The weather and climate are obviously very complex.  Physical influences on the climate occur at all spatial and time scales.  The picture has been greatly complicated in this century by the rapid growth of another influence: man's industrial activities and an unprecedented release of chemicals into the atmosphere.  Fortunately none of the factors that control the climate have yet been so great that they would threaten human survival, but major catastrophic events have happened in the past, long before man appeared.

The theories of Ice Ages and Global Warming are highly dependent on oceanic and atmospheric circulation.  The ocean and atmosphere are critical to the transfer of heat and moisture between regions at different latitudes.  If the atmosphere were motionless, or if the circulation was entirely along bands of constant latitude, the Earth would be a quite different place.  The circulation of both the oceans and atmosphere are fully three dimensional, and involves a multitude of spatial and temporal scales.

The large-scale circulation of the oceans and atmosphere have apparently changed greatly in the past.  The circulation today is conducive to highly seasonal weather.  The location of North and South America, with no breaks between them constrains the circulation in the Atlantic Ocean to a predominantly north-south flow—facilitating exchange of heat between the poles and tropics.  Moreover, the deflection of the atmospheric circulation by the North American Cordillera promotes a highly seasonal climate in the Northern Hemisphere.  In the Southern Hemisphere, where there is a broad gap between Antarctica and the other continents, the circulation is quite different and the weather over the continents less extreme.  The current configuration is a relatively recent development, and quite different from the configuration during theMesozoic and early Cenozoic Eras.

Several other planets, notably Jupiter and Saturn, have mainly east-west, or zonal, circulation.  On the Earth, as anyone who has lived on the Great Plains knows, the circulation can transport cold or hot air across 20 or more degrees of latitude.  The effect is especially important in the winters, when arctic air can sometimes penetrate as far south as Texas.  In understanding how that comes about we must first look at the east-west and vertical transport of air, as illustrated in the figure.  Warm air rising at the equator or sub-solar point drives the Hadley cells, which are rather like huge coils lying along zones of constant latitude.  The air descends again at higher latitudes.  The rotation of the Earth, along with the conservation of angular momentum, causes a Coriolus Force which deflects the descending air toward the east and the ascending air toward the west.  This east-west deflection establishes the high-altitude jet streams that drive much of the weather in mid to high latitudes.

In the southern hemisphere the eastward motion of the air is almost unimpeded between 40 and 70 degrees of latitude, so the winds blow fiercely around the pole.  In the north there are land masses to hinder the flow of air.  The air must rise to cross the North American cordillera, which runs along the entire western side of the continent.  When the air descends again on the lee of the mountains, it is deflected southward by the action of the Coriolus Force.  The jet stream acquires a kink, and the circulation wraps up into the familiar high and low pressure cells.

60 million years ago the distribution of the continents was quite different.  If we are to understand the climate in the Mexozoic and earlier ages, we must understand the effects of the continents.  We know that in the age of the dinosaurs there were many apparently cold-intolerant animals living at very high latitudes.  This might imply a greater degree of cold tolerance than previously suspected for some dinosaurs; and this idea has been taken up by many who believe the dinosaurs were warm blooded.  However, when one finds dinosaur remains in Antarctica, one must infer a much warmer, less seasonal climate than now prevails in that region.  We know from other evidence that the glaciation of Antarctica is a relatively recent development; dating from no earlier than about 40 million years ago.  We must conclude that the climate was not just warmer in the past, but the difference between the equator and poles may have been reduced.

The computations needed to model the oceanic and atmospheric circulation, and the change in the circlation over periods of days to hundreds of years, can only be done with aid of powerful computers.  The methods for accomplishing these computations have evolved rapidly, so that computations that were beyond anyone's wildest dreams several decades ago are now performed routinely at many research centers.

To construct numerical models in order to simulate the changes in the weather and climate, one must first set up computational grids comparable to the scales of the smallest phenomena one wishes to examine.  In order to simulate the daily weather one would rather not have to take account of small-scale effects, such as the flow of air over a building or brief gusts of wind.  Likewise, in order to simulate the evolution of climate over millions of years, clearly it is impractical to include the daily weather patterns, dominated by rapdily moving high and low pressure regions and storms.  Generally the smallest scale is of the order of 1 to 100 km for short-term weather simulations, 100 to 1000 km for simulations of climate over tens to hundreds of years, and 1000 km or greater for simulations of climate over hundreds to thousands of years.

