Contents

Chapter 1

Planetary origins

An understanding of the composition, internal dynamics and evolution of the planets and their satellites demands a historical perspective that extends back to the beginnings of the Solar System and the context afforded by comparison with solar systems in all stages of development circling other stars. The difficulties are magnified by the extraordinary range of scales, both in time and space, over which phenomena occur. From the dance of individual ions in the solar wind to the more sedate movement – atom by atom – of the Earth's mantle as it convects, we are challenged to discern singular events from cyclic behaviour or from long-term trends. On Earth, the readily-apparent cyclicity of the seasons, the tides and even minute changes in the length of the day, must be understood against a backdrop of millennial orbital cycles and changes in solar luminosity over aeons of time. As this chapter shows, geoscientists draw on astronomical sources when tackling these problems, but in return they contribute chronological or chemical tests for competing astrophysical explanations. Later chapters will further illustrate the blurred boundary between astronomy and geoscience in such matters as the traffic of meteorites, comets and interstellar dust between the various planetary bodies, the gravitational forces that bind them together, and the cosmic and solar radiation to which they are exposed. As Marvin (2000) puts it, the 20th century saw geology change from an Earth science to a planetary science.

Architecture of the Solar System

The bodies that make up the Solar System range in size from dust particles a few thousandths of a millimetre in diameter (Fig. 1.1) to the giant outer planets dominated by Jupiter (Fig. 1.2). Between these two extremes, we find a spectrum of objects that reflect the way in which the Solar System formed from its original cloud of cold gas. At all scales, however, there exists information that can provide us with insights into the origin and evolution of our planetary system, and ultimately our own origins. Before we go on to discuss these pieces of evidence, it is important to provide some definitions, particularly in the light of recent changes made by the International Astronomical Union (IAU). In 2006, the IAU redefined the term 'planet' to mean a celestial body that (1) is in orbit around the Sun, (2) is sufficiently massive to assume a spherical shape by virtue of gravitational selfcompression, and (3) has cleared the region around its orbit of large debris. A second class of objects, named dwarf planets, was conceived; these share the first two properties of a planet, but not the last. According to the new criteria, Pluto – by virtue of the discovery of many similar-sized icy worlds (termed Trans-Neptunian Objects or TNOs) – became just one of a handful of icy dwarf planets. Ceres, on the other hand, was promoted from the multitudinous ranks of the asteroids to become the only dwarf planet inside the orbit of Uranus. Today, the IAU recognizes five such objects (Table 1.1) and there are several other candidates awaiting consideration.

Table 1.1 Bulk properties of the major planets, dwarf planets and selected satellites. 1 AU = 149,597,870 km. Mean radii of the gas giant planets correspond to the 1 bar level in their atmospheres. Categories are assigned according to the most recent IAU classification, as described in the text.

On length scales between the dwarf planets and microscopic dust particles we find the bulk of the residue from the planet-formation process. These are the aggregated 'lumps' of solar nebula condensate – planetesimals – which have subsequently undergone varying degrees of thermal or chemical processing or both. They include the small rocky bodies of the asteroid belt and a massive reservoir of residual icy material (Fig. 1.3). The observable part of this icy reservoir, which includes the TNOs, is known as the Kuiper belt although most of us will be best acquainted with the members of this population that follow highly elliptical orbits into the inner Solar System and emit enormous extended clouds of gas and dust – the comets.

The Minor Planet Center's catalogue lists just over 620,000 minor planets (including rocky asteroids, icy TNOs and comets, as of 17 May 2013), of which 363,000 have a numerical designation and approximately 17,600 are named. Amongst this remarkable list are 9800 near- Earth asteroids, including 25143 Itokawa (Fig.1.4), of which almost 10% are larger than 1 km across and could inflict devastation in the event of a collision with the Earth. The catalogues document the orbits of over 5000 comets (e.g., Fig. 1.5), which appears superficially to be a relatively small number. However, comets typically only survive a few hundred close encounters with the Sun; their average lifetime is therefore only around 500 kyr. This observation implies that comets are replenished from some larger reservoir in the distant reaches of the outer Solar System. The possible nature of this reservoir was postulated by Ernst Öpik in 1932 and built upon by Jan Oort in 1950. What we now call the Oort cloud is thought to be a roughly spherical cloud of icy objects (perhaps 5–10 Earth masses of material) reaching up to 50,000 astronomical units (AU) from the Sun, grading inwards to a thinner disk of material that is ultimately contiguous with the Kuiper belt.

Theories on the origin of the Solar System

The nature and distribution of matter amongst these myriad small objects, as well as the larger planets and dwarf planets (including their satellites), and the uncountable grains of dust, feed directly into our search for models that account for the origin and evolution of the Solar System itself. The search provides a framework for the more focused viewpoint of the planetary geologist, but any successful Solar System model has to account for first-order issues that include the very existence of the planets and their division into terrestrial (i.e., rocky) and icy outer groups, and the distribution of angular momentum between the Sun and the planets, as well as diverse second-order features such as the tilt of planetary spin axes relative to the orbital planes and the evidence that meteorites experienced often substantial thermal modification (Woolfson 1982).

