|
Past Cycles: Ice Age Speculations
To understand climate change, the obvious
first step would be to explain the colossal coming and going of ice
ages. Scientists devised ingenious techniques to recover evidence from
the distant past, first from deposits left on land, then also from sea
floor sediments, and then still better by drilling deep into ice. These
paleoclimatologists succeeded brilliantly, discovering a strangely
regular pattern of glacial cycles. The pattern pointed to a surprising
answer, so precise that some ventured to predict future changes. The
timing of the cycles was apparently set by minor changes in sunlight
caused by slow variations of the Earth's orbit. Just how that could
govern the ice ages remained uncertain, for the climate system turned
out to be dauntingly complex. One lesson was clear: the system is
delicately poised, so that a little stimulus might drive a great
change. (There is a separate essay on shorter-term climate
fluctuations, lasting a few years to a century or so, possibly related
to Variations of the Sun.)
"The origin
of these climatic trends... is a difficult subject: by long tradition
the happy hunting ground for robust speculation, it suffers because so
few can separate fact from fancy." — G.S. Callendar(1)
Evidence and Speculations (to 1954)
| It was an
incredible claim, yet the evidence was eloquent. The scraped-down rock
beds, the boulders perched wildly out of place, the bizarre deposits of
gravel found all around northern Europe and the northern United States,
all these looked exactly like the effects of Alpine glaciers
— only far, far larger. By the late 19th century, after
passionate debate, most scientists accepted the incredible. Long ago
(although not very long as geological time went, for Stone Age humans
had lived through it), northern regions had been buried kilometers deep
in continental sheets of ice. This Ice Age stood as evidence of a
prodigious climate change. |
| Toward the end of the 19th century, field studies
by geologists turned up another fact, almost as surprising and
controversial. There had been not one Ice Age but several. The
stupendous ice sheets had slowly ground south and retreated, time and
again. The series of glacial periods had alternated with times of
warmer climate, each cycle lasting many tens of thousands of years.
German geologists, meticulously studying the scars left by ancient
rivers on what were now hillsides in the Alps, worked out a scheme of
four major cycles.(1a)
|
| Geologists turned up evidence that the past few
million years, during which the ice sheets cycled back and forth, was
an unusual time in the Earth's history. They gave it a name of its own,
the Pleistocene epoch. Before that there had been long eras of less
turbulent climate, when fossils of tropical plants and animals had been
deposited in regions that were now frigid. Much farther back there had
been a few other relatively brief epochs of glaciation, revealed by
very ancient ice-scraped rocks and gravel deposits. Most geologists
concluded that the planet’s climate had at least two possible
states. The most common condition was long temperate epochs, like the
balmy times of the dinosaurs. Much rarer were glacial epochs like our
own, lasting a few millions of years, in which periods of glaciation
alternated with warmer "interglacial" periods like the present. This
essay does not cover studies of the very remote past, before the
Pleistocene. |
| What could explain the change from a warm to a
glacial epoch, and the cycling of ice ages within a glacial epoch? A
solution to the puzzle would bring deep satisfaction and eternal fame
to whoever solved it. Perhaps the solution would also tell when the
next ice age might descend upon humanity. |
| Many theories were offered from the mid-19th to
the mid-20th century. None amounted to more than plausible hand-waving.
Most favored were ideas about how the uplift of mountain ranges, or
other reconfigurations of the Earth's surface, would alter the
circulation of ocean currents and the pattern of winds. Other theories
ranged from the extraterrestrial, such as a long-term cyclical
variation of solar energy, to the deep Earth, such as massive volcanic
eruptions. All these theories shared a problem. Given that something
had put the Earth into a state conducive to glaciation, what made the
ice sheets grow and then retreat, over and over again? None of the
theories could readily explain the cycles. |
| Many things in
the natural world come and go in cycles, so it was natural for people
to suppose that there was a regular pattern to the ebb and flow of ice
sheets. After all, there was evidence — convincing to many
meteorologists, although doubted by as many more — that
temperature and rainfall varied in regular cycles on human timescales
of decades or centuries. The glacial periods of the ice ages likewise
seemed to follow a cyclical pattern, on a far grander timescale. A
series of repeated advances and retreats of the ice was visible in
channels carved by glacial streams and in the fossil shorelines of
lakes in regions that were now dry. If the pattern of advance and
retreat could be measured and understood, it would give a crucial clue
to the mystery of ice ages. |
| Simple observations of surface features were
joined by inventive methods for measuring what a region's climate had
been like thousands or even millions of years ago. In particular, from
the early 20th century forward, a few scientists in Sweden and
elsewhere developed the study of ancient pollens ("palynology"). The
tiny but amazingly durable pollen grains are as various as sea shells,
with baroque lumps and apertures characteristic of the type of plant
that produced them. One could dig up soil from lake beds or peat
deposits, dissolve away in acids everything but the sturdy pollen, and
after some hours at a microscope know what kinds of flowers, grasses or
trees had lived in the neighborhood at the time the layer of lake-bed
or peat was formed. That told scientists much about the ancient
climate. We had no readings from rain gauges and thermometers 50,000
years ago, but pollen served as an accurate "proxy." |
| Studying ancient pollens,
scientists found again a sequence of colder and warmer spells, glacial
and interglacial periods. The most recent ice age had ended ten
thousand years or so ago. Other ingenious studies showed that a
particularly warm period had followed. For example, fossil hoards of
nuts collected by squirrels revealed that five thousand or so years
ago, hazel trees had grown farther north in Sweden than at present.
Were we drifting toward another ice age?(2)
|
| The problem that
researchers set themselves was to find a pattern in the timing of the
changes. Unfortunately, there were no tools to accurately determine
dates so far in the past; any figure might be wrong by thousands of
years. That did not stop people from seeing regular patterns. An
example was a 1933 study of ancient beach deposits by W.M. Davis. As
the continental ice sheets formed and then melted, they had locked up
and then released so much water that the oceans had dropped and risen
many tens of meters. Wave-carved fossil shores stood as testimony of
the different sea levels. Davis believed he saw a pattern, in which the
warm interglacial periods were long. Our own time seemed near to the
preceding ice age, so he concluded that the Earth ought to get warmer
for a while before it cooled again. When this was added to reports that
the climate of the 1930s was measurably getting warmer, predictions
appeared in Science magazine and in the public
press that "The poles may become useful and inhabited places."(3) |
| The pattern of past changes, no matter how
accurately geologists might measure it, would always be suspect until a
plausible theory explained it. Of all the proposed theories, only one
was bound by its very nature to give regular cycles of change. This
theory promised, moreover, to give the timing of past changes precisely
from basic physical principles, and to predict future ice ages. The
history of the measurement of ancient climates is inseparable from the
history of this "astronomical" theory. |
| In the mid 19th century, James Croll, a
self-taught British amateur scientist, published calculations of how
the gravitational pulls of the Sun, Moon, and planets subtly affect the
Earth's motion and orientation. The inclination of the Earth's axis and
the shape of its orbit around the Sun oscillate gently in cycles
lasting tens of thousands or hundreds of thousands of years. During
some periods the Northern Hemisphere would get slightly less sunlight
during the winter than it would get during other centuries. Snow would
accumulate. Croll argued that this would change the pattern of trade
winds, leading to the deflection of warming currents like the Gulf
Stream, and finally a self-sustaining ice age. The timing of such
changes could be calculated exactly using classical mechanics (at least
in principle, for the mathematics were thorny). Croll believed that the
timing of the astronomical cycles, tens to hundreds of thousands of
years long, roughly matched the timing of ice ages.(4*)
|
| Most scientists found Croll's ideas unconvincing,
and his timing of the ice ages wholly wrong.(5)
Nevertheless a few enthusiasts
pursued his astronomical theory. It became almost plausible in the
hands of the Serbian engineer Milutin Milankovitch. Working in the
1920s and 1930s, he not only improved the tedious calculations of the
varying distances and angles of the Sun's radiation, but also came up
with an important new idea. Suppose there was a particular time when
the sunlight falling in a high-latitude zone of a given hemisphere was
so weak, even in the summer, that the snow that fell in winter did not
all melt away? It would build up, year after year. As others had
pointed out, a covering of snow would reflect away enough sunlight to
help keep a region cold, giving an amplifying feedback. Under such
circumstances, a snowfield could grow over the centuries into a
continental ice sheet.(6) |
| "The possibility of
dating the varying episodes of the Pleistocene ice ages by correlating
them with the [Milankovitch] radiation curve appealed to a number of
workers," a meteorologist reported in 1940. "Correlations with the
radiation curve were found everywhere."(7)
It was also encouraging that even
the tiny changes in solar radiation that came with the eleven-year
sunspot cycle had some effect on weather — at least according
to some studies. By the 1940s, some climate textbooks were teaching
that Milankovitch's theory gave a plausible solution to the problem of
timing the ice ages.(8) |
| Supporting evidence came from "varves," a Swedish
word for the pairs of layers seen in the mud covering the bottom of
northern lakes. Each year the spring runoff laid down a thin layer of
silt followed by a settling of finer particles. From bogs and outcrops
where the beds of fossil lakes were exposed, or cores of slick clay
drilled out of living lakes, researchers painstakingly counted and
measured the layers. Some reported finding a 21,000-year cycle of
changes. That approximately matched the timing for a wobbling of the
Earth's axis which Milankovitch had calculated as a crucial element
(namely, the precession of the equinoxes, in fact a combination of
19,000- and 23,000-year
cycles).(9) |
| Most geologists, however, dismissed the
astronomical theory. For they could not fit Milankovitch’s
timing to the accepted sequence of four ice ages. A generation of
geologists had laboriously constructed this sequence from studies
around the world of surface features, such as the gravel deposits
(moraines) that marked where glaciers had halted, and hillside terraces
that showed the level of ancient rivers. The Milankovitch theory, wrote
one authority condescendingly in 1957, had served a useful function as
"a dogma of faith" that had stimulated research, but compared with the
actual glacial record, the orbital chronology "must be stamped as
illusory." Another problem lay in the fact that ice sheets had spread
at the same time in the Northern and Southern Hemispheres. Since the
astronomical theory relied upon an increase in the sunlight falling on
one hemisphere along with a decrease on the other hemisphere, many
experts considered the world-wide pattern of ice ages a devastating refutation.(10)
Finally, there was a basic physical argument against the theory which
seemed insurmountable. |
| One reviewer —
who had himself seen 21,000-year variations in lake deposits
— explained at a 1952 conference that it was a problem of
magnitudes. The computed variations in the angle and intensity of
incoming sunlight were only tiny changes, "insufficient to explain the
periods of glaciation."(11)
Meanwhile, the studies that had
found correlations between sunspot cycles and weather had all turned
out wrong, giving an air of cranky unreliability to every connection
between solar radiation variations and climate. That same year, a
leading American planetary scientist wrote a European colleague to ask
how the astronomical theory stood over there, remarking that "People I
have consulted in this country... are not impressed by this work." His
correspondent replied, "I have discussed the question of the appraisal
of Milankovitch's theory with colleagues here. They are of the opinion
that the theory cannot account for past changes. The effects are too
small and the chronology of the occurrence of glaciation is so
uncertain that any correspondence... appears fortuitous."(12)
So what had caused the ice ages? That was still anybody's guess. |
Reliable
Dates and Temperatures (1955-1971)
|
| The tool that
would unlock the secret was constructed in the 1950s, although it took
scientists a decade to make full use of it. This tool was radiocarbon
dating. It could tell with surprising precision the age of features
like a glacial moraine. You only needed to dig out fragments of trees
or other organic material that had been buried thousands of years ago,
and measure the fraction of the radioactive isotope carbon-14 in them.
