricane, for its course at sea can be traced from re- ports of the Australind | |
| 1901„ | Feb. | 7th | Cossack | Approaching coast from W. | Port Hedland, 13″; Peak Hill, 3″, 8th; Eastern Gold-fields, 3″, on 9th |
Barometer at Condon, 28.78. Ap- parently same storm as above |
| 1902 | Feb.„ | 9th, at 12.30 p.m. |
Cossack„ | Travelled along the coast from Wyndham to Cossack, thence inland to Coolgardie and Eyre |
Hall's Creek, 4″, 2nd to 6th; La Grange, 7″, 4th-8th; Condon, 7″, 9th-10th; Peak Hill, 3″, 10th-11th; Men- zies, 3″, on 11th |
Barometer, 29.03, was the minimum re- corded at Cossack. "Willy-willy" did a good deal of damage at Cossack. Prior to the rains the clouds in the in- terior were moving from N. to N.W. |
| 1903 | Jan. | 10th | Off East Kimber- ley Coast |
Recurved north of Broome on 16th; passed inland just touching east of gold-fields; thence to South Australia |
Wyndham, 25″, 9th to 13th; Turkey Creek, 10″; Hall's Creek, 6″; Quanbun, 2″; Menzies, 1.7″, on 19th |
Wind not so severe as usual. Barometer only fell to 29.65 |
| 1904 | Apl. | 17th | Off Broome, 30 hours |
Broome, 5.5″; Yeada, 8″ | Schooner Star of the West wrecked | |
| 1905 | Feb. | 8th | Onslow | From N.W. coast across in- terior to the Bight |
Onslow, 1.5″; Fortescue, 2.7″; Carnarvon, 2″; New Forest, 7″; Field's Find, 2″ |
Barometer at Onslow, 29.11; damage to shipping and loss of life |
| 1907 | Jan. | 14th | La Grange | Wallal, 8.4″; La Grange, 4″; Broome, 6″ |
Barometer at Broome, 29.51. Heavier gale than April, 1908 | |
| 1907„ | Mar. | 13th, 14th | From Kimberley down coast to Shark's Bay; thence across continent to the Bight |
Cossack 13.5″, 11th to 15th; Chinginana, 7″, 9th to 16th; Winning Pool, 3.5″; Clif- ton Downs, 4.2″, 15th; general heavy rain through- out North-west Gascoyne and gold-fields, reaching Eyre, 2.3″, on 17th |
Mildura wrecked. Barometer at Ons- low, 29.30, on 14th; and at Yalgoo, 29.24, on 15th. No heavy winds re- ported | |
| 1908 | Apl. | 26th and 27th (midnight) |
La Grange Bay (Broome) |
Broome to Walal (180 miles to S.W.) |
Derby, 3″; Broome, 8½″; La Grange, 5½″; Bohemia Downs, 5.6″ |
Barometer, 29.54. Pearling fleet wrecked; damage, £40,000; 50 lives lost. Apparently a local hurricane |
| 1908„ | Dec. | 8th | La grange to Wal- lal |
La Grange, 3.9″; Wallal, 8.5″; Eel Creek, 2.5″; Red Hill, 2.6″ |
Cutty Sark wrecked | |
| 1909 | Jan. | 19th 11.45 p.m. |
Onslow | Moving S.E. over Onslow | Onslow, 8.8″; Leonora, 5″, on 21st |
Barometer at Onslow, 28.945. E.N.E wind at maximum force of gale. Fout luggers and 24 men lost |
| 1909„ | Apl. | 6th, noon | Moving towards S.S.E. over Onslow; thence overland to South Australia |
Onslow, 5.4″; Yaurie, 6″; Bangomal, 3.5″; Abbotts, 2.6″ |
Barometer at Onslow, 29.188 | |
| 1910 | Nov. | 19th, at 2 p.m. |
Broome | On the 19th a marked "low" moved over Broome from the sea; on 20th over La Grange; on 21st dissi- pated |
Broome, 11″; Derby 5.7″; La Grange, 7.4″; Rollah, 10″ |
Twenty-six luggere and 40 lives lost. This hurricane did not give a well- marked track. Barometer at Broome, 28.64 |
| 1911 | Feb. | 7th, at 4.30 a.m. |
Onslow | Moving down the coast from N.E. to Carnarvon |
Whim Creek, 6″, on 7th; Winning Pool, 3″, on 8th |
Barometer, minimum 28.77, at Onslow, and wind E.S.E. Glenbank wrecked off Legendre Island at 9 p.m. on 6th. All the crew drowned but one man |
| 1912 | Mar. | 6th | Broome | Broome, 7″, from 4th to 6th; La Grange, 8.7″ |
Barometer 29.33 | |
| 1912„ | Mar.„ | 21st, at 11 p.m. |
Cossack (see map) | From N.E. to Cossack; and thence to Murchison and Coolgardie Gold-fields |
Cossack, 9″; Nullagine, 4″; Tambrey, 6.6″ |
Barometer, mininum 28.86. Wind changed from east through south to west during the hurricane at Cossack; Koombana wrecked off Condon; all lives lost |
The hurricanes are thus most numerous in the hottest months. Of the thirty-one tabulated above—
| January | has nine; | April | has six; | |
| February | has„ six; | December | has„ three; | |
| March | has„ six; | November | has„ one. |
