HUMAN INFLUENCE ON THE WEATHER*

B. J. Mason

INTRODUCTION

Weather plays so important a role in the life and activity of mankind that the question of the possibility of controlling it has occupied human thought since time immemorial. Primitive man, attributing the caprices of the weather to the mockery of the gods, tried to bring about changes in their mood by prayers and sacrifices. The development of scientific knowledge and pressing economic needs were bound inevitably to prompt man to attempt a more direct influence on nature. In this undertaking, however, enormous difficulties are encountered; not the least of them is insufficient understanding of the natural behavior of the atmosphere. Although from time to time it became fashionable to attribute anomalous changes in the weather to disturbances produced by man (artillery fire, wireless telegraphy, aviation, hydrogen bombs, etc.), in reality we are still very far from being able to control the atmospheric mechanism to any significant extent. It is true that in recent years some success has been achieved, as will be discussed below, but its importance is exaggerated. Even the most striking results must be interpreted with great caution. Even the discovery of a high degree of correlation between two series of data in meteorology is by no means proof of a causal connection between them, because the atmosphere does not behave like a controlled laboratory experiment in which most parameters are fixed and the rest are varied under control, so that one may hope to find direct relations between them. In the atmosphere we have a system of continuously changing parameters interacting with one another in a very complex manner. This constitutes the chief obstacle to a complete description and understanding of the behavior of the atmosphere and to the prediction and control of the weather.

Owing to this extraordinary complexity and variability of meteorological phenomena, and to our incomplete understanding of the physical processes occurring in the atmosphere, it is impossible to predict with absolute certainty the possible consequences of artificial influences on the weather. Any prediction, like a weather forecast, must be expressed in the language of probability; true, this does not mean that one assumption is as good as another: at present there is already sufficient information about the scales of atmospheric processes and about their dependence on time to permit a critical analysis.

Thus, for example, the probability of disturbing a large region of the atmosphere by the arbitrary application of brute force appears vanishingly small,

* An account of a public lecture delivered at the meeting of the British Association for the Advancement of Science in September 1955. Translated by S. A. Kamenetskii.

B. J. MASON

when we recall that the atmospheric mechanism is driven by the radiant energy of the Sun, the amount of which, averaged over all latitudes and over a day, is equivalent to the explosion of one hydrogen bomb every second. Even a small thunderstorm releases energy equal to ten atomic bombs, and the energy dissipated by a typical hurricane is equivalent to several hundred hydrogen bombs. If we consider still larger meteorological disturbances, for example depressions, then the amount of kinetic energy associated with them is colossal, since masses of air of the order of a trillion \((10^{12})\) tons must be displaced in order to create a pressure fall.

It therefore seems extremely improbable to hope to influence the behavior of the atmosphere (apart from a very limited local influence) by using, against the enormous forces of nature, the amount of energy at our disposal. If we want to have any chance of success, we must direct our attack toward finding weak points in nature, and not attempt to carry out a direct frontal assault.

THE ACTION OF THE ATMOSPHERIC MECHANISM

Only when we obtain a considerably more complete description of the atmospheric machine and understand more deeply the mechanism that sets it in motion shall we be able to discover those elements in the atmospheric cycle on which man might successfully act. Nevertheless, the general principles of the action of the atmospheric mechanism are sufficiently well understood.

Labels in the diagram:

• Tropopause
• \(23 \tfrac{1}{2}\%\) reflected back by clouds
• Incoming \(100\%\)
• \(2 \tfrac{1}{2}\%\) reflected by the Earth
• \(9\%\) scattered back by air, dust, and water vapor
• \(65\%\) long-wave radiation emitted into space
• Earth
• \(9\%\) absorbed by water vapor
• \(6\%\) absorbed by clouds
• Average cloudiness \(0.52\)
• \(30\%\) direct solar radiation
• \(14\%\) diffuse radiation from clouds
• \(6\%\) diffuse radiation of air, dust, and water vapor
• \(45\%\) long-wave radiation from the atmosphere
• \(20\%\) long-wave radiation from soil
• Latent heat
• \(23\%\)
• Sensible heat
• \(7\%\)