In the 17th century, the observation that the planets orbited very close to a common plane (the plane of the ecliptic; Pluto had yet to be discovered) led to the suggestion that they had grown by condensation of material in a ‘protoplanetary’ disk, either within the same cloud of dust and gas from which the Sun originally condensed (Table 1.2), or from a filament drawn out by another star passing the Sun. The former idea was developed by Pierre Simon de Laplace in 1796, who suggested it as the nebular hypothesis, whereby a rotating primordial nebula contracted under its own gravitational pull. Centrifugal and viscous forces caused the nebula to form a disk, and the planets were built up from planetesimals formed by the accretion of dust particles, while the central part of the disk continued to contract to form the Sun. The hypothesis is seriously hampered by the problem that the Sun rotates so slowly that it only possesses 2% of the Solar System's angular momentum yet embodies 99.8% of its mass; one would expect most of the angular momentum to be concentrated near the centre of mass of the rotating system. Later versions postulated the existence of turbulent instabilities in the nebular disk, which would go on to produce planets and the distant Oort cloud.

Table 1.2 Abundances of the ten most common elements (by mass fraction) in a variety of astronomical settings. (After Anderson 1989, Moore 2003, Fairbridge 1967 and Mason 1966)

A second group of theories, sometimes labelled 'catastrophic', hinge on tidal interaction between the Sun and another star. An early version, advanced in the 1880s, proposed that a 'close stellar passage' drew a filament from the Sun from which the planets were formed. The theory was developed as the tidal theory by James Jeans in 1916, but was found wanting by the geophysicist Harold Jeffreys a decade later, notably as regards the planetary spin rates that were likely to result. It was also shown by others that, although the theory did not need to account for the Sun's low angular momentum, it could not explain the high angular momentum of the planets. Nor could it explain the relatively high abundance of light elements such as lithium, beryllium and boron on Earth (and presumably the other planets), if indeed it had formed from material at solar temperatures.

Despite the persisting problem over angular momentum, the consensus favours some version of the nebular model. The pattern of the planetary orbits remains a powerful argument for their origin in a circumsolar disk, and there is much new observational support for disks around young stars in our galaxy. Figure 1.6 is a recently published collection of images acquired by the Hubble Space Telescope; these show protoplanetary disks in a part of the Orion Nebula. A review by Mamajek (2009) found that protoplanetary disks are very common, occurring around 80% of stars younger than 2 Myr in nearby stellar clusters. However, the abundance declines rapidly so that hardly any stars older than 10 Myr have such disks, placing important constraints on the timescale of planet formation. A few examples are known of older stars in which we can detect gas-poor debris disks that may represent the last stages of planet formation. Around β-Pictoris (distance 19.4 parsecs, age 8 – 20 Myr), there is a well-defined debris disk within which a roughly 10 Jupiter-mass planet orbits at a distance of around 8 AU (Fig.1.7). Recent spectroscopic analysis by de Vries et al. (2012) has shown that dust between 15 – 45 AU from the star contains almost pure Mg-olivine grains, with a relative abundance of ~ 4% of the total dust mass. The similarity in these observations to cometary dust grains in our own Solar System suggests that we are observing a proto-Kuiper belt around β-Pictoris.

Since the mid-1990s, astronomers have begun to identify planets around other stars at an ever increasing rate using a range of techniques. At the latest count (http://exoplanet.eu/catalog), planets have been detected around 693 stars, of which 133 are multiplanet systems, making a total of 888 confirmed planets. Hitherto, our understanding of Solar System formation was predicated upon just one exemplar – our own planetary system – but today we have a wealth of statistical information on the abundance of planetary systems, the distribution of planetary masses and the surprising range of possible solar system architectures. Initially, the observational techniques were heavily biased towards detection of extremely massive planets orbiting very close to their parent star. As a result, it appeared for several years as though the cosmos was dominated by solar systems with gas giants (often many multiples of the mass of Jupiter) orbiting in a matter of days barely outside the photosphere of their star, making the architecture of our own Solar System appear unusual by comparison. However, the transit technique employed more recently is better suited to detection of smaller planets in larger orbits, and consequently the number of solar systems that look reassuringly familiar has increased. Indeed, there are now candidate extrasolar planets that look as though they may possess that most sought-after quality – habitability. Table 1.3 ranks several extrasolar planets according to the Earth Similarity Index, a quantity based upon various properties of the planet, such as mass and estimated surface temperature; on this scale, Earth has an ESI = 1 and Mars has an ESI = 0.66. Note that most of these objects orbit red dwarf stars (spectral classes M and K) that are relatively old and relatively deficient in heavy elements (low metallicity) compared to our own Sun.

Table 1.3 List of known or suspected extrasolar planets with high Earth Similarity Indices, making them plausible candidates in the search for life beyond our Solar System (see Ch. 12). Data compiled from exoplanets.eu/catalog.