In these months there is a very hot region situated between Nullagine and Winning Pool, Western Australia. This area of high temperature means also an area of low pressure, and this is no doubt a factor in swinging the hurricane to the south-east. The cold Westralian current may act as a buffer to the passage of these cyclones along the coast beyond Onslow; the general easterly drift of the upper air also undoubtedly helps to swing them inland from Cossack. Moreover, their destructive violence rarely penetrates far beyond the coast, for this can only be maintained by large supplies of ascending moist air. This can only take place over the ocean, so that though an individual cyclone can be traced to Coolgardie, to Eucla, and even to Tasmania, it rarely inflicts much damage inland, but, as noted in the preceding table, rather brings beneficial heavy rains.
XI.(b).—Hurricanes in the South-Western Pacific.
In the warm, moist, and calm region of the north-east coast of Queensland, hurricanes develop just as in the similar area north of Western Australia.
Owing, however, to the fact that their parabolic paths do not approach the land so closely as in Western Australia, we have very little data of their behaviour. They originate generally between Fiji and New Guinea, and move westward to Queensland. They usually hit the coastline near Cairns, and then "rebound" to the south-east—being carried thither by the dominant upper currents as in the western examples. Occasionally their path can be traced from shipping reports; but very often it is not possible to obtain mainland data, for the telegraphic and postal services are nearly always greatly disorganized for some time after the disaster. Two examples are briefly described, which will give some idea of their character.
The Fiji Hurricane, March, 1910.
(Figs. 93-96.)
FIJI HURRICANE.
March 1910
yards away. One well-found house of four large rooms and verandah all round was shifted 52 feet. Another 3 feet would have put it over the hill. A new two-story concrete house was cut in half, the upper portion going."
Apparently the hurricane came down upon the group without any of the usual barometric warnings. Captain Wooley, the harbor-master at Suva, reports that "it was a most extraordinary and most unexpected happening; the barometer showed no signs to speak of, as we often have a very much larger drop than shown on the 24th at 4 p.m. I was on a trip, but put back at 5.30 p.m. owing to the excessive heat, for although we were having a fresh south-east breeze, it was beyond anything I had ever experienced in similar weather. On the evening of the 24th there were no signs of dirty weather, except this heated wind; from 6 to 10 p.m. there was a moderate easterly wind with some rain; at midnight, barometer falling and wind increasing; 2 a.m., barometer 29.000 inches, increasing wind from east with heavy rains; 3 a.m., blowing with hurricane force, house started to go, impossible to venture out, sheets of iron flying from all directions from east to north-east; at 4 a.m., the barometer started to rise and the wind fell; 5 a.m., back to only a north-north-east gale, with heavy torrents of rain." At Suva, the storm was at its maximum about 4 a.m., when the barometer read 28.50 inches; at Levuka, the lowest reading was 28.64; and at Nausori, 28.40. At Levuka, there was a calm of ten minutes or more, then a complete change of wind from south-east to north-west, showing that the eye of the storm passed right over the town, but at Suva there was only a steady easterly blow backing north north-east as the barometer rose.
The storm swept over Bau in Fiji, leaving only four houses out of 100 standing. Hence it pursued a westerly course through the New Hebrides. Here the storm centre passed near the island of Tanna. In Mare, one of the Loyalty Islands, three storm waves destroyed houses and plantations. On Easter Sunday, it swept down on Noumea, New Caledonia. Here small coasting boats, barges, and the steam launch Obus were sunk in the bay of Moselle, and the town of Noumea was considerably damaged.
Norfolk Island.
After passing New Caledonia and recurving to the south-west of the island, the storm pursued a southerly course to Norfolk Island, covering the intervening distance in about 26 hours. Dr. P. Herbert Metcalf, the Deputy Chief Magistrate, reports that the barometer fell to 28.882 inches, with a gale of hurricane violence in the east, but, except to trees and gardens, there was little damage done. No houses or persons were injured. The following are his meteorological notes before and after the passage of the storm:—
- 28th March, 9 a.m.—Dull, damp, and drizzly. Barometer, 29.538 (uncorrected 383 feet above sea-level). Wind, east; force, 7.
