With the ships of the U. S. Navy actually, if not literally, spanning the seven seas in increasing numbers, knowledge of severe tropical storms becomes of increasing importance. Thus the geographical sources of origin, the seasonal periods of occurrence and seasons of maximum and minimum intensity, general path and areas affected, all become factors of prime interest to the Captain and the Navigator.
Heretofore information relative to these matters has been contained in widely scattered volumes and publications and, in some instances, only in periodicals dealing with a specific area. Therefore, in order to simplify the task of the individual mariner, the pertinent data have been compiled from the divergent sources and are presented in this article so as to present a concise account of what is known about tropical cyclones, which attain hurricane force, throughout the globe. No attempt has been made to restate the law of storms, the premonitory signs of hurricane development, or the navigation of ships out of the “dangerous semicircle” (the eastern semicircle for a north-bound storm in the Northern Hemisphere, and the western half for a south-bound cyclone in the Southern Hemisphere). It is believed that this information is generally known or else may be readily found in standard books such as Bowditch or Knight’s Seamanship.
Various names, depending upon the locality in which they originate, are given to these meteorological blitzkriegs. The most common designations, according to oceanic region of formation, are as follows:
(1) North Atlantic—Cape Verde Islands, the West Indies, Gulf of Mexico, and the Atlantic Seaboard—hurricanes.
(2) South Atlantic—none.
(3) North Pacific Ocean.
(a) Philippines, China Sea, and Japan— typhoons or “bagnios”.
(b) Central America and Mexican coast— hurricanes or “ cordonazos”.
(4) South Pacific Ocean—Taumotu Islands to the Queensland coast (Australia)—hurricanes or cyclones.
(5) North Indian Ocean—Bay of Bengal and Arabian Sea—cyclones.
(6) South Indian Ocean.
(a) Cocos or Keeling Islands to Madagascar—cyclones.
(b) Sea of Arafura and Northwestern Australian coast—willy-willies.
Certain widely accepted concepts regarding these storms are recounted here. For instance, it is a generally accepted belief that tropical hurricanes originate in well- defined areas of the globe and have preferred seasonal periods of occurrence and intensity (Fig. 1).1 Thus the initial development of a tropical cyclone is to be found in the doldrum belt of the oceans bordering the equator, especially in the western reaches of these water areas, the South Atlantic Ocean being the only exception to the above rule. Hurricanes also develop in the southeast North Atlantic (Cape Verde region) and in the southeast North Pacific (west coast of Central America and Mexico).
The doldrum belt, previously mentioned, is the low-pressure area near the equator which separates the prevailing winds of the north and south hemispheres (Fig. 2). In common with pressure and wind belts of the world, its mean position fluctuates with the declination of the sun, since essentially the reason for the low pressure is caused by the differences in the solar heating of the earth’s surface. This seasonal migration leads to preferred periods of the year for the formation of intense tropical storms in each oceanic area, known as the hurricane season. Within the hurricane season itself for a particular region there is some evidence to further postulate that the time of maximum hurricane activity coincides with the time during which the doldrum belt is the greatest distance from the equator.
Tropical cyclones form over the warm ocean regions during certain months of the year. For any particular area this period is known as the hurricane or typhoon season. The seasonal distribution of tropical cyclone development throughout the world is conveniently set forth by the table on the next page.
Tropical cyclones display preferred directions of travel for each hemisphere. In the Northern Hemisphere they normally move west and then northwest, whereas in the Southern Hemisphere their movement is west and then southwest. This general information does not carry too much weight in resolving the exact course and speed of a specific storm, as each is a separate phenomenon which is guided and affected by the controlling meteorological factors existent at the particular time. However, the chief benefit to be derived from mean tracks is the over-all knowledge that, based upon past records, the probability of the storm progressing along a certain path is greater than along any other path. Accordingly the recognition that in a particular ocean area tropical cyclones are more apt to proceed west and thence northwest and not west and then southwest, will be a salient factor in setting a course to clear the hurricane area.