Of course researchers had to devise and test a number of laboratory
techniques before they could get trustworthy results. Once that was
done, they could assign a reliable timescale to the climate
fluctuations that had previously been sketched out by various
traditional means. The best of these means, in the 1950s, was pollen
science. The study of ancient climates (as manifested by changes of
vegetation) had turned out to be invaluable for identifying strata as
an aid to oil exploration. That had paid for specialists who brought
the technique to a high degree of refinement.(13)
Carbon-14 measurements could
now assign accurate dates to the palynologists’ tables of
cool and warm periods in northern regions. For example, dating of lake
deposits in the Western United States showed surprisingly regular
cycles of drought and flood — which seemed to match the
21,000-year cycle predicted by Milankovitch. But other carbon-14 dates
seemed altogether out of step with the Milankovitch timetable. |
| The swift postwar
development of nuclear science meanwhile fostered another highly
promising new technique. In 1947, the nuclear chemist Harold Urey
discovered a way to measure ancient temperatures. The key was in the
oxygen built into fossil shells pulled up from the sea floor. The
amount of heavier or lighter oxygen isotopes that an organism took up
from sea water varied according to the water's temperature at that
time, so the ratio (O18/O16 ) served as a proxy
thermometer.(14) This ingenious
method was taken up by Cesare Emiliani, a geology student from Italy
working in Urey's laboratory at the University of Chicago. Emiliani
measured the oxygen isotopes in the microscopic shells of foraminifera,
a kind of ocean plankton. Tracking the shells layer by layer in long
cores of clay extracted from the seabed, he found a record of
temperature variations. Emiliani's 1955 paper, a landmark of
paleoclimatology, provided the world's first high-quality record of ice
age temperatures.(15) |
| Historians usually treat techniques as
a stodgy foundation, unseen beneath the more exciting story of
scientific ideas. Yet techniques are often crucial, and controversial.
The stories of two especially important cases are explored in short
essays on Uses of Radiocarbon Dating
and Temperatures from Fossil Shells. |
| Emiliani tentatively identified the rises and
dips of temperatures with the geologists' traditional chronology of the
past three glacial periods. His efforts were motivated largely by a
desire to learn something about the evolution of the human race, which
had surely been powerfully influenced by the climate shocks of the ice
ages. But his results turned out to tell less about the causes of human
evolution than about the causes of climate change. To get a timescale
connecting the temperature changes with depth down the core, he made
carbon-14 measurements covering the top few tens of thousands of years
(farther back there was too little of the isotope left to measure).
That gave him an estimate for how fast sediments accumulated on the
seabed at that point. Emiliani now found a rough correlation with the
varying amount of sunlight that, according to Milankovitch's
astronomical calculations, struck high northern latitudes in summer. To
get the match he had to figure in a lag of about five thousand years.
That seemed reasonable, considering how long it would take a mass of
ice to react. "A causal connection is suggested but not proved,"
Emiliani concluded.(16*) |
| Carbon-14 could date the more recent layers in
the deep-sea cores, pinning down the chronology of temperature changes
with unprecedented precision. The chemist Hans Suess, another graduate
of Urey's lab, took the lead in improving the chronology. He reported,
among other things, that the last ice age had come to a surprisingly
abrupt end, starting sometime around 15,000 years ago. Looking farther
back, Suess found hints of a roughly 40,000-year cycle, which sounded
like the 41,000-year cycle that Milankovitch had computed for slight
variations in the inclination of the Earth's axis.(17)
Emiliani too, reporting a cycle
of roughly 50,000 years, was increasingly confident that orbital
changes set the timing of ice ages.(18)
His curves, however, did not
match up with the canonical four ice ages. |
| To resolve the issue, Emiliani began urging
colleagues to launch a major program and pull up truly long cores, a
hundred-meter record covering many hundreds of thousands of years. But
for a long time the drillers' crude techniques were incapable of
extracting long, undisturbed cores from the slimy ooze. As one of them
remarked ruefully, "one does not make wood carvings with a butcher's
knife."(19) |
| Meanwhile suggestive evidence
was dug out (literally) by the geochemist Wallace Broecker and
collaborators. Ancient coral reefs were perched at various elevations
above the present sea level on islands that geological forces were
gradually uplifting. The fossil reefs gave witness to how sea level had
risen and fallen as ice sheets built up on the continents and melted
away. The coral could be dated by hacking out samples and measuring
their uranium and other radioactive isotopes. These isotopes decayed
over millennia on a timescale that had been accurately measured in
nuclear laboratories. Unlike carbon-14, the decay was slow enough so
there was still enough left to measure after hundreds of thousands of
years. As a check, the sea level changes could be set alongside the
oxygen-isotope temperature changes measured in deep-sea cores. Again
the orbital cycles emerged, plainer than ever. At a conference on
climate change held in Boulder, Colorado in 1965, Broecker announced
that "The Milankovitch hypothesis can no longer be considered just an
interesting curiosity."(20)
People at the conference began
to speculate on how the calculated changes in sunlight, although they
seemed insignificantly small, might somehow trigger ice ages. That
could happen if the climate system were so delicately balanced that a
small push could prompt it to switch between different states. |
| Emiliani improved his
measurements, thanks to a fine set of cores that reached back more than
400,000 years. He announced he could not make the data fit the
traditional ice ages timetable at all. He rejected the entire scheme,
painstakingly worked out around the end of the 19th century in Europe
and accepted by generations of geologists, of a Pleistocene epoch
comprising four major glacial advances alternating with long and
equable interglacial periods. Emiliani said the interglacials had been
briefer, and had been complicated by irregular rises and falls of
temperature, making dozens of ice ages.(21)
Many other scientists found his
chronology dubious, but he defended his position tenaciously. Most
significantly, he believed the sequence correlated rather well with the
complex Milankovitch curve of summer sunlight at high northern
latitude. Calculating how the cycle should continue in the future, in
1966 Emiliani predicted that "a new glaciation will begin within a few
thousand years."(22) It was a
step toward what would
soon become widespread public concern about future cooling. |
| Seldom was such work straightforward. Geologists
defended their traditional chronology passionately and skillfully. For
a few years they held their ground, for it turned out that Emiliani's
data on oxygen isotopes taken up in plankton shells did not directly
measure ocean temperatures after all. Emiliani fiercely defended his
position, but other workers in the late 1960s convinced the scientific
community that he was mistaken. When water was withdrawn from the
oceans to form continental ice sheets, the heavier and lighter isotopes
evaporated and fell as rain or snow in different proportions. The way
plankton absorbed oxygen at a given temperature did not matter so much
as what proportion of each isotope was left in the sea water as ice
sheets came and went. |
| Yet in a deeper sense Emiliani was vindicated.
Whatever the forces that changed the isotope ratio, its rise and fall
did represent the coming and going of ice ages. "Emiliani's
'paleotemperature' curve," the new findings revealed, "...may be
renamed a 'paleoglaciation' curve."(23)
|
| These changes did turn out to correlate with ocean
surface temperatures. New evidence for that came from scientists who
took a census of the particular species of foraminifera, recognizing
that the assemblage of different species varied with the temperature of
the water where the animals had lived. The data confirmed that there
had been dozens of major glaciations during the past couple of million
years, not the four or so enshrined in textbooks. Corroborating
evidence came from a wholly different type of record. In a brick-clay
quarry in Czechoslovakia, George Kukla noticed how wind-blown dust had
built up into deep layers of soil (what geologists call "loess").
Although Kukla could not get dates that matched Emiliani's, the
multiple repetitions of advance and retreat of ice sheets were
immediately visible in the colored bands of different types of loess.
It was one of the few cases in this story where traditional field
geology, tramping around with your eyes open, paid a big dividend. |
| In 1968, still more complete and convincing
evidence came from an expedition that Broecker and a few others took to
Barbados. Terraces of ancient coral covered much of the island, rising
to hundreds of meters above the present sea level. The dates for when
the coral reefs had been living (125,000, 105,000, and 82,000 years
ago) closely matched dates from Milankovitch cycles for times when the
ice sheets should have been melted and the seas at their highest
(127,000, 106,000, and 82,000 years ago). The dating matched, that is,
so long as one looked for the times when the maximum amount of sunlight
struck mid-northern latitudes during the summer. "The often-discredited
hypothesis of Milankovitch," declared Broecker and his collaborators,
"must be recognized as the number-one contender in the climatic
sweepstakes."(24*) |
| Since the Milankovitch cycles could be computed
directly from celestial mechanics, one could project them forward in
time, as Emiliani had done in 1966. In 1972, presenting more Caribbean
cores, he again advised that "the present episode of amiable climate is
coming to an end." Thus "we may soon be confronted with... a runaway
glaciation." However, he added, greenhouse effect warming caused by
human emissions might overwhelm the orbital shifts, so we might instead
face "a runaway deglaciation."(25)
|
| Some other scientists
agreed that the current interglacial warm period had peaked 6,000 or so
years ago, and should be approaching its natural end. A prominent
example was Kukla, continuing his study of loess layers in
Czechoslovakia. He could now date the layers thanks to a new technique
provided by other scientists. Geological and oceanographic studies had
shown that over the course of millions of years, now and then the
Earth's entire magnetic field flipped: the North magnetic pole became
the South magnetic pole and vice-versa. These reverses were recorded
where layers of sediment or volcanic lava had entombed the direction of
the magnetic field at the time. Geologists had worked out a chronology
in lava flows, dated by the faint radioactivity of an isotope of
potassium that decayed very slowly.(26)
If even one magnetic-field
reversal could be identified in any set of layers, it pinned down the
timing of the entire sequence. When the loess layers were dated in this
fashion, Milankovitch cycles turned up. Extrapolating the cycles into
the future, Kukla thought the next shift to an ice age "is due very
soon."(Link
from below)(27) |
| If the climate experts of the time seem to have
been a bit preoccupied with ice ages, that fitted their training and
interests. For a hundred years their field had concerned itself above
all with the ice ages. Their techniques, from pollen studies to sea
floor drilling, were devoted to measuring the swings between warm and
glacial epochs. Home at their desks, they occupied themselves with
figuring how glacial-period climates had differed from the present, and
attacking the grand challenge of explaining what might cause the
swings. Now that they were beginning to turn their attention from the
past to the future, the most natural meaning to attach to "climate
change" was the next swing into cold.(28)
|
| In 1972, a
group of leading glacial-epoch experts met at Brown University to
discuss how and when the present warm interglacial period might end. A
large majority agreed that "the natural end of our warm epoch is
undoubtedly near." Unless there were impacts from future human
activity, they thought that serious cooling "must be expected within
the next few millennia or even centuries."(29)
But many other scientists
doubted these conclusions. They hesitated to accept the Milankovitch
theory at all unless they could get definitive proof from some entirely
different kind of evidence. |
| Theories
Confirmed (1971-1988) |
| The Greenland ice
sheet is a daunting sight. Most investigators first come to it by air,
past colossal bare cliffs where unimaginable quantities of ice pour
down to the sea in a slow-motion flood. Beyond that the landscape rises
and rises, over entire mountain ranges hidden under ice, to a limitless
plain of gently undulating white. Greenland had played an important
role in the 19th-century controversy over the ice ages. A few
geologists had dared to postulate the existence, in the distant past,
of seas of solid ice kilometers thick. Then astonished explorers of
Greenland found just such a thing beneath their skis. |
| In the late 1950s, scientists came back to
Greenland, hoping to find the key to the history of climate change. The
logistics were arduous, but there was good support thanks to the
International Geophysical Year — backed up by the United
States government's concern to master the Arctic regions that lay on
the shortest air routes to the Soviet Union. At Camp Century,
Greenland, workers drilled short cores from the ice to demonstrate that
it could be done. An improved drill, brought onto the ice in 1961,
produced glistening cores 5 inches in diameter in segments several feet
long. This was no small feat in a land where removing your gloves for a
few minutes to adjust something might cost you the skin on your
fingertips, if not entire fingers. After another five years of
difficult work, organized by the U.S. Army's Cold Regions Research and
Engineering Laboratory, the drill at Camp Century reached bedrock. The
hole reached down some 1.4 kilometers (7/8 of a mile), bringing up ice
as much as 100,000 years old.(30*)
Two years later, in 1968,
another long core of ancient ice was retrieved from a site even colder
and more remote: Byrd Station in West Antarctica.(31)
|
| Much could be read from these cores. For
example, individual layers with a lot of acidic dust pointed to past
volcanic eruptions. Individual eruptions could be assigned dates simply
by counting the annual layers of ice.(32)
(Known eruptions like the
destruction of Pompeii in the year 79 gave a check on the counts.)