The problem of computational scales is partly a practical one; no one wishes to perform a computation that takes longer than the actual event.  There is also a fundamental problem with computational scales in the solution of highly nonlinear problems such as weather and climate simulations.  Even the tiniest of errors at the smallest scale can build up and negate the results.  In order to ensure that those tiny errors do not spoil the solutions, it has been necessary to develop many "tricks" to ensure stability.  This usually means that the computational models may not reveal some real instabilities that could cause the real circulation to change relatively suddenly.  Such sudden changes, or "tipping points," have been the focus of much concern in recent years.

Another scale problem is a result of the vastly different properties of the ocean, lower atmosphere, and upper atmosphere.  Ocean currents are relatively slow, and great amounts of computation time would be wasted if the ocean circulation had to be updated every time the atmospheric computatation is stepped forward.  There are several approaches to this problem, one of which is to run two essentially separate simulations.  In the atmospheric simulation the ocean is treated as being an essentially static boundary that changes very slowly over time.  The feedback from the atmosphere to the ocean can usually be averaged over relatively long times, so this approach is usually quite safe.

During the ice ages the atmospheric circulation was profoundly affected by the cold masses of ice.  The ice built up to depths comparable to mountains—several kilometers in North America.  The cold air blew down the slopes and produced katabatic winds which were deflected toward the right by the coriolus force.  The result was a continuous clockwise, or anticyclonic, wind blowing around the margins of the ice sheet; average wind velocities may have been as great as 100 km per hour.  These winds swept up huge amounts of soil and sand from the barren regions at the margin of the ice, and deposited them far from the ice.

The large scale disturbance of the surface flow was accompanied by a diversion of the jet stream.  Simulations of the flow indicate that the jet stream may have split into two parts, flowing to the north and to the south of the ice (the North American and European ice caps did not extend the whole way to the North Pole).  The effects of the continental ice caps reduced the average temperatures, and may have stabilized the circulation at low altitudes.  This could have resulted in reduced seasonal variations for regions far to the south of the ice.

Though we experience the weather mainly though the influence of the winds and moisture in the atmosphere, the oceanic circulation is crucial to the Earth's climate.  Both the atmosphere and the oceans transport heat and moisture to the continents, but the thermal capacity of the oceans is far greater than the atmosphere.  The water in the northern oceans circulates in huge swirls that cool the eastern shores of the continents and warm the western shores.  The gulf stream in the Atlantic Ocean brings relatively warm water northeastward toward Europe.  The water is not hot, but it is warm enough to give Bergen, Norway, at 60 degrees Latitude a climate as mild as Portland, Maine at 45 degrees Latitude.

El Niño is a phenomenon related to the circulation of water near the equator, and mainly affects the west coast of South America, though its effects are felt as far north as British Columbia.  Usually the oceanic circulation along the equator is in a sequence of horizontally aligned rolls.  The El Niño Southern Oscillation (ENSO) is a breakup and reorganization of those rolls at intervals of 5 to 8 years.  In one phase of the ENSO the surface currents off the South American coast flow toward the coast, bringing warm water.  In the other phase the surface currents flow away from the South American coast, accompanied by the upwelling of cold water from deep in the Pacific Ocean.  The circulation system can be thought of as an unstable oscillator, with two states.  The precise mechanism that causes the switch from state to the other is not understood.  Enough is known, however, to empirically model the effects on the atmosphere and climate once either state is established.

A drought is a sparsity of moisture sufficient to retard the growth of vegetation throughout a widespread regions.  Because of the complicated flow of moisture in the atmosphere, some regions, such as the Sahara Desert are in a more-or-less permanent state of drought.  The Sahara and many other arid deserts were not always dry. Some 12,000 to 6,000 years ago, at the end of the last Ice Age, the Sahara supported abundant vegetation.  But it was probably susceptible to droughts, and eventually the droughts became perpetual.  Droughts lasting from a year to several decades occur frequently at the margins of deserts and in continental interiors.  The American Great Plains and much of western North America have been subject to long periods of drought, sometimes lasting many decades.

