[Footnote 60: Chas. Schuchert: Review of Knowlton's Evolution of Geological Climates, in Am. Jour. Sci., 1921.]

[Footnote 61: G. R. Wieland: Distribution and Relationships of the Cycadeoids; Am. Jour. Bot., Vol. 7, 1920, pp. 125-145.]

[Footnote 62: D. T. MacDougal: Botanical Features of North American Deserts; Carnegie Inst.i.t. of Wash., No. 99, 1908.]

[Footnote 63: _Loc. cit._]

[Footnote 64: H. H. Clayton: Variation in Solar Radiation and the Weather; Smiths. Misc. Coll., Vol. 71, No. 3, Washington, 1920.]

[Footnote 65: B. h.e.l.land Hansen and F. Nansen: Temperature Variations in the North Atlantic Ocean and in the Atmosphere; Misc. Coll., Smiths.

Inst., Vol. 70, No. 4, Washington, 1920.]

[Footnote 66: The climatic significance of ocean currents is well discussed in Croll's Climate and Time, 1875, and his Climate and Cosmogony, 1889.]

[Footnote 67: F. J. B. Cordeiro: The Gyroscope, 1913.]

[Footnote 68: W. W. Garner and H. A. Allard: Flowering and Fruition of Plants as Controlled by Length of Day; Yearbook Dept. Agri., 1920, pp.

377-400.]

[Footnote 69: Report of Committee on Sedimentation, National Research Council, April, 1922.]

CHAPTER XI

TERRESTRIAL CAUSES OF CLIMATIC CHANGES

The major portion of this book has been concerned with the explanation of the more abrupt and extreme changes of climate. This chapter and the next consider two other sorts of climatic changes, the slight secular progression during the hundreds of millions of years of recorded earth history, and especially the long slow geologic oscillations of millions or tens of millions of years. It is generally agreed among geologists that the progressive change has tended toward greater extremes of climate; that is, greater seasonal contrasts, and greater contrasts from place to place and from zone to zone.[70] The slow cyclic changes have been those that favored widespread glaciation at one extreme near the ends of geologic periods and eras, and mild temperatures even in subpolar regions at the other extreme during the medial portions of the periods.

As has been pointed out in an earlier chapter, it has often been a.s.sumed that all climatic changes are due to terrestrial causes. We have seen, however, that there is strong evidence that solar variations play a large part in modifying the earth's climate. We have also seen that no known terrestrial agency appears to be able to produce the abrupt changes noted in recent years, the longer cycles of historical times, or geological changes of the shorter type, such as glaciation.

Nevertheless, terrestrial changes doubtless have a.s.sisted in producing both the progressive change and the slow cyclic changes recorded in the rocks, and it is the purpose of this chapter and the two that follow to consider what terrestrial changes have taken place and the probable effect of such changes.

The terrestrial changes that have a climatic significance are numerous.

Some, such as variations in the amount of volcanic dust in the higher air, have been considered in an earlier chapter. Others are too imperfectly known to warrant discussion, and in addition there are presumably others which are entirely unknown. Doubtless some of these little known or unknown changes have been of importance in modifying climate. For example, the climatic influence of vegetation, animals, and man may be appreciable. Here, however, we shall confine ourselves to purely physical causes, which will be treated in the following order: First, those concerned with the solid parts of the earth, namely: (I) amount of land; (II) distribution of land; (III) height of land; (IV) lava flows; and (V) internal heat. Second, those which arise from the salinity of oceans, and third, those depending on the composition and amount of atmosphere.

The terrestrial change which appears indirectly to have caused the greatest change in climate is the contraction of the earth. The problem of contraction is highly complex and is as yet only imperfectly understood. Since only its results and not its processes influence climate, the following section as far as page 196 is not necessary to the general reader. It is inserted in order to explain why we a.s.sume that there have been oscillations between certain types of distribution of the lands.