Farther down the layers became blurred, but approximate dates could
still be assigned. Deep in the ice there were large amounts of mineral
dust, evidence that during the last ice age the world had been windier,
with storms carrying dust clear from China. Still better, ancient air
had been trapped and preserved as bubbles in the ice, a million tiny
time capsules packed with information about past climates. However, for
a long time nobody could figure out how to extract and measure the air
reliably. |
| In the early
years, the most useful work was done from the ice itself. The method
had been worked out back in 1954 by an ingenious Danish scientist,
Willi Dansgaard. He showed that the ratio of oxygen isotopes (O18/O16)
in the ice measured the temperature of the clouds at the time the snow
had fallen — the warmer the air, the more of the heavy
isotope got into the ice crystals.(33)
It was an exhilarating day for
the researchers at Camp Century, making measurements along each
cylinder of ice after it was pulled up from the borehole, when they saw
the isotope ratios change and realized they had reached the last ice
age. The preliminary study of the ice cores, published in 1969, showed
variations that indicated changes of perhaps 10°C (that is,
18°F). Some cycles were tentatively identified, including one
with a 13,000 year length.(34)
Comparison of the Greenland and
Antarctic cores showed that the climate changes were truly global,
coming at essentially the same time in both hemispheres. That put a
strict constraint on theories about the cause of cycles.(35) |
| There is a supplementary site on the
History of Greenland Ice Drilling, with some documentation of the US
"GISP" projects of the 1980s. |
| Ice core studies
also confirmed a feature that researchers had already noticed in
deep-sea cores: the glacial cycle followed a sawtooth curve. In each
cycle, a spurt of rapid warming was followed by a more gradual,
irregular descent back into the cold over tens of thousands of years. A
closer look showed that temperatures tended to cluster at the two ends
of the curve. It seemed that the climate system had two fairly stable
modes, brief warmth and more enduring cold, with relatively rapid
shifts between them. Warm intervals like the past few thousand years
normally did not last long.(36)
Beyond such fascinating hints,
however, the Greenland ice cores could say little about long-term
cycles. They were too short to reach past a single glacial cycle. And
the ice flowed like tar at great depths, confusing the record. In the
1970s, despite the arduous efforts of
the ice drillers, the most reliable data were still coming from
deep-sea cores. |
| That work too was strenuous and hazardous,
manhandling long wet pipes on a heaving deck. Oceanographers (like ice
drillers) lived close together for weeks or months at a time under
Spartan conditions, far from their families. The teams might function
smoothly — or not. Either way, the scientists labored long
hours, for the problems were stimulating, the results could be
exciting, and dedication to work seemed normal with everyone around
them doing the same. |
| To make it worthwhile, scientists had to draw on
all their knowledge and luck to find the right places to drill on the
ocean bed. In these few places, layers of silt had built up unusually
swiftly and steadily and without disturbance. Meanwhile, drilling
techniques were finally worked out that could extract the continuous
hundred-meter cores of clay that Emiliani had been asking for since the
1950s. Improved techniques for measuring the layers gave data good
enough for thorough analysis. |
| The most prominent feature turned out to be a
100,000-year cycle — evidently a key to the entire climate
puzzle. Several earlier studies had tentatively identified this
long-term cycle. Corroboration was in hand from Kukla's loess layers in
Czechoslovakia, at the opposite end of the world from some of the
deep-sea cores. Here too the 100,000-year cycle stood out.(37) |
| Yet nobody could be entirely sure. Radiocarbon
decayed too rapidly to give dates going back more than a few tens of
thousands of years. A deeper timescale could only be estimated by
measuring lengths down a core, and it was uncertain whether the
sediments were laid down at a uniform rate. For a decade controversy
had smoldered between Emiliani, as usual sticking by his original
position, and other scientists who felt that his chronology was
seriously in error. According to their data, the prominent cycle he had
seen and attributed to the 41,000-year orbital shifts was actually the
100,000-year cycle.(38) Here
again Emiliani had been
bogged down by erroneous assumptions, yet somehow had muddled through
to the fundamental truth that Milankovitch cycles were real. |
| In 1973, Nicholas
(Nick) Shackleton nailed it all down for certain. What made it possible
was the new magnetic-reversal dates established by radioactive
potassium, plus Shackleton's uncommon combination of technical
expertise in different fields. A splendid deep-sea core had been pulled
— "one of the best and most complete records of the entire
Pleistocene that is known" — the famous core Vema 28-238
(named after the Lamont Observatory's oceanographic research vessel, a
converted luxury yacht). It reached back over a million years, and
included the most recent reversal of the Earth's magnetic field, which
geologists dated at a bit over 700,000 years ago. This calibrated the
chronology for the entire core. As a further benefit, Shackleton
managed to extract and analyze the rare foraminifera that lived in the
deep sea, and which reflected basic oceanic changes independent of the
fluctuating sea-surface temperatures. The deep-sea forams showed the
same isotopic variations as surface ones, confirming that the
variations gave a record of the withdrawal of water to form ice sheets.
When Shackleton showed his graph of long-term change to a roomful of
climate scientists, a spontaneous cheer went up.(39*)
|
| The core Vema
28-238 and a few others contained such a long run of consistent data
that it was possible to analyze the numbers with a mathematically
sophisticated "frequency-domain" calculation, a well-established
technique for picking out the lengths of cycles with unimpeachable
techniques.(40)
Detailed measurements and numerical calculations found a set of favored
frequencies, a spectrum of regular cycles visible amid the noise of
random fluctuations. The first unimpeachable results (well, almost
unimpeachable) were achieved in 1976 by James Hays, John Imbrie and
Shackleton. The trio not only analyzed the oxygen-isotope record in
selected cores from the Indian Ocean, but checked their curves against
temperatures deduced from the assemblage of foraminifera species found
in each layer. |
| The long cores
proved beyond doubt what Emiliani had stoutly maintained —
there had been not four major ice ages, but dozens. The analysis showed
cycles with lengths roughly 20,000 and 40,000 years, and especially the
very strong cycle around 100,000 years, all in agreement with
Milankovitch calculations.(41)
Extrapolating the curves ahead,
the group predicted cooling for the next 20,000 years. As Emiliani,
Kukla, and other specialists had already concluded several years
earlier, the Earth was gradually (or perhaps not so gradually?) heading
into a new ice age (see
above). |
| These results, like
so many in paleoclimatology, were promptly called into question.(42)
But the main results withstood all criticism. Confirmation came from
other scientists who likewise found cycles near twenty and forty
thousand years, give or take a few thousand. The most impressive
analysis remained the pioneering work of Hays, Imbrie, and Shackleton.
They could even split the 20,000 year cycle into a close pair of cycles
with lengths of 19,000 and 23,000 years — exactly what the
best new astronomical calculations predicted. By the late 1970s, most
scientists were convinced that orbital variations acted as a
"pacemaker" to set the timing of ice ages.(43)
Science
magazine reported in 1978 that the evidence for the Milankovitch theory
was now "convincing," and the theory "has recently gained widespread
acceptance as a factor" in climate change.(44)
|
| Yet the cause
of the ice ages remained more a mystery than ever. How could the
"pacemaker" possibly work? The variation in the intensity of sunlight
that was computed for the 100,000-year astronomical cycle came from a
minor change in orbital eccentricity — a slight stretching of
the Earth's path around the sun out of a perfect circle. It was a
particularly tiny variation; the changes it caused should be trivial
compared with the shorter-term and larger orbital shifts, not to
mention all the other influences on climate. Yet it was the
100,000-year cycle that dominated the record. Scientists began to turn
from hunting down cycles to searching for the physical mechanisms that
could make the climate system respond so dramatically to subtle changes
in sunlight. As a reviewer admitted, "failures to support the
Milankovitch theory may only reflect the inadequacies of the models."(45)
A number of people took up the challenge, devising elaborate numerical
models that took into account the sluggish dynamics of continental ice
sheets. It seemed likely that eventually the modelers would produce a
suite of feedbacks that would entirely explain the schedule of the ice
ages. |
| During the 1980s,
the work advanced steadily with few surprises. Ocean drilling in
particular, pursued on an international scale, produced ever better
cores. A costly project dedicated to "spectral mapping" (SPECMAP)
yielded a spectrum of cycles that matched the astronomical calculations
with gratifying precision going back hundreds of millennia. Five
separate cores confirmed that variations in the Earth's orbit drove the
coming and going of ice ages.(46)
But an unexpected finding
brought in a new complication. The prominent 100,000-year cycle (due to
changes in the orbit's eccentricity) had dominated climate change only
during the most recent million years. During a long earlier phase of
the Pleistocene epoch, the rise and decay of ice sheets had followed
the 41,000-year cycle (due to shifts in the inclination of the Earth's axis).(47)
Milankovitch and his followers had originally expected that this cycle
would have a much stronger effect than the feeble 100,000-year shifts.
They had recognized, however, that the 41,000-year variations in
sunlight might still have been too small to cause ice ages without some
kind of amplification. The experts understood that "the response
characteristics of the Earth's climate system have themselves evolved,"
so that the details of cycling could well change.(48)
The shift in the dominant cycle
surely gave a clue, if an enigmatic one, to the variety of feedback
mechanisms at work. |
| Meanwhile ice drillers, reaching ever farther
into the past at locations where the flow of ice in the depths did not
introduce too much confusion, joined the deep-sea drillers as a main
source of information. The ice and seabed climate curves were found to
go up and down in fine agreement, and researchers began to combine data
from both sources in a single discussion. The most striking news from
the ice was evidence that the level of CO2
in the atmosphere had risen and fallen more or less in time with the
temperature. |
|
The
outstanding record was extracted by a French-Soviet team at the
Soviets' Vostok Station in Antarctica. It was a truly heroic feat of
technology, wrestling with drills stuck a kilometer down, at
temperatures so low that a puff of breath fell to the ground in
glittering crystals. Vostok was the most remote spot on the planet,
supplied once a year by a train of vehicles that clawed across hundreds
of kilometers of ice. Underfunded and threadbare, the station was
fueled by the typically Russian combination of cigarettes, vodka, and
stubborn persistence. ("What do you do for recreation?" "Wash... you
have a bath once every ten days.")(49)
|
| The effort paid off. While
the Greenland record reached into the most recent ice age, by 1985 the
Antarctic team had pulled up cores of ice stretching clear through the
cold period and into the preceding warm period—a complete
glacial cycle.(50)
During the cold period, the CO2
level had been much lower than during the warm periods before and
after. Indeed the curves of gas level and temperature tracked one
another remarkably closely. Measurements in ice cores of an even more
potent greenhouse gas, methane, showed a similar rise and fall that
matched the rise and fall of temperature.(51)
This work fulfilled the old
dream that studying the different climates of the past could be almost
like putting the Earth on a laboratory bench, switching conditions back
and forth and observing the consequences. |
| The Vostok team pointed out that the swings in
greenhouse gas levels might play an important role in amplifying the
relatively weak orbital effects. The changes in the atmosphere could
also answer the old persuasive objection to Milankovitch's theory
— if the timing of ice ages was set by variations in the
sunlight falling on a given hemisphere, why didn't the Southern
Hemisphere get warmer as the Northern Hemisphere cooled, and
vice-versa? The answer was that changes in atmospheric CO2 and methane physically linked the
two hemispheres, warming or cooling the planet as a whole.(52*) |
| After 1988 |
|
Looking at the rhythmic curves of past cycles, one
could hardly resist the temptation to extrapolate into the future. By
the late 1980s, most calculations had converged on the familiar
prediction that the natural Milankovitch cycle should bring a mild but
steady cooling over the next few thousand years. As climate models and
studies of past ice ages improved, however, worries about a swift
descent into the next great glaciation — what many in the
1970s had tentatively expected — died away. Improved
calculations said that the next ice age would not come naturally within
the next ten thousand years or so. The conclusion was backed up in 2004
by a heroic new ice core from Antarctica that brought up data spanning
the past eight glacial
cycles.(53*)
|
| The scientists
who published these calculations always added a caveat. In the
Antarctic record, atmospheric CO2
levels over the past 750,000 years had cycled between about 180 and 280
parts per million. The level in the late 20th century had now climbed
above 370 and kept climbing. (The other main greenhouse gas, methane,
was soaring even farther above any level seen in the long ice record.)