- 28th March, 9 p.m.—Barometer, 29.192, falling fast. Wind, force, 10. Rain pouring down.
- 28th March, 9.45 p.m.—Barometer, 29.062.
- 29th March, 1.30 a.m.—Barometer, 28.882. Awful gale from east. (The barometer fell no lower with me; was up many times in the night to look.)
- 29th March, 9 a.m.—Barometer, 28.970. Wind, north-east; force, 11. Rain recorded, 2.65 inches.
- Then came a rapid rise in the barometer; the evening of the 29th was fine and fairly calm.
- 30th March, 9 a.m.—Barometer, 29.542. Wind, westerly; force, 4. Trees fell chiefly from north-east to south-west.
This storm system was remarkable in that it retained its cyclonic form and energy over so long a period. Usually tropical disturbances spread out over a wider area on reaching higher latitudes and lose their vigour, but in this instance, as shown by the following table giving the lowest barometer readings along its track, the disturbance retained its intensity in a very remarkable degree throughout its whole path:—
| Barometer. | |||
| March 24 | Levuka (Fiji) | 28.64 inches | |
| March„ | Nausori (Fiji), 10 p.m. | 28.80 inches„ | |
| March„ | Nausori (Fiji), 3 a.m. | 28.40 inches„ | |
| March„ | Suva (Fiji), 2 a.m. | 29.00 inches„ | |
| March„ | Suva (Fiji), 4 a.m. | 28.50 inches„ | |
| March„ | New Hebrides, 3-4.30 a.m. | 28.28 inches„ | |
| March„ | Noumea (N.C.), 11.30 p.m. | 28.42 inches„ | |
| March„ | Norfolk Island, 1.30 a.m., &c. | 28.88 inches„ | |
| March„ | Russell (N.Z.), 9 a.m. | 29.00 inches„ | |
The series of isobar charts which accompany this report shows that the disturbance followed the usual track of storms generated in the southern tropical zone, and first moved in a west and south-westerly direction, until caught in the easterly atmospheric drift of mid-latitudes, when it curved off to the south-east. They also clearly confirm what has been before established, that storms originating in the vicinity of the Fiji and other groups of islands east of New Caledonia are likely to affect the northern parts of New Zealand later. New Zealand has thus ample warning of the coming of the cyclone, and the Dominion Meteorologist, the Rev. D. C. Bates, was enabled to advise all the North Island coast stations some two or three days beforehand of the storm approaching.
The large monsoonal storm system which crossed the continent in the early part of the month, the continued monsoonal disturbance and heavy rains in Queensland later, and the development of the Fiji hurricane, are evidence of unusual activity in the southern monsoonal low pressure belt during March, and it would be interesting to ascertain whether a similar activity was shown in other parts of the Southern Hemisphere.
Port Douglas, March, 1911.
On the North Queensland coast, March is perhaps the month most liable to tropical hurricanes. In 1911 (16th March), a disastrous storm struck Port Douglas, killing several people and almost destroying the town.
A brief description of the hurricane is as follows :—
Until the 13th, a tongue-shaped shallow depression remained persistently over the north-western quarter of the State and the Gulf of Carpentaria without developing unsettled characteristics of any definite kind. On the 14th data showed that the trough was undergoing a deepening process and under the impetus of vigorous high-pressure waves from the westward, was moving slowly across the Peninsula (Figs. 97—98). This deepening was maintained, and on the 16th a disturbance was evolved with its centre approximately a little to the north-east of Cairns. It suddenly moved towards the
Queensland Hurricane
Figs. 97—98.
coast, between Cooktown and Cardwell, and its worst effects were experienced in the Port Douglas District, which it swept with disastrous results to property and life.
During this hurricane the fine steamer Yongala was lost, with all on board.
It subsequently moved on a westerly course, rapidly decreasing in intensity, and remained over the Gulf of Carpentaria until the 21st, causing unsettled conditions in the north-western interior.
This storm obviously did not follow the parabolic path characteristic of the more typical tropical hurricanes. It, however, caused very heavy precipitation all along its track, as the following 24 hours' totals show:—
| 16th and 17th | Innisfail | 6.05 inches |
| Atherton | 6.20 inches„ | |
| Thornborough | 4.78 inches„ | |
| 17th and 18th | Normanton | 4.05 inches„ |
| Donor's Hill | 3.08 inches„ | |
| Atherton | 1.26 inches„ |
XI.(c). Southerly Bursters.
