The eventful ice age

Jupsat Pro Astronomy Software

Secrets of the Deep Sky

Get Instant Access

Two key questions arise from the study of the Quaternary ice age: Why was it so eventful, displaying alternations between glacial and interglacial stages and stadial and interstadial shifts? And, what caused it?

The orbital forcing hypothesis

The jostling of the planets, their satellites, and the Sun leads to medium-term orbital variations occurring with periods in the range 10,000 to 500,000 years that perturb Earth's climate. These orbital forcings do not change the total amount of solar energy received by the Earth during the course of a year, but they do modulate the seasonal and latitudinal distribution of solar energy. In doing so, they wield a considerable influence over climate (Table 4.1). Orbital variations in the 10,000-500,000-year frequency band appear to have driven climatic change during the Pleistocene and Holocene. Orbital forcing has led to climatic change in middle and high latitudes, where ice sheets have waxed and waned, and to climatic change in low latitudes, where water budgets and heat budgets have marched in step with high-latitude climatic cycles. Quaternary loess deposits, sea-level changes, and oxygen-isotope ratios of marine cores record the 100,000-year cycle of eccentricity. The precessional cycle (with 23,000- and 19,000-year components) and the 41,000-year tilt cycle ride on the 100,000-year cycle. They, too, generate climatic changes that register in marine and terrestrial sediments. Oxygen isotope ratios (SO18) in ocean cores normally contain signatures of all the Earth's orbital cycles, though the tilt cycle, as it affects season-ality, has a stronger signature in sediments deposited at high latitudes.

The connection between orbital cycles and climate has a long and interesting history that shows how the germ of an idea may not fully develop until long after its inception. In the seventeenth century, some commentators suggested that Earth's orbital variations might influence climate. Monsieur de Mairan, writing in 1765, remarked on the effect of

Table 4.1 Orbital forcing cycles.

Cycle Approximate Examples in climatic data period (years)

Tilt 41,000

Precession 19,000 and 23,000

Short eccentricity and 100,000

orbital plane inclination1

Long eccentricity 400,000

Note: 1 See Muller and MacDonald (1995).

Oxygen-isotope records from deep-sea cores Oxygen-isotope records from deep-sea cores; magnetic susceptibility variations in deep-sea cores; loess deposits

Diatom temperature records in deep-sea cores

Diatom temperature records in deep-sea cores the distance of the Sun from the Earth in apogee and perigee (cited in Croll 1875, 528). Charles Lyell, in the first edition of his Principles of Geology (1830-33, vol. I, 110), commented on the effect of precession of the equinoxes on the receipt of 'solar light and heat' in the two hemispheres. The French mathematician, Joseph Alphonse Adhemar (1842) thought that the differences in the seasons between the hemispheres brought about by pre-cessional changes would be large enough to have caused the Ice Age. John Frederick William Herschel (1835) raised the possible effects of eccentricity on climate. However, the first detailed discussion of the matter was due to James Croll (1864, 1867a, 1867b) in a series of papers and later in his book (1875) Climate and Time in Their Geological Relations: A Theory of Secular Changes of the Earth's Climate. Croll argued that an ice age would occur when an elongated orbit combined with a winter solstice occurring near aphelion. The effects of precession, ellipticity, and obliquity on the seasonal and latitudinal distribution of radiation were studied in depth by the Yugoslavian mathematician and engineer, Milutin Milankovitch (1920, 1930, 1938). Milankovitch's chief conclusions were threefold. First, orbital eccentricity and precession produce effects large enough to cause ice sheets to expand and contract. Second, the climatic effects of obliquity are far greater than Croll had presumed. Third, astronomical variations in eccentricity, precession, and obliquity were sufficient to produce ice ages by changing the seasonal and geographical distribution of solar radiation (Figure 4.1). In his 1920 publication, Milankovitch suggested that small variations in the orbital variables drive a great cycle of climate, which takes roughly 100,000 years to go through one round. By analogy with the annual march of the seasons, the march of the great seasons runs through a 'great winter', when the Earth is gripped in an ice age, a 'great spring', when there is a great thaw, a 'great summer', when interglacial conditions prevail, and a 'great autumn', when conditions start to deteriorate presaging the coming of the next 'great winter'.