The fraction of incoming solar radiation reflected by the Earth as a planet is

\[ 23 \tfrac{1}{2} + 2 \tfrac{1}{2} + 9 = 35\%. \]

The fraction absorbed by the soil and atmosphere is \(65\%\). The fraction absorbed at the Earth’s surface (direct radiation \(30\% +\) diffuse radiation \(20\% = 50\%\)). The fraction expended on the evaporation of water from the Earth \((2 \tfrac{3}{4}\ \text{mm per day}) = 23\%\).

The total average of all incoming radiation is

\[ = 0.5 \ \frac{\text{cal}}{\text{cm}^{2}\,\text{min}} = 720 \ \frac{\text{cal}}{\text{cm}^{2}\,\text{day}}. \]

The source of energy for the atmosphere is solar radiation. What happens to this radiant energy as it passes through the troposphere, containing an average amount of clouds, is shown in the figure. Of the entire average flux of solar energy arriving at the tropopause and reaching

approximately 0.5 cal/cm² min, about 35% is reflected back into space by clouds, atmospheric constituents, and the Earth's surface, and 65% is absorbed by the atmosphere (15%) and by the Earth (50%). The portion absorbed by the Earth, consisting both of direct solar radiation and of radiation scattered downward by the atmosphere and clouds, is either radiated back in the form of long-wave (infrared) radiation (after part of it has been used to evaporate water from the Earth's surface), or is carried away in the form of sensible heat by convection and by all the systems that make up the weather. Part of the energy absorbed by the atmosphere is converted into the kinetic energy of winds, and this energy, in turn, is again converted into heat through friction, occurring mainly near the Earth's surface.

Since the atmosphere as a whole neither heats up nor cools down, the amount of heat radiated into space in the form of long-wave radiation must be balanced by the radiation received from the Sun. However, if we calculate the heat balance for different latitudes, we find that between the equator and latitude 40° the outgoing radiation is less than the incoming radiation, while at higher latitudes the reverse is the case. Thus the tropics and subtropics may be regarded as sources of heat, and the poles as sinks of heat; and since the mean temperatures of the tropics and the poles do not change, heat must be continuously transported from the equator toward the poles. It can also be shown that the sink of heat must be at lower pressure (at greater height) than the source of heat, and it follows from this that heat must be transported not only toward the poles, but also upward.

The temperature difference between the equator and the poles is maintained at higher levels as well, and associated with it is a pressure difference that increases with height. This pressure gradient, together with the effect of the Earth's rotation, gives rise to a planetary belt of westerly winds, whose intensity increases with height up to the tropopause but then decreases again in the stratosphere. Such a system, however, is dynamically unstable to disturbances having horizontal dimensions on the order of 1000 km or more, so that if this flow is disturbed, the circumpolar vortex breaks up into a series of disturbances (depressions), each with dimensions on the order of 1000 km across. The air in these depressions moves from low levels and low latitudes upward and toward the poles along slightly inclined surfaces, while air from the upper layers moves downward and toward the equator. Thus heat is transported in those directions in which this is necessary for the radiative balance of the atmosphere. It is precisely these upward-directed vertical air currents in such large cyclones that generate the extensive belts of clouds and rain which are the principal characteristics of the weather.

Thus we see that the radiation field determines the large-scale features and the mean intensity of atmospheric circulation, which, owing to its dynamic instability, may assume various forms capable of passing from one into another through an internal redistribution of local sources and sinks of heat and leading to various meteorological situations. Thus the radiation pattern and the dynamic behavior of the atmosphere are closely interconnected.