The Sun’s position suggests that the primitive solar nebula, as the primordial dust–gas cloud is now called, consisted of 98.4% hydrogen and helium, 1.3% volatile ices, and 0.3% rock. At the prevailing low pressures (about 0.01% of Earth's present atmospheric pressure at sea level) the gases could condense directly from the solid state without passing through a liquid state, but the chondrules, which give the chondrites their name (Fig.1.8), show that melting happened soon after. The first minerals to condense, assuming a progressive decrease in temperature, were the refractory group, which is characterized by high melting points and low vapour pressures; it includes perovskite (CaTiO₃), which occurs as small inclusions in some meteorites. These were followed by the group that includes olivines and nickel–iron alloys (Table 1.4).

The composition of the nebula must have varied from place to place; indeed, there is some evidence that the inner nebula was depleted in volatiles before any planets had formed (Humayun & Clayton 1995). Where it was highly oxidized, all of the available carbon was present as carbon monoxide (CO), leaving little oxygen to condense in water ices and resulting in icy bodies with rock to ice ratios in the region of 70:30. Likewise, all the available nitrogen would be present as N₂. On the other hand, if the primitive solar nebula was highly reduced, much of the nitrogen would have been present as ammonia (NH₃) and the carbon as methane (CH₄), almost all of the oxygen being free to combine with hydrogen and thus to condense as ice, and the rock to ice ratio would have been nearer 50:50.

The accretionary model is consistent with several observations. First, there is a large range of sizes in the solid constituents of the Solar System, and their relative proportions continue to change as the larger bodies sweep up material by their gravitational pull.

The episode of asteroid bombardment that affected the Moon (and doubtless its neighbours) some 3.9 Gyr ago (see Chapter 8) may appear heavy simply because it involved most of the larger missiles that were available. Secondly, bodies continue to interact through gravity, so that some orbits are stabilized and others disturbed. For instance, comets with periods <200 yr (identified by the prefix P/) typically move in unstable orbits perturbed by Jupiter.

The major planets, dwarf planets and satellites

The Greek astronomers distinguished planets by their movement against the pattern of the stars; the term 'planet', as we saw earlier, now has a very specific meaning according to IAU convention, and there are additional classifications of objects into dwarf planet and minor planet categories.

At the parochial level of the inner planets – Mercury, Venus, Earth and Mars – divergences remain that planetary geoscience may help to explain, notably why Earth displays a rich range of life forms, and how Venus acquired its dense atmosphere when Mars has very little and the Moon practically none. Their different masses and thus their potential for retaining their ancestral gases provide only part of the explanation.

The inner group came to be known as the terrestrial planets because, unlike the outer giants, they are predominantly rocky in composition. Pluto qualifies neither by location nor composition, as it is at least superficially an icy body and was sometimes viewed as an asteroid by virtue of its large eccentricity and inclination. The giant outer planets are divided into the gas giants Jupiter and Saturn, and the ice giants Uranus and Neptune.

The terrestrial planets have bulk densities of 3300– 5550 kg m^–3, depending on their metal content, and typically have silicate rock mantles around nickel–iron cores. Their current volatile inventories are highly variable, and it is perhaps as much a function of the rate of infall of comets as of their original composition, as well as loss mechanisms acting over geological time. They currently have internally generated magnetic fields that are predominantly dipolar or show that they formerly possessed active fields. The gas giants are composed almost entirely of hydrogen and helium, some of which is in the liquid metallic state. Their cores are probably rocky, although the familiar distinctions between different forms of matter cease to have significance at terapascal pressures (measured in millions of atmospheres).

They have relatively regular magnetic fields, similar to Earth's though far more powerful.

The ice giants appear to have mixed rock and ice cores overlain by oceans of super-critical ammonia–water fluid and deep atmospheres of hydrogen and helium. Methane is responsible for the blue-green coloration of their atmospheres, but it may be broken down into pure carbon in the form of diamond under the high pressures of the planets' deep interiors. Both ice giants have highly unusual and complex magnetic fields.

Improved observation from Earth and from spacecraft has greatly enlarged the catalogue of satellites. The present tally of planetary bodies is 8 major planets, 5 dwarf planets and 179 satellites (e.g., Fig. 1.9) distributed as follows: Earth (1); Mars (2); Jupiter (66); Saturn (62); Uranus (27); Neptune (13); Pluto (5); Haumea (2); and Eris (1). There are ring systems around Jupiter, Saturn, Uranus, Neptune and possibly Pluto, each distinctive in character, composition and dynamics (Fig. 1.10). Not surprisingly, a comprehensive theory of origin is available for only one of these 179 satellites, the Earth's Moon, as it alone has benefited from the sort of direct geochemical sampling that allows us to rule out some earlier hypotheses. Thus, co-accretion from the same part of the solar nebula as Earth is inconsistent with the very different iron content of the two bodies; capture of a passing body by Earth conflicts with the very similar oxygen isotope ratios of Earth and the Moon; and spontaneous fission requires an unknown source of energy. Current opinion favours the suggestion of Hartmann and Davis (1975) that a planetesimal the size of Mars struck Earth after its iron core had formed and blasted off iron-poor mantle material, some of which accreted into a satellite, and this is supported by continued analysis of various isotope fractionation signatures (e.g., Paniello et al. 2012) and modelling (Cuk and Stewart 2012).