Greenhouse warming and other human influences seemed strong enough to
overwhelm any natural trend. We might not only cancel the next ice age,
but launch our planet into an altogether new climate regime. The ice
cores themselves gave convincing evidence of the threat, according to
analyses published in the early 1990s. The "climate sensitivity"
— the response of temperature to changes in carbon dioxide
— could be measured for the last glacial maximum. The answer
was in the same range that computer models were predicting for our
future, raising confidence that the models were not far wrong.(53a) |
| In climate science, where everything is subtle
and complex, it is rare for an issue to be settled. By the late 1980s,
it did seem to be an established fact that ice ages were timed by
orbital variations. The chief question that remained in the minds of
most scientists was what kind of feedbacks amplified the effect. Yet
some people challenged whether any of this was really understood. The
feedbacks that helped drive glacial cycles remained uncertain. The
cycles, most scientists now agreed, involved not only orbital
variations in solar irradiation, but also a variety of geological
effects. First came the massive settling and flow of continental ice
sheets, but large-scale physical and chemical changes in the oceans
might be important too. New evidence gave a particularly crucial role
to changes in CO2 and other
greenhouse gases. These changes were apparently driven not just by
geochemistry and ocean circulation, but still more by changes in
biological activity. And of course the biosphere depended in turn on
climate — and not just temperature, but also trickier matters
like fertilization of the seas by minerals eroded from glacial era
deserts. Further peculiar influences were added to the list of
possibilities almost every
year.(54) It would take much more study
to determine just what combination of effects determined the shape of
glacial cycles. |
| In 1992, a more fundamental challenge was raised
by the ingenious exploitation of a novel source of data: layers of
calcite laid down in the desert oasis of Devils Hole, Nevada. The
layers showed glacial and interglacial periods much like those seen in
the ice cores. The dating (using uranium isotopes) failed to agree with
Milankovitch calculations. The authors suggested that the timing of ice
ages followed no regular cycle at all, but was driven wholly by
"internal nonlinear feedbacks within the atmosphere-ice sheet-ocean
system."(55)
A vigorous controversy followed, but in the end most climate scientists
stuck by the Milankovitch theory. The Devils Hole measurements looked
solid, but didn't they represent only a strictly local effect? |
|
Apparently the
story (like most climate stories) was not simple. As two experts
reviewing the problem put it, "climate is too complicated to be
predicted by a single parameter."(56)
The faint variations of summer
sunlight were effective only because the astronomical schedule somehow
resonated with other factors — ice sheet and ocean dynamics,
the bio-geochemical CO2
system, and who knew what else. The more precise the data got, the less
precise seemed the match between Milankovitch and ice age cycles.
Evidently when orbital effects served as a pacemaker, it was only by
partially adjusting the timing of greater forces working through their
own complex cycles. As one reviewer said, "The sheer number of
explanations for the 100,000-year cycle... seems to have dulled the
scientific community into a semipermanent state of wariness about
accepting any particular explanation."(57)
As researchers extracted more precise data from
the distant past, they discovered that the weak 100,000-year orbital
cycle had not always dominated the ice ages after all. Go back more
than a million years, and it was the 40,000-year cycle that ruled. The
reason for the switch was obscure. The grand puzzle of the ice ages
stood unsolved — except insofar as scientists now understood
that nobody would ever jump up with a neat single solution. There would
instead be a long collective trudge through the intricacies of field
data and models, gradually increasing our knowledge of all the
interacting forces that drive climate cycles. The invaluable fruit of a
century of ice ages research was the recognition of how complex and
powerful all the feedbacks could be.
|
|
Among these
feedbacks, the most obvious and momentous was the close connection
between global temperature and greenhouse gas levels through the ice
age cycles. Relatively straightforward analysis of the data showed that
a doubled level of CO2 had
always gone along with a rise of a few degrees in global temperature.
It was a striking verification, with entirely independent methods and
data, of what computer models had been predicting for the
planet’s greenhouse future
|
|
1.
Callendar (1961),
p. 1. BACK
1a. The
history is reviewed by Imbrie and
Imbrie (1986); the scheme of four ice ages was propounded by
Albrecht Penck, Penck and
Brückner (1901-1909). BACK
2.
Nuts: G.
Andersson in 1902 as cited in Lamb
(1977), p. 397; for the history overall, see Lamb pp. 193,
378ff. and Webb (1980). BACK
3.
Davis (1933). BACK
4.
Croll (1864); Croll (1875); Croll predicted
glaciation when the Earth was at aphelion in winter. But summer
aphelion (with the distant Sun less likely to melt the snow away) was
more likely to do it, as pointed out by Murphy
(1876); Croll (1886)
defends his views; Imbrie and
Imbrie (1979), pp. 77-88. BACK
5.
For example, Arrhenius (1896), p. 274; Brooks (1922),
p. 18-19. BACK
6.
Milankovitch (1920) ; Milankovitch (1930), see pp.
118-21 for additional history; Milankovitch
(1941) ; for this history I have used Imbrie and Imbrie (1986). BACK
7.
Simpson (1939-40), p. 203. BACK
8.
E.g., Landsberg (1941, rev. ed. 1947, 1960),
pp. 191-92. BACK
9.
Bradley (1929) ; Zeuner (1946 [4th ed., 1958]). BACK
10.
"dogma...
illusory," Öpik (1957);
"This theory has been answered devastatingly by... Sir George Simpson,"
Wexler (1952), p.
74. BACK
11.
van Woerkom (1953); see also Science Newsletter (1952);
similarly, "the changes of solar radiation due to changes in the
Earth's orbit are always too small to be of practical importance,"
Simpson (1939-40), p. 209. BACK
12.
Kuiper to H.
Sverdrup, 28 May 1952, and reply, 11 June 1952, Box 11, G.P. Kuiper
files, Special Collections, U. Ariz., kindly reported to me by Ron
Doel.; similarly, the theory "has failed utterly," Humphreys (1920),
pp. 564-66,
quote p. 568, on the Croll theory, but repeated without change in the
3rd (1940) edition, p. 586, without reference to Milankovitch. BACK
13.
Faegri et al. (1964); Manten (1966). BACK
14.
Urey (1947). BACK
15.
Emiliani (1955); see Emiliani (1958). BACK
16.
The effect
was never expected to correlate with sunlight in the Southern
Hemisphere, which is mostly ocean where snow would never accumulate.
Emiliani (1955), p. 509; see
also Emiliani (1958); on
evolution, Emiliani (1958)
p. 63. BACK
17.
Suess (1956). BACK
18.
Emiliani and Geiss (1959). BACK
19.
Hsü (1992), pp. 30-32,
220. BACK
20.
Quote: Broecker (1968), p. 139; for
early work, see Broecker (1966). BACK
21.
Emiliani (1966). BACK
22.
Emiliani (1966). BACK
23.
Dansgaard and Tauber (1969). BACK
24.
"sweepstakes":
Broecker et al. (1968)
p. 300; as 125, 105, and 82,000 in Mesolella
et al. (1969); see also summary in Broecker
and van Donk (1970); an important confirmation, using
boreholes drilled in Barbados reefs now drowned, was Fairbanks and
Matthews (1978);
the objection that the sea level changes might be due to local uplift
in Barbados, and not a world-wide phenomenon, was refuted by an
expedition to another fine set of coral terraces on a rarely visited
coast of New Guinea, Bloom et al.
(1974); for discussion Berger
(1988). BACK
25.
Emiliani (1972). BACK
26.
Glen (1982). BACK
27.
Kukla and Kocí (1972),
p. 383. BACK
28.
Chambers and Brain (2002), p.
239. BACK
29.
Kukla et al. (1972), p. 191; Kukla and Matthews (1972);
"large majority" according to Flohn
(1974), p. 385. BACK
30.
The first
long core (411m), using a drill developed by B. Lyle Hansen, was
extracted at another site in Greenland in 1956: Dansgaard et al.
(1973); for
brief history and references, see also Langway
et al. (1985); Levenson
(1989) pp. 40-41; for a firsthand account, Alley (2000). BACK
31.
Epstein et al. (1970). BACK
32.
Hamilton and Seliga (1972). BACK
33.
Dansgaard (1954); Dansgaard (1964); for further
bibliography on gases in ice, see Broecker
(1995), pp. 279-84. BACK
34.
Dansgaard et al. (1969).
Exciting day: oral history interview of Klaus Hammer by Finn Aaserud,
1993, GISP interviews, records of Study of Multi-Institutional
Collaborations, AIP. BACK
35.
Epstein et al. (1970). BACK
36.
Newell (1974); using results of
Johnsen et al. (1972). BACK
37.
Kukla and Kocí (1972);
see Schneider and Londer (1984),
p. 53. BACK
38.
Broecker and van Donk (1970);
cf. Ericson and Wollin (1968),
using foram temperatures. BACK
39.
Later
revised to 780,000. Shackleton and
Opdyke (1973), quote p. 40. They determined temperatures by
oxygen isotopes. Opdyke did the magnetic work. Cheer: John Imbrie, oral
history interview by Ron Doel, 1997, AIP; see Imbrie
and Imbrie (1979), p. 164. BACK
40.
For history
and comments, see Imbrie (1982);
Imbrie and Imbrie (1979). BACK
41.
Hays et al. (1976); for other
work, see Imbrie et al. (1975). BACK
42.
Evans and Freeland (1977). BACK
43.
Hays et al. (1976); Berger (1977); other data: Berger
(1978); see review, Berger (1988). BACK
44.
Kerr (1978). BACK
45.
Shift of
emphasis: paraphrase of Imbrie
(1982), p. 408; for example, see North
and Coakley (1979); review: North
et al. (1981), p. 107. BACK
46.
Imbrie et al. (1984). The
definitive "SPECMAP" chronology was published by Martinson et. al.
(1987) BACK
47.
Pisias and Moore (1981); Ruddiman et al. (1986). BACK
48.
Imbrie (1982), p. 411. BACK
49.
Quote: J.-R.
Petit in Walker (2000). BACK
50.
Lorius et al.(1985); Barnola et al.(1987); Genthon et
al. (1987). BACK
51.
Stauffer et al. (1988). BACK
52.
E.g., Pisias and Shackleton (1984);
"The existence of the 100-kyr [kiloyear] cycle and the synchronism
between Northern and Southern Hemisphere climates may have their origin
in the large glacial-interglacial CO2
changes." Genthon et al. (1987),
p. 414. BACK
53.
E.g., Berger (1988), p. 649; see Falkowski et al.
(2000); Berger and Loutre (2002)
discusses a long interglacial. The new Antarctic "Dome C" record of
climate went back 750,000 years through a previous cycle where the
orbital elements had been similar to those in our own cycle. EPICA
community members (2004).
On the drilling see Flannery (2005),
p. 58. See also reports in Science (November 25,
2005): 1285-87, 1313-21. BACK
53a. Lorius et al. (1990); Hoffert
and Covey (1992). BACK
54.
A review of
ice sheets (which added yet another factor, permafrost melting beneath
a sheet) is Clark et al. (1999). BACK
55.
Winograd et al. (1992), p. 255;
Ludwig et al. (1992). BACK
56.
Karner and Muller (2000). BACK
57.
Crowley (2002), p. 1474.
Changing
Sun, Changing
Climate?
Since it is the Sun's energy that drives the weather
system, scientists naturally wondered whether they might connect
climate changes with solar variations. Yet the Sun seemed to be stable
over the timescale of human lifetimes. Attempts to discover cyclic
variations in weather and connect them with the 11-year sunspot cycle,
or other possible solar cycles ranging up to a few centuries long, gave
results that were ambiguous at best. These attempts got a well-deserved
bad reputation. Jack Eddy overcame this with a 1976 study that
demonstrated that irregular variations in solar surface activity, a few
centuries long, were connected with major climate shifts. The mechanism
remained uncertain, but plausible candidates emerged. The next crucial
question was whether a rise in the Sun's activity could explain the
global warming seen in the 20th century? By the 1990s, there was a
tentative answer: minor solar variations could indeed have been partly
responsible for some past fluctuations... but future warming from the
rise in greenhouse gases would far outweigh any solar effects.
| The Sun
so greatly dominates the skies that the first scientific speculations
about different climates asked only how sunlight falls on the Earth in
different places. The very word climate (from Greek klimat,
inclination or latitude) originally stood for a simple band of
latitude. When scientists began to ponder the possibility of climate
change, their thoughts naturally turned to the Sun. Early modern
scientists found it plausible that the Sun could not burn forever, and
speculated about a slow deterioration of the Earth's climate as the
fuel ran out.(1) In 1801 the
great astronomer William Herschel introduced the idea of more transient
climate connections. It was a well-known fact that some stars varied in
brightness. Since our Sun is itself a star, it was natural to ask
whether the Sun's brightness might vary, bringing cooler or warmer
periods on Earth? As evidence of such a connection, Herschel pointed to
periods in the 17th century, ranging from two decades to a few years,
when hardly any sunspots had been observed. During those periods the
price of wheat had been high, he pointed out, presumably reflecting
spells of drought.(2) |
| Speculation increased in the mid-19th century,
following the discovery that the number of spots seen on the Sun rose
and fell in a regular 11-year cycle. It appeared that the sunspots
reflected some kind of storminess on the Sun's surface —
violent activity that strongly affected the Earth's magnetic field.