The extent of the earth's contraction may be judged from the shrinkage indicated by the shortening of the rock formations in folded mountains such as the Alps, Juras, Appalachians, and Caucasus. Geologists are continually discovering new evidence of thrust faults of great magnitude where ma.s.ses of rock are thrust bodily over other rocks, sometimes for many miles. Therefore, the estimates of the amount of shrinkage based on the measurements of folds and faults need constant revision upward.

Nevertheless, they have already reached a considerable figure. For example, in 1919, Professor A. Heim estimated the shortening of the meridian pa.s.sing through the modern Alps and the ancient Hercynian and Caledonian mountains as fully a thousand miles in Europe, and over five hundred miles for the rest of this meridian.[71] This is a radial shortening of about 250 miles. Possibly the shrinkage has been even greater than this. Chamberlin[72] has compared the density of the earth, moon, Mars, and Venus with one another, and found it probable that the radial shrinkage of the earth may be as much as 570 miles. This result is not so different from Heim's as appears at first sight, for Heim made no allowance for unrecognized thrust faults and for the contraction incident to metamorphism. Moreover, Heim did not include shrinkage during the first half of geological time before the above-mentioned mountain systems were upheaved.

According to a well-established law of physics, contraction of a rotating body results in more rapid rotation and greater centrifugal force. These conditions must increase the earth's equatorial bulge and thereby cause changes in the distribution of land and water. Opposed to the rearrangement of the land due to increased rotation caused by contraction, there has presumably been another rearrangement due to tidal r.e.t.a.r.dation of the earth's rotation and a consequent lessening of the equatorial bulge. G. H. Darwin long ago deduced a relatively large r.e.t.a.r.dation due to lunar tides. A few years ago W. D. MacMillan, on other a.s.sumptions, deduced only a negligible r.e.t.a.r.dation. Still more recently Taylor[73] has studied the tides of the Irish Sea, and his work has led Jeffreys[74] and Brown[75] to conclude that there has been considerable r.e.t.a.r.dation, perhaps enough, according to Brown, to equal the acceleration due to the earth's contraction. From a prolonged and exhaustive study of the motions of the moon Brown concludes that tidal friction or some other cause is now lengthening the day at the rate of one second per thousand years, or an hour in almost four million years if the present rate continues. He makes it clear that the r.e.t.a.r.dation due to tides would not correspond in point of time with the acceleration due to contraction. The r.e.t.a.r.dation would occur slowly, and would take place chiefly during the long quiet periods of geologic history, while the acceleration would occur rapidly at times of diastrophic deformation. As a consequence, the equatorial bulge would alternately be reduced at a slow rate, and then somewhat suddenly augmented.

The less rigid any part of the earth is, the more quickly it responds to the forces which lead to bulging or which tend to lessen the bulge.

Since water is more fluid than land, the contraction of the earth and the tidal r.e.t.a.r.dation presumably tend alternately to increase and decrease the amount of water near the equator more than the amount of land. Thus, throughout geological history we should look for cyclic changes in the relative area of the lands within the tropics and similar changes of opposite phase in higher lat.i.tudes. The extent of the change would depend upon (a) the amount of alteration in the speed of rotation, and (b) the extent of low land in low lat.i.tudes and of shallow sea in high lat.i.tudes. According to Slichter's tables, if the earth should rotate in twenty-three hours instead of twenty-four, the great Amazon lowland would be submerged by the inflow of oceanic water, while wide areas in Hudson Bay, the North Sea, and other northern regions, would become land because the ocean water would flow away from them.[76]