The giant impact suggestion helps to explain why the oldest rocks on the Moon are 4.51 Gyr old, little different from those on Earth. The theory benefits from a new climate of opinion as well as new data. Impacts, once seen as ad hoc solutions to impossible problems, are now treated as commonplace items of planetary processing, and they appear to offer plausible solutions to some of the second-order puzzles of Solar System history, such as the 98° obliquity and retrograde motion of Uranus. It may be that impacts will help to resolve other, apparently intractable puzzles, including the current distribution of angular momentum in the Solar System.

The other satellites, as Buratti (1997) has shown, broadly conform in composition to the radial pattern displayed by the planets, namely the predominance of silicates, iron and other materials with high melting points near the Sun, ice–silicate mixtures in the satellites of Jupiter – apart from Io, which appears to have lost its water – and ice, silicate, ammonia and methane mixtures in the satellites of Saturn and Uranus.

Even lower temperatures on the satellites of Neptune (and on Pluto) favour N₂, CO and CO₂ in solid form. There is also the suggestion that the zone beyond the orbit of Mars saw the condensation of carbon, together with silicates and organic molecules characteristic of C-type asteroids and also reported from Phoebe and some Uranian satellites. But, as noted earlier, the pattern breaks down within any one satellite system. Whether or not it ever was in place there are many reasons why it should not persist.

Some of the satellites, such as Phoebe and Nereid, may be relatively recent captures, as they have not been coerced by tidal forces into synchronous rotation, whereby a satellite keeps the same face towards its primary. Others, such as Io and Triton, are being heated and distorted by tidal forces, whereas their neighbours, including the Moon and Ganymede, underwent active geological processing driven by gravitational and radioactive energy and impacts early in their evolution before becoming effectively inert.

Dust, asteroids and comets

Interplanetary dust particles (Fig. 1.1), asteroids (Fig. 1.4) and comets (Fig. 1.5) come directly from the primitive solar nebula and therefore provide samples that may be suitable for dating the formation of the Solar System. Dust is collected from balloons, high-flying aircraft and satellites, and from terrestrial collecting sites such as the bottom of a well at the South Pole. About 4 ×10⁷ kg of stony dust accretes to Earth every year, some of it cosmic in origin, some of it resulting from meteoritic and cometary collisions and break-up. Asteroids have been the object of dedicated missions and will doubtless soon supply samples for return to Earth. Remote sensing of comets, supplemented by direct mass spectrometry of Halley during flybys, has revealed many chemical species. Cometary nuclei are dominated by water ice, with perhaps 15% composed of CO, CO₂, N₂ and various organic constituents, notably H₂CO and CH₃OH (formaldehyde and methanol, respectively). The measurements bear on the possible role of comets as sources of planetary water and organic molecules, as well as on the composition of the primitive nebula.

Meteorites remain the richest source of Solar System specimens open to analysis, and enthusiastic collecting is rapidly enlarging the holdings around the world. Although the majority of meteorites are probably derived from the asteroid belt, a significant number are now known to originate from the Moon and Mars. The ages obtained by analysis of meteorites range from 4.53 to 4.57 Gyr. A few appear to be derived from asteroids or planets that had undergone some differentiation: they are the cosmic equivalent of stratigraphical samples.

The classification of meteorites is complex, because it takes into account composition, degree of alteration and source, and incorporates subgroups for which there may be very few examples (Table 1.5). The key distinction is between undifferentiated meteorites and those that reflect melting of the parent body. Chondrites contain spherical bodies called chondrules (see Fig. 1.8), which are 0.1–2.0 mm in diameter and have solidified from a melt, inclusions rich in calcium and aluminium, and grains apparently older than the Solar System. The chemical composition of chondrites is very similar to that of the Sun if its gas components are disregarded. They are thought to have formed within 20 Myr of the origin of the Solar System: chondrules from the Allende meteorite yield ages of 4.563 Gyr (Fig. 1.11).

Table 1.5 Categories of meteorites (mainly after McSween 1999). Percentages indicate the relative abundance of falls, which are a more accurate reflection of the meteorite flux onto the Earth's surface. Shading identifies undifferentiated meteorites from those that are differentiated. (Data from the Meteoritical Bulletin Database, http://www.lpi.usra.edu/meteor)

Stones account for 94% of meteorite falls, comprising chondrites (86%) and achondrites (8%); however, finds of stones (as distinct from falls) total 70% (chondrites 68%, achondrites 2%). The detection of falls might seem to be a matter of luck, especially as it is limited to a few centuries of reliable record, but finds are also accidental, and biased in favour of sandy and icy deserts, and the proportion of stones is distorted by the relative resistance of iron meteorites to weathering.

Although the isotopic composition of meteorites is relatively uniform, variations in oxygen-isotope ratios may show that a subset, such as the howardites, eucrites and diogenites (HED meteorites), originated in a single parent body; these are now widely believed to be fragments of the asteroid Vesta. By the same token, differences in composition may rule out a common origin, although there is always the possibility that the parent body had not reached isotopic equilibrium by the time it shed the meteorites in question.

