Popular Science Monthly/Volume 15/September 1879/The Birth, Life, and Death of a Storm


By ROBERT H. SCOTT, M. A., F. R. S., Etc.

WHEN we are asked to give an account of the birth of a storm, we are reluctantly compelled to admit that our storms are, almost without exception, foundlings, and that, as the precise conditions to which they owe their origin are, for the most part, shrouded in uncertainty, warm discussions at times arise as to the parish whence they have set out on their wanderings.

Dove said long ago that storms were due to the interference of the polar current or the east wind with the equatorial current or west wind. He gave the winds these names, because on his views the east winds really consisted of air flowing from the north or south pole toward the equator, which was modified in the direction of its motion by its change of latitude; while west winds were really due to air endeavoring to make its way back to the pole from the equator, whose coarse was in its turn modified by its moving from lower to higher latitudes. To the conflict of these two grand currents, east and west winds, Dove attributed all our storms; but he did not attempt to explain how the currents came into collision.

These views, however correct on their cosmical principles, have been superseded, of late years at least, as regards the explanation of our winds, by the modern views of the relation between the wind and the distribution of barometrical pressure; but, unfortunately, we still remain in comparative ignorance of the ultimate causes to which this distribution of pressure, or the rise and fall of the barometer, are due. To give some conception of the existing difference of opinion on these fundamental principles of our science, I may say that while some authorities maintain that the force of the wind in a hurricane is caused by the amount of barometrical disturbance which accompanies it, others hold that the fall of the barometer at the center is itself, in great measure, due to the centrifugal force of the revolving mass of air.

Of the various theories which have been propounded to account for storms, which are generally more or less cyclonic in their character, I shall only mention four:

1. Some authorities, and among them our own countryman the Rev. Clement Ley, attribute the formation and subsequent progress of a storm to the condensation of moisture, but they apparently ignore the fact that many of our very heaviest rains do not give rise to cyclonic disturbances of serious character. For instance, when on April 10 and 11, 1878, 4·6 inches of rain fell at Haverstock Hill, we had no storm of wind at all. In partial confirmation of this view, Professor Mohn, of Christiania, points to the accidental condensation of moisture caused by the contact of a mass of damp air with the surface of an extensive snow-field as a possible cause of a storm. About the sixty-first parallel of latitude the glacier region of Justedal stretches for several miles along the coast of Norway, and this has occasionally been known to exert an influence in increasing the intensity of an existing cyclone, and even in some instances has appeared as the center of a newly formed depression.

These gentlemen, moreover, rely greatly on the fact that the rain area which accompanies every cyclonic system is roughly oval in shape, with its longer axis extending in the direction in which the system is advancing, and that by far the greatest amount of rain falls in front of the storm. They do not, however, explain the fact that very heavy rain frequently occurs on the northern side of a depression, where the wind is easterly, and that this circumstance does not indicate a northward motion of the system.

The most serious objection to this theory is, however, that first stated, that not only do the heaviest rains not come with the severest storms, but that frequently they are observed in times of nearly absolute calm.

2. The second theory to which I shall refer is the mechanical one, most strongly urged by Mr. Meldrum, of the Mauritius, whose investigations into the weather over the Indian Ocean have led him to the belief that every cyclone is generated in the intervening space between two oppositely flowing currents of air, of which the easterly moving stream, speaking in the most general terms, lies on the polar side of the westerly wind. Such a disposition of the currents would be that which would naturally arise were the cyclone once formed.

This view is called seriously in question by Messrs. Blanford and Eliot in their discussion of recent cyclones in the Bay of Bengal, which they have been able to study from very early stages, and in which they fail to see evidence of the preexistence of two, and only two, determinate currents.

Another serious objection to this theory is that it does not assign a vera causa sufficient to give the first impetus to the barometrical fall and the rotatory movement of the air.

3. A third theory of the origin of these storms is that which is strongly urged by M. Faye, in Paris, and is to the effect that, as interfering currents in rivers give rise to vortices which extend from the surface downward into the water, so all our water-spouts, trombes, and even the largest tropical hurricanes must be all formed in the upper regions of the atmosphere, and extend downward to the earth: the force which gives them their onward motion being supplied by the upper currents.