Following the prompt equatorward movement of water which would occur as the speed of rotation increased, there must also be a gradual movement or creepage of the solid rocks toward the equator, that is, a bulging of the ocean floor and of the lands in low lat.i.tudes, with a consequent emergence of the lands there and a relative rise of sea level in higher lat.i.tudes. Tidal r.e.t.a.r.dation would have a similar effect. Suess[77] has described widespread elevated strand lines in the tropics which he interprets as indicating a relatively sudden change in sea level, though he does not suggest a cause of the change. However, in speaking of recent geological times, Suess reports that a movement more recent than the old strands "was an acc.u.mulation of water toward the equator, a diminution toward the poles, and (it appears) as though this last movement were only one of the many oscillations which succeed each other with the same tendency, i.e., with a positive excess at the equator, a negative excess at the poles." (Vol. II, p. 551.) This creepage of the rocks equatorward seemingly might favor the growth of mountains in tropical and subtropical regions, because it is highly improbable that the increase in the bulge would go on in all longitudes with perfect uniformity. Where it went on most rapidly mountains would arise. That such irregularity of movement has actually occurred is suggested not only by the fact that many Cenozoic and older mountain ranges extend east and west, but by the further fact that these include some of our greatest ranges, many of which are in fairly low lat.i.tudes. The Himalayas, the Javanese ranges, and the half-submerged Caribbean chains are examples. Such mountains suggest a thrust in a north and south direction which is just what would happen if the solid ma.s.s of the earth were creeping first equatorward and then poleward.

A fact which is in accord with the idea of a periodic increase in the oceans in low lat.i.tudes because of renewed bulging at the equator is the exposure in moderately high lat.i.tudes of the greatest extent of ancient rocks. This seems to mean that in low lat.i.tudes the frequent deepening of the oceans has caused the old rocks to be largely covered by sediments, while the old lands in higher lat.i.tudes have been left more fully exposed to erosion.

Another suggestion of such periodic equatorward movements of the ocean water is found in the reported contrast between the relative stability with which the northern part of North America has remained slightly above sea level except at times of widespread submergence, while the southern parts have suffered repeated submergence alternating with great emergence.[78] Furthermore, although the northern part of North America has been generally exposed to erosion since the Proterozoic, it has supplied much less sediment than have the more southern land areas.[79]

This apparently means that much of Canada has stood relatively low, while repeated and profound uplift alternating with depression has occurred in subtropical lat.i.tudes, apparently in adjustment to changes in the earth's speed of rotation. The uplifts generally followed the times of submergence due to equatorward movement of the water, though the buckling of the crust which accompanies shrinkage doubtless caused some of the submergence. The evidence that northern North America stood relatively low throughout much of geological time depends not only on the fact that little sediment came to the south from the north, but also on the fact that at times of especially widespread epicontinental seas, the submergence was initiated at the north.[80] This is especially true for Ordovician, Silurian, Devonian, and Jura.s.sic times in North America.

General submergence of this kind is supposed to be due chiefly to the overflowing of the ocean when its level is slowly raised by the deposition of sediment derived from the erosion of what once were continental highlands but later are peneplains. The fact that such submergence began in high lat.i.tudes, however, seems to need a further explanation. The bulging of the rock sphere at the equator and the consequent displacement of some of the water in low lat.i.tudes would furnish such an explanation, as would also a decrease in the speed of rotation induced by tidal r.e.t.a.r.dation, if that r.e.t.a.r.dation were great enough and rapid enough to be geologically effective.

The climatic effects of the earth's contraction, which we shall shortly discuss, are greatly complicated by the fact that contraction has taken place irregularly. Such irregularity has occurred in spite of the fact that the processes which cause contraction have probably gone on quite steadily throughout geological history. These processes include the chemical reorganization of the minerals of the crust, a process which is ill.u.s.trated by the metamorphism of sedimentary rocks into crystalline forms. The escape of gases through volcanic action or otherwise has been another important process.

Although the processes which cause contraction probably go on steadily, their effect, as Chamberlin[81] and others have pointed out, is probably delayed by inertia. Thus the settling of the crust or its movement on a large scale is delayed. Perhaps the delay continues until the stresses become so great that of themselves they overcome the inertia, or possibly some outside agency, whose nature we shall consider later, reenforces the stresses and gives the slight impulse which is enough to release them and allow the earth's crust to settle into a new state of equilibrium. When contraction proceeds actively, the ocean segments, being largest and heaviest, are likely to settle most, resulting in a deepening of the oceans and an emergence of the lands. Following each considerable contraction there would be an increase in the speed of rotation. The repeated contractions with consequent growth of the equatorial bulge would alternate with long quiet periods during which tidal r.e.t.a.r.dation would again decrease the speed of rotation and hence lessen the bulge. The result would be repeated changes of distribution of land and water, with consequent changes in climate.