|
May |
June |
July |
Aug. |
Sept. |
Oct. |
Nov. |
Dec. |
Jan. |
Feb. |
Mar. |
April |
North Atlantic hurricanes |
|
X |
X |
X |
X |
X |
X |
|
|
|
|
|
North Pacific hurricanes (Mexican coast) |
|
X |
X |
X |
X |
X |
X |
|
|
|
|
|
North Pacific typhoons |
X |
X |
X |
X |
X |
X |
X |
X |
X |
|
|
|
South Pacific hurricanes |
|
|
|
|
|
|
|
X |
X |
X |
X |
X |
North Indian cyclones (Bay of Eengal) |
X |
|
|
|
|
X |
X |
|
|
|
|
|
North Indian cyclones (Arabian Sea) |
X |
X |
|
|
|
|
X |
|
|
|
|
|
South Indian cyclones |
X |
|
|
|
|
X |
X |
X |
X |
X |
X |
X |
NW Australia willy- willies |
|
|
|
|
|
|
X |
X |
X |
X |
X |
X |
X Tropical cyclones occur during these months.
Therefore, with the above considerations in mind, mean tracks together with the areas likely to be affected are set forth as specifically as the past observational data will permit. In some of the regions a seasonal fluctuation has been noted and is contained in the plates detailing the mean paths. It is to be again emphasized, however, that the use of the mean tracks should be tempered by the consideration that each storm will exercise a certain amount of deviation from the mean path portrayed.
Forewarned is forearmed and, consequently, knowing the regions and seasons of occurrence the mariner is able to intelligently assess his possibilities of encountering a tropical cyclone in whatever part of the globe he may be. Should the nature of the voyage be such that nothing can be done to avoid traversing a hurricane area in season, and should a hurricane or typhoon be reported, reference to the track charts will provide an initial estimate of future movement based upon the mean course which a greater number of previous storms have followed. Frequent plot of the storm’s position as given by radio broadcast will serve as a welcome check on the preliminary assumption of future movement.
In passing it should be mentioned that during the prenatal stages of tropical cyclone formation incipient vortices will temporarily make up, only to die out as rapidly as they were formed owing to a lack of sufficient energy which is necessary to continue their cyclonic circulation. If the atmosphere continues to be unstable, this formation and deformation persists until such time as a particular storm center emerges with the strength and intensity necessary to overcome the factors tending to dampen out the perturbation. When this occurs, the cyclonic center then becomes the controlling factor and, in a capitalistic manner, tends to swallow up the lesser disturbances. Evidence of the above prelude to hurricane or typhoon formation is manifest by sudden torrential downpours and rapid changes in wind direction and velocity, accompanied by brief intense gustiness. Therefore, whenever such instability exists, caution should be exercised in assessing the exact location of the principal storm center.
Tropical cyclones according to ocean areas. —In the following paragraphs tropical storms attaining hurricane force are discussed for each ocean area with regard to source region, seasonal occurrence, and mean path. Each part is complete in itself so that it is necessary to refer only to the area under consideration for ready reference purposes.
North Atlantic hurricanes.—The derivation of the name “hurricane,” as applied to the tropical storms which frequent the North Atlantic, according to Tannehill2, is said to be of Carib Indian origin and means “big wind.” It has come into general use as the name for any tropical cyclone of the Atlantic Ocean, Gulf of Mexico, Caribbean Sea, Eastern Pacific (off Central America and Mexico), and the Southern Pacific, which develops into a storm of great violence. As such, the particular type of storm it denotes is identical with the typhoon of the China Sea, the bagnio of the Philippines, and the cyclone of the Bay of Bengal and other portions of the Indian Ocean.
Source regions.—North Atlantic hurricanes originate in the Atlantic doldrum belt, which extends from the African coast to the north eastern corner of South America (sec Fig. .1). The position of the Atlantic doldrum belt varies in accordance with the intensity and strength of the converging trades. On some days it will be nonexistent, hence its reputation as the belt of equatorial calms, only to suddenly form as a band of squally weather as a marked surge of air sweeps up from the south or down from the north. The distribution of the air currents on either side of the above zone is graphically depicted in Fig. 3, which indicates the winds over the Atlantic Ocean and the confluence of the prevailing trades in the belt of equatorial calms or doldrums.