The Croll-Milankovitch theory of climatic forcing was popular up to about 1950, after which time most Quaternary geologists ignored it or rejected it. During the late 1960s and early 1970s, researchers rediscovered Milankovitch's cycle of great seasons. Evidence for a 100,000-year cycle was unearthed independently in loess sequences exposed in a quarry in Czechoslovakia (Kukla 1968, 1975), in sea-levels (Broecker et al. 1968; Mesolella et al. 1969; Chappell 1973), and in the oxygen-isotope ratios of marine cores (Broecker and van Donk 1970; Ruddiman 1971). Moreover, both the terrestrial and

Ice Age Ice Age Ice Age Ice Age

600 500 400 300 200 100 0

Time (thousands of years before present)

Figure 4.1 Orbital forcing and the occurrence of ice ages according to Milutin Milankovitch. £ is the obliquity of the ecliptic; e is the eccentricity of the Earth's orbit; n is the longitude or perihelion. Source: Adapted from Koppen and Wegener (1924).

600 500 400 300 200 100 0

Time (thousands of years before present)

Figure 4.1 Orbital forcing and the occurrence of ice ages according to Milutin Milankovitch. £ is the obliquity of the ecliptic; e is the eccentricity of the Earth's orbit; n is the longitude or perihelion. Source: Adapted from Koppen and Wegener (1924).

marine records attested to long periods of glacial expansion (climatic cooling) abruptly ended by rapid deglaciations (climatic warming). The short eccentricity cycle was a strong contender for explaining the 100,000-year signal, leading to the revivification of James Croll's argument of a century before: when the Earth's orbit is unusually elongate, pre-cessional effects are amplified, producing more contrasted seasons and allowing an ice age to start. Later, researchers demonstrated that the precessional and tilt cycles explained climatic oscillations superimposed on the 100,000-year cycle. This demonstration required a finely calibrated calendar of Pleistocene events. Largely owing to the endeavours of the members of the CLIMAP project, a suitably detailed calendar emerged. Nicholas J. Shackleton and Neil D. Opdyke (1973), by making oxygen-isotope and magnetic measurements of a Pacific deep-sea core, and establishing that Marine Isotope Stage 19 occurs at the boundary between the Bruhnes and Matuyama epochs, gave the first accurate chronology of late Pleistocene climate. Confirmation of the Croll-Milankovitch theory was eventually forthcoming when John D. Hays found suitable cores from the Indian Ocean, which recorded climatic change over the last 450,000 years, for subjecting to spectral analysis. The results of the analysis revealed cycles of climatic change at all frequencies corresponding to orbital forcings (Hays et al. 1976). In addition to the 23,000-year precessional cycle, a 19,000-year precessional cycle component was present. The Belgian astronomer André Berger (1978) confirmed this minor precessional cycle theoretically. The publication of these findings convinced most scientists that the motion of the Earth around the Sun did drive the world climate system during the late Pleistocene, that orbital variations were the 'pacemaker' of the ice ages.

Problems with orbital forcing

Variations in orbital parameters do not explain all aspects of Quaternary climatic change. Maya Elkibbi and José Rial (2001) identified five challenges to the astronomical theory of ice ages. Three relate to the '100,000-year problem'. First, 100,000-year variations of insolation forced by eccentricity changes are too small (less than 1 per cent) to drive the great ice ages. Second, 100,000-year oscillations have dominated the last 900,000 years but 41,000-year oscillations dominated the late Tertiary and early Quaternary, the switch being known as the mid-Pleistocene transition. The third challenge is the '400,000-year problem', which is the absence of a 413,000-year signal in oxygen isotope ratios from marine cores over the past 1.2 million years, despite that being the largest component of eccentricity forcing. Fourth, over the last 500,000 years, the length of glacial stages ranges from about 80,000-120,000 years, which variation cannot correlate linearly with insolation changes. The fifth challenge is the presence of signals for climatic cycles that appear unrelated to insolation forcing, which indicate non-linear responses of the climate system. In addition to these five problems is the finding that a number of palaeoclimatic records, when subjected to re-examination, have a variance attributable to orbital changes never exceeding 20 per cent (Wunsch 2004).