Primitive achondrites have compositions similar to that of chondrites but lack their distinctive texture. The remaining achondrites reflect differentiation from irons and stony irons in planetary bodies (presumably asteroids) that had undergone melting. The ureilites contain carbon minerals, sometimes in the form of diamond, as well as silica minerals. The SNC group (Shergottites–Nakhlites–Chassignites; Table 1.5, Figure 1.12) is sometimes broadened to include ALH84001, a meteorite over 4 Gyr old found in Antarctica, which is thought by some investigators to bear evidence of life on its parent planet Mars (see Chapter 12). A few achondrites originate on the Moon.

Irons (4% of falls, 27% of finds) are thought to originate in the cores of asteroids that broke up, but a few are suspected of being the product of impact melting. They consist of iron–nickel alloys, and can be subdivided on a textural basis into hexahedrites (< 6% Ni) octahedrites (6–17% Ni) and ataxites (> 17% Ni). The last two groups, and most spectacularly the octahedrites, display an arrangement of plates of the nickel–iron alloys kamacite (~6% Ni) and taenite (~30% Ni) known as the Widmannstätten pattern (Fig. 1.13). More recently, the iron meteorites have been classified on the basis of geochemical signatures rather than petrographic textures (see Table 1.5). The eucrites are sometimes compared with mantle rocks from Earth and the Moon, and the diogenites with mantle rocks from Earth.

Planetary bodies are held to have formed by accretion of grains that condensed from the nebula. The notion of homogeneous accretion requires that condensation ceased – somewhat implausibly – prior to the onset of accretion, producing cold homogeneous bodies that become heated by the kinetic energy of large impactors, gravitational energy, and the decay of short-lived radionuclides. Heterogeneous accretion postulates, more plausibly, simultaneous condensation and accretion; this produces a layered body containing condensates at high temperature in the core and at low-temperature around it.

Centimetre-size particles probably took ~ 4 kyr to grow, metre-size objects a further 2 kyr, and kilometresize planetesimals about 10 kyr. Further accretion was probably inhibited by turbulence, but, as the solar wind asserted itself, the nebula was dispersed and planet formation could begin. Initially, there would have been millions of kilometre-size bodies competing for space. Collisions between these resulted in both destruction and further accretion, gradually shifting the peak of the size distribution towards larger and larger radii and smaller populations. After many tens of millions of years the inner Solar System may have been populated by several tens of objects over 1000 km across, but only a few bodies more than 6000 km in diameter.

A more violent process has recently been proposed (Zhang 2002). It has two variants. In the first, the history of the Earth between 4.55 and 4.45 Gyr was deleted by collision with the Mars-sized impactor discussed above and differentiation had to begin afresh: in Zhang's words, the Earth was 'rehomogenized and reborn'. In the second version there was continuous accretion and core formation, which was interrupted by impacts too weak to restart the process; however, the last of these occurred ~ 4.45 Myr ago and was substantial enough to strip away the Earth's atmosphere and re-melt the crust. The proposal is in tune with analyses of biological evolution which postulate that a major impact destroyed the bulk of terrestrial life, thereby resetting the evolutionary clock, at least once.

The accretionary process itself provided enough energy to melt much of the Earth, especially if, as is now thought, many of the bodies in question were more than 10 km in diameter and in some cases (including the impactor held to be responsible for ejecting the Moon) similar in mass to Earth. The Earth's constituents were thereby mobilized and rearranged primarily on the basis of their chemistry and density. Following the Goldschmidt scheme formulated in 1912, the elements in question can be grouped into lithophile, which tend to occur with oxygen in oxides and silicates, atmophile, which are gaseous under the conditions found at the surface of the Earth, chalcophile, which are concentrated as sulphides, and siderophile, which tend to be metallic and have an affinity for iron.

Each of these groups of minerals had a different range of densities. The siderophiles were the densest and gradually accumulated in the metallic core we see today. The atmophile substances were the lightest of all, and rose to the surface to form an atmosphere. The sulphide minerals fell between the silicates and metals in density. The differentiation is not complete, although some of the sulphides and siderophiles in the crust were probably brought in by impactors.

The energy released by the formation of the core amounted to about 1.3 × 10³⁰ J and would have raised the Earth's internal temperature by around 1000 K, hastening differentiation. Radioactive decay is another major heat source, whose discovery early in the 20th century reconciled the geological and biological evidence for an old Earth with its present temperature.

In contrast, the composition of Mercury is thought to be 60% iron, whence the suggestion that its rocky mantle was lost to an impact. However, the low levels of titanium and iron at the surface may indicate that deep-seated, widespread volcanism stopped early in the planet's history, so that the core cooled slowly enough to remain partly molten (de Pater & Lissauer 2001). Venus is thought to have formed in a warmer part of the nebula than did the Earth and may consequently contain less sulphur, an inference consistent with the somewhat lower average density of Venus, although some workers prefer to explain the difference by variations in the composition of the founding planetesimals.