The axis of the Atlantic belt in June occupies a mean position of 8° north on the African coast and extends west-southwest to 5° north on the northern corner of South America. In the latter part of July and during August it reaches its northern limits extending from Bathurst west-southwest to the vicinity of Trinidad and occasionally extending farther north to the Leeward Islands. During this season of the year hurricanes originate in the eastern extremity of the doldrum belt in the vicinity of the Cape Verde Islands or farther to the west between 40° and 55° west. Occasionally tropical cyclone centers are first reported along the Barbados, Martinique, Guadeloupe Islands, but in all probability their actual inception took place farther to the east. It is common belief that hurricanes rarely, if ever, germinate in the region between 50° west and the South American coast.
Hurricanes also break out in the western Caribbean during June and October when the Pacific doldrums extend into that area. On the other hand, there is no definite extension of the Atlantic doldrum belt into the eastern Caribbean and consequently this area is relatively free from hurricane genesis.
Seasonal frequency.—A number of compilations of occurrence by months have been made and that of Mr. Mitchell, U. S. Weather Bureau, for the period 1887-1941 is reproduced in the following table:
|
May |
June |
July |
Aug. |
Sept. |
Oct. |
Nov. |
Dec. |
Total |
No. of storms |
5 |
25 |
27 |
85 |
131 |
105 |
26 |
2 |
406 |
No. of storms of hurricane intensity |
0 |
10 |
14 |
58 |
74 |
39 |
7 |
0 |
|
Percentage |
0 |
40 |
52 |
68 |
56 |
37 |
27 |
0 |
|
From the above table it is apparent: (1) The hurricane season extends from June through November; (2) September is the month with the maximum number of hurricanes; (3) tropical storms occurring in August have a greater tendency to develop into full-fledged hurricanes than those in October by a percentage of 68 to 37.
Mean path.—Examination of a chart of West Indian hurricane tracks for September (Fig. 4) indicates the difficulty in attempting to detail mean paths for this area. During the beginning of the season, i.e., in June and the first part of July, hurricanes originating in the western Caribbean progress northwestward into the Gulf of Mexico and thence inland into Mexico or the Gulf Coast of the United States. Those which form later in the season in the eastern reaches of the Atlantic doldrum belt follow a protracted westward course across the Atlantic prior to veering to a northwesterly course and final recurvation to the north and northeast. Hurricanes occurring near the end of the season, namely, in October and November, normally appear in the western Caribbean. These storms assume a northwesterly course with rapid re- curvature to the northeast. Figure 4a, which is the hurricane plot of positions for the 1944 season, shows close agreement with the above ideas. Although no hurricanes were formed in June, the majority of those occurring during the middle of the season followed a protracted westerly course from the eastern portion of the source region; whereas the last hurricane of the season, namely, that of October 13-21, was initially located in the western Caribbean and progressed in a northwesterly thence north and northeasterly direction.
Pacific Ocean typhoons and “bagnios.’’— The word “typhoon” is said to be of Chinese origin and, similar to the word “hurricane,” it means “strong wind.” Thus typhoon has come to be the name applied to the intense' cyclonic disturbances of tropical origin which occur in the western North Pacific Ocean, the China Sea, and the Sea of Japan. The same phenomena, when it reaches the Philippines, is known locally as a bagnio.
Source regions.—The particular ocean area presently under discussion produces more severe storms than any other region on earth, with the South Indian Ocean and the West Indian regions next in order of storm-producing capacity. The majority of the storms of the southwestern North Pacific Ocean form to the east of the Philippine Islands with only two or three per year originating in the China Sea. More specifically, the primary source region of the majority of the typhoons is the vast area contained between the 5th and 20th parallels of north latitude and from 170° east longitude to the Philippine Islands. Although typhoons may form anywhere in this broad expanse from the Marshalls to the Philippines, the most prolific sector is the region of the Caroline Islands.
A secondary source region is located in the middle of the China Sea in the Paracels (Paracels-Maccesfield Shoals). In a large number of cases typhoon inception in the China Sea occurs concurrently with the existence of another typhoon near or in the Pacific in the Balingtang Channel between Luzon and Formosa.
Although the number of typhoons will vary from year to year, the following record for those occurring during 1901-25 will give some idea of monthly frequency as observed during the 25 years.