Figs. 99-100.
These winds have been described as perhaps the most remarkable of the "squall" winds which characterize various parts of the earth.
They are cold winds, succeeding a period of hot weather. They blow from the south, usually arriving as a sudden squall after calms or northerly winds.
Although not unknown along the southern shores of Australia, it is the coast of New South Wales, from Port Macquarie to Cape Howe, where they are especially noteworthy. Here the topographic conditions are peculiar, and are undoubtedly contributing causes to the local intensity of the burster.
Some 50 or 80 miles from the coast extends a belt of highlands from 3,000 to 7,000 feet high. They are parallel to the coast and therefore at right angles to the general easterly atmospheric drift. Between this cooler belt of highlands and the ocean is a "hinterland" which is considerably warmer than the ocean or the mountains during the summer months when these winds chiefly occur.
The monthly averages for the bursters are as follows:—
| January | 17.3 | per cent. |
| February | 13.2 | per cent.„ |
| March | 9.1 | per cent.„ |
| April | 4.1 | per cent.„ |
| May | .4 | per cent.„ |
| August | .5 | per cent.„ |
| September | 6.3 | per cent.„ |
| October | 14.0 | per cent.„ |
| November | 16.8 | per cent.„ |
| December | 18.3 | per cent.„ |
So that the spring and summer half-year accounts for 86 per cent. of the bursters.
The average of their maximum velocities is 42.7 miles per hour, though 153 miles was recorded in 1877.
(The maximum velocity occurs usually about twelve hours after the burst.)
The weather preceding the burster may be summarized as follows:—For a period varying from three hours to three days before high temperatures prevail, northerly winds, with a tendency to east in summer, and to west in early and late summer, are the rule.
The clouds are characteristic. Usually cirro-cumulus and thunder clouds appear in the south-west, then a long dark cumulus roll approaches with a front of 30 miles or more. As this approaches the wind drops completely, and then a whirl of dust ushers in the "burst." The windvane flies round to the south, and the southerly in a few minutes may reach gale force.
Rain may accompany the burst, but it is generally due to the electrical disturbances. The warm antecedent north winds are, however, occasionally accompanied by some rain.
The southern swing in wind direction is accompanied by a very rapid fall in temperature on occasions amounting to as much as 37° Fah.; 18° is about the average total fall, but on the 30th December, 1891, the thermometer dropped this amount in the first five minutes.
Origin.
There is no doubt that the burster is merely an intensification of the normal southerly wind ushering in an anti-cyclone in New South Wales. This intensification is at least partly due to the relation of topographic and temperature conditions noted previously.
The following are some of the factors which have to be considered in discussing this phenomenon:—
(1.) The effect of the friction of the land mass, combined with differences of latitude, in gradually tilting the axis of the Λ-depression, so that as it passes New South Wales, it tends to lie N.W. and S.E.
(2.) The differences between the wind forces over land and water for the the same barometric gradients.
(3.) The convectional effects of the evaporation from the warmer waters off the East coast of Australia.
(4.) The effect of the ridges of the Kosciusko Ranges, which attain a height of over 5,000 feet for nearly 100 miles, in delaying the Eastward advance of the "highs" and "lows."
That "lows" coming from the West are intensified as they reach the warm coastal waters of New South Wales is a matter of common observation, as is the tendency for cyclonic storms to form there, and persist for days at a time. This is probably due to factor (3) above. If we assume that the Northern portion of a passing Antarctic depression undergoes a decided deepening from this cause, and that the axis of the depression is tilted slightly towards the N.W., as suggested in (1), the winds in rear of the trough will not only be strengthened by the fall in the barometer, but will come from a point sufficiently near Southerly to blow mostly over water. The two factors will, therefore, assist one another in increasing the force of the Southerlies.
As regards the influence of the Kosciusko Ranges, it may be premised that they have a greater retarding influence upon the "highs" than the "lows," since the increase of atmospheric density downwards near the surface is greater in "highs" than in "lows," and this may cause the change to anticyclonic high pressures along the New South Wales coast to be delayed slightly, and then to be more sudden than elsewhere, causing a temporarily steep barometric gradient on the rear side of the trough with corresponding wind force. Some influence may be traceable also to the contrast between the cooling effects of the Alpine summits, partly covered with melting snows, and the abnormal warmth of the strip of country lying between the mountains and the coast.[2]
- ↑ Meteorological Report, Western Australia, 1902, p. 6.
- ↑ For fuller discussion of this subject see "Essay on Southerly Bursters," by H. A. Hunt, Proceedings of the Royal Society of New South Wales, 1894.