Popular Science Monthly/Volume 66/April 1905/Sunspots and Weather


By Professor ERNEST W. BROWN,


IT is perhaps not far from the truth to say that the most pressing problem, as far as the daily needs of humanity are concerned, consists in finding some method of predicting the weather. Of the great value which a solution would have it is not necessary to say a word. Yet, in spite of many attempts, the chief problem is as yet unsolved. One may go a step further and say that there is at present no indication of any approach to a solution in the work which has hitherto been published. It is possible to forecast the weather with fair accuracy for a day or two in advance, but the methods by which this is done do not appear to have any bearing on the problem of predicting what the weather will be next week, next month or next year. The latter question must be approached in a quite different manner, and it is the object of this article to show the degree of success attained by one attempt to disentangle the wider fluctuations of climate which nearly every region of the earth's surface shows from year to year.

Weather and climate, like all other phenomena of nature, are nothing more than particular cases of the interaction of certain laws. Properly speaking, chance plays no part in their variations. That the motions of the atmosphere can be classed under well-known mechanical principles, there can be little doubt; that the various types of weather and climate are capable of being deduced from those principles with a sufficiently powerful method of analysis is equally certain; but whether such a method has yet been invented or is in process of being discovered at the present time is open to doubt, if we may judge from the very little that has hitherto been done in the direction of deducing the observed phenomena from the laws which govern them. The difficulties presented by this natural and logical process have caused meteorologists to turn to some other method of restoring order out of the chaos. Instead of deducing the phenomena from the general laws, attempts are being made to bridge the chasm which separates them by starting with the observations and trying to find out if some kind of order can be discovered—a support on which it may be possible to continue the bridge towards the opposite bank.

The procedure to be adopted when the latter method is used is sufficiently simple. Large numbers of observations of temperature, barometric pressure, rainfall, etc., are made and recorded. We may use one of two methods to examine them. The first is to group the observations in various ways until we discover some group or groups into which the observations appear to fall. This method, however, entails such an enormous amount of labor that it has rarely been adopted. It is only successful in the few cases where the general nature of a group appears without much trouble. The second method is to assume that the observations must run in series which are repeated after definite intervals of time. These intervals of time are in reality obtained from the general physical principle that if any force is periodic, that is, if its fluctuations are repeated after a definite period of time, then certain of its effects are also repeated after a definite period of time, which is the same as that of the force. The first method is an attempt to construct the bridge by starting from one bank and building piers or supports on which the structure may be gradually extended across the chasm; the second method resembles a series of cables thrown across in the hope that they may be attached firmly enough on each bank to bear the weight of the bridge.

There is no difficulty whatever in attaching two of these cables to both sides. Whatever other causes operate, there can be little doubt that the sun must play some part in governing the climate of any particular place. The day and the year are therefore marked out beforehand as periods into which observations of temperature can be grouped, and these periods are fully confirmed. If the temperatures be recorded every hour out of the twenty-four for a large number of days, and the average temperature for each hour be formed by adding all the observations for that hour and dividing by the number of them, we obtain a series of twenty-four average temperatures. When the number of days on which observations are recorded is great enough, these averages will show a regular change rising to a maximum, for most places, in the early afternoon and descending to a minimum in the early morning hours. A similar method followed with the average temperature, say, for each day, will show a yearly period in the observations.

We thus get a series of average temperatures for each hour of the day and for each day of the year. But these are only averages and they only present regularity when very large numbers of observations have been used in forming them. At any particular time the difference of the actual temperature from the average temperature for that time may be as large or larger than the greatest difference of the averages during the period. It is evident then that other cycles must be sought which, by their combined effect, will give the actual temperature at any particular time. And what has been said about temperature can be said to a greater or less degree about the other phenomena, such as rainfall or barometric pressure, which it has been customary to record. We find periods of a day and a year, but the differences of the averages from the actual observations are very considerable.

The next step consists in trying to find out whether cycles can be discovered in these differences, and for this purpose the same method is followed. But the periods to be examined are by no means obvious. The time of revolution of the moon, under the supposition that the moon has any influence, is not found in meteorological observations, or at any rate the effect is so small as to be negligible in comparison with what we want to find, and no other astronomical period seems very likely to find a counterpart in the records. There is, however, one period which at first sight appears to be a possible one in meteorological observations, namely, that known as the sunspot period. Before going into the evidence for its existence as a climate factor, it is necessary to say a word or two about the effects of periodic changes.