Jan. |
Feb. |
Mar. |
Apr. |
May |
June |
July |
Aug. |
Sept. |
Oct. |
Nov. |
Dec. |
Total |
13 |
4 |
6 |
7 |
20 |
28 |
76 |
111 |
89 |
75 |
38 |
24 |
49 |
When complications for other periods are taken into consideration a yearly expectation of typhoons by month is:
Feb. |
Mar. |
Apr. |
May |
June |
July |
Aug. |
Sept. |
Oct. |
Nov. |
Dec. |
Jan. |
0.4 |
0.4 |
0.4 |
1.0 |
1.2 |
3.2 |
3.9 |
3.9 |
3.4 |
2.0 |
1.2 |
1.0 |
Obviously those months listed as below 1.0 have no practical significance except to serve as a warning that typhoons have infrequently occurred during those periods. Consequently, it may be seen that the typhoon season sets in late in May and continues through January. Maximum activity occurs from July through October; minimum activity in February, March, and April.
Mean paths.—From the wealth of old records and studies made by various weather services it is possible to specify seasonal fluctuation of mean paths and areas likely to be affected. This is done by the accompanying monthly charts for the typhoon season (Plates I-1X).
By far the greater number conform to the expected tropical cyclone directional pattern and move initially from the source region westward and then gradually veer into the northeast upon reaching higher latitudes. Others travel almost due westward and dissipate over the Philippines and China coast. However, as a word of caution regarding phenomena of this nature regardless of source region, large departures for this particular area have been recorded. For example, in one specific case a typhoon approached the northern end of Formosa in a west-north- westerly direction and then turned southward, not northwestward as might be expected, and consequently passed through the Formosa Straits from north to south. Hence, the course of each individual storm is dependent upon existing meteorological factors influencing its movement. Therefore the progress of each typhoon should be closely followed on all weather broadcasts in order not only to determine its current position but also to be forewarned as soon as possible of abnormal movement, either in course or speed, of the original center or the formation of secondary typhoon centers, equally violent, which may affect the ship’s safety. In some instances as many as four typhoons have existed simultaneously within the typhoon area.
Central American and Mexican hurricanes or “cordonazos."—Although not as well known as the West Indian hurricanes which frequent the south and east coast of North America, a counterpart of the tropical cyclone is to be found in the southeastern part of the North Pacific Ocean which affects the west coast of (he American continent from Costa Rica to the middle of Lower California. These storms are called hurricanes or cordonazos. Regarding the history of the local name given to this phenomenon, Mr. Willis Hurd of the U. S. Weather Bureau furnishes this interesting account:
The most violent storm known to the early weather historians of this coast was designated by the natives as El Cordonazo de San Francisco— “The lash of the whip of St. Francis”—since it was popularly connected with the time of occurrence of the Feast of St. Francis on October 4. Commodore (later Admiral) Dewey, U.S.N., writing of the weather of the lower part of the Gulf of California during the rainy season, refers thus to these storms: “Occasionally a local hurricane, known as El Cordonazo, blows with great violence.” A hurricane of this character occurred at Santa Rosalia, well up the east coast of Lower California, in October, 1890. It was of brief duration, as are some of the intense winds of these storms, but in a half hour or less two of the three ships at anchor in the harbor were driven on shore.
More historic than the foregoing is the often quoted but classic reference to the terrific hurricanes of the same character of 1839 and 1840, each on November 1 of its respective year. The former burst upon the harbor of Mazatlan and caused the destruction of the greater number of 12 ships stationed there, with the loss of all on board. The latter resulted in the drowning of a number of people and the loss of three vessels in the roadstead of San Bias. Perhaps some of the severe local windstorms of these early days received the same common term of El Cordonazo, but today the name is recognized as that given to the southeasterly to southwesterly gales which blow on the coast during the passage of a tropical cyclone very close to sea.
Source regions.—No orderly record of storms occurring along the west coast of Mexico was maintained until 1920, during which year the Mexican Government assumed the responsibility and control of all meteorological stations on both coasts of Mexico. The data thus obtained show that nine-tenths of the storms which occur are formed in a triangular area beginning at Cape San Eugenia in the middle of Lower California, thence southwesterly to 10° north, 130° west, and thence easterly to the Costa Rican coast. Of the remaining one-tenth, some are known to originate nearer the equator and work their way westward into the open sea. It is possible that these are the storms which are infrequently encountered between 130° west longitude and the Hawaiian Islands (Plate X).
