In the mid 1960s, a young New Zealand geologist named Vic McGregor (1940–2000) was commissioned by the Copenhagen-based Geological Survey of Greenland to make a geological map of a large region of complex ancient rocks in the southern part of West Greenland, in and around Greenland's capital, Godthaab (later renamed Nuuk). This scenic wonderland of mighty fjords and mountains forms part of a strip roughly 130 km wide at this latitude between Greenland's coast and the edge of the huge inland ice-cap (figure 1).
For several years McGregor explored the geology of this region from a tiny, partly open boat, with just enough room for himself, two local crew, an occasional guest, and essential camping, hunting, fishing and geological equipment.
Before his work in Greenland, McGregor was strongly influenced by the work of the husband-and-wife team John Sutton and Janet Watson (both later elected FRS) of Imperial College London, who had earlier pioneered modern field-based methods for clarifying the geological evolution and structure of the ancient rocks of the so-called Lewisian Complex of the northwest Highlands of Scotland, as described in their landmark paper of 1950.1 By careful and detailed geological mapping, they recognized a whole series of varied geological events that had affected these complex rocks, most of which had been strongly altered, deformed and recrystallized (‘metamorphosed’) at high pressures and temperatures. Sutton and Watson showed that the Lewisian Complex comprises two main groups of geological events, which they conjectured were widely separated in geological time and which they termed ‘Scourian’ (the older) and ‘Laxfordian’ (the younger), after prominent localities. However, no absolute ages for these rocks were available at that time.
In the late 1950s, my colleagues and I at Oxford began to set up the first geological dating laboratory in the UK, using recently developed isotopic techniques based on radioactive schemes rubidium–strontium (87Rb–87Sr), potassium–argon (40K–40Ar) and uranium–lead (238U–206Pb and 235U–207Pb), as well as the powerful, time-dependent combination lead/lead (207Pb/206Pb). This led to a concerted attack on the Lewisian Complex, providing age data for a range of Lewisian events, including the oldest and youngest. This confirmed Sutton and Watson's prediction of two widely separated groups of events, at ca. 2800 million years (Myr) and 1700 Myr for the Scourian and Laxfordian, respectively.2 At first, the Scourian rocks remained some of the oldest known, reliably dated, rocks on Earth. By this time, the work of such great pioneers as Arthur Holmes, Fritz Houtermans and Clair Patterson on Pb isotopes in terrestrial rocks and meteorites had already demonstrated that the Earth and Solar System were close to 4.5 billion years (Gyr) old.,3
My own activities in the isotopic dating of rocks arose in line with the growing international realization in the mid twentieth century that geology (sensu lato) must in future deal with the origin, structure and evolution of the whole Earth, involving the quantitative evaluation of all physical, chemical and biological processes that have shaped our planet throughout its entire history. The continuing technical and interpretational improvement of dating methods since the 1950s, using radioisotopes with widely differing half-lives, has produced a timescale for numerous geological, palaeobiological, palaeoanthropological and palaeoenvironmental processes, not to mention the study of meteorites and planetary material as well as the accretion and early differentiation history of the Earth. As the result of huge international effort, all these fields of enquiry have been placed on a truly realistic historical-evolutionary footing, although much work still remains to be done.
Results of fieldwork and isotopic dating
Taking up the story in Greenland, by 1970 McGregor had single-handedly made a detailed geological map of the Godthaabsfjord region. Using field techniques analogous to those that Sutton and Watson had earlier applied in the Lewisian, he came up with a complex sequence of two main groups of 10 successive magmatic (that is, igneous), supracrustal (that is, rocks originally laid down at the surface of the Earth) and metamorphic (affecting existing rocks) events.4 The latest major rock-forming event in this sequence had already given a potassium–argon date of 2710±130 Myr in some unconnected reconnaissance research.,5 Hence McGregor reasoned that if the youngest event in such a complex sequence gave 2.7 Gyr, then the very oldest event that he recognized at that time (which has since been surpassed by even older events) could be ‘very old indeed’ and, possibly, the oldest rocks known on Earth. Indeed, he hazarded an inspired guess of ca. 3.6 Gyr,6 which seemed very courageous at the time. McGregor termed these potentially oldest rocks the Amîtsoq gneisses after a local place name.