Zebulon Pike's expedition to the Western Plains in 1810 returned reports of an arid region, which came to be known as "The Great American Desert".  Pike's reports have often been cited as an example of how early explorers were misled by their own expectations, based on the needs of agriculture.  We know now that Pike went west during a period of drought.  Moreover, he visited some of the most arid parts of the Great Plains.

Two other explorers, Isaac I. Stevens and John Palliser, visted the western plains in the 1850's.  Palliser also reported severe aridity on the Candian Plains, in what is now Alberta and Saskatchewan.  Palliser was accompanied by trained naturalists, so the veracity of his reports should not be doubted.  The arid region he described, later known as Palliser's triangle, extends from the U.S.  border north to the North Saskatchewan River, and is now the heart of one of the world's greatest wheat producing regions.

Stevens' reports were somewhat more equivocal.  He traveled on U.S. Great Plains, not far south of the regions visited by Palliser.  The official reports of that expedition made claims of high agricultural potential; but a careful reading reveals that those conclusions were perhaps too optimistic.  Everywhere we find observations of lack of vegetation, such as rivers that dried up in the summer.

Palliser and Stevens, too, had the ill luck to come west during one of the worst droughts to hit the Great Plains in the past several hundred years.  That drought, which marked the end of the Little Ice Age, began in 1840 and lasted until the mid 1860's.

Prolonged droughts, lasting more than several years, have struck the Great Plains at intervals of 75 years since the end of the Little Ice Age.  After the drought of 1840 – 1865 there was the "dustbowl" drought of the 1920's and 1930's. The most recent major drought on the northern Plains began about 1996, and has continued into the 2000's.

In North America, with such short written records, we have had to rely on scientific reconstructions of the climates of the past.  The analysis of tree rings has been very useful in determining when the trees had a surplus or a shortage of moisture.  There are some handicaps of using such "proxy" data, as with any method of reconstructing past climates, but tree ring analysis remains the best means of determining when droughts made it difficult for plants to flourish.  Several records for the past 400 years are shown below [taken from the work of Fritz and Shao].  The first is for the northern Great Plains, which are especially sensitive to the occurrence of prolonged droughts.  There have been many periods of reduced moisture lasting as long as several decades.  The droughts of 1805 – 1815 and 1840 – 1865 show up very clearly.  The southern Great Plains provide a similar record.  The two records from the plains are similar, and suggest that droughts on the Plains tend to last for ten or more years once they have begun.  The record from the southwest deserts is subtly different, showing more rapid variations as the moisture fluctuated from year to year.

The rainfall records are accompanied by reconstructed temperatures, indicating that droughts on the plains may be associated with periods of elevated temperatures.  That association is weaker in the southwest deserts.

The effects of the great drought in the middle of the 19th century must have been perceptible until late in the century.  Both the composition of the vegetation and the establishment of trees were affected.  Explorers and settlers from 1860 to the end of the century encountered far fewer mature trees than are present today.  The scarcity of trees may be partly due to the prevalence of fires, which were encouraged by drought conditions; but this can hardly explain why the trees everywhere were less abundant than today.  It is likely that the scarcity of trees is partly attributable to the failure of young trees to become established.

Early photographs of the Great Plains generally show a nearly treeless region.  A particularly instructive example is from the Little Bighorn Battlefield in southern Montana, where, because of the historical associations, there is a long series of photographs of an otherwise unintersting location on the prairies.  The earliest photographs, from about 1877 to 1880, show far fewer trees along the Little Bighorn River in the distance.  Compare an early photograph with a modern photograph taken from exactly the same spot.  Both photographs were made in the spring, so the foliage and vegetation should be approximately the same.  That there are considerable numbers of older cottonwood trees, which would have been quite vulnerable to fires, in the earlier photograph is evidence that the scarcity of vegetation is not entirely due to fires.

Most talk of drought cycles is of a non-scientific kind, and merely reflects the observation that droughts come and go over periods of many years.  Any convincing evidence that there is actually a regular periodicity in the occurrence of droughts would be extremely valuable.

