lating or pumping of the mercury caused by the disturbed condition of the atmosphere. During the season of tropical storms any interruption in the regularity of West Indian hurricanes. August, 1900 to 1921. West Indian hurricanes. September, 1900 to 1921. West Indian hurricanes. October, 1900 to 1921. FIG. 77. the diurnal oscillation of the barometer characteristic of low latitudes should be considered an indication of a change of weather. The barometer is by no means an infallible guide as a warning much in advance, but after the beginning of a storm it will more or less accurately indicate the rapidity of approach and distance from the center. Its indications should not be disregarded. A long swell evidently not caused by the winds blowing at the place of observation is another warning that should never be overlooked. Frequently a swell from the direction of the storm sets in before any other indication becomes marked. Such swell has in some instances given warning of a tropical cyclone days in advance of its arrival. As the cyclone comes nearer the sky becomes overcast and remains so, at first with a delicate cirrus haze, which shows no disposition to clear away at sunset, but which later becomes gradually more and more dense until the dark mass of the true hurricane cloud appears upon the horizon. From the main body of this cloud portions are detached from time to time and drift across the sky, their progress marked by squalls of rain and wind of increasing force. Rain, indeed, forms one of the most prominent features of the storm. In the outer portions it is fine and mistlike, with occasional showers, these later increasing in frequency and in copiousness. In the neighborhood of the center it falls in torrents. The rain area extends farther in advance of the storm than in the rear. Surrounding the actual storm area is a territory of large extent throughout which the barometer reads a tenth of an inch or more below the average, the pressure diminishing toward the central area, but with no such rapidity as is noted within that area itself. Throughout the outer ring unsettled weather prevails. The sky is ordinarily covered with a light haze, which increases in density as the center of the storm are approaches. Showers are frequent. Throughout the northern semicircle of this area (in the Northern Hemisphere) the wind rises to force 6 or 8-the "reinforced trades"-and is accompanied by squalls; throughout the other semi after the appearance of cirrus clouds, sometimes before, the barometer shows an unmistakable although gradual decrease in pressure. As the clouds grow thicker and lower and the wind increases the fall of the barometer usually becomes more vations to determine the location of its center and its direction of movement should be begun. The average tracks of the different classes of typhoons are the result of a study of 244 of these storms which occurred during the period 1884-1897, and are taken from the report of the director of the Hongkong Observatory for 1897. The relative frequency of each class and the period during which it is apt to occur are given in the following table: IIId 15 IIIe 12} IV aa 8 IVb IVc April and December. October and November. October. July, August, and September. June to October. Most frequent in August and September. May to December. Rare in August. Beginning and end of typhoon season. 4 September 1 to December 1. Most common in November. Beginning and end of typhoon season. Most frequent in May. The appearance of the clouds and their value as warnings of tropical cyclones is described as follows by Faura in Cyclones of the Far East, by Jose Algue, of the Manila Observatory: Long before the least sign of bad weather is noticeable and in many cases when the barometer is still very high-being under the influence of a center of high pressure, which generally precedes a tempest-these small isolated clouds (cirri, little clouds of a very fine structure and clear opal color resembling elongated feathers) appear in the upper regions of the atmosphere. They seem to be piled up on the blue vault of heaven and drawn out in the direction of some point on the horizon toward which they converge. The first to present themselves are few in number but well defined and of the most delicate structure, appearing like filaments bound together but whose visibility is lost before they reach the point of radiation. We often had an opportunity to watch them at the observatory of Manila, when the center was still 600 miles distant. The best times for observing the cirri are sunrise and sunset. If the sun is in the east and very near the horizon, the first clouds which are tinged by the solar rays are the cirro-strati which precede the cyclone, and they are also the last to disappear at sunset, inasmuch as they overspread the horizon. Such times are the best for determining the radiant point of the cloud streaks and at the same time for ascertaining the direction in which the center lies. Later on the delicacy of form, which characterizes this class of clouds in its earliest stages, is lost, and the clouds appear in more confused and tangled forms, like