Astronomers also found that some stars, which otherwise seemed quite
similar to the Sun, went through very large variations. By the end of
the century a small community of scientists was pursuing the question
of how solar variability might relate to short-term weather cycles, as
well as long-term climate changes.(3)
Attempts to correlate weather patterns with the sunspot cycle were
stymied, however, by inaccurate and unstandardized weather data, and by
a lack of good statistical techniques for analyzing the data. Besides,
it was hard to say just which of many aspects of weather were worth
looking into. |
| At the end
of the 19th century, most meteorologists held firmly that climate was
stable overall, about the same in one century as in the last. That
still left room for cycles within the overall stability. A number of
scientists looked through various data hoping to find correlations, and
announced success. Enthusiasts for statistics kept coming up with one
or another plausible cycle of dry summers or cold winters or whatever,
in one or another region, repeating periodically over intervals ranging
from 11 years to several centuries. Many of these people declined to
speculate about the causes of the cycles they reported, but others
pointed to the Sun. An example was a late 19th-century British school
of "cosmical meteorology," whose leader Balfour Stewart grandly
exclaimed of the Sun and planets, "They feel, they throb together."(4) |
| Confusion persisted in the early decades of the
20th century as researchers continued to gather evidence for solar
variation and climate cycles. For example, Ellsworth Huntington,
drawing on work by a number of others, concluded that high sunspot
numbers meant storminess and rain in some parts of the world, resulting
in a cooler planet. The "present variations of climate are connected
with solar changes much more closely than has hitherto been supposed,"
he maintained. He went on to speculate that if solar disturbances had
been magnified in the past, that might explain the ice ages.(5) |
| Meanwhile an Arizona
astronomer, Andrew Ellicott Douglass, announced a variety of remarkable
correlations between the sunspot cycle and rings in trees. Douglass
tracked this into past centuries by studying beams from old buildings
as well as Sequoias and other long-lived trees. Noting that tree rings
were thinner in dry years, he reported climate effects from solar
variations, particularly in connection with the 17th-century dearth of
sunspots that Herschel and others had noticed. Other scientists,
however, found good reason to doubt that tree rings could reveal
anything beyond random regional variations. The value of tree rings for
climate study was not solidly established until
the 1960s.(6*) |
| Through the 1930s the most persistent advocate
of a solar-climate connection was Charles Greeley Abbot of the
Smithsonian Astrophysical Observatory. His predecessor, Samuel Pierpont
Langley, had established a program of measuring the intensity of the
Sun's radiation received at the Earth, called the "solar constant."
Abbot pursued the program for decades. By the early 1920s, he had
concluded that the solar "constant" was misnamed: his observations
showed large variations over periods of days, which he connected with
sunspots passing across the face of the Sun. Over a term of years the
more active Sun seemed brighter by nearly one percent. Surely this
influenced climate! As early as 1913, Abbot announced that he could see
a plain correlation between the sunspot cycle and cycles of temperature
on Earth. (This only worked, however, if he took into account temporary
cooling spells caused by the dust from volcanic eruptions.)
Self-confident and combative, Abbot defended his findings against all
objections, meanwhile telling the public that solar studies would bring
wonderful improvements in weather
prediction.(7*) He and a few others at the Smithsonian pursued the
topic single-mindedly into the 1960s, convinced that sunspot variations
were a main cause of climate change.(8)
|
| Other scientists were
quietly skeptical. Abbot's solar constant variations were at the edge
of detectability if not beyond. About all he seemed to have shown for
certain was that the solar constant did not vary by more than one
percent, and it remained an open question whether it varied anywhere
near that level. Perhaps Abbot was detecting variations not in the
solar constant, but in the transmission of radiation through the
atmosphere.(9) Still, if that
varied with the sunspot cycle, it might by itself somehow change the
weather. |
| Despite widespread skepticism,
the study of cycles was popular in the 1920s and 1930s. By now there
were a lot of weather data to play with, and inevitably people found
correlations between sunspot cycles and selected weather patterns.
Respected scientists and over-enthusiastic amateurs announced
correlations that they insisted were reliable enough to make
predictions. |
| Sooner or later,
every prediction failed. An example was a highly credible forecast that
there would be a dry spell in Africa during the sunspot minimum of the
early 1930s. When that came out wrong, a meteorologist later recalled,
"the subject of sunspots and weather relationships fell into disrepute,
especially among British meteorologists who witnessed the discomfiture
of some of their most respected superiors." Even in the 1960s, he said,
"For a young [climate] researcher to entertain any statement of
sun-weather relationships was to brand oneself a crank."(10) Specialists in solar physics
felt much the same. As one of them recalled, "purported connections
with... weather and climate were uniformly wacky and to be
distrusted... there is a hypnotism about cycles that... draws all kinds
of creatures out of the
woodwork."(11) By the 1940s, most meteorologists and astronomers had
abandoned the quest for solar cycles in the weather. Yet some respected
experts continued to suspect that they did exist, lurking somewhere in
the data.(12) |
| Less prone to crank enthusiasm and scientific
scorn, if equally speculative, was the possibility that the Sun could
affect climate on much longer timescales. During the 1920s, a few
people developed simple models that suggested that even a modest change
in solar radiation might set off an ice age, by initiating
self-sustaining changes in the polar ice. A leading British
meteorologist, Sir George Simpson, believed the sequence of ice ages
showed that the Sun is a variable star, changing its brightness over a
cycle some 100,000 years long.(13)
"There has always been a reluctance among scientists to call upon
changes in solar radiation... to account for climatic changes," Simpson
told the Royal Meteorological Society in a Presidential address of
1939. "The Sun is so mighty and the radiation emitted so immense that
relatively short period changes... have been almost unthinkable." But
none of the terrestrial causes proposed for ice ages was at all
convincing, he said, and that "forced a reconsideration of
extra-terrestrial causes."(14) |
| Such thinking was still in circulation two
decades later. The eminent astrophysicist Ernst Öpik wrote
that none of the many explanations proposed for ice ages was
convincing, so "we always come back to the simplest and most plausible
hypothesis: that our solar furnace varies in its output of heat."
Öpik worked up a theory for cyclical changes of the nuclear
reactions deep inside the Sun. The internal fluctuations he
hypothesized had a hundred-million-year timescale that seemed to match
the major glacial epochs. Manwhile,within a given glacial epoch "a kind
of 'flickering' of solar radiation" in the Sun's outer shell would
drive the expansion and retreat of ice sheets.(15)
In the 1950s, when reviews and textbooks listed various possible
explanations of ice ages and other long-term climate changes, ranging
from volcanic dust to shifts of ocean currents, they often invoked
long-term solar variation as a particularly likely cause. As a U.S.
Weather Bureau expert put it, "the problem of predicting the future
climate of Planet Earth would seem to depend on predicting the future
energy output of the sun..."(16)
|
| Meanwhile some people continued to pursue the
exasperating hints that minor variations in the sunspot cycle
influenced present-day weather. Interest in the topic was revived in
1949 by H.C. Willett, who dug out apparent relationships between
changes in the numbers of sunspots and long-term variations of wind
patterns. Sunspot variations, he declared, were "the only possible
single factor of climatic control which might be made to account for
all of these variations." Others thought they detected sunspot cycle
correlations in the advance and retreat of mountain glaciers. Willett
admitted that "the physical basis of any such relationship must be
utterly complex, and is as yet not at all understood." But he pointed
out an interesting possibility. Perhaps climate changes could be due to
"solar variation in the ultraviolet of the sort which appears to
accompany sunspot activity." As another scientist had pointed out a few
years before, ultraviolet radiation from the explosive flares that
accompany sunspots would heat the ozone layer high in the Earth's
atmosphere, and that might somehow influence the circulation of the
lower atmosphere.(17) |
| In the 1950s and 1960s, instruments on rockets
that climbed above the atmosphere managed to measure the Sun's
ultraviolet radiation for the first time. They found the radiation did
intensify during high sunspot years. However, ultraviolet light does
not penetrate below the stratosphere. Meteorologists found it most
unlikely that changes in the thin stratosphere could affect the layers
below, which contain far more mass and energy. Still, the hypothesis of
atmospheric influence remained alive, if far from healthy. |
| A few scientists
speculated more broadly. Maybe weather patterns were affected by the
electrically charged particles that the Sun sprayed out as "solar
wind." More sunspots throw out more particles, and they might do
something to the atmosphere. More indirectly, at times of high sunspot
activity the solar wind pushes out a magnetic field that tends to
shield the Earth from the cosmic rays that rain down from the universe
beyond. When these rays penetrate the upper reaches of the atmosphere,
they expend their energy producing sprays of charged particles
— so more sunspots would mean fewer of these particles.
Either way there might be an influence on the weather. Meteorologists
gave these ideas some credence.(18*)
But the solar wind and ultraviolet carried only a tiny fraction of the
Sun's total energy output. If they did influence weather, it had to be
through a subtle triggering mechanism that remained altogether
mysterious. Anyway variations connected with sunspots seemed likely to
bear only on temporary weather anomalies lasting a week or so (the
timescale of variations in sunspot groups themselves), not on long-term
climate change.(19) |
| People continued to report weather features that
varied with the sunspot cycle of 11 years, or with the full solar
magnetic cycle of 22 years (the magnetic polarity of sunspots reverses
from one 11-year cycle to the next). There were also matches to
possible longer solar variation
cycles.(20) It was especially scientists in the Soviet Union who
pursued such correlations. In the lead was a team under the Leningrad
meteorologist Kirill Ya. Kondratyev, who sent balloons into the
stratosphere to measure the solar constant. In 1970 his group claimed
that the Sun's output varied along with the number of sunspots by as
much as 2%. This drew cautious notice from other scientists. As the
authors admitted, the conclusion would remain in doubt unless it could
be verified by spacecraft entirely above the atmosphere.(21) |
| Another tentatively credible study came from a team led
by the Danish glaciologist Willi Dansgaard. Inspecting layers of
ancient ice in cores drilled from deep in the Greenland ice sheet, they
found cyclical variations. They supposed the Sun was responsible. For
the cycle that they detected, about 80 years long, had already been
reported by scientists who had analyzed small variations in the sunspot
cycle.(22*) Another cycle with a
length of about 180 years was also, the group suspected, caused by
"changing conditions on the Sun." The oscillations were so regular that
in 1970 Dansgaard's group boldly extrapolated the curves into the
future. They began by matching their results with a global cooling
trend that, as others reported, had been underway since around 1940.
The group predicted the cooling would continue through the next one or
two decades, followed by a warming trend for the following three
decades or so.(23) |
| The geochemist Wallace
Broecker was impressed. He "made a large leap of faith" (as he later
put it) and assumed that the cycles were not just found in Greenland,
but had a global reach.(24) He
calculated that the global cooling trend since around 1940 could be
explained by the way the two cycles both happened to be trending down.
His combined curve would bottom out in the 1970s, then quickly head up.
Greenhouse effect warming caused by human emissions of carbon dioxide
gas ( CO2) would come on top
of this rise, making for a dangerously abrupt warming.(25)
|
| (Later studies failed to find Dansgaard's cycles
globally. If they existed at all, the cause did not seem to be the Sun,
but quasi-cyclical shifts in the North Atlantic Ocean's surface warmth
and winds. This was just another case of supposed global weather cycles
that faded away as more data came in. It was also one of several cases
where Broecker's scientific instincts were sounder than his evidence.
The downturn in temperature since the 1940s, whether due to a variation
in the Sun's radiation or some other natural cause, could indeed change
to a natural upturn that would add to greenhouse warming instead of
subtracting from it. In fact that happened, beginning in the 1970s.) |
| The 1970s also
brought controversial claims that weather data and tree rings from
various parts of the American West revealed a 22-year cycle of
droughts, presumably driven by the solar magnetic cycle. Coming at a
time of severe droughts in the West and elsewhere, these claims caught
some public attention.(26*)
Scientists were beginning to understand, however, that the planet's
climate system could go through purely self-sustaining oscillations,
driven by feedbacks between ocean temperatures and wind patterns. The
patterns cycled quasi-regularly by themselves on timescales ranging
from a few years (like the important El Niño Southern
Oscillation in the Pacific Ocean) to several decades. That might help
to explain at least some of the quasi-regular cycles that had been
tentatively associated with sunspots. |
| All this helped to guarantee that scientists
would continue to scrutinize any possibility that solar activity could
influence climate, but always with a skeptical eye. If meteorologists
had misgivings, most astronomers dismissed outright any thought of
important solar variations on a timescale of hundreds or thousands of
years. Surface features like sunspots might cycle over decades, but
that was a weak, superficial, and short-term effect. As for the main
energy flow, improved theories of the nuclear furnace deep within the
Sun showed stability over many millions of years. Alongside this sound
scientific reasoning there may have been a less rational component. "We
had adopted a kind of solar uniformitarianism," solar physicist John
(Jack) Eddy suggested in retrospect. "As people and as scientists we
have always wanted the Sun to be better than other stars and better
than it really is."(27) |
| Evidence was accumulating,
however, that the Sun truly does change at least superficially from one
century to another. Already in 1961 Minze Stuiver had moved in the
right direction. Stuiver was concerned about peculiar variations in the
amount of radioactive carbon-14 found in ancient tree rings. Carbon-14
is generated when cosmic rays from the far reaches of the universe
strike the atmosphere. Stuiver noted how changes in the magnetic field
of the Sun would change the flux of cosmic rays reaching the Earth.(28) He had followed this up in
collaboration with the carbon-14 expert Hans Suess, confirming that the
concentration of the isotope really had varied over past millennia.
