Shortening budgets and the role of continental subduction during the India–Asia collision

The India–Asia collision continues to attract the attention of earth scientists from many disciplines concerned with all levels upwards from the deep structure to the atmosphere above the Himalaya and its Hinterland (Fig. 1). Models devised for the origin of crustal thickening in the Himalaya and SE Asia (Tibet, Tien Shan, etc.; Fig. 2) fall into two main categories (Fig. 3): (1) those in which the Indian lithosphere is confined to the south of the Tsangbo suture within the Himalaya, and (2) those in which the Indian lithosphere—either mantle or crust or both mantle and crust—extends to the north of the suture so as to lie under part or all of Tibet. This is an oversimplification because few would now maintain that there has been no underthrusting by Indian lithosphere. Rather the controversy centres on how much underthrusting has occurred. Within each category, there is great diversity of mechanism. Examples of the first category are:

(a) the indentation model, with India acting as a rigid indenter, resulting in continental expulsion (Molnar and Tapponnier, 1975) or homogenous thickening in Tibet and areas to the north Dewey et al., 1988, England and Houseman, 1988, Platt and England, 1993;

(b) mantle delamination and roll-back (Willett and Beaumount, 1994);

(c) migratory subduction of continental lithosphere (Meyer et al., 1998).

Examples of the second category are (a) underthrusting of India beneath all of Tibet Argand, 1924, Powell, 1986, Powell and Conaghan, 1973, Beghoul et al., 1993, Matte et al., 1997; (b) injection of Indian crust into Tibetan crust, either in the form of a solid ‘piston’ (Zhou and Morgan, 1995) or as a viscous fluid (Westerway, 1995). Zhou and Morgan's hypothesis seemingly begs the question: in order to provide the ‘weak’ lower crust needed for injection in Tibet presumably it is first necessary to achieve major crustal thickening, a feature which they are attempting to explain! However, the weakening could have been provided by the injection of magma by Andean-type granite intrusion, at the base of the Tibetan crust thereby increasing the temperature there.

An Yin and Harrison (2000) have given a recent comprehensive review of the geological evolution of the Himalayan–Tibetan orogen. The present review considers the evidence for the amount and the rate of the convergence between India and Asia taking place after initial collision of these continents at roughly 55–50 Ma. This topic is still highly controversial and the principal theme of this review is to reappraise the landmark papers of Dewey et al., 1988, Dewey et al., 1989 in the light of later developments. Evidence for the total convergence comes from magnetic anomaly, palaeomagnetic and volumetric balancing studies. The results are consistent in showing about 1800–2100 km convergence between India and Asia in the western sector, 2475 km in the central and 2750–2800 km in the eastern sector. The rate of convergence between India and Asia over the period since initial collision is roughly 5 cm/year. However the possible long term rate of convergence within the Himalaya is considerably less, i.e. 1.5–2.0 cm/year, a fraction of the total convergence.

The shortening budget

How much shortening has taken place across the orogen (Besse et al., 1984)? Patriat and Achache (1984) and Klootwijk et al. (1992) considered that the time available for the shortening of Indo-Eurasian lithosphere since the completion of contact between the continents was 55–45 ma. Patriat and Achache (1984) suggested that 700±300 km of shortening had taken place over this time-span but this was based on an imprecise estimation of the latitude of the south margin of Asia at the time of

Rates of convergence

Earlier reference was made to estimates of shortening rates across the Himalaya. The oft-quoted rate is 15–20 mm/year based on the analysis of on-lap in the foredeep (Lyon-Caen and Molnar, 1985). Fortunately, their calculations are supported by other workers using other lines of evidence (Yeats and Thakur, 1998; Table 2). In addition, Avouac et al. (1998) and Wesnousky et al. (1999) quoted rates of 21±1 and 13.8±3.6 mm/year, respectively, for slip on the active Himalayan Main Frontal Thrust.