streamers of feather work, with central nuclei, which still maintain this direction, so that the point of radiation can still be detected. In order to ascertain approximately the direction in which the center is advancing in its movement of translation, it is necessary to determine the changes of the radiant point at equal intervals of time and to compare them with the movements of the barometer. If the point of convergence does not perceptibly change its position, but remains fixed and immovable for a long time, even for several consecutive days, it is almost certain that the tempest will break over the position of the observer. In this case the barometer begins to fall shortly after the first cirrus clouds have been observed and sometimes even before. At first it falls slowly, without completely losing the diurnal and nocturnal oscillatory movements, but changing somewhat the hours of maximum and minimum. The daily reading is observed to be each day less than that of the preceding day. That part of the horizon in the direction of the storm begins to be covered by a cirrus veil, which increases slowly until it forms an almost homogeneous covering of the sky. This veil is known by the name "cirro-pallium" of Poëy, and is that which causes the solar and lunar halos, which are never absent when a storm approaches. Beneath the veil a few isolated clouds, commonly called "cotton," appear. They are much more numerous and larger on the side lying toward the storm, where they soon appear as a compact mass. At such times the sunrises and sunsets are characterized by the high red tint which the clouds assume, resembling a great fire, especially in the direction of the cyclone. The wind remains fixed at one point, showing only a few variations, which are due principally to the squalls, which continually exert their force within the limits of the storm. The low, or "cotton," clouds successively and from time to time a lull ensues, the cirrus veil remaining, and likewise the hurricane bank of clouds, which seems fixed to the same spot in the direction of the storm. This state of the atmosphere continues until the bank of clouds invades the point of observation, in which case the squalls will be continuous and the wind will increase in violence each moment. 480. TO FIX THE BEARING OF THE STORM CENTER.-It is very important to determine as early as possible the location and direction of travel of the center. While this can not be done with absolute accuracy with one set of observations, a sufficiently close approximation can be arrived at to enable the vessel to maneuver to the best advantage. Since the wind circulates counterclockwise in the Northern Hemisphere, the rule in that hemisphere is to face the wind, and the storm center will be on the right hand. If the wind traveled in exact circles, the center would be eight points to the right when looking directly into the wind. We have seen, however, that the wind follows more or less a spiral path inward, which brings the center from eight to twelve points to the right of the direction of the wind. The number of points to the right may vary during the same storm, and as the wind usually shifts in squalls its direction should not be taken during a squall. Ten points to the right (left in south latitude) when facing the wind is a good average allowance to make if in front of the storm, but a larger allowance should be made when in the rear. If very near the center the allowance should be reduced to eight or nine points in the front quadrants. Based on the average, the following rules will enable an observer to fix approximately the bearing of the storm center: In the Northern Hemisphere, stand with the face to the wind; the center of the cyclone will bear approximately ten points to the observer's right. In the Southern Hemisphere, stand with the face to the wind; the center of the cyclone will bear approximately ten points to the observer's left. It may be noted here that the storm center almost always bears very close to eight points from the direction of movement of the lower clouds of the cyclone. Therefore, when the direction of movement of the lower clouds can be observed it may serve as a more accurate indication of the bearing of the center than does the direction of surface wind. Further assistance in locating the approximate position of the storm center may be obtained in some instances by observations of the clouds. When the sky first becomes overcast with the characteristic veil of cirrus the storm center will most probably lie in the direction of the greatest density of the cloud. Later when the hurricane cloud appears over the horizon it will be densest at the storm center. The hurricane cloud, sometimes called the "bar of the cyclone," is a dense mass of rain cloud formed about the center of the storm, giving the appearance of a huge bank of black clouds resting upon the horizon. It may retain its form unchanged for hours. It is usually most conspicuous about sunrise or sunset. When it is possible to observe this cloud the changes in its position at intervals of a few hours will enable the observer to determine the direction