It is also necessary to consider the interaction with the hydrological cycle, which uses part of the solar radiation absorbed by the Earth's surface for the evaporation of water from the land and ocean surface. Mean planetary evaporation is estimated at approximately 2¾ mm per day, or about 1 m per year, which, of course, is equal to the mean amount of precipitation over the entire planet. Water vapor is carried upward by ascending air currents (arising dynamically or by convection), cools aloft and condenses, and then forms clouds and falls as snow, rain, or hail. But the formation of clouds, which is the result of absorption

radiation at the earth’s surface and air motions, in turn affects the radiation balance, since clouds, which on average reflect back about half of the solar radiation incident upon them, exert a strong influence on the amount of radiation remaining to set the atmospheric mechanism in motion.

Thus, we have a mechanism with multiple feedback, which is difficult to understand as a whole and whose component parts cannot easily be isolated from one another for separate study. This means that it is usually impossible to trace directly the connection between cause and effect, owing to the presence of links between various intermediate stages.

FUNDAMENTAL POSSIBILITIES FOR INFLUENCING THE WEATHER

Having briefly considered the principles of operation of the atmospheric mechanism, we can now turn to an investigation of the possibilities of disrupting its activity. In principle, the weather can be changed by intervening either in the radiation balance, or in the development of dynamical systems, or, finally, in the hydrological cycle. Any major change introduced into one of these factors can apparently affect the others to extents that depend on the degree of interaction between them.

a) Changing the radiation balance

An obvious method of changing the amount of solar radiation entering the atmosphere would be to reflect a considerably larger fraction of this radiation back into space by creating an artificial mirror. One can imagine two possible types of mirror: the creation of a dense layer of dust in the upper atmosphere, or an artificial increase in the amount of cloud, since clouds are effective reflectors of solar radiation.

If a layer of dust or smoke in the lower stratosphere were sufficiently dense for several months to cut off an appreciable fraction of the solar radiation, there can hardly be any doubt that, as a result, temperatures at the earth’s surface would be lowered. Changes might also occur in the north–south temperature gradient and in the intensity of the atmospheric circulation, which would cause accompanying changes in winds and precipitation. However, in estimating the amount of dust that would be required in order to produce substantial changes, one is struck by the fact that the enormous quantities of dust thrown into the atmosphere by the eruption of the volcano Krakatoa in 1883, which, according to some estimates, reduced the amount of direct solar radiation by 10% and was certainly the cause of various optical phenomena during the following three years, did not produce substantial decreases in surface temperatures or changes in the amount of precipitation. Such changes are not easy to detect if they are not sufficiently large; even at the present time, with a considerably improved observing network, it is difficult to discover small general temperature tendencies because of the large regions of positive and negative anomalies that occur on the earth every year. On the other hand, the exceptionally dense smoke haze formed as a result of forest fires in Canada and passing over the eastern part of the United States in September 1950 (it was the cause of the appearance in Europe of the so-called “blue sun”) lowered the maximum daytime temperatures by 5°.

Another possible method of creating a reflecting mirror would be the formation of an artificial cloud layer in the upper layers of the atmosphere. Even a thin layer of cirrus clouds would screen 15% of the incoming solar radiation. It has been supposed that such a solar screen could be created with the help of aircraft leaving condensed trails behind them, and that—

...a similar method can be applied in semiarid regions to reduce the very high rate of soil evaporation that occurs under a clear sky. However, in order to cover a large area, the number of airplanes would have to be so great that the proposal proves uneconomical.