It is sufficient to say that this theory has not met with acceptance from any practical meteorologist, while it is directly controverted by recent investigations into the motion of cirrus clouds, which show beyond a doubt that the motion of the upper currents of air over a cyclone is outward, and not inward, as the descending theory would demand.

Moreover, some of our readers may have noticed, in "Nature" of January 16th, a notice, copied from the "Times," of the formation on the Lake of Geneva, on January 2d, of a veritable small water-spout, forty feet high and ten yards in circumference, by the meeting of two winds, known locally as the Föhn and the Bise, on the surface of the lake. Here the water-spout was raised, and did not descend from the clouds.

4. The last theory we shall notice is that of the late Mr. Thomas Belt, who seeks for the origin of the disturbance on the ground, and, like M. Faye, assigns the same explanation to the smallest dust-whirl eddies and the largest storms which sweep over the earth.

This theory assumes as the first cause the heat of the sun. The heat-rays pass through the atmosphere without warming the upper strata, and so Mr. Belt supposed that over a sandy soil a mass of air close to the ground might rise in temperature much higher than the superincumbent layers of the atmosphere. The lower strata would therefore become lighter, and a condition of unstable equilibrium would arise. This, however, could not last for ever, and, sooner or later, the heated lower air would burst up, and the ascending column thus produced would be the nucleus of the nascent cyclone.

The difficulty in accepting this explanation is, that we should like some ocular evidence of such a sequence of conditions. The supporters of the theory, however, point to accredited instances of the formation of whirlwinds over volcanoes like Santorin, and over extensive fires like those of Carolina canebrakes.

In confirmation of these views of the effect of solar heat in producing a depression, I may cite an investigation by Dr. Hamberg, of Upsala, who has found that in July, 1872, after a prevalence of intensely warm weather in southern Sweden, pressure gave way over the heated area, the isobaric lines following the trend of the coast; and a rotatory movement was thereby generated in the atmosphere above it, resulting in a perfectly formed cyclone which passed on over northern Finland. It would appear, therefore, that the production of a cyclonic disturbance may be attributable to more than one agency, as all the theories mentioned have some facts in their favor.

Leaving, then, this abstruse and imperfectly understood line of inquiry, let us proceed to a subject which yields us results of more immediate practical utility: the character and history of the storms when they have once started on their travels. I shall commence by saying that a greater mistake can not be made than to assert that all storms are distinctly connected with cyclonic disturbances.

The force of the wind depends on differences of atmospherical pressure over a given area, and the only reason why storms are generally associated with cyclones is that these systems afford us the most serious instances of disturbances of atmospheric equilibrium, and consequently of differences of pressure, which are met with on the globe.

At any place where an area of relatively high pressure comes into close proximity to an area of relatively low pressure, a gale will result, and so a storm may be due just as much to the rise of the barometer in one region as to its fall in an adjacent district. For the same physical reason, however, that the eddies in a river extend downward, and the water does not pile itself up in a peak, the normal disturbance of atmospherical equilibrium is the appearance of one of these vortices with pressure decreasing rapidly toward the center. Wherever there is a rapid decrease there is a steep gradient, and consequently a strong wind.

Defining the term cyclone, in its very widest acceptation, as indicating a region of diminished pressure, round and in upon which the air is moving along paths which are more symmetrical all round the center the more perfect is the circular form of the system, we must at once see that not every cyclone is accompanied by a storm. The fact is, that the direction and force of the wind are regulated by the difference of barometrical pressure over a given distance, and not in any way by the actual height of the barometer at the station at which the storm is felt, or by the distance of that station from the point where the barometrical reading for the time being is the lowest.

This explanation of wind-motion is almost the only new principle which has been recognized in our science during the present generation, and its practical importance is daily forcing itself more and more into public notice with the development of weather telegraphy. It is usually known under the name of Buys Ballot's Law, and is stated as follows: "Stand with your back to the wind, and the barometer will be lower on your left hand than on your right." The truth of this law is evident to any one who looks at a weather chart; but the Dutch Professor, after whom it is named, though he justly claims the credit of having persistently advocated the acceptance of this relation of the wind to the distribution of pressure, was not by any means the first to discover it.