On Mars, the large shield volcanoes point to early heat loss by low-viscosity (high FeO) magmas. The planet has a higher content of relatively volatile material (including sodium, potassium, phosphorus and rubidium) than the Earth, a difference again attributed to the planetesimals from which it formed. There is isotopic evidence to suggest that Mars accreted much more quickly than the Earth, perhaps in as little as 100 Myr (Stevenson 2001). Its small size means that both heating up and cooling down were much faster than on Earth (Zuber 2001). As regards the gas giants, initial core formation by planetesimals would have been followed by rapid gas accretion. Additional planetesimals contributed to the solid content of the gas envelopes. The planets then contracted until the limit of incompressibility was reached, and the gaseous envelope began to cool.

In sum, the early evolution of the planets has left a clear imprint in their composition, it has dominated their thermal histories and thus their internal differentiation, and it strongly coloured the extent and character of the contribution to their morphology and surface composition made by impacts. As we shall see in the next chapter, it continues to play an active role in geological and atmospheric change through the orbital geometry, dynamics and relative position of the various planetary bodies. Planetary surfaces, as Gould (1991) puts it, lie in the domain of complex historical sciences, where modes of explanation differ from the stereotypes of simple and well-controlled laboratory experiments.

Chapter 2

Orbits and cycles

Discussion of planetary orbits is commonplace, if often cursory, in the context of climate history, thanks to the general acceptance in the 1960s of the Milankovitch thesis, first advanced in 1941, that periodic fluctuations in the Earth's axial tilt and precession, and the departure from circularity of its orbit, control its receipt of solar radiation and hence the timing and severity of glacial cycles. Despite these regularities, however, there is a wide range of orbital inclinations, eccentricities, periods of revolution and periods of rotation among the planets which reflect the vagaries of accretion, complicated by gravitational interaction between them and the outcome of major impacts, and which impinge on recent and current geological processes.

Once Copernicus' heliocentric model had gained general acceptance the door was opened to objective orbital observation and thence the dynamic interpretation of planetary motion. The key stepping stones were provided by the three laws of planetary motion advanced by Johannes Kepler in 1609 and Isaac Newton's law of universal gravitation (1687). Kepler broke away from the idealized circular and epicyclic models of antiquity to state (1) that all planets move in elliptical orbits, with the Sun at one focus, (2) that a line joining a planet to the Sun sweeps out equal areas in equal times (Fig.2.1), and (3) that the square of the orbital period of a planet is directly proportional to the cube of the semi-major axis of its orbit.

The first law is clearly vital to our understanding of annual variations in a planet's receipt of solar energy and, if the eccentricity (ε) of the ellipse varies, in longterm climatic change. The eccentricity of the planets, defined as the ratio of the distance between the two foci to the length of the major axis, ranges from 0.007 for Venus to 0.206 for Mercury (Table 2.1). The Earth's is currently 0.017 but has varied over the millennia from 0.003 to 0.058. Pluto's is 0.248; for most asteroids it is 0 – 0.35; for comets it is generally close to 1 (Halley's is 0.967). Of course, strictly speaking the Sun and Earth orbit their common centre of mass or barycentre, which is 450 km from the Sun's centre rather than midway between the two bodies. For the Earth–Moon system the centre of mass is 4700 km away from the Earth's centre. As shown by Newton, the Sun moves around the Solar System's barycentre under the gravitational influence of the planets (Fig.2.2), with an approximate periodicity of 179 yr (Mackey 2007). Some see this value as a component of numerous geological chronologies which hinge on changes in solar irradiance (e.g., Fairbridge and Shirley 1987).

Table 2.1 Orbital data for the major planets, dwarf planets and largest satellites (various sources).

¹ With respect to the plane of the ecliptic, unless indicated otherwise

² With respect to the object's orbital plane

* With respect to the equatorial plane of the object's primary

R = retrograde

synchr = orbital period synchronous with spin period

+ internal rotation period derived from modulation of magnetic field

The second law is a picturesque way of saying that the planet's orbital velocity is greater near the Sun (i.e., at perihelion) than when furthest from the Sun (aphelion) and, as Newton showed, a consequence of the principle of the conservation of angular momentum. Newton modified the law to obtain the familiar universal law of gravitation F = Gm₁m₂/r², and this allows the mass of distant objects to be determined. The third law (Fig.2.3) has acquired topical interest in current research on exoplanets. KEPLER, a NASA mission launched in 2009, is dedicated to finding Earth-like planets in the habitable zone of other stars; although its primary tool is a photometer, which can detect the periodic dimming of a star caused by the passage of an orbiting planet before it, Kepler's formula yields the distance between a star and its satellite. By 2012 KEPLER had identified 2321 planet candidates of which 48 appeared to occupy habitable zones. In due course we will know enough about the composition of exoplanets to justify their inclusion in the repertoire of planetary geology.

More generally, the three laws and their derivatives led to the exquisite precision of modern orbital calculations and their role in executing space travel and in calculating the perturbing effect on each other of planets and other bodies. The subject, as we saw, is generally investigated in Newtonian gravitational terms despite the key role played by relativity in cosmology, although corrections need to be applied when relativistic factors come into play over large distances or at high velocities. Similarly, although most of the calculations treat planetary bodies as a whole, we may need to consider differential response to gravitational effects by their constituents, as this provides valuable clues to subsurface composition and geological processes.