Subduction of Indian lithosphere

Dewey et al. (1988) gave several arguments against the under thrusting of India. For example, they cited the preservation of the Mesozoic succession on the Indian continental margin. However, this neglects the possibility that the Indian upper/middle crust has been detached or scraped-off from the lower part of the Indian lithosphere along the surface later taken over by the South Tibetan Detachment. Another possibility is that the Mesozoic margin rocks were involved in a rebound in the manner

What are the tests?

(1) Molnar (1988) showed that the P-wave and S-wave velocities in the uppermost mantle under most of Tibet are relatively high and typical of those of Precambrian shields. However, travel times and waveforms of S-waves passing through the uppermost mantle of much of Tibet require a much lower average velocity than in shield mantle. As he points out, this does not necessarily disprove the existence of Indian mantle there because shield mantle may have been modified during the collision process.

How much subduction is required?

Table 3 shows the amount of subduction that could be achieved at various rates and assuming that subduction operated since the time of initial collision. Rates of convergence used are: overall rate for India–Asia convergence (top) and upper and lower limits for internal convergence given by Lyon-Caen and Molnar (1985). Two thousand five hundred kilometers can be ruled out because it would involve lithospheric thickening on the scale envisaged by Argand (1924). Also it fails to explain why the

Partitioning of compressive strain in the Himalaya and SE Asia

Dewey et al., 1988, Dewey et al., 1989 suggested a partitioning of the shortening by “thrust regimes” following the India–Asia collision as follows: in the east/central Himalaya: about 1000 km; in Tibet ca. 50% shortening, that is: 660 km in central Tibet and 960 km in eastern Tibet. According to them, relatively small-scale wrench fault regimes make up the total convergence (2475 km, Central and 2750 km, East). Two problems with this budget are:

(1) the shortening in the Himalaya appears be

Crustal thickening: mechanisms, timing and partitioning—a new consensus?

If there is a problem with previous shortening budgets then solutions other than those involving major homogenous thickening of Tibetan lithosphere must be entertained, e.g. loss of crust downwards or upwards or sideways. Let us consider underthrusting of Indian lithosphere first. Fig. 6 shows that volume balancing requires such underthrusting if the basal detachment, including the Main Central thrust, takes place in mid-crustal levels. On this diagram the displacements on thrusts are uncertain

Lateral expulsion

Turning now to the possibility that lateral expulsion has taken place on a much larger scale than was held by Dewey et al., 1988, Dewey et al., 1989. Peltzer and Tapponnier (1988) presented evidence for large slip on the Altyn Tagh and Red River faults but the amount of slip on strike–slip faults in SE Asia is still controversial. Harrison et al. (1998) gave a full discussion of the evidence including the following estimates for displacement on strike–slip faults in Tibet and elsewhere (see

Indentation

Finally a comment on the indentation model first put forward by Molnar and Tapponnier (1975). If the Himalayan detachment continues under Tibet as suggested by Owens and Zandt (1997) then the concept of Indian lithosphere acting as a rigid indenter is unrealistic. Indian middle/upper crust cannot have been rigid throughout the Cenozoic—it failed by thrusting and it was weakened by high-grade metamorphism in this time. Furthermore, if all or part of the Indian lithosphere has underthrust Asia

Conclusions

(1) Models requiring major underthrusting of Tibet by Indian lithosphere are difficult to test. Equally difficult is discrimination between underthrusting models that advocate entirely low angled subduction and those proposing rotation of the underthrust slab and thus underplating of Tibet. However the volumetric evidence suggests that only mid/upper crust appears to be conserved in the Himalayan thrust sheets while Indian lower crust and mantle has been detached along successive roughly

Acknowledgements

My special thanks are due to Alastair Robertson and Mike Searle for encouragement, advice and information. In addition, I am grateful to Grahame Oliver for helpful comments and references. I am grateful to Giuliano Panza for directing me to geophysical data on the nature of the Indian lithosphere. Of course all errors of fact and interpretation are my responsibility.