Seasonal frequency.—From the Mexican Meteorological Office information, plus the weather reports forwarded to the U. S. Weather Bureau by ships at sea, a study of 177 tropical cyclones occurring during the period 1910-40 has been made. The results of this work summarized in tabular form is as follows:
|
May |
June |
July |
Aug. |
Sept. |
Oct. |
Nov. |
Dec. |
Total |
No. of storms |
4 |
25 |
22 |
31 |
59 |
32 |
3 |
1 |
177 |
No. attaining hurricane force |
3 |
7 |
5 |
15 |
21 |
15 |
1 |
0 |
67 |
Percentage |
— |
28 |
23 |
48 |
36 |
47 |
33 |
0 |
|
The above figures point out that the hurricane season extends from June to October similar to the West Indies season, with occasional storms in May and November. The greatest number of hurricanes occur in September and the percentage of tropical cyclones attaining hurricane force is approximately equal for August and October.
Mean paths.—Generally speaking, the hurricanes of the eastern North Pacific Ocean area may be classified into broad geographical classes insofar as mean paths are concerned, namely: (1) the coastwise hurricanes; (2) the Revilla Gigedo Islands and vicinity hurricanes (Plate X).
The coastwise cyclones are those which originate in the vicinity of the Gulf of Tehuantepec and may be either near or moderately far oil the coast. In any event, the normal course parallels the coast line, the point of going inland normally being north of latitude 20° north. They may roar up the Gulf of California or remove inland across Mexico in the vicinity of Mazatlan.
The Revilla Gigedo Islands class start their trek farther to the westward and.move west-northwest, turning to the north in the vicinity of the islands.
A third and less frequent type are those hurricanes which are generated in low latitudes and move northward recurving and striking the Mexican coast perpendicularly without ever having paralleled it.
South Pacific hurricanes.—Directly in contrast to the South Atlantic, in whose tropical regions no hurricanes are known to exist, the South Pacific enjoys a fairly active hurricane season. This activity becomes most pronounced in January and February when the Pacific doldrum belt has progressed to its southernmost limit. Consequently, activation of hurricanes in the South Pacific is coincidental with cessation of typhoon activity in the North Pacific.
Source regions.—The South Pacific is unusual in another respect, namely, in the extensiveness of the “breeding grounds” of its severe tropical storms. This source region extends from the Tuamotu Archipelago (140° W.) westwards to approximately 160° east longitude and from 5° south to 20° south.
The portion of the South Pacific from 140° west eastward to the South American coast is either remarkably free of these cyclonic trouble makers or else their existence is unknown. As evidenced from the accompanying charts delineating the areas of activity, hurricanes occurring during the first of the season (December) are formed to the cast of 180° longitude. As the season progresses the source region expands westward and by February hurricanes occasionally reach the Queensland coast.
A secondary source region, not shown on the accompanying charts, is often stated as existing in the Coral Sea. However, it is believed these storms have originated in the primary zone described above and have moved unobserved westward into the Coral Sea before their existence has been reported.