They were not suggesting that changes in carbon-14 (or cosmic rays)
altered climate; rather, they were showing that the isotope could be
used to measure solar activity in the distant past. For the
development of this important technique, a good example of laboratory
work and its attendant controversies, see the supplementary essay on
Uses of Radiocarbon Dating. |
| In 1965
Suess tried correlating the new data with weather records, in the hope
that carbon-14 variations "may supply conclusive evidence regarding the
causes for the great ice ages." He focused on the bitter cold spell
that historians had discovered in European writings about weather from
the 15th through the 18th century (the "Little Ice Age"). That had been
a time of relatively high carbon-14, which pointed to low solar
activity. Casting a sharp eye on historical sunspot data, Suess noticed
that the same centuries indeed showed a low count of sunspots. In
short, fewer sunspots apparently made for colder winters. A few others
found the connection plausible, but to most scientists the speculation
sounded like just one more of the countless correlations that people
had announced over the past century on thin evidence.(29*)
|
| Meanwhile carbon-14 experts refined their
understanding of how the concentration of the isotope had varied over
past millennia. They could not decide on a cause for the shorter-term
irregularities. Solar fluctuations were only one of half a dozen
plausible possibilities.(30) The
early 1970s also brought claims that far slower variations in the
Earth's magnetic field correlated with climate. In cores of clay drawn
from the seabed reaching back a million years, colder temperatures had
prevailed during eras of high magnetism. The magnetic variations were
presumably caused by processes in the Earth's interior rather than on
the Sun, but the correlation suggested that cosmic rays really did
influence climate. As usual the evidence was sketchy, however, and it
failed to convince most scientists.(31)
|
| In 1975, the respected meteorologist Robert
Dickinson, of the National Center for Atmospheric Research (NCAR) in
Boulder, Colorado, took on the task of reviewing the American
Meteorological Society's official statement about solar influences on
weather. He concluded that such influences were unlikely, for there was
no reasonable mechanism in sight — except, maybe, one.
Perhaps the electrical charges that cosmic rays brought into the
atmosphere somehow affected how aerosol particles coalesced. Perhaps
that somehow affected cloudiness, since cloud droplets condensed on the
nuclei formed by aerosol particles. This was just piling speculation on
speculation, Dickinson hastened to point out. Scientists knew little
about such processes, and would need to do much more research "to be
able to verify or (as seems more likely) to disprove these ideas." For
all his frank skepticism, Dickinson had left the door open a crack. One
way or another, it was now at least scientifically conceivable that
changes in sunspots could have something to do with changes in climate.
Most experts, however, continued to consider the idea discredited if
not preposterous.(32*) |
|
In 1976, Eddy tied all the threads together in
a paper that soon became famous. He was one of several solar experts in
Boulder, where a vigorous community of astrophysicists, meteorologists,
and other Earth scientists had grown up around the University of
Colorado and NCAR. Yet Eddy was ignorant of the carbon-14 research
— an example of the poor communication between fields that
always impeded climate studies. He had won scant success in the usual
sort of solar physics research, and in 1973 he lost his job as a
researcher, finding only temporary work writing a history of NASA's
Skylab. In his spare time he pored over old books. Eddy had decided to
review historical naked-eye sunspot records, with the aim of
definitively confirming the long-standing belief that the sunspot cycle
was stable over the centuries.
Read the details in our Interview with
Eddy
|
| Instead, Eddy found evidence that the Sun was by
no means as constant as astrophysicists supposed. Especially intriguing
was evidence suggesting that during the "Little Ice Age" of the
16th-17th centuries, sky-watchers had observed almost no sunspot
activity. People clear back to Herschel had noticed this prolonged
dearth of sunspots. A 19th-century German astronomer, G.W.
Spörer, had been the first to solidly document it, and a
little later, in 1890, the British astronomer E. Walter Maunder drew
attention to the discovery and its significance for climate. Other
scientists, however, thought this was just another case of dubious
numbers at the edge of detectability. Maunder's publications sank into
obscurity. It was only by chance that while Eddy was working to prove
the Sun was entirely stable, another solar specialist told him about Maunder's work.(33*) |
| "As a solar astronomer I felt certain that it
could never have happened," Eddy later recalled. But hard historical
work gradually persuaded him that the early modern solar observers were
reliable — the absence of sunspot evidence really was
evidence of an absence. Digging deeper, he found the inconstancy
confirmed by historical sightings of auroras and of the solar corona at
eclipses (both of which reflected heightened activity on the Sun's
surface). Once his attention was drawn to the carbon-14 record, he saw
that it too matched the pattern. All the evidence pointed to
long-sustained minimums and at least one maximum of solar activity in
the past two thousand years. It was "one more defeat in our long and
losing battle to keep the Sun perfect, or, if not perfect, constant,
and if inconstant, regular. Why we think the Sun should be any of these
when other stars are not," he continued, "is more a question for social
than for physical science."(34) |
| As it happened,
the ground had already been prepared by developments in astrophysics in
the early 1970s. Physicists had built a colossal particle detector
expressly to observe the elusive neutrinos emitted by the nuclear
reactions that fueled the Sun. The experiment failed to find anywhere
near the flux of neutrinos that theorists insisted should be reaching
the Earth. Was it possible that deep within the Sun, production of
energy was going through a lull? Perhaps the output of stars like the
Sun really could wander up and down, maybe even enough to cause ice
ages? The anomaly was eventually traced to neutrino physics rather than
solar physics. Meanwhile, however, it called into doubt the theoretical
reasoning that said the Sun could not be a variable star.(35) |
| Eddy's announcement of a solar-climate connection
nevertheless met the customary skepticism. He pushed his arguments
vigorously, stressing especially the Little Ice Age, which he memorably
dubbed the "Maunder Minimum" of sunspots. The name he chose emphasized
that he was not alone with his evidence. It is not unusual for a
scientist to make a "discovery" that others had already announced
fruitlessly. A scientific result cannot flourish in isolation, but
needs support from other evidence and ideas. Eddy had gone some
distance beyond his predecessors in historical investigation. More
important, he could connect the sunspot observations with the carbon-14
record and the new doubts about solar stability. It also mattered that
he worked steadily and persuasively to convince other scientists that
the thing was true. |
| Pushing farther, Eddy drew attention to a spell of
high carbon-14, and thus low solar activity, during the 11th-12th
centuries. Remarks in medieval manuscripts showed that these centuries
had been unusually warm in Europe. It was far from proven that those
were times of higher temperatures all around the globe. However,
scientists were (as usual) particularly impressed by evidence from the
North Atlantic region where most of them lived and where the historical
record was best known. Especially notable was the mild weather that had
encouraged medieval Vikings to establish colonies in Greenland
— colonies that endured for centuries, only to perish from
starvation in the Little Ice Age. Eddy warned that in our own times,
"when we have observed the Sun most intensively, its behavior may have
been unusually regular and
benign."(36) |
|
Decades later, after painstaking studies developed
much fuller series of data covering the entire globe, these data showed
a complex variety of periods of warmth and periods of cold. The
so-called "Medieval Warm Period" when Iceland and Greenland were
settled was a group of regional variations, significant but not as
universal and extreme as the steep temperature rise felt around the
world since the 1980s. The "Little Ice Age" was more definite (although
the cooling may have been partly caused by a spate of volcanic
eruptions), but it too had many local variations, not everywhere as
important as in the North Atlantic region. As one pair of experts
remarked in 2004, "If the development of paleoclimatology had taken
place in the tropical Pacific, Africa,... or Latin America, the
paleoclimatic community would almost certainly have adopted other
terminology." Instead of a Little Ice Age and Medieval Warm Period,
scientists of the 1970s might have talked, for example, about great
periods of drought. Still, Eddy's central point would stand: regional
climates were more susceptible to perturbing influences, including
small changes on the Sun, than most scientists had imagined.(36a)
Eddy worked hard to "sell" his findings. At a 1976
workshop where he first presented his full argument, his colleagues
tentatively accepted that solar variability might be responsible for
climate changes over periods of a few hundreds or thousands of years.(37) Eddy pressed on to turn up more
evidence connecting temperature variations with carbon-14, which he
took to measure solar activity. "In every case when long-term solar
activity falls," he claimed, "mid-latitude glaciers advance and climate
cools."(38)
|
| Already while Eddy's sunspot figures were in
press, other scientists began to explore how far his idea might account
for climate changes. Adding solar variability to the sporadic cooling
caused by dust from volcanic eruptions did seem to give a better match
to temperature trends over the entire last millennium.(39)
Peering closer at the more accurate global temperatures measured since
the late 19th century, a group of computer modelers got a decent match
using only the record of volcanic eruptions plus greenhouse warming
from increasing carbon dioxide, but they improved the match noticeably
when they added in a record of solar variations. All this proved
nothing, but gave more reason to devote effort to the question.(40) |
| Meanwhile Stuiver and others confirmed the
connection between solar activity and carbon-14, and this became a
standard tool in later solar-climate studies.(41)
An example was a study that reported a match between carbon-14
variations and a whole set of "little ice ages" (indicated by advances
of glaciers) that had come at random over the last ten thousand years.(42) Other studies, however, failed
to find such correlations. As a 1985 reviewer commented, "this is a
controversial topic... the evidence relating solar activity and
carbon-14 variations to surface temperatures is equivocal, an
intriguing but unproven
possibility."(43) |
| Scientists continued to report new phenomena at the
border of detectability. In particular, Ronald Gilliland (another NCAR
scientist) followed Eddy's example in analyzing a variety of old
records and tentatively announced slight periodic variations in the
Sun's diameter. They matched not only the 11-year sunspot cycle but
also the 80-year cycle that had long hovered at the edge of proof.
Adding these solar cycles on top of greenhouse warming and volcanic
eruptions, Gilliland too found a convincing match to the temperature
record of the past century. He calculated that the solar cycles were
currently acting opposite to the rise in carbon dioxide, so as to give
the world an equable climate until about the year 2000. This might lead
to complacency about greenhouse warming, he feared, which "could be
shattered" when the relentlessly increasing carbon dioxide added onto a
solar upturn. Most of his colleagues awaited more solid proof of the
changes in diameter and the long-term cycle (and they continue to await
it).(44) |
| Yet how could changes in the number of sunspots
affect climate? The most direct influence would come if the change
meant a rise or fall in the total energy the Sun radiated upon the
Earth, the so-called "solar constant." The development of highly
accurate radiometers in the 1970s raised hopes that variations well
below one percent could be detected at last. But few trusted any of the
measurements from the ground or even from stratospheric balloons.