I first heard of McGregor's geological model for the Godthaabsfjord region in 1969, when I contacted him. Quite early on, I sensed his frustration in trying to convince his colleagues in the Survey of the correctness of his proposed geological succession. As a result, we decided at Oxford to attempt to date some crucial rock units in McGregor's complex sequence, using methods previously applied to the Lewisian Complex of Scotland (see above).
One memorable day in 1970, we first analysed isotopically a sample of Amîtsoq gneiss (see above) sent to Oxford by McGregor, representing the oldest member of his sequence. We found that it had a more ‘unradiogenic’ lead isotope composition (that is, lower 206Pb/204Pb and 207Pb/204Pb ratios)7 than any terrestrial ore or rock lead ever reported previously. On any plausible terrestrial lead-isotope growth models, which were already fairly well understood by that time,,3 the analysed lead, and hence the rock, must be at least 3.7 Gyr old. By that time the gneiss had already become highly depleted in uranium (a common observation in strongly metamorphosed gneisses), so that radiogenic evolution of their lead from uranium had virtually ceased.
During the early 1970s I had the good fortune of visiting the mapped Godthaabsfjord localities with McGregor, while amassing a large collection of Amîtsoq gneisses and younger rocks (figure 2). Using the dating methods 87Rb–87Sr, 207Pb/206Pb and (later) 147Sm–143Nd, it was confirmed that the Amîtsoq gneisses were ca. 3.65–3.75 Gyr old, far older than any reliably dated rocks on Earth at that time.8
From the age data it was also possible to calculate the initial Sr, Pb and Nd isotope ratios8 of the ancient gneisses at the time of their formation. These results showed that the magmatic precursors of the Amîtsoq gneisses had formed by partial melting of a mantle-like source region, and that they represented a juvenile addition of continental-crust-like material at or near the Earth's surface. Indeed, later work showed that much of southern West Greenland is made up of three great, separate episodes of mantle-crust differentiation at ca. 3.7, 2.8 and 1.7 Gyr,9 all of which were tectonically quite independent, and were not produced by remelting or reworking of earlier continental crust (,figure 3). Radiogenic isotope research has always strongly supported the view that continental crust has grown discontinuously throughout geological time as a result of huge, widely separated (in time) episodes of mantle differentiation.,10
In 1971 McGregor and I visited the remote and, at that time, virtually unknown Isua region, some 150 km northeast of the already mapped Godthaabsfjord region and close to the edge of the inland ice. It proved quite an arduous trip to sail in McGregor's tiny boat up to the head of the iceberg-packed Godthaabsfjord (where, in the Middle Ages, the resident Vikings practised domestic agriculture). There we were picked up by a helicopter belonging to the Kryolit-Marcona Mining Company; the Company had started to explore a huge iron-ore deposit at an altitude of 1240 m at the very edge of the inland ice (latitude 65° N), which had been discovered from a major aeromagnetic anomaly (figure 4).
Aided by locality sketch-maps provided by the Company geologists, McGregor and I made the first geological interpretation of the now well-known Isua supracrustal belt (ISB), sometimes termed the Isua greenstone belt, which has since been closely studied by many workers. Right from the start, we regarded the ISB as older than the adjacent gneisses, which we provisionally equated (correctly, as it later turned out; see below) with the ca. 3.7 Gyr-old Amîtsoq gneisses some 150 km to the southwest, near the coast. These gneisses are enclosed within the broadly circular outcrop of the ISB, which is ca. 20 km in diameter and ca. 35 km in strike (that is, parallel to the outcrop), with a variable outcrop width of 1–3 km. The ISB rocks form an incredible contrast with the bordering gneisses of deep-seated, magmatic origin. Despite strong, but variable, metamorphic recrystallization and deformation, the original ISB components are mostly identifiable as a complex series of chemical (that is, precipitated from solution) and clastic (that is, eroded as detrital components from older rocks) sediments, as well as mafic (basaltic, silica-poor) and felsic (silica-rich) volcanics.11 Almost all the rocks were deposited by recognizably uniformitarian processes (that is, those active throughout subsequent geological time) in a marine environment. Of particular interest are the voluminous basaltic pillow lavas, which were erupted directly into seawater. The banded iron-formation (BIF), which comprises the major iron-ore deposit at Isua, with its alternating bands of chert (SiO2) and magnetite (Fe3O4), was also deposited under marine conditions.