Insolation

Comparison with the observations made by the astronomer Timocharis about 150 years previously led Hipparchus (190–120 BC) to infer that the Earth's axis traced out a circle on the sky with a period of just under 25.8 kyr. Whereas Earth's axis points almost exactly to the star α-Ursa Minoris (Polaris), 5000 years ago it pointed to α-Draconis (Thuban), and in 12,000 years it will be pointing at α-Lyrae (Vega). The effect of this change is to reverse the hemisphere experiencing summer at perihelion over a 13 kyr period. Thus, whereas perihelion currently occurs in January, 13,000 years ago it would have occurred in July, coinciding with northern hemisphere summer.

The second major cyclical source of change in the amount and character of the solar energy received by a planet is the eccentricity of its orbit. For Earth it changes from almost perfect circularity to maximum ellipticity with a period of ~96 kyr; perihelion is currently in January, and aphelion in July. Moreover, because the Earth moves faster in its orbit at perihelion, and more slowly at aphelion, northern winter is approximately seven days shorter than northern summer, and vice versa in the southern hemisphere.

The obliquity of the Earth's axis provides an additional orbital parameter. The value of the Earth's obliquity is currently 23° 27' 54.16', but varies with a period of 41 kyr between 21.6° and 24.5°. The narrower the angle, the smaller the insolation contrast between seasons, and hence winters are milder and summers generally cooler.

The three orbital variables (Fig.2.4) are now named collectively after the Serbian Milutin Milankovitch (1879–1958), who put on a firm footing an astronomical explanation for the ice ages that had been under discussion for several decades. For some years the variations he invoked were dismissed as inadequate, but analysis of isotopes from deep-sea cores in the 1960s eventually demonstrated that repeated glacial/interglacial cycles corresponded in timing with the Milankovitch insolation cycles and led to widespread acceptance of the model. The 96 kyr cycle appears to dominate the climate record, at least during the Pleistocene, and this may be the result not only of changes in insolation as the orbit changes shape, but also of the Earth's dipping into a large dust cloud probably formed by collisions between asteroids (Muller & Macdonald 2000).

It was soon realized that orbital variations could also help to explain the climatic history of Mars (Sagan 1971) even though the variables are slightly different in magnitude and period (Mutch et al. 1976). As shown in Table 10.1, eccentricity varies with two periodicities, resulting in insolation changes at the poles of ~1% and at the subsolar point by as much as 30% at perihelion. The obliquity of the Martian axis varies from 15° to 35° and perhaps as much as 0–60°. In addition to variations in the tilt and direction of the spin axis, the inclination and direction of Mars' orbital plane also change. The inclination of the orbital plane of Mars varies with a period of 1.2 Myr (the corresponding value for the Earth being ~1 Myr) and the plane precesses with a period of ~ 69 kyr.

The precession of the Martian rotation axis takes ~ 175 kyr compared with 26 kyr for Earth because Mars does not have a large moon to accelerate the precession. In 1881 Alfred Russel Wallace reported short-term changes in the extent of the snow caps of Mars and he used the observations to test the significance of precession at times of moderate orbital eccentricity. For example, on 23 June the southern Martian cap had a diameter of 11° 30'; by 9 July it had shrunk to 5° 46'. In recent years there has been much interest in the inverse effect, that is the influence of the changing ice caps on Mars' obliquity. It has been shown that, if the ice caps at some stage in the cycle contain 10¹⁷–10¹⁸ kg of water, the obliquity may oscillate correspondingly by tens of degrees (Bills 1999; Rubincam 1999).

The present combination of orbital and spin elements for Mars produces a short hot summer in the southern hemisphere and a longer cooler summer in the northern hemisphere. The calculated variation in solar illumination with change in axial inclination shows an overall evening out of the latitudinal distribution of insolation at the current obliquity of ~25°. The annual average insolation at the equator is ~1275 W m^-2 and at the poles it is ~75 W m^-2. When the obliquity increases to > 50°, the equatorial insolation falls to 155 W m^-2, but the polar insolation increases dramatically to 148 Wm^-2. Modelling of the effects of these changes is very complicated, since the models must account for changes in the size, composition, and albedo of the polar caps, thermal response and outgassing of the regolith, and changes in the thickness, opacity to infrared (IR) radiation, and dust content of the atmosphere.

The solar constant is about 1370 W m^–2 at the mean distance of the Earth from the Sun and drops to about 1/900 at the distance of Neptune but it cannot therefore be dismissed as irrelevant. For example, auroral displays (Fig.2.5) demonstrate that the solar wind interacts with the magnetospheres of the outer planets, and hitherto unexplained changes in brightness on Uranus and Neptune have been correlated with the 11-year solar cycle (Lockwood 1986). Short-term, irregular changes, such as solar flares, do not yet lend themselves to geological investigation, although isotopes in lunar samples collected by the Apollo missions appear to be the product of flares over the last 10⁷ yr. Geological data can make a more substantial contribution to the study of long-term trends in solar output and in particular 'the faint young Sun' paradox discussed in Chapter 9.