I. We shall now consider the climatic effect of the repeated changes in the relative amounts of land and water which appear to have resulted from the earth's contraction and from changes in its speed of rotation.

During many geologic epochs a larger portion of the earth was covered with water than at present. For example, during at least twelve out of about twenty epochs, North America has suffered extensive inundations,[82] and in general the extensive submergence of Europe, the other area well known geologically, has coincided with that of North America. At other times, the ocean has been less extensive than now, as for example during the recent glacial period, and probably during several of the glacial periods of earlier date. Each of the numerous changes in the relative extent of the lands must have resulted in a modification of climate.[83] This modification would occur chiefly because water becomes warm far more slowly than land, and cools off far more slowly.

An increase in the lands would cause changes in several climatic conditions. (a) The range of temperature between day and night and between summer and winter would increase, for lands become warmer by day and in summer than do oceans, and cooler at night and in winter. The higher summer temperature when the lands are widespread is due chiefly to the fact that the land, if not snow-covered, absorbs more of the sun's radiant energy than does the ocean, for its reflecting power is low. The lower winter temperature when lands are widespread occurs not only because they cool off rapidly but because the reduced oceans cannot give them so much heat. Moreover, the larger the land, the more generally do the winds blow outward from it in winter and thus prevent the ocean heat from being carried inland. So long as the ocean is not frozen in high lat.i.tudes, it is generally the chief source of heat in winter, for the nights are several months long near the poles, and even when the sun does shine its angle is so low that reflection from the snow is very great. Furthermore, although on the average there is more reflection from water than from land, the opposite is true in high lat.i.tudes in winter when the land is snow-covered while the ocean is relatively dark and is roughened by the waves. Another factor in causing large lands to have extremely low temperature in winter is the fact that in proportion to their size they are less protected by fog and cloud than are smaller areas. The belt of cloud and fog which is usually formed when the wind blows from the ocean to the relatively cold land is restricted to the coastal zone. Thus the larger the land, the smaller the fraction in which loss of heat by radiation is reduced by clouds and fogs. Hence an increase in the land area is accompanied by an increase in the contrasts in temperature between land and water.

(b) The contrasts in temperature thus produced must cause similar contrasts in atmospheric pressure, and hence stronger barometric gradients. (c) The strong gradients would mean strong winds, flowing from land to sea or from sea to land. (d) Local convection would also be strengthened in harmony with the expansion of the lands, for the more rapid heating of land than of water favors active convection.

(e) As the extent of the ocean diminished, there would normally be a decrease in the amount of water vapor for three reasons: (1) Evaporation from the ocean is the great source of water vapor. Other conditions being equal, the smaller the ocean becomes, the less the evaporation.

(2) The amount of water vapor in the air diminishes as convection increases, since upward convection is a chief method by which condensation and precipitation are produced, and water vapor removed from the atmosphere. (3) Nocturnal cooling sufficient to produce dew and frost is very much more common upon land than upon the ocean. The formation of dew and frost diminishes the amount of water vapor at least temporarily. (f) Any diminution in water vapor produced in these ways, or otherwise, is significant because water vapor is the most essential part of the atmosphere so far as regulation of temperature is concerned.

It tends to keep the days from becoming hot or the nights cold.

Therefore any decrease in water vapor would increase the diurnal and seasonal range of temperature, making the climate more extreme and severe. Thus a periodic increase in the area of the continents would clearly make for periodic increased climatic contrasts, with great extremes, a type of climatic change which has recurred again and again.