Rockets launched above the atmosphere provided brief observations that
seemed to show variation over time, but it was hard to rule out
instrumentation errors. Nor were many convinced by Peter Foukal when he
applied modern statistical methods to Abbot's huge body of old data,
and turned up a faint connection between sunspots and the amount of
solar energy reaching the Earth. Even if that were accepted, was it
because the Sun emitted less energy? Or was it because ultraviolet
radiation from solar storms somehow changed the upper atmosphere, which
in turn somehow influenced climate, and thus affected how much sunlight
Abbot had seen at the surface?(45)
|
| To try to settle
the question, NASA included an instrument for measuring the solar
constant on a satellite launched in 1980. The amazingly precise device
was the work of a team at the Jet Propulsion Laboratory led by Richard
C. Willson. Soon after the satellite's launch, they reported distinct
if tiny variations whenever groups of sunspots passed across the face
of the Sun. Essential confirmation came from an instrument that John
Hickey and colleagues had previously managed to insert in the Nimbus-7
satellite, a spacecraft built to monitor weather rather than the Sun.(46) Both instruments
proved stable and reliable. In 1988, as a new solar cycle got underway,
both groups reported that total solar radiation did vary slightly with
the sunspot cycle.(47) |
| After 1988 |
| Satellite measurements pinned down precisely how
solar brightness varied with the number of sunspots. The radiation
varied by only about one part in a thousand; measuring such tiny
wiggles was a triumph of instrumentation.(48)
A single decade of data was too short to support any definite
conclusions about long-term climate change, but it was hard to see how
such a slight variation could matter much.(49)
Since the 1970s, rough calculations on general grounds had indicated
that it should take a bigger variation, perhaps half a percent, to make
a serious direct impact on global temperature. However, if the output
could vary a tenth of a percent or so over a single sunspot cycle, it
was plausible to imagine that larger, longer-lasting changes could have
come during the Maunder Minimum and other major solar variations. That
could have worked a real influence on climate. |
| Some researchers carried on with the old quest for
shorter-term connections. Sunspots and other measures plainly showed
that the Sun had grown more active since the 19th century. Was that not
linked somehow to the temperature rise in the same decades? Some people
persevered in the old effort to winkle out correlations between
sunspots and weather patterns. For example, according to a 1991 study,
Northern Hemisphere temperatures over the past 130 years correlated
surprisingly well with the length of the sunspot
cycle (which varied between 10 and 12 years). This finding was
highlighted the following year in a widely publicized report issued by
a conservative group. The report maintained that the 20th-century
temperature rise might be entirely due to increased solar activity. The
main point they wanted to make was less scientific than political: "the
scientific evidence does not support a policy of carbon dioxide
restrictions with its severely negative impact on the U.S. economy."(50) |
|
Critics of the report pointed out that the new
finding sounded like the weary old story of sunspot work — if
you inspected enough parameters, you were bound to turn up a
correlation. As it happened, already by 2000 the correlation of climate
with cycle length began to break down. Moreover, a reanalysis published
in 2004 revealed that from the outset the only pattern had been a
"pattern of strange errors" in the key study's data. Little more could
be said without further decades of observations, plus a theory to
explain why there should be any connection at all between the sunspot
cycle and weather. The situation remained as an expert had described it
a century earlier: "from the data now in our possession, men of great
ability and laborious industry draw opposite conclusions."(51)
The most straightforward correlation, if it could
be found, would connect climate with the Sun's total output of energy.
Hopes of finding evidence for this grew stronger when two astronomers
reported in 1990 that certain stars that closely resembled the Sun
showed substantial variations in total output. Perhaps the Sun, too,
could vary more than we had seen in the decade or so of precise
measurements? In fact, studies a decade later showed that the varying
stars were not so much like the Sun after all. Still, it remained
possible that the Sun's total luminosity had climbed enough since the
19th century to make a serious impact on climate — if anyone
could come up with an explanation for why the climate should be highly
sensitive to
such changes.(51a)
|
| A more
promising approach pursued the possibility of connections between
climate shifts and the slow changes in the Sun's magnetic activity that
could be deduced from carbon-14 measurements. A few studies that looked
beyond the 11-year sunspot cycle to long-term variations turned up
indications, as one group announced, of "a more significant role for
solar variability in climate change... than has previously been supposed."(52) In 1997 a pair
of scientists drew attention to a possible explanation for the link.
Scanning a huge bank of observations compiled by an international
satellite project, they reported that global cloudiness increased
slightly at times when the influx of cosmic rays was greater. Weaker
solar activity apparently meant more clouds. A later reanalysis of the
data found severe errors, but the study did serve to stimulate new
thinking. |
| The proposed mechanism roughly resembled the
speculation that Dickinson had offered, with little confidence, back in
1975. It began with the fact that in periods of low solar activity, the
Sun's shrunken magnetic field failed to divert cosmic rays from the
Earth. When the cosmic rays hit the Earth's atmosphere, they not only
produced carbon-14, but also sprays of electrically charged molecules.
Perhaps this electrification promoted the condensation of water
droplets on aerosol particles? If so, there was indeed a mechanism to
produce extra cloudiness. A later study of British weather confirmed
that at least regionally there was "a small yet statistically
significant effect of cosmic rays on daily cloudiness."(53) |
| Other studies meanwhile revived the old idea that
increased ultraviolet radiation in times of higher solar activity might
affect climate by altering stratospheric ozone. While total radiation
from the Sun was nearly constant, instruments in rockets and satellites
found the energy in the ultraviolet varying by several percent over a
sunspot cycle. Plugging these changes into elaborate computer models
suggested that even tiny variations could make a difference, by
interfering in the teetering feedback cycles that linked stratospheric
chemistry and particles with lower-level winds and ocean surfaces. By
the end of the 1990s, many experts thought it was possible that changes
in the stratosphere might affect surface weather after all.(54) |
| Whatever the exact form solar influences took,
most scientists were coming to accept that the climate system was so
unsteady, in general, that many kinds of minor external change could
trigger a shift. With somewhat plausible mechanisms to back up the
evidence for a solar-climate connection, the long-wavering balance of
scientific opinion tilted. Many experts now thought the connection
might be real. |
| When a 1999 study
reported evidence that the Sun's magnetic field had strengthened
greatly since the 1880s, it brought still more attention to the key
question: was increased solar activity the main cause of the
temperature rise over that period? Whether any of the proposed solar
mechanisms did in fact produce a noticeable effect on global climate
was still no more than speculation. But as the 21st century began, most
experts thought it plausible that the Sun might have driven at least
part of the previous century's warming. Most convincingly, the warming
from the 1880s to the 1940s had come when solar activity had definitely
been rising, while the carbon dioxide buildup had not yet been large
enough to matter much. A cooling during the 1950s and 1960s followed by
the resumption of warming also correlated loosely with solar activity.
How far the solar changes had influenced climate, however, remained
speculative. An increase in smoggy haze, dust from farmlands, and other
aerosols had probably had something to do with the cooling. It was also
possible that the climate system had just swung randomly on its own.(55*) |
| By now it was
evident that the old dream of predicting climate change directly from
solar variations was hopeless. Even if solar physicists could predict
long-term changes of the Sun (which they could not), so many other
interactions pushed and pulled the climate system that no single force
would explain it. One senior solar physicist insisted, "We will have to
know a lot more about the Sun and the terrestrial atmosphere before we
can understand the nature of the contemporary changes
in climate."(56) |
|
However, rough limits could be set on the extent of the
Sun's influence. Average sunspot activity did not increase after 1980,
and overall solar activity during the period 1950-2000 looked little
different from earlier periods. The satellite measurements of the solar
constant found it cycling within narrow limits (less than one part in a
thousand). Yet the global temperature rise that had resumed in the
1970s was accelerating at a record-breaking pace. It seemed impossible
to explain that using the Sun alone, without invoking greenhouse gases.(57*) The consensus of most
scientists, arduously hammered out in a series of international
workshops, flatly rejected the argument that the global warming of the
1990s could be dismissed as a mere effect of changes on the Sun. For
example, in 2004 when a group of scientists published evidence that the
solar activity of the 20th century had been unusually high, they
nevertheless concluded that "even under the extreme assumption that the
Sun was responsible for all the global warming prior to 1970, at most
30% of the strong warming since then can be of solar origin."(57a)
|
| Some experts persevered in arguing that slight
solar changes (which they thought they detected in the satellite
record) had driven the extraordinary warming since the 1970s. Most
scientists expected that these correlations would follow the pattern of
every other subtle solar-climate correlation that anyone had reported
— fated to be disproved by the following decade or two of
data. Even if the contrarians were right, however, greenhouse warming
was bound to swamp the solar effects as humanity emitted ever more
gases. Willson, the leader of the satellite experts, explained that in
the future,"solar forcing could be significant, but not dominant."(58*) |
| The import
of the claim that solar variations influenced climate was now reversed.
Critics had used the claim to oppose regulation of greenhouse gases.
But what if the planet really did react with extreme sensitivity to
almost imperceptible changes in the radiation arriving from the Sun?
The planet would surely also be sensitive to greenhouse gas
interference with the radiation once it entered the atmosphere. A U.S.
National Academy of Sciences panel estimated that if solar radiation
were now to weaken as much as it had during the 17th-century Maunder
Minimum, the effect would be offset by only two decades of accumulation
of greenhouse gases. As one expert explained, the Little Ice Age "was a
mere 'blip' compared with expected future
climatic change."(59)
For recent work on temperature changes
over the past millenium or so, probably related to solar variations
plus volcanic eruptions and perhaps other factors as well as the recent
rise of greenhouse gases, see the conclusion and figure captions in the
essay on The Modern Temperature Trend.
|
1. Feldman (1993);
Fleming (1990). BACK
2. Herschel (1801), pp. 313-16; see Hufbauer (1991). BACK
3. Notably, for variations related to the evolution of the Sun and
stars, Dubois (1895); for sunspot cycles Czerney (1881) . BACK
4. See for example, Brückner (1890), chapter 1; translated in
Stehr and von Storch (2000), pp. 116-121; Stewart: Gooday (1994). BACK
5. Huntington (1914), quote p. 480; Huntington (1923); summarized in
Huntington and Visher (1922) . BACK
6. Webb (2002), chapter 3; Webb (1986); Fritts (1976) notes the
skepticism (page v) and shows how it was overcome; climate periods of
11-12 years as well as longer cycles also appeared in annual layers of
clay laid down in lake beds (varves), Bradley (1929); for references
and summary, see Brooks (1950). BACK
7. Abbot and Fowle (1913); similarly A. Ångström,
using
Abbot's data, said the solar constant varied with sunspot number,
although decades later he retracted. Ångström
(1922);
Ångström (1970); historical studies are Hufbauer
(1991), p.
86; DeVorkin (1990). BACK
8. Abbot (1967); Aldrich and Hoover (1954). BACK
9. Fröhlich (1977). BACK
10. Lamb (1997), pp. 192-93. BACK
11. J. Eddy, interview by Weart, April 1999, AIP, p. 6. BACK
12. Nebeker (1995), p. 95. BACK
13. Simpson (1934); Simpson (1939-40). Simpson cited A. Penck, who
argued that the entire world had cooled and only solar changes could
explain this. BACK
14. Simpson (1939-40), p. 210. BACK
15. Öpik (1958); "flickering" (due to uncertain convective
changes): Öpik (1965), p. 289. BACK
16. E.g., Brooks (1949), ch. 1; Shapley (1953); Wexler (1956), quote p.
494, adding that turbidity (from volcanoes) was equally important. BACK
17. Willett (1949), pp. 34, 41, 50; see Lamb (1997), p. 193; the
earlier hypothesis (not cited by Willett) is in Haurwitz (1946);
glacier papers are cited by Wexler (1956), p. 485. BACK
18. A possible connection between cosmic rays and clouds was already
established at the end of the 19th century by the inventor of the cloud
chamber, Wilson (1899); it was admittedly "speculation" that ionization
in the upper troposphere affected storminess. Ney (1959); the ideas
found some favor with, e.g., Roberts (1967), pp. 33-34. BACK
19. Sellers (1965), pp. 220-23. BACK
20. Lamb (1977), pp. 700-704. BACK
21. Kondratyev and Nikolsky (1970); Fröhlich (1977). BACK
22. Johnsen et al. (1970); similarly, Dansgaard et al. (1971), same
quote p. 44; the period they reported was precisely 78 years, and
Schove (1955) had reported a 78-year variation between long and short
sunspot cycles as well as a possible 200-year period; in addition, not
noted by the glaciologists, a roughly 80-year modulation in the
amplitude of the sunspot cycle was reported by Gleissberg (1966);
weather correlations with the 80-year cycle were reported in 1962 by
B.L. Dzerdzeevski as cited by Lamb (1977), p. 702. BACK
23. Johnsen et al. (1970); see also Dansgaard et al. (1971); Dansgaard
et al. (1973). BACK
24. Broecker (1999). BACK
25. Broecker (1975). BACK
26. Roberts and Olson (1975) (admitting that "A mere coincidence in
timing... will not, of course, constitute proof of a physical
relationship"); Mock and W.D. Hibler (1976) (a "pervasive" but only
"quasi-periodic" 20-year cycle); Mitchell et al. (1979) (tree-ring data
analysis "strongly supports earlier evidence of a 22 yr drought
rhythm... in the U.S.... in some manner controlled by long-term solar
variability..." ). BACK
27. Eddy (1977), p. 92. BACK
28. Stuiver (1961). BACK
29. Suess (1968), p. 146; in the best review of sunspot history
available to Suess at this time, D.J. Schove took no notice of any
anomaly such as the early-modern minimum, although it is visible in his
data. Schove (1955); a tentative longer-term correlation of climate
(glacier advances) with C-14 was shown by Denton and Karlén
(1973), who suggest that "climatic fluctuations, because of their close
correlation with short-term C14 variations, were caused by varying
solar activity," p. 202; for the Little Ice Age, see Fagan (2000); Lamb
(1995), ch. 12. BACK
30. Ralph and Michael (1974). BACK
31. Wollin et al. (1971); Gribbin (1982), ch. 7. BACK
32. Dickinson (1975); a similar speculation, connecting cosmic rays
with storminess, was offered by Tinsley et al. (1989); see also the
solar activity-atmosphere connection reported by Wilcox et al. (1973).