The search for orbital influences on environmental change is being extended to other moons and planets. On Triton and Pluto atmospheric pressure is controlled by the distribution of nitrogen and methane frosts at their surfaces, and their global climates are consequently subject to extreme seasonal variations. On Pluto there is a severe change in insolation due to the high orbital eccentricity, the net effect of which appears to be the almost total condensation of the atmosphere as surface frosts. Indeed, detectable albedo changes between 1950 and 1995 could be modelled as the redistribution of frosts associated with condensation of the atmosphere following perihelion. Pluto is moving farther away from the Sun, and the present atmosphere may have completely condensed by 2025. On Triton (Fig.2.6) the combination of rapid orbital precession, the period and inclination of Triton's orbit, and the period and inclination of Neptune's orbit, yield vary large variations in the latitude of the subsolar point. Seasonal variations with a period of 165 years (Neptune's orbital period) are modulated by an oscillation with a period of 688 years (Triton's orbital precession period), and because of this the subsolar point may remain within 10° of the equator during some parts of the cycle, yet stray as far as ±52° away from the equator at others (Cruikshank 1999).

In August 1989 Voyager 2 flew past Triton as it was approaching the height of its most intense summer since the middle of the 14th century. It was therefore anticipated that the atmosphere should be thickening rapidly as more and more nitrogen frost sublimated. Recent observations by the Hubble Space Telescope have confirmed that Triton's atmosphere has thickened considerably since 1989 and that by 1997 the atmospheric pressure at the surface had increased from ~14 μbar to ~40 μbar, and the temperature has increased from 37.5 K to 39.3 K. The atmosphere currently appears to be doubling, or even tripling, in bulk every decade, with some indication that the increase is not linear. This inferred global warming of Triton is consistent with increased insolation on a permanent N₂ cap at the south pole (Elliot et al. 1998), and orbiters being considered for the second decade of the present century may see a quite different Triton from that which greeted Voyager 2 in 1989.

Besides helping to account for climatic history, orbital variables are valuable clues to mass distribution within the planetary body in question. The modern version of Hipparchus' technique is to compare the orientation of the Earth's pole in the sky relative to fixed extragalactic radio sources over periods of many years. The smaller the value, the more matter is condensed towards the core. A uniform mass distribution yields a moment of inertia factor M = 0.4; the terrestrial planets have M approximately equal to 0.33, and the gas giants have M closer to 0.2. On Mars, the presence of the Tharsis bulge prevents our finding the polar moment of inertia from the gravity field, and thus our making inferences about the planet's interior to make up for the lack of seismic data, but estimates of the precession rate can be based on observations of the pole positions recorded twenty years apart by the Viking landers (1976) and by the Pathfinder lander (1996).

Interaction between planetary bodies

Early in the Earth's history, variations in its obliquity may have been chaotic and extended from 0° to about 85° until acquisition of the Moon stabilized the obliquity at its present value of 23.44° with variations of ±1.3° (Laskar et al. 1993). On the other hand, spin–orbit resonances between the spin of the Earth and the orbital period of the Moon can produce extreme oscillations of the terrestrial spin axis to over 60°. The significance of this is that tilts greater than 54° yield higher solar insolation at the poles than at the equator. However, at the present rate of decrease in lunar mean motion the principal resonances will not be encountered for another 1.5 to 2 Gyr.

The Moon is of course primarily responsible for the Earth's tides, which are most prominent in the oceans but also detectable in the solid Earth. A related effect is the progressive change in the Earth's rate of rotation and the distance between it and the Moon. Edmund Halley (1656–1742) noted that the dates of eclipses reported in antiquity did not agree with those calculated on the basis of the current rate of rotation of the Earth and suggested that the Earth had slowed down. Modern calculations on the same basis indicate deceleration by 2 milliseconds per 100 yr.

Tidal friction was soon identified as the likely cause. In 1963 J. W. Wells showed that the growth bands of corals dating from the Devonian, 370 million years ago, reflected a year composed of about 400 days. Later studies on corals and molluscs showed a corresponding decrease in the number of days in the lunar month.

Results consistent with these rates come from tidal sediments (rhythmites), which are perhaps more reliable than fossil growth layers for this kind of analysis. Late Proterozoic and Palaeozoic deposits from the USA and Australia dating from 300 to 900 Myr ago suggest that the Moon's rate of retreat has been nearly constant since the late Precambrian. They point to about 456 days in the year (481 if a solar component of tidal friction is included) for the oldest units, but, unlike the palaeontological evidence, the rhythmite record did not require that the Moon was close to the Earth late in geological time.

By the principle of the conservation of momentum, any deceleration of the Earth has to be balanced by an acceleration of the Moon. Hence the gradual increase in its distance from the Earth, which can now be checked by laser reflection on devices left by astronauts on the Moon for this purpose. The measured rate is 3.5 cmyr^-1, corresponding to a deceleration of the Earth's rotation of 1.48 ms per century, close to the estimated average.