Indeed, each great glaciation accompanied or followed extensive emergence of the lands.[84]

Whether or not there has been a _progressive_ increase from era to era in the area of the lands is uncertain. Good authorities disagree widely.

There is no doubt, however, that at present the lands are more extensive than at most times in the past, though smaller, perhaps, than at certain periods. The wide expanse of lands helps explain the prominence of seasons at present as compared with the past.

II. The contraction of the earth, as we have seen, has produced great changes in the distribution as well as in the extent of land and water.

Large parts of the present continents have been covered repeatedly by the sea, and extensive areas now covered with water have been land. In recent geological times, that is, during the Pliocene and Pleistocene, much of the present continental shelf, the zone less than 600 feet below sea level, was land. If the whole shelf had been exposed, the lands would have been greater than at present by an area larger than North America. When the lands were most elevated, or a little earlier, North America was probably connected with Asia and almost with Europe. Asia in turn was apparently connected with the larger East Indian islands. In much earlier times land occupied regions where now the ocean is fairly deep. Groups of islands, such as the East Indies and Malaysia and perhaps the West Indies, were united into widespreading land ma.s.ses.

Figs. 7 and 9, ill.u.s.trating the paleography of the Permian and the Cretaceous periods, respectively, indicate a land distribution radically different from that of today.

So far as appears from the scattered facts of geological history, the changes in the distribution of land seem to have been marked by the following characteristics: (1) Accompanying the differentiation of continental and oceanic segments of the earth's crust, the oceans have become somewhat deeper, and their basins perhaps larger, while the continents, on the average, have been more elevated and less subject to submergence. Hence there have been less radical departures from the present distribution during the relatively recent Cenozoic era than in the ancient Paleozoic because the submergence of continental areas has become less general and less frequent. For example, the last extensive epeiric or interior sea in North America was in the Cretaceous, at least ten million years ago, and according to Barrell perhaps fifty million, while in Europe, according to de Lapparent,[85] a smaller share of the present continent has been submerged since the Cretaceous than before.

Indeed, as in North America, the submergence has decreased on the average since the Paleozoic era. (2) The changes in distribution of land which have taken place during earth history have been cyclic.

Repeatedly, at the close of each of the score or so of geologic periods, the continents emerged more or less, while at the close of the groups of periods known as eras, the lands were especially large and emergent.

After each emergence, a gradual encroachment of the sea took place, and toward the close of several of the earlier periods, the sea appears to have covered a large fraction of the present land areas. (3) On the whole, the amount of land in the middle and high lat.i.tudes of the northern hemisphere appears to have increased during geologic time. Such an increase does not require a growth of the continents, however, in the broader sense of the term, but merely that a smaller fraction of the continent and its shelf should be submerged. (4) In tropical lat.i.tudes, on the other hand, the extent of the lands seems to have decreased, apparently by the growth of the ocean basins. South America and Africa are thought by many students to have been connected, and Africa was united with India via Madagascar, as is suggested in Fig. 9. The most radical cyclic as well as the most radical progressive changes in land distribution also seem to have taken place in tropical regions.[86]

[Ill.u.s.tration: _Fig. 9. Cretaceous Paleogeography._ (_After Schuchert._)]

Although there is much evidence of periodic increase of the sea in equatorial lat.i.tudes and of land in high lat.i.tudes, it has remained for the zoologist Metcalf to present a very pretty bit of evidence that at certain times submergence along the equator coincided with emergence in high lat.i.tudes, and vice versa. Certain fresh water frogs which carry the same internal parasite are confined to two widely separated areas in tropical and south temperate America and in Australia. The extreme improbability that both the frogs and the parasites could have originated independently in two unconnected areas and could have developed by convergent evolution so that they are almost identical in the two continents makes it almost certain that there must have been a land connection between South America and Australia, presumably by way of Antarctica. The facts as to the parasites seem also to prove that while the land connection existed there was a sea across South America in equatorial lat.i.tudes. The parasite infests not only the frogs but the American toads known as Bufo. Now Bufo originated north of the equator in America and differs from the frogs which originated in southern South America in not being found in Australia. This raises the question of how the frogs could go to Australia via Antarctica carrying the parasite with them, while the toads could not go. Metcalf's answer is that the toads were cut off from the southern part of South America by an equatorial sea until after the Antarctic connection between the Old World and the New was severed.