How old are the Isua rocks? The first age determinations, performed at Oxford, gave rubidium–strontium and lead/lead ages of close to 3700–3750 Myr for the gneisses and for the BIF.12 The BIF dating proved the unexpected presence of liquid water on the Earth's surface as early as ca. 3.7 Gyr ago. The gneiss dating showed that our provisional correlation of the Isua gneisses with the Amîtsoq gneisses of the Godthaabsfjord region was valid. In addition, McGregor had been correct right from the start of our visit in proposing that the Isua region rocks as a whole represented a more pristine, somewhat less deformed and metamorphosed version of the Godthaabsfjord region rocks, and that one could look even further back in time because of the presence of the ISB rocks. (Later, McGregor reported the discovery of rocks from the Godthaabsfjord region that were almost certainly equivalent in type and age to the ISB rocks, but they were in not nearly such a good state of preservation.)13
Since those early years, the ISB has yielded numerous uranium–lead dates in the range 3.7–3.8 Gyr ago, particularly using the accessory mineral zircon (ZrSiO4), which contains trace amounts of uranium and which lends itself to very precise dating.8 Published zircon dates strongly suggest that the depositional age of at least part of the ISB could be closer to 3.8 Gyr than to 3.7 Gyr.,14
Bordering directly on the southern margin of the ISB is a still little-explored terrain of at least 100 km2 of well-exposed, low-deformation gneisses of magmatic origin. Zircon uranium–lead dates range up to 3.82 Gyr, and these gneisses contain abundant inclusions of older rocks, easily identified as varied volcanic and sedimentary rocks.15 These inclusions may be of the same age as the ISB rocks to the north, or even slightly older.
Other sites of ancient rocks (more than 3.5 Gyr old) have been reported from around the world,16 although the age evidence for some of them is still preliminary. None is as extensive, well preserved and well exposed as the Godthaabsfjord–Isua region, with its overall age range of ca. 3.82–3.65 Gyr (perhaps locally up to 3.85 Gyr). Here we assume (arguably) that the ancient Greenland rocks are representative of global terrestrial processes at ca. 3.8 Gyr old.
I conclude with a brief summary of what type of scientific insights these oldest rocks can provide.
Scientific insights from the oldest rocks
The most important observation is that there is no direct evidence within the Greenland rocks for any exposed primordial crust significantly older than ca. 3.8 Gyr. All rocks in the Isua region are of secondary origin in that they were produced from varied source rocks by identifiable uniformitarian geological processes. Chemical sediments were precipitated in warm ocean water fed with iron-rich chemicals derived from hydrothermal vents discharging massive volumes of basaltic lavas into the water from the hot mantle below. Clastic sediments were produced close to the shorelines of a low-relief volcanic landscape. In the 3.8 Gyr sediments there is no sign of detrital fragments of any older continental type crust of granitic character, such as are frequently seen elsewhere in younger rock assemblages of this type that are known to postdate the existence of continental crust. Nevertheless, there is positive evidence that the types of deep-seated magmatic rock of broadly granitic composition that form the backbone of modern continents were already in production only slightly later, by 3.7–3.8 Gyr, as judged from their substantial presence in the Godthaabsfjord–Isua region. Such rocks (termed Tonalite–Trondhjemite–Granodiorite, or TTG gneisses) were generated by well-understood processes of global tectonics, involving plate tectonics, ocean-floor mobility, subduction and partial melting of mafic crust, leading to growth of continental crust throughout geological time. It seems from the Greenland rocks that these global tectonic processes might already have commenced at ca. 3.7–3.8 Gyr ago, soon leading to the formation of the Earth's first true continental crust.17
Intense debate and controversy persists over whether life already existed on Earth by 3.7–3.8 Gyr ago. Such claims have been strongly voiced for the Isua sediments, and also for possibly related rocks some 150 km to the southwest on the coast near Nuuk.18 Because of the relatively high degree of deformation and metamorphism of these rocks, the presence of morphological fossil evidence is most unlikely. The claims for life are therefore based on certain carbon isotope ratios in graphite grains, which occur in some of these sediments. These grains show the low 13C/12C ratios (13C depletion) characteristic of all biological material. However, it is now known that low 13C/12C ratios in graphite can be produced by thermochemical reactions involving the decomposition of non-biological carbonate rocks, thus greatly decreasing the diagnostic value of carbon isotope ratios in ancient rocks for the past presence of life.19 I am highly sceptical of all claims so far made for the presence of biogenic material in the oldest Greenland rocks. Now that we know more about their depositional environment, there is no a priori reason why life should not have got going by Isua times, but there is simply no convincing evidence for it. The search in this region continues for more reliable diagnostic criteria, but the best evidence so far for the oldest life comes from almost unmetamorphosed, undeformed 3.4–3.5 Gyr sediments from Western Australia and South Africa that contain widely accepted morphological evidence for cellular life.20