Locality |
July |
Aug. |
Sept. |
Oct. |
Nov. |
Dec. |
Jan. |
Feb. |
Mar. |
Apr. |
May |
June Total |
|
Fiji Islands |
|
— |
— |
— |
1 |
13 |
21 |
15 |
20 |
4 |
— |
— |
74 |
Tonga Islands |
— |
— |
— |
— |
3 |
3 |
16 |
7 |
14 |
6 |
1 |
— |
50 |
Samoan Islands |
— |
— |
— |
— |
1 |
6 |
10 |
2 |
8 |
2 |
— |
— |
29 |
New Hebrides |
— |
— |
— |
1 |
3 |
1 |
11 |
9 |
9 |
2 |
1 |
— |
37 |
New Caledonia |
1 |
— |
1 |
2 |
— |
4 |
11 |
11 |
10 |
— |
— |
— |
40 |
Norfolk Island |
— |
1 |
— |
2 |
— |
— |
4 |
7 |
7 |
2 |
— |
2 |
25 |
Low Archipcl. |
— |
— |
1 |
— |
— |
— |
3 |
2 |
1 |
— |
— |
— |
7 |
Society Islands |
— |
— |
1 |
— |
— |
3 |
3 |
1 |
— |
— |
— |
— |
8 |
Solomon Islands |
— |
— |
— |
— |
— |
1 |
2 |
1 |
2 |
— |
— |
— |
6 |
Cook Islands |
— |
— |
— |
— |
— |
5 |
2 |
5 |
2 |
2 |
— |
— |
16 |
Total |
1 |
1 |
3 |
5 |
8 |
36 |
83 |
60 |
73 |
18 |
2 |
2 |
292 |
Seasonal frequency.—Based upon the reports from South Pacific Islands prior to the World War II, the following table3 for the period 1789-1922 will present some idea as to the months of maximum activity. The data are based on a table of 246 individual hurricanes, but as some of the storms affected more than one group of islands, the sum total is 292.
December through April is the hurricane (cyclone) season, the greatest activity occurring during January and again in March. Hurricanes sometimes occur in October and November but very rarely in May, June, July, August, and September.
Infrequently storms affect the Coral Sea and the Queensland coast and Northern Territory of Australia, as depicted by the below table of 159 hurricanes for the period 1839- 1922.
It is to be noted that these storms occur during the same season as the other South Pacific hurricanes previously mentioned, namely, from December through April, with maximum activity occurring in January and again in March. They may occur in all the remaining months with the exception of August, during which month no tropical storm has been recorded for the interval 1839-1922. It is the above similarity to hurricanes passing to the west of Noumea which leads some authorities to believe that a secondary source region exists in the Coral Sea. However, it may well be that they are storms which have formed in the western portion of the primary source region and have been unobserved until their later stages in the Coral Sea and Queensland coast.
Locality |
July |
Aug. |
Sept. |
Oct. |
Nov. |
Dec. |
Jan. |
Feb. |
Mar. |
Apr. |
May |
June |
Total |
Queensland |
5 |
— |
4 |
4 |
2 |
8 |
29 |
22 |
29 |
12 |
6 |
6 |
127 |
Northern Territory |
— |
— |
— |
— |
4 |
7 |
7 |
4 |
8 |
2 |
— |
— |
32 |
Total |
5 |
— |
4 |
4 |
6 |
15 |
36 |
26 |
37 |
14 |
6 |
6 |
159 |
Mean paths.—Tropical storms in this part of the world travel in a westerly to a southwesterly direction, then recurve to the southeast decreasing in intensity and losing their identity with the storms of the temperate latitudes. Their mean paths and areas likely to be affected are as depicted on the accompanying charts (Plates XI to XV).
North Indian Ocean tropical cyclones.—The continuity of the North Indian Ocean is rudely interrupted by the projection of the Indian continent into the very center of this body of water. Therefore, as might be expected, the behavior of tropical storms in this region is dependent upon their relative position to this physical barrier. Naturally, then, it is not surprising to learn that two distinct families of storms frequent this portion of the glob, each following their own particular life history dependent upon whether they originate in the Bay of Bengal or in the Arabian Sea. Of the two, those of the Bay of Bengal are much more numerous by the ratio of about 7 to 1.
Source regions (Bay of Bengal cyclones).— The majority of the tropical cyclones affecting the Bay of Bengal originate in the bay itself, commencing late in April in the Sea t)f Andaman near the Nicobar Islands. This zone gradually moves northward so that by the end of June the southern half of the bay is free of storms and the majority of the activity takes place north of 16° north. However, as the monsoon surges southward, the storm “breeding grounds” also recesses southward and late in November is south of 12° north latitude.
Quite often cyclonic storms appearing in the bay arc visitors, typhoons which have moved westward from the China Sea across Burma or Indo-China into the Indian Ocean. This typhoon gateway is open from July to November and is particularly used in September and October, when an estimated one out of every five or six cyclones in the Bay of Bengal is of typhoon origin.
Seasonal frequency.—May through November is the cyclone season, with storms infrequently occurring late in April and December, but with no activity in January, February, and March.