The final result of all the inquiries into the question is, that on the mean of all winds the angle between their direction and the tangent to the isobar at the place is about 20°.

These principles of wind-motion have a most important bearing on the theory of the motion of the air in hurricanes and typhoons. The old popular idea of these phenomena is, that the air blew round and round the central calm in circles, so that any sailor caught in one of these storms could at once know that when he was hove-to, if he looked in the wind's eye, the center bore eight points to the right in the northern hemisphere, and to the left in the southern; or, what is the same thing, if he was scudding before the wind, the center would lie exactly on the starboard beam in the northern and on the port beam in the southern hemisphere.

Modern meteorologists, however, almost with one voice, declare for a spirally incurving movement as the most probable behavior of the wind, as would be indicated by the angle which its direction makes with the isobars as just explained; but this view presents no novelty, for it was first stated about forty years ago, and Piddington, in his "Sailor's Hornbook," says that even Redfield, when propounding his "Law of Storms," stated:

"I have never been able to conceive that the wind in violent storms moved only in circles. On the contrary, a vortical movement, approaching to that which may be seen in all lesser vortices, aërial or aqueous, appears to be an essential element of their violent and long-continued action, of their increased energy toward the center of axis, and of the accompanying rain. In conformity with this view, the storm-figure on my chart of the storms of 1830 was directed to be engraved in spiral or involute lines, but this point was yielded for the convenience of the engraver."

We see, therefore, that when we trace back to its origin the belief that any storms are really circular, we find that it was "the convenience of an engraver" which decided the question.

It may be safely asserted that there does not exist, for a single instance of a West Indian hurricane or China Sea typhoon, a sufficiency of evidence to convince any unprejudiced investigator as to what was the true path of the air in the storm. To show this path beyond the possibility of doubt, we require a considerable number of simultaneous observations taken on different sides of the storm center. These, however, were not forthcoming in the case of a single storm described by Redfield, Reid, or Piddington, so that the authority of the founders of the law of storms can not be cited as decisive of the question.

This suggestion of spiral motion must of course modify the simple rule for a ship scudding, of looking in the wind's eye, and taking eight points on the starboard or port side for the storm center, and indicates the probability that the true position of that spot will be at least two or three points ahead of the bearing given by that rule, so that the ship, if scudding, may be gradually approaching the most dangerous part of the storm.

The recent investigations of Mr. Meldrum, which have been thoroughly confirmed by Captain Toynbee's examination of the Nova Scotia storm of August 24, 1873, lead to the suspicion, not to use a stronger word, that these cyclonic storms are not symmetrical at all, and that at some parts of the system the wind blows directly toward the center, so that for a ship in such a situation, and scudding before the wind, the center would lie right ahead.

This is a subject which requires most careful study, in order to see whether or not the time-honored rules for handling ships in rotating storms require modification.

I shall now leave the subject of the air-motion, and proceed to describe the phenomena of a cyclonic disturbance when it passes over us. In the first place, very few of them, in these latitudes, exhibit much approach to a circular shape, as regards the course of the inner isobars, and we may say that none of them develop equal violence in all segments. The reason of these differences in the force of the wind is to be found in the distribution of pressure in the vicinity of the storm area, for if on any side of that area there exists a region of high barometer readings, on that side steep gradients will be produced, and of course proportionably great violence of the wind. The actual weather phenomena of a typical cyclonic disturbance, if plotted on a diagram, show very clearly how cloud and rain prevail over the whole front of the system, and how in the rear, where the wind is northwesterly, the sky clears up. There is one fact worth remembering about these storms, and that is, that just before the sky clears a very smart squall of rain frequently comes on; so that we get this practical hint: if, during a westerly gale, we find the rain becoming exceptionally heavy, we may look for the weather speedily to clear up.

Such a diagram also shows us that it is quite a mistake to consider all east winds as dry ones, for in a cyclonic system the cloud area extends on the northern side, where the wind is easterly, nearly as much as on the southern, where the wind is from the westward. In fact, many of our wettest days occur with easterly winds, when one of these depressions passes to the south of the station where we may be.