b) Bringing dynamic disturbances into action

The only proposal for possible direct intervention in the dynamics of the atmosphere that one can imagine is based on the fact that large meteorological phenomena, for example depressions, can grow rapidly out of small disturbances—small perturbations of a dynamically unstable large-scale flow—and that, as a result, it would seem possible to initiate a sequence of weather disturbances with the help of some initiator, perhaps in the form of a bomb. This consideration is apparently based on an insufficient understanding of the fact that large regions of the atmosphere are often so delicately balanced that a small disturbance at almost any point begins to grow rapidly and may develop in a direction quite different from that expected. In fact, because of the large number of continuously arising natural disturbances, the application of a suitable artificial stimulus at a precisely chosen moment and at a precisely chosen point so that it would develop predominantly before the others would represent an utterly exceptional piece of luck. Moreover, it would be difficult to assess the results of such an experiment, since on a weather map there is apparently no possibility of distinguishing between effects caused by an artificial stimulus and the results of natural disturbances. In any case, the effects caused by any special initial disturbance are apparently short-lived—the most active meteorological systems tend to degenerate within several days and are replaced by others arising in some other place. Therefore, there is apparently no hope at present of controlling the weather, for example, over an entire season by means of an appropriate “launching” of the atmosphere.

It is conceivable that changes in the general circulation of the atmosphere might be produced by redistributing the sources and sinks of heat in the atmosphere or in the oceans, since the latter act not only as large reservoirs of heat but also participate in the transfer of heat toward the poles. Some redistribution of temperature in the oceans (which might occur, for example, with the partial melting of polar ice) would probably cause changes in weather and climate; however, such possibilities do not go beyond speculative reasoning.

c) Changing the hydrological cycle

In principle, the rate of the evaporation—condensation—precipitation cycle could be changed in various ways. The rate of evaporation of water from the earth’s surface could be altered by changing the amount of incoming solar radiation, as already indicated, by applying irrigation on a large scale, or by preventing water molecules from escaping from open water surfaces.

Schemes have been proposed for creating large artificial lakes in semiarid regions on the assumption that an increase in the rate of evaporation due to the new water surfaces will cause an increase in the amount of precipitation, and that the additional precipitation may lead to an even larger evaporating surface, and so on. In other words, after the creation of a large water reservoir, water may be distributed over the surrounding locality by the atmosphere, and the whole system will become self-sustaining. Unfortunately, the atmosphere,

apparently, not cooperate in such an undertaking, and it may quite likely happen that a considerable part of the evaporated water will fall as precipitation very far from the place of evaporation, and only a small share will reach the desired place. If the application of the method described pursues only this aim, then it would probably be far more effective and economical to deliver water to the consumer by ordinary irrigation methods than to make it circulate through the atmosphere.

At present experiments are being carried out in Australia to determine to what extent the evaporation of natural bodies of water can be reduced by covering them with very thin chemical films. It turns out that monomolecular layers of certain straight-chain alcohols can reduce evaporation by approximately 40% under normal summer conditions in South Australia, and that the action of such films is little affected by wind and dust. It has been calculated that about 3 centners of such a substance in crystalline form is sufficient to treat \(2.5\ \mathrm{km}^2\) of water surface and that the action of this substance will last for at least several months. Large-scale experiments are now being carried out, and it will be very interesting to follow them.

THE HYDROGEN BOMB AND THE WEATHER

In assessing the bad summer of 1954, the hydrogen bomb was usually cited as the “scapegoat.”

According to one theory, it turned out that the bomb could have affected the weather by increasing the concentration of fine dust in the atmosphere, which served as the cause of the cold summer. However, the bomb produced far less dust than Krakatoa, when the layer of dust was visible to the naked eye, and nevertheless it did not produce noticeable changes in the mean temperature or in the average amount of precipitation at the Earth’s surface. The effect of the bomb explosion, which was probably a thousand times less powerful than the eruption of Krakatoa, should, of course, have been correspondingly smaller.

In estimating the possibility that the bomb could set in motion large dynamic disturbances, one can only repeat that the chances of this are extremely small, if one takes into account the multitude of natural disturbances. But even if one were to admit such an improbable occurrence, it should be said that its influence would disappear without a trace within a few days.