Tides are not usually thought of as influencing the weather, but over many years the slight changes induced in the oceans' circulation could have significant effects.  If the Moon's orbit were absolutely stable the effect would be lost in all the larger influences.  But the lunar orbit is tilted at 5° to the ecliptic plane, and the orbit precesses with a period of about 18.6 years.  This precession is responsible for the Saros cycle of eclipses, which repeat every 18.03 years, and for the 18.6 year tidal cycle.  The two cycles should not be confused.  Both eclipses and tides are affected by the relative declination of the Moon and Sun.  Every 18.6 years the tides in the northern ocean basins experience a maximum amplitude, when the Moon is highest in the sky in the spring or fall.

As we look at longer periods, of more than 100 years, we have to consider the possibility of variations of the Sun's output.  The Earth's climate is highly dependent on the receipt of solar energy at the surface.  Being immersed in the highly variable atmosphere of the Earth, it would be easy to overlook the possibility that part of the perceived climate variations may be due to changes in the energy output of the Sun.

As geologists began to unravel the history of previous ice ages, it soon became clear that there is a distinct periodicity to episodes of continental glaciation.  During the past million years, vast glaciers have covered large parts of North America and Europe at regular intervals of about 100,000 to 125,000 years.  An explanation for cyclic ice ages was put forth early in the twentieth century by Milutin Milankovitch, a Serbian meteorologist, in terms of a coupling of the wobble and change of eccentricity of the Earth's orbit.  It provides a convincing explanation of the major climatic changes over the past 3 million years, with 125,000 year glacial cycles and other cycles at shorter periods.  The Milankovitch hypothesis was not accepted for many years because the subtle changes in the heating of the earth were thought to be too slight to cause major ice ages.  Within the past several decades the theory has been revived, tested, and accepted by nearly all workers in climate history.

Three major periodic cycles are at work.  The longest is a change in the ellipticity of the Earth's orbit, with periods of 95,800 years and about 400,000 years.  The ellipticity is small, and the greatest value only permits a difference of several percent between the closest distance and the farthest from the sun; but it can lead to a perceptible variation in the solar energy received during the summer, the critical time for melting ice and snow.  There is also a variation of the inclination of the Earth's axis, with a period of 41,000 years.  The extreme values of the inclination are 21.4° and 24.4°.  The axis also wobbles at period of 21,700 years.  This "precession of the equinoxes" cause a change in the phase of the seasons, relative to the location in the orbit.

The three cycles combine to produce variations in the amount of heating and the length of the seasons.  The effect is most pronounced when the Earth is farthest from the Sun during the northern winter.  The northern hemisphere is critical to the formation of large glaciers because most of the land is concentrated there.  When ice accumulates on the land, it can spread far southward and remain over many years.  The greater the extent of the ice, the greater the feedback effect due to solar heat reflected back to space.  The glaciers grow not because of overall temperature decreases, but because there is insufficient heating during the summer to melt the accumulated ice.  Of course the overall temperatures will decrease due to the increased reflection of solar radiation back into space by snow covering much of the northern hemisphere.  The figure shows how the increased eccentricity (exaggerated) of the orbit leads to longer winters.

The increased orbital eccentricity during ice ages has two effects.  The first is to take the Earth further from the Sun during the northern winter.  The other is a subtle decrease in the length of the summer.  Due to Kepler's Second Law, the Earth moves more slowly when it is further from Sun, so the spring equinox is retarded and the autumn equinox is advanced—thus shortening the summer.  The heat received from the sun during the summer is critical because the growth of polar ice caps is largely due to reduced melting during the summer.  The temperature during the winter is not very important, just so it remains below the freezing point. 

Once the ice begins to accumulate there is a feedback effect as the ice reflects the sunlight and leads to further cooling.  The Milankovitch theory gives a convincing explanation of the triggering of cycles of glaciation, but it does not explain why the cycles did not appear in earlier ages, when the Earth experienced millions of years of warm climates.  Apparently there are other conditions which must be met before the Milankovitch mechanism can be effective.  The picture of ice ages must also include changes in the composition and circulation of the atmosphere and oceans.