The '100,000-year problem'

This is part of a wider issue concerning the sensitivity of the climate system to really rather modest changes in the seasonal and latitudinal pattern of insolation receipt. The astronomer Fred Hoyle believed that, given the vast amount of heat stored in the oceans, which buffers the climate system against perturbations, the changes involved are far too tiny to have any significant impact on climate. He rejected the astronomical theory of ice ages with gusto: 'If I were to assert that a glacial condition could be induced in a room liberally supplied during winter with charged night-storage heaters simply by taking an ice cube into the room, the proposition would be no more likely than the Milankovitch theory' (Hoyle 1981, 77). Hoyle's condemnation is rather extreme, and later work suggests that he perhaps underestimated the degree of the seasonal insolation anomalies. Anomalies of insolation during caloric half years reach a maximum of up to about 6 kcal/cm2; they decrease towards the winter poles but are still not small: an anomaly of about 4 kcal/cm2 could melt a 2.5-km-thick ice sheet in 5,000 years. No, the world climate system is sensitive to the seasonal changes of climate resulting from Earth's orbital variations. The effects of orbitally induced changes of summer temperatures during the Holocene epoch, for example, are clearly recorded in melt layers in high-Arctic ice cores: the warmest summers occurred from 10,000 to 8,000 years ago and the coldest 150 years ago, as would be expected based on Croll-Milankovitch forcing (Koerner and Fisher 1990). General circulation models have also highlighted the sensitivity of the climate system to orbital forcing (e.g. Kutzbach 1981; Kutzbach and Otto-Bliesner 1982; Kutzbach and Guetter 1986). Studies made with general circulation models have revealed that changes in solar radiation receipt brought about by variations in the Earth's orbital characteristics elicit a different thermal response in the sea and on land, and so cause major changes in monsoons and the global water cycle.

Another possible explanation for the strength of the 100,000-year cycle of eccentricity lies in oscillations of the inclination of the Earth's orbital plane, which bring the Earth into a cloud of interplanetary dust and produce glacial conditions (Muller and

MacDonald 1995). Alternatively, it is possible that an integer number of orbital oscillations paces the 100,000-year cycle (Ridgwell et al. 1999). Every fourth or fifth pre-cessional cycle seems to match the oxygen-isotope spectra the best, but other permutations produce equally good spectral matches so it is not possible to determine the processes that pace the climate cycles. However, Maureen Raymo (1997) had noticed that glacial-interglacial terminations tend to occur when the previous summer insolation maximum was unusually low at mid-Northern Hemisphere latitudes. If it seems reasonable that periods of high summer insolation constrain the build-up of Northern Hemisphere ice, then the episodic occurrence of weak insolation maxima, caused by the superposition of periods of low obliquity and eccentricity, may have produced a substantial build-up of ice volume. At the following precessional high summer insolation maximum, a threshold process based upon excess ice volume, as the critical level of isostatic bedrock adjustment was attained, would explain the observed swift ice-sheet breakdown. Indeed, the most pronounced glaciation terminations are observed to be typically associated with increases in summer insolation at 65°N, which tends to rule out orbital inclination variations and instead bolster the idea of the 100,000-year cycle's being related to the eccentricity modulation of precession (Raymo 1997; see also Imbrie and Imbrie 1980).