As Patagonia let go of Antarctica by subsidence of the intervening land area, there was a probable concomitant rise of land through what is now middle South America and the northern and southern portions of this continent came together.[87]

These various changes in the earth's crust have given rise to certain specific types of distribution of the lands, which will now be considered. We shall inquire what climatic conditions would arise from changes in (a) the continuity of the lands from north to south, (b) the amount of land in tropical lat.i.tudes, and (c) the amount of land in middle and high lat.i.tudes.

(a) At present the westward drift of warm waters, set in motion by the trade winds, is interrupted by land ma.s.ses and turned poleward, producing the important Gulf Stream Drift and j.a.pan Current in the northern hemisphere, and corresponding, though less important, currents in the southern hemisphere. During the past, quite different sets of ocean currents doubtless have existed in response to a different distribution of land. Repeatedly, in the mid-Cretaceous (Fig. 9) and several other periods, the present American barrier to the westward moving tropical current was broken in Central America. Even if the supposed continent of "Gondwana Land" extended from Africa to South America in equatorial lat.i.tudes, strong currents must still have flowed westward along its northern sh.o.r.e under the impulse of the peculiarly strong trade winds which the equatorial land would create. Nevertheless at such times relatively little warm tropical water presumably entered the North Atlantic, for it escaped into the Pacific. At several other times, such as the late Ordovician and mid-Devonian, when the isthmian barrier existed, it probably turned an important current northward into what is now the Mississippi Basin instead of into the Atlantic. There it traversed an epeiric, or mid-continental sea open to both north and south. Hence its effectiveness in warming Arctic regions must have been quite different from that of the present Gulf Stream.

(b) We will next consider the influences of changes in the amount of equatorial and tropical land. As such lands are much hotter than the corresponding seas, the intensity and width of the equatorial belt of low pressure must be great when they are extensive. Hence the trade winds must have been stronger than now whenever tropical lands were more extensive than at present. This is because the trades are produced by the convection due to excessive heat along the heat equator. There the air expands upward and flows poleward at high alt.i.tudes. The trade wind consists of air moving toward the heat equator to take the place of the air which there rises. When the lands in low lat.i.tudes were wide the trade winds must also have dominated a wide belt. The greater width of the trade-wind belt today over Africa than over the Atlantic ill.u.s.trates the matter. The belt must have been still wider when Gondwana Land was large, as it is believed to have been during the Paleozoic era and the early Mesozoic.

An increase in the width of the equatorial belt of low pressure under the influence of broad tropical lands would be accompanied not only by stronger and more widespread trade winds, but by a corresponding strengthening of the subtropical belts of high pressure. The chief reason would be the greater expansion of the air in the equatorial low pressure belt and the consequent more abundant outflow of air at high alt.i.tudes in the form of anti-trades or winds returning poleward above the trades. Such winds would pile up the air in the region of the high-pressure belt. Moreover, since the meridians converge as one proceeds away from the equator, the air of the poleward-moving anti-trades tends to be crowded as it reaches higher lat.i.tudes, thus increasing the pressure. Unless there were a corresponding increase in tropical cyclones, one of the most prominent results of the strengthened trades and the intensified subtropical high-pressure belt at times of broad lands in low lat.i.tudes would be great deserts. It will be recalled that the trade-wind lowlands and the extra-tropical belt of highs are the great desert belts at present. The trade-wind lowlands are desert because air moving into warmer lat.i.tudes takes up water except where it is cooled by rising on mountain-sides. The belt of highs is arid because there, too, air is being warmed, but in this case by descending from aloft.