These changes may be roughly divided into two classes, those of short period and those of long period. From the former we should expect changes from day to day or every few days, constituting what we usually call the 'weather.' Nothing at all is known of the existence of any such period, all attempts to trace one having ended in absolute failure. The long period changes are in general those which affect the character of a season and which show the difference between one season and the next, or the variations which a given season of the year will show in several successive years. In countries where the most important interests are manufacturing and ocean transportation, the 'weather' is the first consideration; in lands where the agricultural interests are predominant, although the weather is important, it is more essential to have an idea of the climatic or seasonal changes. Thus if the sunspot period could be shown to have a large influence on the climate of any region, we should be able to forecast to a corresponding extent the general characteristics of future seasons, and lay plans accordingly. The practical value of such an investigation is obvious, but its scientific interest is not less if it leads to a more accurate knowledge of the laws of nature. The case therefore deserves examination.

In order to establish any connection between sunspots and meteorological changes, two things are necessary. First, the periods in which the sunspots can be grouped must be discovered and their relative importance determined. Second, the meteorological records over long intervals of time must be examined in order to see whether they exhibit similar cycles of change and whether those cycles, supposing that they exist, are sufficiently well-marked to be considered important factors in the variations of the seasons at any place from year to year.

Although the existence of sunspots has probably been known from the earliest times, the fact that they undergo periodic changes in number and size was not recognized until long after the discovery of the telescope, and the law of the change was not determined with anything like accuracy until the middle of the last century. Accurate observations of their number and size were only commenced about 1830, but from about 1750 to 1850 enough material had been gathered to show that a maximum occurred on the average every eleven years. In the last two or three decades the use of the spectroscope has added greatly to our knowledge of their nature and motions, and the photographic camera has enabled astronomers to add permanent records of the state of the sun's surface at any time to the numerical estimates which were the chief contribution of the earlier observers. Such pictures are of special value whenever, as in the case of sunspots, much depends on the personal equation of the observer and still more on the particular method used in forming an estimate of them. Thus there is a period extending over the last seventy years in which continuous observations have been made, a period of another seventy years earlier in which the observations are worth discussing, although much less reliance can be placed on them for accurate deductions, and a still earlier period of about a hundred years from which a little doubtful information can be gleaned.

The net result which has been deduced from this series is an average period of 111/9 years between the maxima or times of greatest sunspot activity. But this average period by no means represents all the facts. The time between two consecutive maxima has been as long as 17 years and as short as 8 years, while some maxima are marked by much larger and more numerous spots than others. The average period between the minima or dates of fewest sunspots is rather less irregular, since it has never been known to differ by much more than two years from the mean period of 11 years. There is some evidence that these stronger and weaker maxima themselves run in a cycle, but the period is very doubtful. Wolf, who devoted most of his life to the investigation of sunspots, deduced a period of 55 years for this longer cycle, while there are later determinations of 60, 351/2 and 61 years. Further, the curves representing the number of spots from day to day or from month to month show many other irregularities which have up to now defied any attempt to group them in any regular order. Thus we have a well-marked average period of about 11 years and doubtful ones of about 35 and 60 years. These are therefore the periods into which the meteorological records are to be grouped in order to inquire whether there is any connection between the two sets of phenomena.

There is good reason to believe that a period of about 35 years is to be traced in certain of the meteorological records, and therefore it is important to know how much dependence can be placed on the evidence for the existence of a similar period in the sunspot observations. In 1900 Dr. Lockyer took Wolf's sunspot numbers from 1833 to 1900; these numbers show six maxima and six minima of the average eleven-year period. But the interval from a minimum to the next following maximum is not the same for each of these six periods; it varies from 31/2 years to about 41/2 years. The six numbers representing these intervals were arranged according to their proper dates and it was noticed that they could be made to fit in with a periodic change of about 35 years in length. Again, the area of the sunspot curve from one minimum to the next was found, and another set of six numbers was obtained, the last five of which would again fit into a 35-year period. The two series corresponded well with each other if we excepted the first of each, the reason for the exception being that there was a doubt as to the observations from which the first number of the second set is obtained having been made on the same plan as the others. Unfortunately, the evidence thus presented is far from conclusive. There is nothing unnatural in the two series agreeing with one another as far as a period is concerned since they were deduced from the same set of observations. The difficulty arises in the attempt to find the length of the period from five or six numbers spread over rather less than two revolutions of the cycle. One would prefer to say that this result was a reason for further investigation rather than that it proved anything definite as to the existence of a 35-year period.

Far more numerous are the difficulties which beset an examination of the conditions on the earth's surface. In the first place, accurate observations are practically confined to the latter half of the last century, and these have been made chiefly in the northern temperate zones where the daily and weekly changes are apt to be very irregular and violent and where the local conditions frequently exercise much influence in determining the weather or climate of a particular place. To obtain averages free from these local and temporary conditions requires the examination of a very large number of observations extended over a long period. In the second place, the observations at one place should be kept separate from those at other places, for it is theoretically possible and even probable that a maximum at one place of observation may occur at the same time as a minimum at another place. For example, the yearly averages might show that a maximum rainfall in one place always occurred with a minimum rainfall in another, and vice versa. If the results from the two places were combined, a part or the whole of the periodic change would be lost. Then again, there is no clear indication what observations should be chosen for examination; average daily temperature, maximum or minimum daily temperatures, rainfall, number of violent storms, number of days when the barometers or thermometers were above or below certain marks, and so on.