Another weather-Sun correlation was laid out in Herman and Goldberg
(1978), which met strong resistance including attempts to suppress
publication, according to Herman (2003), ch. 18. BACK
33. Maunder (1890) attributes the discovery to Spörer; some
authors now refer to a 17th-century Maunder Minimum and a 15th-century
Spörer Minimum. Eddy chose "Maunder" to make a phrase that
would
be memorable: Eddy, interview by Weart, April 1999, AIP, p. 11. For
history and references, see Eddy (1976); examples of neglect of
Maunder: he was cited, but only for other work, in Abetti (1957);
Kuiper (1953); Menzel (1949); the 17th-century paucity of sunspots was
noted without any reference by Willett (1949), p. 35. BACK
34. The first published statement was an abstract for the March 1975
meeting of the American Astronomical Society Eddy (1975); and next at a
Solar Output Workshop in Boulder, Colo., Eddy (1975); the famous
publication was Eddy (1976), "defeat" p. 1200; "felt certain," Eddy
(1977), pp. 80-81. See Eddy, interview by Weart, April 1999, AIP. BACK
35. Hufbauer (1991), pp. 269-78. BACK
36. "benign," Eddy (1977), p. 69. BACK
36a. Jones and Mann (2004), p. 20, see p. 7 and passim. BACK
37. White (1977), see Mitchell p. 21, Hays p. 89; note also the
earlier, more doubting response of Mitchell (1976), p. 491. "Salesman":
Eddy, interview by Weart, April 1999, AIP, p. 14. BACK
38. Eddy (1977), quote p. 173; for more extensive speculations and
reflections, see Eddy (1977). BACK
39. Schneider and Mass (1975); similarly, Schneider and Mass (1975).
BACK
40. Hansen et al. (1981), using what was admittedly a "highly
conjectural" (p. 93) measure of variability by D.V. Hoyt. BACK
41. Stuiver and Quay (1980). BACK
42. Wigley and Kelly (1990). BACK
43. Bradley (1985), p. 69. BACK
44. Gilliland (1981), reporting 11- and 76-year variations in solar
size; Gilliland (1982); Gilliland (1982), quote p. 128. BACK
45. Hufbauer (1991), pp. 278-80; for example, a 1978 workshop concluded
that changes in stratospheric ozone due to ultraviolet radiation might
influence climate McCormac and Seliga (1979), pp. 18, 20. BACK
46. Hickey et al. (1980); Willson et al. (1981); Hufbauer (1991), pp.
280-92. BACK
47. Willson and Hudson (1988); Hickey et al. (1988). BACK
48. Lee et al. (1995). BACK
49. Hoyt and Schatten (1997). BACK
50. Seitz (1992), p. 28, see p. 17. BACK
51. Friis-Christensen and Lassen (1991); Kerr (1991); Young (1895), p.
162. Errors: Damon and Laut (2004). BACK
51a. Baliunas and Jastrow (1990); Foukal (2003). BACK
52. "More significant" (an "admittedly crude" analysis): Cliver et al.
(1998), p. 1035. BACK
53. Svensmark and Friis-Christensen (1997); Friis-Christensen and
Svensmark (1997); the effect was also reported, less convincingly, by
Pudovkin and Verentenenko (1995); Pudovkin and Veretenenko (1996).
Errors: Damon and Laut (2004). Later study: Harrison and Stephenson
(2006). BACK
54. Haigh (1994); Haigh (1996); McCormack et al. (1997); Shindell et
al. (1999); for discussion, see Wallace and Thompson (2002). BACK
55. Lockwood et al. (1999); Marsh and Svensmark (2000). Reviewing
various claims, including some based on observations of variations in
supposedly Sun-like stars, three experts concluded in 2004 that "Any
relationship" between long-term solar variations and climate "must
remain speculative," Foukal et al. (2004). BACK
56. Parker (1999); cf. criticism of Parker by Hoffert et al. (1999).
BACK
57. Tett et al. (1999); moreover, the stratosphere was cooling, which
made sense for the greenhouse effect but was hard to explain through a
solar influence: IPCC (2001), p. 709. BACK
57a. Consensus: IPCC (2001). Quote: Solanki et al. (2004), p. 1087. BACK
58. Willson reported a brightening of 0.04 percent between the two most
recent solar cycles, Willson (1997); this was controversial, see Kerr
(1997); similarly and more recently, Willson and Mordvinov (2003);
discussed by Byrne (2003); quote: Nelson (1997). BACK
59. National Research Council (1994), combining statements on pp. 3 and
4; blip: Wigley and Kelly (1990), p. 558. BACK
Hurricanes and Climate Change
Hurricanes have always bedeviled the Gulf Coast states, but global
warming is making matters worse. Sea level is rising and will continue
to rise as oceans warm and glaciers melt. Rising sea levels means
higher storm surges, even from relatively minor storms, causing coastal
flooding and erosion and damaging coastal properties. In a distressing
new development, scientific evidence now suggests a link between
hurricane strength and duration and global warming.
Understanding the relationship between hurricanes and global warming is
essential if we are to preserve healthy and prosperous coastal
communities for ourselves and our children.
More Intense Storms
Recent research has found that storm intensity and
duration increases as global warming emissions increase in our
atmosphere. Rising sea levels, also caused in part by rising global
temperatures, intensify storm damage along coasts. For hurricanes to
occur, surface ocean temperatures must exceed 80 degrees Fahrenheit.
The warmer the ocean, the greater the potential for stronger storms.
More destructive hurricanes not only inflict billions of dollars in
damage to communities and businesses, but also put thousands of human
lives at risk.
|
| For Hurricanes
to occur, surface temperatures must exceed 80 degrees
Fahrenheit. The ocean conditions were conducive for
hurricanes during the powerful Atlantic hurricane season of 2005. |
Hurricane Behavior
To understand how global warming can affect ocean
storms, it’s important to understand how these storms develop
in the first place. Seasonal shifts in global wind patterns cause
atmospheric disturbancesin the tropics, leading to a local drop in
pressure at sea level and forcing air to rise over warm ocean waters.
As warm, moist air rises, it further lowers air pressure at sea level
and draws surrounding air inward and upward in a rotating pattern
called a vortex. When the water vapor-laden air rises to higher
altitudes, it cools and releases heat as it condenses to rain. This
cycle of evaporation and condensation brings the ocean’s
thermal energy into the vortex, powering the storm. Depending on the
severity, meteorologists call these tropical storms or hurricanes in
the Atlantic Ocean.
Many factors influence storm behavior, including surface
temperatures, humidity, and atmospheric circulation. A sudden change in
wind speed or direction (wind shear), for example, may prevent the
vortex from forming. But as long as conditions are favorable, the storm
will thrive.
Warming Ocean Waters
Natural cycles alone cannot explain recent ocean
warming. Because of human activities such as burning fossil fuels and
clearing forests, today’s carbon dioxide (CO2)
levels in the atmosphere are significantly higher than at any time
during the past 400,000 years. CO2 and other
heat-trapping emissions act like insulation in the lower atmosphere,
warming land and ocean surface temperatures. Oceans have absorbed most
of this excess heat, raising sea temperatures by almost one degree
Fahrenheit since 1970. September sea surface temperatures in the
Atlantic over the past decade have risen far above levels documented
since 1930.
Recent Scientific Developments
A 2004 study published in the peer-reviewed Journal
of Climate explored the relationship between
today’s storms compared with simulated storms under
conditions with increased atmospheric CO2 (the
primary global warming gas). The study simulated storm behavior under a
one percent per year increase in CO2 over 80
years. Nine different global climate models projected that storms
generated under increasing CO2 conditions were
consistently more intense. By the end of the projection, maximum
surface wind speeds increased six percent and rainfall increased on
average 18 percent over present-day conditions.
A 2005 study published in the journal Nature
suggests that storm intensity and duration is linked to the recent
ocean warming trends associated with global warming. Scientists tracked
measurements of the destructive power of storms, termed the Power
Dissipation Index (PDI), since 1950. The study, which combined each
storm’s maximum wind speeds and storm duration, found that
during the last 30 years, the destructive power of storms has doubled
in the Atlantic and Pacific.¹
Most of this has occurred during the past 10 years when
global average surface ocean temperatures were at record levels. Thus
far, scientific evidence does not link worldwide storm frequency with
global warming. Individual ocean basins have multiyear cycles of storm
activity. While the total number of storms in the tropics remained
similar through time, the percentage of category 4 and 5 hurricanes
have increased over the past 30 years, according to a 2005 paper in the
journal Science.
Protecting Coastal Communities
Given the huge price tag from the cleanup of recent
hurricanes such as Andrew ($43.7 billion)², Ivan ($14.2
billion), and Katrina ($125 billion projected), it is essential to do
whatever we can to avoid dangerous warming and preserve healthy and
prosperous coastal communities for ourselves and our children. Because
CO2 can stay in the atmosphere for 100 years or
more, even an aggressive plan to use energy more efficiently and reduce
emissions from power plants and vehicles will not stop warming in it
tracks. Therefore, it is essential that we combine aggressive emission
reduction efforts with improved measures to protect coastal
communities. These measures— including building codes, storm
drainage plans, and preservation and restoration of wetlands, dunes,
and barrier islands— must be designed to cope with increasing
sea level rise and storm intensity due to global warming.
¹Tracking of ocean temperatures has been
relatively accurate over the past 50 years while storm tracking data
have improved significantly in the past 30 years. Both sea surface
temperatures and hurricane intensity increased most rapidly over the
past 15 years.
²Inflation adjusted to the year 2004.
References
Anikouchine, W. and R.W. Sternberg. 1981.
The World
Ocean. Englewood Cliffs, NJ: Prentice-Hall, Inc.
Barnett, T.P., D.W. Pierce, and R. Schnur.
2001.
Detection of Anthropogenic climate change in the world’s
oceans. Science 292: 270–274.
BBC News, September 13, 2005, Big rise in
Katrina cost
forecast.
Blake, E.S., J.D. Jarrell, M. Mayfield,
E.N. Rappaport,
and C. W. Landsea, (2005) The costliest U.S. Hurricanes 1900-2004
(adjusted) table derived from NOAA Technical Memorandum NWS TPC-1.
Online at http://www.nhc.noaa.gov/pastcost2.shtml?.
Accessed September 2005
Committee on the Science of Climate Change,
National
Research Council. 2001. Climate Change Science: An Analysis of Some Key
Questions. Washington, DC: National Academy Press.
Emanuel, K. 2005. Increasing
destructiveness of tropical
cyclones over the past 30 years. Nature 436:686-688.
Knutson T.R. and R.E. Tuleya. 2004. Impact
of
CO2-induced warming on simulated hurricane intensity and precipitation:
Sensitivity to the choice of climate model and convective
parameterization. Journal of Climate 17: 3477–3495.
NOAA. 2005. Global Surface Temperature
Anomalies.
Asheville, NC: NOAA/National Climatic Data Center. Online at
http:/www.ncdc.noaa.gov/oa/climate/research/
anomalies/anomalies.html#index. Accessed July 2005.
Reynolds, R.W. and T.M. Smith. 1995. A high
resolution
global sea surface temperature climatology. Journal of Climate 8:
1571–1583.
Trenberth, K. 2005. Uncertainty in
Hurricanes and Global
Warming. Science 308: 1753–1754.
Webster, P.J., Holland, G.J., Curry, J.A.
and H.-R.
Chang. 2005. Changes in tropical cyclone number, duration, and
intensity in a warming environment. Science 309:1844-1846.
U.S. Department of State. 2002. U.S.
Climate Action
Report – 2002: Third National Communication of the United
States of America under the United Nations Framework Convention on
Climate Change. Washington, DC. Online at
http://yosemite.epa.gov/oar/globalwarming.nsf
/content/ResourceCenterPublicationsUSClimateActionReport.html
|