A compilation of cyclones for the period 1891-1937 reveals the following monthly activity:
|
Jan. |
Feb. |
Mar. |
Apr. |
May |
June |
July |
Aug. |
Sept. |
Oct. |
Nov. |
Dec. |
Total |
No. of storms |
0 |
0 |
0 |
14 |
16 |
8 |
9 |
4 |
10 |
20 |
19 |
8 |
106 |
No. attaining cyclone intensity |
0 |
0 |
0 |
6 |
11 |
1 |
2 |
1 |
5 |
8 |
8 |
3 |
45 |
Percentage |
0 |
0 |
0 |
45 |
69 |
12 |
29 |
25 |
50 |
40 |
44 |
38 |
|
A glance at the above table reveals that the greatest number of intense cyclonic storms occur in May and again in October and November. This double maxima of storm frequency is undoubtedly caused by the visitation of the typhoons from the China Sea during the latter part of the season. Also, 69 per cent of the storms forming in May reach cyclone intensity, as compared to only 44 per cent in November. Therefore, May is the month of most activity and, insofar as the mariner is concerned, the worst time for sea operations.
|
Jan. |
Feb. |
Mar. |
Apr. |
May |
June |
July |
Aug. |
Sept. |
Oct. |
Nov. |
Dec. |
Total |
No. of storms |
3 |
0 |
0 |
3 |
11 |
15 |
4 |
0 |
4 |
13 |
16 |
3 |
72 |
No. attaining cyclone intensity |
0 |
0 |
0 |
3 |
9 |
10 |
0 |
0 |
0 |
4 |
11 |
1 |
38 |
Mean paths.—Cyclones in the Bay of Bengal may affect the entire coast of India, including Ceylon and even Burma. Their most common track is northwest and thence northward to the mouth of the Ganges. However, the storms which originate at the head of the bay in June will necessarily have a relatively short path over water and consequently upon passing inland are soon dissipated (Plate XVI).
The typhoon gateway from the China Sea, previously mentioned, occupies a mean position in the vicinity of 20° north in July and moves south with the advance of the season, occupying in November its southernmost position near 10° north. Therefore typhoons enter farther down the bay with the advance of the season.
Infrequently a Bengal cyclone will cross India and become activated in the Arabian Sea. However, the reverse is not true.
Arabian Sea cyclones.—The Arabian Sea tropical cyclones are more irregular in occurrence than those of the Bay of Bengal and as compared to tropical storms of other regions. Thus during a 57-year period (1881- 1937) no storm of force 10 or greater was reported in 13 of the years, whereas in 1902 five such storms occurred. The over-all frequency is three storms every four years.
Source regions.—Storms of tropical origin originate early in the season in the area 5°-10° north and 65°-75° east in the southeastern portion of the Arabian Sea in the vicinity of the Maidive Islands. This source region moves northward to the Laccadives Islands in June and is then approximately 9-14° north and 70-75° cast.
Seasonal frequency.—May through November is the cyclone season, although it is to be remembered that the yearly average, taken over a long period of time, is less than one storm per year. Distribution of those recorded from 1881 to 1937 is set forth in the below table.
The significant feature brought out by the table, aside from the double maximum in June and November, is the total absence of cyclones not only in January, February, and March, but also in July, August, and September. Byers4 explains this absence by the fact that when the interior of India is the warmest and with the occurrence of the semipermanent low-pressure system in the region of the Persian Gulf, the northerly winds of the lower Nile Valley and the Red Sea continue as northwest to west winds across the Arabian Sea. This air flow effectively blocks cyclones from the Arabian Sea during the summer months.
Mean paths.—The paths of these infrequent storms may vary from a west-northwesterly course toward the southeast coast of Oman to a northerly course entering the continent near the mouth of the Indus River (Plate XVI).
South Indian Ocean cyclones and willy- willies.—In contrast to the minor number of tropical storms which make their appearance in the North Indian Ocean, the number produced in the Southern Hemisphere of this ocean is surpassed only by the typhoons of the Western North Pacific. Consequently, the activity of the doldrum belt in the South Indian Ocean should be closely scrutinized during the cyclone period, October to May, if the mariner is not to be caught off guard by these violent vortices of nature.