Of course, one can imagine that a large explosion may affect the weather in its immediate vicinity. However, it is unlikely that the energy of an almost instantaneous explosion would serve as the direct cause of the formation of a storm, since the processes of condensation and rain formation require 30 minutes or more, during which energy must be supplied continuously. Thus, the explosion of an atomic bomb would probably have small direct consequences, except under such special circumstances as, for example, occurred in the underwater explosion at Bikini in July 1946, when the release of large quantities of water and steam caused a sufficient increase in the humidity of tropical air masses to reveal their instability and lead to the formation of clouds and rain lasting about 30 minutes. Naturally, atomic and nuclear weapons can indirectly cause local storms. After the explosion in Hiroshima, fires broke out in many places, and soon a thunderstorm broke out as a result of the rising of hot air. In the case of a hydrogen bomb, heating of the ground beneath it owing to the radiation of the fireball may cause convection that will persist for some time after the explosion.

It has also been suggested that the ionized particles created by the bomb may act as additional condensation centers and may lead to an increase in precipitation. But although small ions can accelerate

condensation in Wilson’s chamber, the required supersaturation is very great—probably several thousand times greater than that which exists in natural clouds. This is because the atmosphere already contains a sufficient number of fairly large particles for them to serve as centers of condensation for the existing water vapor and to prevent the supersaturation from reaching a value at which small ions could take part in the condensation process.

Perhaps, in summing up, one may say that although it would be unreasonable to assert categorically that the hydrogen bomb could have no influence whatever on the weather, it may nevertheless be regarded as highly improbable that accidental explosions on the scale on which they are being carried out at present could produce more than local and transient effects.

The methods of changing the weather considered above appear to offer little promise. In conclusion let us turn to a more hopeful direction, in which some success has already been achieved.

ARTIFICIAL ACTION ON CLOUDS AND THE PRODUCTION OF RAIN

During the last several years it has been incontrovertibly proved that rain can be produced from suitable clouds by “seeding” them with dry ice, silver iodide, water droplets, salt particles, and the like, under conditions in which no rain falls from similar untreated clouds in the vicinity. It has turned out that there are two principal mechanisms for producing rain. One mechanism consists in the growth of ice crystals, their aggregation into snowflakes, the melting of the latter, and their conversion into raindrops. This mechanism produces the majority of the widespread “continuous” rains falling from large systems of layered clouds. The second mechanism consists in the fact that droplets larger than the average droplets in the cloud continue to grow by collisions and coalescence at the expense of their smaller neighbors; this mechanism is important only for shower clouds.

Experiments on rain production are carried out on the assumption that from some clouds precipitation either does not fall at all, or falls inefficiently, because in the natural state of the cloud there are no ice crystals or comparatively large water droplets (capable of initiating the coalescence process); and that this deficiency can be corrected by artificial “seeding” of clouds with “dry ice” or silver iodide (to obtain ice crystals), or by introducing droplets of water or large hygroscopic nuclei.

One type of experiment consists in dropping, from an aircraft onto the upper part of a supercooled cloud (at a temperature below \(0^\circ\) C), pellets of dry ice about \(1\ \mathrm{cm}\) in diameter. Each pellet produces about 10,000 billion small ice crystals, so that only a few kilograms of dry ice are sufficient to seed a large cumulus (shower) cloud. More than two hundred such experiments, carried out in Australia, Canada, and South Africa, have incontrovertibly proved that from cumulus clouds at a certain stage of development rain was artificially produced, whereas from neighboring clouds not subjected to seeding no precipitation fell.

A more recent development of these experiments consists of attempts to stimulate the coalescence process in shower clouds. By means of aircraft equipped with water tanks and sprayers, droplets approximately \(1/20\ \mathrm{mm}\) in diameter were introduced into the bases of developing clouds; in this way the cloud was supplied with the large droplets necessary to initiate the coalescence process. The first experiments, carried out in Australia, gave encouraging results; in ten experiments out of eleven, ...

rain, and in four cases heavy precipitation was produced. Some success has also been achieved in the most recent experiments with tropical cumulus clouds in the Caribbean Sea.