The mid-Pleistocene transition

The problem of the mid-Pleistocene transition is very interesting. The strongest climatic signal in the marine sedimentary record for the last 900,000 years corresponds to the 100,000-year short cycle of eccentricity, but from about 2,400,000 to 900,000 years ago, the 41,000-year cycle of tilt is the dominant signal in the record (Mix 1987; Ruddiman and Raymo 1988). Other studies have underscored the changing nature of the dominant orbital signal. Spectral analysis of magnetic-susceptibility measurements of terrigenous sediment in deep-sea cores taken from the eastern tropical Atlantic Ocean, spanning the past 3.5 million years, and from the Arabian Sea, spanning the past 3.2 million years, shows that the effect of orbital forcing changed around 2.4 million years ago (Bloemendal and deMenocal 1989). Prior to 2.4 million years ago, both records carry strong 23,000-year and 19,000-year signals of the precessional cycle, suggesting that the summer monsoons bearing the terrigenous sediments were largely modulated by insolation variations during the summer season; but after that date, the 41,000-year tilt cycle signal predominates. This switch coincides with the onset of major glaciation in the Northern Hemisphere and, in the case of the Arabian Sea site at least, is reflected in the supply of terrigenous sediment (carried by monsoon winds) responding to a rapid increase in ice cover in Eurasia and North America. These shifts in dominant pulse suggest that forces additional to orbital variations have influenced Pleistocene climates. A change in the configuration of land, sea, ice, and atmosphere is a possibility (Ruddiman and Raymo 1988). In particular, the rapid tectonic uplift of the Himalayas and parts of western North America during the past few million years has led to a change in the pattern of the jet stream and the growth of cold spots over North America and Europe in the very same places that the Northern Hemisphere ice sheets were located. However, the 100,000-year eccentricity cycle has been detected in the sedimentary record before the onset of the last Ice Age, so the dominance of the tilt cycle between 2,000,000 and 900,000 years ago may be anomalous, and the resumption of the 100,000-year cycle after 900,000 years ago simply a return to normal behaviour (Mix 1987).

This refers to the lack of the long eccentricity pulse in marine records over the last 1.2 million years, but it may have a solution. Rial (2004a) showed that it is possible to tease out the 413,000-year component of eccentricity directly from orbitally untuned deep-sea oxygen-isotope ratio time series. He found that the signal is strong, albeit buried deep in the time series, but masked by frequency modulation (analogous to a carrier electronic signal's being changed in proportion to the amplitude of a lower frequency signal or 'message'). He extracted the 413,000-year signal by numerically demodulating the frequency and phase.

Variable ice-age lengths

Variations in the length of ice ages sit uncomfortably with explanations based on fixed orbital cycles. The oxygen-isotope ratio data from a 36-cm-long vein calcite core in Devil's Hole, Nevada, USA gives proxy temperature changes at odds with the astronomical theory, principally because it registers changes in the duration of the ice ages (Winograd et al. 1992). The standard explanation of climate not sticking to the rigid orbital timetable is that internal, non-linear feedbacks within the atmosphere-ice-sheet-ocean system operate that have little or no connection with orbital cycles. But, Rial (1999) came up with a solution, as with the '400,000-year problem' based on frequency modulation, supportive of the Croll-Milankovitch model and offering a plausible explanation to the variable glacial cycle duration. He found that, in the time domain, frequency modulation of a single frequency signal generates an output with periods varying slowly with time. In particular, frequency modulation of the high frequency 100,000-year short eccentricity signal by the lower frequency 413,000-year long eccentricity cycle accounts for the observed increase in the duration of glacial stages from about 80,000 to 130,000 years.

Non-orbital frequencies

Non-orbital frequencies in the proxy climate record pose difficulties for the Croll-Milankovitch hypothesis. The consensus is that these extra peaks are either harmonics or combination tones of the orbital periods, their presence indicative of non-linearity in the climate's response to orbital forcing. In theory, the creation of new frequencies and coupling among frequency bands characterizes the non-linear response of an oscillator, which is what proxy climatic observations seem to show. Nonetheless, it is unclear what non-linear mechanisms would produce the combination tones of orbital forcing and how they would do so. It is even uncertain if the climate system is capable of generating those frequencies internally, with minimal external influence (Elkibbi and Rial 2001).