In dealing with any complicated problem it is usually best to try and simplify it as much as possible, and if the problem even then appears too difficult, to attack a special case, and this is undoubtedly the best plan to pursue here. In order to avoid the rapid fluctuations which occur in the temperate zones, meteorologists have turned to places in the tropics where the types of weather during the course of a year are apt to be much more permanent and where the changes from one year to another are likely to appear more clearly than in other regions of the earth. It seems to be generally agreed that, owing to its political and geographical situation, India is the country best fitted to satisfy these various conditions. The most marked peculiarity of its climate, taking the country as a whole, is a rainy season extending over some two or three months, with comparatively little precipitation per month during the rest of the year. Independently of the scientific side of the question, the quantity of rainfall and the duration of the rainy season are of immense practical importance, since a deficiency in either may and usually does involve a famine. An examination of Indian records may therefore be of great value; and as a fairly continuous series of observations at several observatories has been obtained for some twenty or thirty years, it may not be too early to commence a systematic examination of them.

Several investigations in this direction have been made lately, unfortunately with but little success in obtaining positive results. Here it is advisable to set forth with a little more detail what is meant by 'success,' since it is necessary to have a clear conception of the value, both scientific and practical, to be attached to the results of investigations based on the theory of averages. A set of observations is recorded and an examination of them is made with a view of finding out the existence of a cycle. Suppose that one cycle is found, but that its effect on any individual observation is very small. For example, if the temperatures at any particular place vary in the course of a year from 20° to 90° and we find a cycle running through all its changes in eleven years and having a maximum effect of 1°, the daily temperatures are scarcely altered if we subtract the amount corresponding to this long-period variation. Thus, though the cycle may be thoroughly well made out, it tells very little about the variations of temperature. It is only when we are able to get a sufficient number of periodic changes whose combined effect will fairly well represent the individual observations that we can be said to be at all successful in our analysis of the latter, and it is only then that a basis is found sufficient to predict the future numbers of the series.

Of the various attempts to trace long-period changes in the meteorological records, only those will be taken in which some possible relation to the sunspot periods may exist. Thus a search is to be made for 11-year periods to establish a good connection; the 35-year period may furnish a possible connection between the two sets of phenomena, although, as we have seen, its existence in the sunspot curves is too doubtful to demonstrate the existence of the connection, while the 55 to 65-year periods, even if they exist in the sunspot curves, are too long to make a demonstration possible within the range of the recorded meteorological observations.

The best proof of the existence of an 11-year period has been furnished by Köppen, who showed that there was such a cycle in the combined mean annual temperature of many places within the tropics. But the total range of variation is only about three-quarters of a degree, while the mean annual temperatures show variations of from five to ten degrees. This appears to be the only well-established quantitative result with respect to the 11-year period. But several qualitative examinations have been made and the recent ones chiefly refer to rainfall and famine in India. Sir Norman and Dr. W. J. S. Lockyer have examined the rainfall statistics in India and Mauritius and they come to the conclusion that India has an increased rainfall near the sunspot maximum and Mauritius one near the sunspot minimum, and further, that the latter gives rise to a smaller pulse of rainfall in India. Thus India has two periods of increased rainfall, a large one near the sunspot maximum and a smaller one near the sunspot minimum. Unfortunately this investigation only refers to a single period of eleven years—from 1877 to 1886. They have also brought forward some evidence which indicates that the famine years occur between these two pulses of rainfall. The results do not seem to be sufficiently well made out to prophesy future famines with any certainty. Last year another writer showed a connection between certain Greenwich temperature records and the 11-year period, but the differences shown were very small. Other investigations on the same lines indicate about the same or a less degree of success in the discovery of a sunspot period.

In view of the doubtful existence of a 35-year period in the solar activity it is unnecessary to say much concerning a similar period in terrestrial phenomena. The most thorough investigation is that of Brückner, who examined a very large number of temperature records from all parts of the earth and deduced from the annual means a period of this length. But the amount of the change is only half a degree Fahrenheit, while the annual averages differ amongst themselves by from 5° to 10°.

In summing up briefly the results of the evidence hitherto presented, it must be admitted that no case has been made out in favor of a definite connection between the number and size of the spots on the sun's surface and the weather or climate on the earth. On the other hand, theoretical reasons would lead us to expect some connection, and the evidence is rather in favor of it than otherwise; but it is highly probable that the direct effect of the spotted area is unimportant compared with the effects produced in our atmosphere by other causes.