These storms fall naturally into two categories according to their geographical source regions, and they will be discussed independently of each other as cyclones in the Cocos or Keeling Islands to Madagascar group and as “willy-willies” in the northwestern Australia group.
Source regions.—The primary source region for cyclones of the first category is contained within the area lying between 6-15° south latitude and from the Cocos or Keeling Islands westward to the Seychelles group. This vast expanse of sea area exhibits a seasonal, if not monthly, preference of cyclonic storm formation. Thus at the beginning of the season, in October, the few storms which have been recorded have occurred to the west of the Cocos and to the east of Rodriguez, i.e., between 70-90° east longitude. In November and December the cyclones first become apparent between 50- 70° east longitude, that is, the source region has shifted westward. During the height of the season, in January, February, and March, the source regions apparently contract to an even smaller area between 60-70° east longitude. During this period very few storms occur in the doldrum belt eastward of 90° east. However, at the end of the season, April and May, occasional storms are again to be found in the area between 80-90° east (Plate XVII).
Seasonal frequency.—The period of foul weather resulting from tropical cyclone activity commences late in October and extends through May, with cyclones occurring most frequently from December through March. The statistical record of storms by months for the 67 years, 1847-1917 (three years being omitted owing to lack of reports), is as follows:
Sept. |
Oct. |
Nov. |
Dec. |
Jan. |
Feb. |
Mar. |
Apr. |
May |
June |
July |
Aug. |
0 |
7 |
33 |
58 |
113 |
98 |
68 |
25 |
3 |
2 |
0 |
|
Quantitatively, then, the real cyclone season commences in November, reaches a maximum intensity in January, February, and March, and gradually diminishes during May. Infrequent cyclones make their appearance in October, June, and July; none have been reported in August and September. Based upon the over-all number occurring during the 67 years, seven to eight cyclones may be expected per year.
Mean paths.—Much might be conjectured about the vagaries of these storms of the Southern Hemisphere, since the reports concerning their existence, development, and movement do not possess either amplitude or continuity. However, bearing in mind the apparent monthly fluctuation in the source regions, the mean paths have been depicted on the accompanying chart with the normal initial movement to the west-southwest curving gradually to the southwest and final re- curvature to the southeast. At the beginning of the season the infrequent October storms normally recurve well to the east of Rodriguez; those in November and December following a general path between Reunion and Madagascar and in any event between Rodriguez and Madagascar. During the height of the season, January through March, Reunion and Mauritius bear the brunt of the storms and lie in the mean path. Again late in the season, in April and May, the storm tracks recurve well to the east of Rodriguez.
Infrequently, from December to April, cyclones enter Madagascar or the Mozambique channel, following a southward course and affecting the East African mainland. However, their occurrence is rare and should be treated as individual cases.
At Seychelles cyclones are of rare occurrence and pass 200 miles south of Mahe. Only one cyclone has been recorded over the Chagos Archipelago, in January, 1891. One has been recorded at Zanzibar, in April, 1872.
Northwest coast of Australia willy-willies.— Another portion of the South Indian Ocean inhabited by tropical cyclones is near the northwest coast of Australia. These storms, known as “willy-willies,” originate in the Sea of Arafura near the island of Timor during the months of November through April. They normally travel from east to west south of Timor, recurving to the southwest. They may also originate in the ocean area in the vicinity of the Cambridge Gulf, traveling in a southwesterly direction with continually increasing force (Plate XVIII).
Insofar as may be deduced from land observations, willy-willies display their greatest force in the vicinity of Cossack and Onslow, Australia. Some of the storms change their direction after reaching the coast and pass southward along it. Others pass inland and either disintegrate or, in the case of intense storms whose core reaches to high levels, continue across the continent to the southern coast near the eastern side of the Great Australian Bight, where they may regenerate as a lesser storm.
1. The Hurricane, by I. R. Tannchill, U. S. Weather Bureau.
2. The Hurricane, by I. R. Tannehill, U. S. Weather Bureau.
3. South Pacific Hurricanes, by H. Keeton, U, S, Weather bureau.
4. Byers—General Meteorology, page 424.