The fact that sea spray is an important natural source of large droplets suggests that it would be more economical to spray salt crystals instead of water droplets; a salt crystal with a diameter of \(0.01\) mm grows into a droplet five times as large when passing through the lower layer of a cloud \(0.5\) km thick, so that instead of \(4.5\) liters of water it is sufficient to use \(100\) g of salt. In England several experiments were carried out in 1952 in which salt was seeded from an aircraft. Later, in East Africa, experiments were conducted with balloon kites equipped with bombs filled with a mixture of gunpowder and finely ground salt; and in Pakistan the dry climate made it possible to carry out experiments in dispersing salt dust from the ground surface. Although the results of the experiments cannot be regarded as final, they are nevertheless encouraging, and it may be assumed that this method will prove sufficiently effective for inducing showers from warm cumulus clouds.

However, since it is uneconomical to try, with the aid of aircraft, to conduct large-scale experiments aimed at causing precipitation from extensive cloud systems spread over thousands of square kilometers, it has become common practice for this purpose to spray silver iodide from the ground in the form of smoke, on the assumption that rising air currents will carry this substance into the supercooled regions of the cloud. Silver iodide was used because of the similarity of its crystalline structure to that of ice crystals; however, the special effectiveness of crystals of this substance as nuclei for condensation and ice formation requires more detailed further investigation. With present methods of use, which are obviously not the best, silver iodide is ineffective at temperatures above \(-5^\circ\) C, but below \(-15^\circ\) C \(1\) g of silver iodide can, when sprayed, form \(1000\) billion condensation nuclei. Of course, with such a spraying method it is impossible to control the subsequent transport of the smoke. We cannot make a reliable estimate of the concentration of nuclei that have reached the cloud level, nor can we know how long silver iodide retains in the atmosphere its properties of forming condensation nuclei. These unknown factors, and the impossibility of making an exact estimate of how much precipitation would have fallen naturally in the absence of seeding, greatly complicate the planning of such operations on a large scale and the assessment of their results. Despite the large claims of commercial firms engaged in rainmaking, the published data do not contain convincing evidence that this method can produce a significant increase in precipitation over long periods and over large areas. Although it is quite admissible that, under suitable conditions, additional precipitation of the order of \(10\)—\(20\%\) can be produced. As a rule, in no region have these operations been carried out for a sufficiently long time, nor have they been designed in such a way that sufficiently reliable and sensitive methods of evaluation could be applied to them, making it possible to separate effects of the indicated order from random fluctuations of natural precipitation. In order to obtain reliable data, it would probably be necessary to conduct systematic experiments for several years and subject their results to thorough statistical processing. Much should also be said about the choice of favorable conditions and the reduction of the number of variables. Thus, for example, it is reasonable to seed for a long time supercooled clouds that form when moist air rises over a large mountain range, since in this case there is reason to expect that silver iodide released on the windward slopes will be picked up by continuously forming clouds and will cause an intensified snowfall on the leeward—

… side. In addition, in this case there are real chances of detecting the results visually.

In considering the possibility of seeding clouds, one should not lose sight of the fact that a total increase in precipitation on the scale of the entire Earth means an acceleration of the whole hydrological cycle; therefore, perhaps the more realistic assumption is the possibility of some redistribution of precipitation by anticipating the natural fall of precipitation. The possibility of reducing precipitation and preventing hail and thunderstorms by “reseeding” clouds has also been discussed; however, the practical difficulties here are such that these ideas still require serious verification.

Conclusion

Although a number of possible ways of changing the behavior of the atmosphere have been proposed, only some of them can meet the criterion of practice. There is some hope of influencing the weather on a local scale, perhaps by intervening in the cycle of evaporation and precipitation, but it still remains to be proved that the technique employed is also suitable for successful use on large scales. However, further intensive study of atmospheric processes may yet reveal hitherto unknown weak links in the chain of natural phenomena that can be used for the benefit of humanity.