I shall now proceed to give a slight sketch of what we have learned of the movement of storms. This, as far as we can see, is regulated by the position of the areas of high pressure, or, as they are called, the anticyclones. This is a term introduced about fifteen years ago by Mr. Francis Galton, to indicate an area of excess of pressure out from which the air is slowly whirling with a motion opposite to that which it has in cyclones. If we find an anticyclonic area existing over any region, we know that the cyclonic disturbances will skirt round it and develop their strongest wind on the side which lies closest to the district of high pressure.

Thus if the anticyclone lies over France, the cyclonic disturbances will move from west to east over the British Isles. If the area of high pressure lies over England, the depressions will sweep outside the Scotch coast, and reach Norway north of the sixtieth parallel. If the anticyclone lies to the westward, and the pressure is higher in Ireland than in Great Britain, there is danger of northerly gales on the east coast of England, from cyclonic disturbances traveling southward over the North Sea.

In every case the cyclone moves with the prevailing wind along its track.

Unfortunately, we know very little about the rate at which these storms advance, some of them moving at the extraordinary speed of fifty or sixty miles an hour, as for instance that of March 12, 1877; while others, like the West India hurricanes, do not attain one fourth of that rapidity of translation. It is remarkable that the rate of progress bears no relation to the intensity of the storm, the slow-moving tropical hurricanes being infinitely more violent than many of our rapidly-moving disturbances; although the storm already mentioned in March, 1877, was severe enough, at least in the north of France, to satisfy any requirements.

As regards the distance which storms have been known to travel, I may cite a very long-lived storm, which lasted nearly a fortnight in August, 1873, and which was traced along its course by my friend Captain Toynbee, by means of the logs of two hundred and sixty ships which were in the Atlantic during its continuance. Its history will be found in the last published work of the Meteorological Office, "The Weather over the Atlantic Ocean during August, 1873." This particular storm wrought immense damage on the coast of Nova Scotia. It did not, however, travel as far as Europe, having disappeared in the neighborhood of Newfoundland. In fact, very few storms have really been proved to maintain their individuality during their transit. Professor Loomis, an American meteorologist, who has devoted much attention during the last twenty years to the connection between European and American weather, has very recently published a paper on the results of discussion of two years' daily synoptic charts of the Atlantic. During that interval thirty-six areas of depression were traceable across the Atlantic, that is, at the rate of eighteen a year. Testing these by wind reports from England alone, he finds that the chance that a storm center coming from the United States will strike England is only one in nine; of its causing a gale anywhere near the English coast it is one in six; while the chance of its causing a strong breeze is an even one.

This brings us to a subject which has attracted an immense amount of public attention in this country and in France: the practical value of the warnings which have been sent over by the "New York Herald" during the last two years. By "practical value" I mean the value to our fishermen and coasting sailors, for whose benefit, more than for that of seagoing men in large vessels, the whole system of storm-warnings has been called into being. It is evident that a warning which is locally unfulfilled may mean a loss of some hundreds of pounds to a fishing fleet; and although the storm to which it referred may have reached some parts of the coasts of Europe, yet if it did not visit the precise district where the fishing was being prosecuted at the time, the fishermen in that district were not benefited by the warning. On the contrary, they were the worse for having received it, on the old principle that "Wolf! wolf!" should not be cried too often.

Of course, every word that I here say as to the usefulness of warnings is just as true with reference to warnings issued by our own office in London as to those of the "New York Herald," but these latter are often very general in their scope. They speak occasionally of a storm reaching the British Isles and France, and affecting Norway. This haul of the net embraces 25° of latitude, from 4°5 to 70°, and it is an unheard-of thing that a gale should prevail simultaneously over such an immense tract of coast, so that on each occasion the seamen in many harbors can not derive immediate benefit from the publication of so vague an announcement.

It is one thing for a scientific man to say that he can recognize the presence of the predicted cyclone on our coast—Professor Loomis admits that the chances are even that he should do so—but it is a totally different matter to prove that a gale which begins two days before or two days after the time of a predicted storm, is really the very disturbance which left the American coasts.