Despite the secondary nature of the Isua rocks themselves (see above) there is ample evidence from their lead isotope ratios that some of the rocks, or more probably their immediate precursors, were derived from an unexposed source region with an age of up to 4.2–4.3 Gyr, which most probably had a mafic, mantle-like composition.21 This must have been a part of the original so-called ‘Hadean’ crust, which has not yet been found exposed anywhere at the Earth's surface. In addition, the Isua rocks and surrounding gneisses are so far the only known terrestrial rocks that preserve positive anomalies in the abundance of the isotope 142Nd, produced by α-decay of 146Sm, with a half-life of 103 Myr. Since virtually all 146Sm produced at the formation of the Solar System would have decayed by some 4.1–4.2 Gyr ago, it means that 142Nd heterogeneities in the source of the 3.7–3.8 Gyr Greenland rocks must have been preserved from a much earlier period in Earth history. When the 146Sm–142Nd data are used in conjunction with Greenland age and isotope data obtained with the 147Sm–143Nd system (the half-life of 147Sm is 106 Gyr), which is one of the principal dating methods used in geochronology, it can be shown that major global differentiation into the primary layers of the Earth occurred ca. 4.45 Gyr ago, only some one hundred million years after the accretion of the Earth itself.22 The search for 142Nd isotope anomalies elsewhere on Earth is in progress, but it is unlikely that this anomaly can be detected in rocks younger than about 3.5 Gyr.
It may not be fortuitous that an age of ca. 3.82 Gyr, or just above, for the oldest supracrustal rocks and associated gneisses in southern West Greenland is only a few tens of millions of years (40±20 Myr?) younger than the probable age of the so-called Late Heavy Meteorite Bombardment (LHMB), which is widely held to have affected Moon and Earth simultaneously, although no geological evidence for it has yet been found on Earth. Detailed measurements of isotopic age on lunar samples in combination with impact crater counts indicate a very rapid decline in the rate of lunar basin formation during the period 3.90–3.85 Gyr ago.23 The only terrestrial evidence so far found for early meteorite bombardment has been the discovery of tungsten isotope anomalies based on the 182Hf–182W system (the half-life of 182Hf is 9 Myr) in some of the sedimentary rocks from the Isua belt.24 Because tungsten isotope heterogeneities cannot have been preserved in the Earth's dynamic crust–mantle environment from a time when short-lived 182Hf was still preserved, it was concluded that the sediments contain a component derived from the meteorites.
All of this creates an enigma concerning the apparent disappearance of the primordial crust of the Earth, including the large number of impact craters that must once have covered the surface. One may conclude that there must have been an immense basaltic resurfacing produced by global melting and volcanism, which may itself have been initiated by the massive impacts.25 Much more work is required in the oldest terrains, including Isua, to track down exactly what happened on Earth not long (in geological terms) before recognizable geological processes started.
A simplified summary of environments and dated events on the early Earth is presented in figure 5. Although this closely represents my own views, it is pointed out that there is still much controversy about such important matters as the nature and composition of the Hadean crust.
In conclusion, both the nature and environment of the oldest exposed rocks as far back as 3.8 Gyr ago have proved amenable to study by conventional geological approaches. In addition, these oldest rocks have provided a limited amount of significant chemical, isotopic and mineralogical information on the previous 700 Myr of Earth history, from which no bulk rocks survive.
At any rate, we can now begin to compare and contrast the early development of the Earth's surface with that of our planetary neighbours the Moon, Mars and Venus, and contemplate in surprise and amazement their increasingly divergent evolution from early on in their respective histories. Although the planets themselves have grown apart, the sciences of geology and planetology seem to be coming ever closer together!
The early stages of the age and isotope work at Oxford on the Lewisian Complex of northwest Scotland were carried out in particular collaboration with B. J. Giletti, R. St J. Lambert and H. J. Welke, and on the ancient Greenland rocks with R. K. O'Nions, R. J. Pankhurst and P. N. Taylor. I am deeply indebted to them and, in particular, to our Chief Technician Roy Goodwin (1942–2002), who was an indispensable member of the Geological Age and Isotope Research Group for 43 years.
- © 2009 The Royal Society