of movement of the storm. Although the approximate bearing of the storm center is a comparatively easy matter to determine, and the direction in which the center is moving may be estimated with fair accuracy from the charted paths of similar storms (see figs. 77 and 78) it is by no means an easy matter for the observer to estimate his distance from the storm center. The following old table from Piddington's "Horn Book" may serve as a guide, but it can only give an imperfect estimate of the distance and too much reliance must not be placed upon it: This table assumes that the vessel is hove-to in front of the storm and that With storms of varying area and different intensities the lines of equal barometric pressure (isobars) must lie much closer together in some cases than in others, so that it is possible only to guess at the distance of the center by the height of the mercury or its rate of fall. A further source of error arises because storms travel at varying rates of progression. In the Tropics this ranges from 5 to 20 miles per hour, generally decreasing as the storm track turns poleward and recurves, increasing again as it reaches higher latitudes. In the North Atlantic its rate of progression may amount to as much as 50 miles per hour. Within the Tropics the storm area is usually small, the region of violent winds seldom extending more than 150 miles from the center. The unsettled state of the barometer described heretofore is usually found in the area between 500 and 1,000 miles in advance of the center. This gives place at a distance of 300 or 400 miles to a slow and steady fall of the mercurial column. When the region of violent winds extending about 150 miles from the center is reached, the barometer falls rapidly as the center of the storm comes on, this decrease within the violent area sometimes amounting to 2 inches. Because of this very steep barometric gradient the winds blow with greater violence and are more symmetrically disposed around the center of a tropical cyclone than is the case with the less intense cyclones of higher latitudes. After a tropical cyclone has recurved it gradually widens out and becomes less severe, and its velocity of translation increases as its rotational energy grows more moderate. Its center is no longer a well-defined area of small size marked by a patch of clear sky and near which the winds blow with the greatest violence. Out of the Tropics the strongest winds are often found at some distance from the center. 481. HANDLING THE SHIP WITHIN THE STORM AREA.-If from the weather indications given above and such others as his experience has taught him, the navigator is led to believe that a tropical cyclone is approaching, he should at onceFirst. Determine the bearing of the center. (See art. 480.) Second. Estimate its distance. (See art. 480.) Third. Plot its apparent path. The first two of the above determinations will locate the approximate position of the center, which should be marked on the chart. The relation between the position of the ship and the position and prospective track of the center will indicate the proper course to pursue (a) to enable the vessel to keep out of or escape from the dangerous semicircle and to avoid the center of the storm; (b) to enable the vessel to ride out the storm in safety if unable to escape from it. Should the ship be to the westward of the storm center before the path has recurved, it may be assumed that the latter will draw nearer more or less directly. It then becomes of the utmost importance to determine its path and so learn whether the vessel is in the right or left semicircle of the storm area. The right and left semicircles lie on the right and left hands, respectively, of an observer standing on the storm track and facing in the direction the center is moving. Prior to recurving, the winds in that semicircle of the storm which is more remote from the Equator (the right-hand semicircle in the Northern Hemisphere, the left-hand semicircle in the Southern) are liable to be more severe than those of the opposite semicircle. A vessel hove-to in the semicircle adjacent to the Equator has also the advantage of immunity from becoming involved in the actual center itself, inasmuch as there is a distinct tendency of the storm to move away from the Equator and to recurve. For these reasons the more remote semicircle (the right hand in the Northern Hemisphere, the left hand in the Southern Hemisphere) has been called the dangerous, while that semicircle adjacent to the Equator (the left hand in the Northern Hemisphere, the right hand in the Southern Hemisphere) is called the navigable. In order to determine the path of the storm and consequently in which semicircle the ship finds herself, it is necessary to wait until the wind shifts. When this occurs, plot a new position of the center 10 points to the right of the new direction of the wind as before, and the line joining these two positions will be the probable path of the storm. If the ship has not been stationary during the time between the two sets of observations (as will indeed never be the case unless at anchor), allowance must be |