Although a negative answer must at present be given to the question whether sunspots have any considerable share in the variations of climate, it would be leaving a wrong impression if no mention were made of a well-established 11-year period in the variations of another set of terrestrial phenomena. The state of the earth's magnetism has for many years been carefully recorded by suitable instruments and there is no doubt that its fluctuations correspond quite closely with the state of the sun's surface, not only in the cycle of long period, but also with respect to the minor variations. It can not as yet be stated that the appearance of a group of sunspots always causes a disturbance of the magnetic needle, but investigations have been published and others are still under way which appear to show a very intimate connection between them. Nothing is known as to the causes of this connection.

We have been dealing hitherto with only one exhibition of the solar activity. It is one which catches the eye and is perhaps well adapted to show the main features of this activity in somewhat the same way as the presence of snow and ice would indicate the higher peaks of a mountain range. Pursuing the simile, while it is true that we get an idea of the more prominent outlines of the mountainous region by noting the white places on the dark background, the latter would include many elevations and depressions which are never touched by snow, and we can not have a correct idea of the range without finding some method of observing them also. The case of sunspots is not very different. They probably form the bolder outlines, but, as Professor Bigelow has pointed out, they constitute but a sluggish register of the solar activity. There is another set of variable phenomena, known as the prominences, which are now supposed to furnish a much better index of the state of the sun at any time. These are elevations of the material of which the sun is composed, visible round the edge of the disc, and they appear to be huge masses of liquid or gaseous matter violently ejected from the main body. Unlike the sunspots, prominences are rarely absent and some of them can nearly always be seen on the edge of the sun's disc, so that by recording the number and height of those observable from day to day a much more continuous and accurate series showing the activity of the sun can be obtained. Moreover, it has been shown that the variations in the series correspond much better with those in the earth's magnetic force than do the variations of the sunspot records. As yet, the number of years in which a continuous account of the prominences has been kept is too small to establish any good connection with the long-period or climatic changes on the earth's surface. It was only in 1868 that Lockyer and Janssen devised a method of observing them at times other than those rare ones when the moon comes in front of the sun's disc and leaves anything protruding from the edge of the latter in plain view for a few seconds. In general, large and numerous prominences are associated with the presence of sunspots, but prominences are frequently observed when no sunspots are visible in their neighborhood.

While this new method of noting the prominences in preference to the sunspots will probably improve our knowledge of what is going on in the sun, there is some tendency towards a change in the methods of observing the terrestrial atmosphere. The possibility of obtaining observations at heights of 2,000 feet or more above the earth's surface has directed attention to the fact that we are much more likely to get results free from local influences in this way, and perhaps a continuous series of such observations may show better any regular changes in the atmospheric phenomena. At the same time, the opinion has been strongly expressed that the observations already at hand should be sufficient to give the main features of the weather and climate if they can tell anything at all, and that it is time to stop the huge accumulation of records of temperature, rainfall, etc., and to undertake the thorough examination of those in our possession. There is undoubtedly great need for this: a plenteous harvest, with perhaps many tares and but few laborers.

In conclusion, while it is not my purpose to go into the reasons why more or fewer sunspots and prominences should affect our atmosphere, a few words may be said on the subject in order to counteract a widely spread misconception. Because the sunspots are dark areas, it is supposed—and one sees the statement frequently in the popular prints—that a portion of the sun's heat is screened off and the immediate deduction is made that 'cold waves' are the result. The first idea is very probably the exact opposite of the truth, and there is no theoretical or observed foundation for the second. As a matter of fact, it is generally agreed that sunspots and the associated prominences are evidences of increased activity, and therefore that they should denote a greater instead of a less output of solar heat. The amount of the change in the solar heat from time to time is as yet unknown; investigations in this direction are only in their infancy. Moreover, it is probable that the changes are too small to very materially affect terrestrial conditions, and even if they do so, the nature of the effect is quite doubtful. For example, an increase of heat from the sun may produce increased evaporation from certain water areas of the earth's surface, and the moisture thus drawn up being carried away by the movements of the atmosphere, may be deposited elsewhere in the form of snow or rain. It is almost impossible without a much more intimate knowledge of the laws governing the movements in our atmosphere to say in advance what the effect of a slow and periodic change in the sun's heat will be. Although the change in any one year may be small, its effect on the earth in the course of a number of years may be cumulative and thus become very evident; or other circumstances, for example, a corresponding periodic change in the radiation of heat from the earth, may so counterbalance the change in the heat received from the sun that the effect of the latter is scarcely perceptible. The conclusion of the whole matter has been well expressed by Professor Cleveland Abbe: that the key to the weather problem is not to be found in the sun or indeed in any external influence, but that the solution is to be worked out by the conditions which hold in the atmosphere itself—conditions which can only be discovered by a thorough examination of the internal laws of motion, quite apart from any external forces which may modify the results.