Non-orbital frequencies include the swift switches from glacial to interglacial conditions. The terrestrial and marine records both register long-lasting periods of glacial expansion ending suddenly with rapid deglaciations throughout the Pleistocene. The Younger Dryas termination, which marks the end of the last glacial stage, is a splendid example of such abrupt climatic mode switches, but other rapid climatic changes occurred during the Pleistocene. For instance, during the last glaciation, oxygen isotope ratios display 24 alternations between relatively high and low values (Grootes et al. 1993). Each high-low interval lasted between several hundred and a few thousand years and involved variations of 4-6 per thousand, implying a temperature change of 7-8°C. The higher values correspond to stadials when full glacial conditions prevailed and indicate temperatures

10-13°C lower than during the Holocene. The lower values testily to warmer interstadi-als. These lasted some 500-2,000 years. The switch from stadial to interstadial climates was remarkably quick, perhaps taking place in as little as a few decades (Johnsen et al. 1992). The return to stadial conditions was less abrupt, commonly consisting of a gradual cooling followed by a more rapid slide into stadial conditions (Grootes et al. 1993). The repeated alternations of stadial and interstadial climates, which display a saw-tooth pattern (fast warming and more gradual cooling), are Dansgaard-Oeschger cycles. The sudden shifts to warmer conditions during the last glaciation may be associated with the periods of maximum North Atlantic iceberg production (Bond et al. 1993). North Atlantic sediments record several rapid episodes of iceberg production, debris rafting, and sediment deposition - Heinrich events - during the last glaciation (Heinrich 1988; Dowdeswell et al. 1995). These events appear to have involved the rapid discharge of icebergs and the melting out and sedimentation of debris held within them, probably from an ice stream lying within the Hudson Strait and draining much of the central Laurentide ice sheet. Detailed studies of the last two events, based on analysis of more than 50 North Atlantic cores, indicate that the most likely duration of a Heinrich event is 250-1,250 years (Dowdeswell et al. 1995).

During the Eemian interglacial, some climatic changes occurred that defy explanation through orbital forcing. Ice cores from Greenland - the Greenland Ice-Core Project (GRIP) and Greenland Ice Sheet Project 2 (GISP2) cores (Greenland Ice-Core Project (GRIP) Members 1993) - reveal a series of climatic 'mode switches': cold snaps alternate warm spells, each lasting 70-5,000 years (Figure 4.2(a)). The records of stable isotopes (oxygen), atmospheric dust (as measured by calcium content), methane content, and electrical measures suggest these climatic changes. The transitions between cold and warm states took place in as little as a decade, and involved a lowering or raising of temperature of up to 10°C. In addition, marine isotope stage 5e (MIS-5e) in the GRIP Summit core contains a spectacular series of climatic oscillations (Figure 4.2(c-d)). Details of two such oscillations show the enormous speed with which large climate shifts have taken place. 'Event 1' took place around 115,000 years ago, at the culmination of the Eemian interglacial. Oxygen isotope levels rose to mid-glacial levels, acidity fell rapidly, and atmospheric dust content shot up. The event appears to have lasted about seventy years (given a calculated annual ice-layer thickness of 2.5 mm). 'Event 2' is one of a lengthy series of huge and sustained oscillations that characterized the first 8,000 years of the Eemian interglacial, and the end of the previous deglaciation, sequence. Stable isotope values rose to Younger Dryas levels and were held there for about 750 years. Oxygen isotope data suggest temperature decreases of 14-10°C!

Episodes of abrupt climatic changes during the Pleistocene indicate a non-linear response of the climate system to internal or external forcing, but the nature of the physical processes involved and the character of the non-linearities themselves remain hard to pin down. Mathematical models have allowed the first steps towards the elucidation of the problem. José Rial (2004b) developed a logistic-delayed differential equation (LODE) model that successfully simulated some details of Pleistocene climatic changes. This model comprised two coupled equations with a small number of adjustable parameters but had complex dynamics. It predicted the saw-tooth pattern seen in the Dansgaard-Oeschger cycles, which according to the model appears to result from the difference between the thermal inertia of the ice and the thermo-mechanical feedback response of the ice cap. An internal origin, therefore, seems likely, but external forcing may play an enhancing role. If the LODE model does satisfactorily mimic the climate system, then it appears that the climate system trans-

Was this article helpful?

0 0
Telescopes Mastery

Telescopes Mastery

Through this ebook, you are going to learn what you will need to know all about the telescopes that can provide a fun and rewarding hobby for you and your family!

Get My Free Ebook

Post a comment