The experience of those who have studied cyclone tracks in northern Europe shows that in winter, on an average, a cyclonic disturbance visits some parts of those regions every fourth clay, so that, if a warning were announced once a week regularly, there would be nearly a certainty of some sort of a fulfillment.

The results of a most careful comparison of these warnings with the weather experienced by us during the years 1877-'78 are given by the following percentage figures:

1877 1878
Absolute success 17·5 42·5 27·0 45·0
Partial success 25·0 18·0
Partial failure 15·0 57·5 10·0 55·0
Absolute failure 42·5 45·0

In order to obtain so favorable a result as forty-five per cent, of general success, great allowances have been made. Thus it has been considered an absolute success if a gale was felt on any part of the coast, whereas the prediction was for all parts; and when three separate storms were predicted in one telegram, none of which arrived, only one failure has been counted.

It is, therefore, pretty clear that these warnings have not, as yet, proved themselves to be of much practical utility to our coasting trade and our fishermen. The question is a most interesting one, and although a satisfactory solution of it has not been attained, we need not despair; but we should attack it from the scientific side, and discuss the results in a calm, dispassionate spirit, and through some other medium than that of letters to newspapers.

Let us now leave these American warnings, and see what we know about the movement of storms over western Europe, which is the problem which most immediately concerns us here. The illustration has often been used that meteorologists, in issuing storm-warnings, and having to estimate the direction and rate of motion of every storm the instant it shows itself in their neighborhood, are in the position of astronomers expected to assign the path of a comet from the first glimpse they get of it through a break in a cloud—a problem which all will allow to be impossible of solution. Accordingly, great interest attaches to the attempts made from time to time to lay down principles for forecasting the motion of the disturbance.

I have already stated that, as a general rule, the cyclones move round the anticyclones; but this principle requires for its application to storm-warning purposes, access to charts embracing a very considerable extent of the earth's surface. These are very difficult for Englishmen to obtain, as our own daily charts are very limited in area, and frequently do not exhibit even the whole extent of a single cyclonic depression, much less its relation to the distribution of pressure all about it. For those, however, who can consult such charts it is possible, so to speak, to take their stand at a higher point of view and survey the conditions prevailing, say over Europe, on any given day.

If the amount of change in the pressure or of rise and fall of the barometer during the preceding night be plotted every morning on such a chart, it is found that the path of the system for the day does not lie directly toward the region where the greatest fall has occurred during the night, but is regulated to a certain extent by the direction of the line drawn from the point of greatest fall to that of greatest rise.

Another theory of storm-motion, strongly held by those who attribute all our storms to condensation of vapor, is that the track of the depression is always directed toward the region where the air is dampest. This principle, like that just noticed, can hardly be turned to account in this country for our own practical benefit, inasmuch as the whole of these islands appear to be almost equally damp, owing to the proximity of most of our telegraphic reporting stations to the sea.

Other suggestions have been made in various quarters, with the view of throwing light on this very important subject; but we can not say that the results have met with general acceptance, and the matter urgently demands further study.

I must now come to the final portion of my theme—the death of a storm; and on this subject, unfortunately, I have very little to say. As we have not been able to produce evidence of the birth of a storm, so have we never been lucky enough to find any one who was in at the death. In fact, some French meteorologists have hazarded the statement that storms can travel all round the world until at last they travel off it.

Storms have been traced from the Pacific coast of North America across the Atlantic; but these instances are necessarily rare, and, as far as European experience goes, no storm arriving from the Atlantic ever travels far into Russia. This fact is, of course, very much in favor of the condensation theory of storm generation, which has already been noticed. The advocates of this view plead very plausibly that, as the moisture in the air is the food of the storm, so, where that moisture is deficient, the storm dies of starvation.

We may, however, point out to them that eddies in a river and dust-whirls at street corners waste and wane without any assistance from vapor condensation.

In conclusion, though it is a humiliating confession for us to make, meteorologists are as yet entirely in the dark as to the reasons why one depression fills up while another becomes deeper. As I have already stated, no meteorologist is able to give a straightforward answer to the simple question, What causes the barometer to rise or fall?—Popular Science Review.

  1. Founded on a lecture delivered by the author at the London Institution, February 3, 1879.