Kohn and Parkinson (2002) proposed a conceptual model invoking Eocene slab breakoff to explain the generation of Eocene eclogites and Miocene leucogranites in the Himalaya and “late Eocene K-rich magmas” in southeastern Tibet. Kohn and Parkinson further argued that the slab breakoff and its resultant events have no direct implications for Tibetan topography. In comparison with the Himalayan eclogites and leucogranites that have ages and distribution that are both well documented, the occurrence of the late Eocene K-rich magmas in the Lhasa terrane, southeastern Tibet is now found to be problematic. Thus, the slab breakoff model that Kohn and Parkinson (2002) proposed satisfies only Himalayan geology and should be applied with caution when interpreting Tibetan magmatic and tectonic evolution.
Kohn and Parkinson's model is essentially based on two independent observations: (1) the Eocene eclogites from different parts of the Himalaya and (2) the late Eocene K-rich magmas from southeastern Tibet. The latter is solely based on our previous work (Chung et al., 1998), in which we dated 40 K-rich lavas from northeastern Tibet to define a magmatic duration ca. 40–30 Ma. We then correlated this late Eocene K-rich magma suite with its potential counterpart that appears to occur in southeastern Tibet; such a correlation, however, was made using only literature data (Bureau of Geology and Mineral Resources of Xizang Autonomous Region, 1993). Our recent work from southeastern Tibet indicated that igneous rocks there consist of Cretaceous to early Paleogene granitoids (the Gangdese Batholith) and associated volcanics (Lee et al., 2001; Lee et al., 2003). Those coined to be “late Eocene” or younger in the literature are virtually deformed and/or sheared granitoids, which are all confined to the dextral Jiali fault zone. The Gangdese Batholith represents an Andean-type magmatic arc resulting from northward Neo-Tethyan subduction. This arc magmatism ceased ca. 40 Ma. After an ~15 m.y. igneous quiescence period from the late Eocene to Oligocene, ultrapotassic and high-K calc-alkaline lavas formed in the Lhasa terrane between ca. 25 and 10 Ma (Coulon et al., 1986; Miller et al., 1999; Williams et al., 2001). Hence, there are no late Eocene K-rich magmas in southeastern Tibet.
Based on a more comprehensive observation of the tempospatial variations in Tibetan postcollisional magmatism, we propose two geodynamic events to have occurred in southern Tibet. These were break-off of the Neo-Tethyan slab ca. 45 Ma and delamination of thickened Lhasa lithospheric root ca. 25 Ma. The former, which is constrained by onset of the hard collision between India and Asia (Lee and Lawver, 1995), corresponds to the petrologic evidence and timing discussed by Kohn and Parkinson (2002). However, we argue that the slab breakoff did not cause any Eocene K-rich magmas and resulted in substantial uplift in southern Tibet, although southern Tibet did not attain its present elevation until ca. 25 Ma, when thickened Lhasa lithospheric mantle was removed and replaced by the asthenosphere so that the ultrapotassic lavas formed by small degree melting of the remaining, enriched lithospheric mantle (Miller et al., 1999; Williams et al., 2001). The slab breakoff, moreover, may have accounted for the prevailing asthenospheric mantle signature observed in the youngest phase of the Gangdese granitoids, which, as depicted by Davies and von Blanckenburg (1995), is owing to replacement of the oceanic lithosphere by hotter asthenosphere, which could have induced stronger mantle convection and elevated mantle heat flow. Loss of the slab pull would have led to not only rapid exhumation of the Greater Himalayan rocks but also uplift of the entire orogen; i.e., where is now southern Tibet even though a high (3–4 km) but relatively narrow (~300 km) mountain range, similar to the central Andes, may have already existed in the region since Cretaceous time (Fielding, 1996). Voluminous sediments therefore eroded appear to have been deposited not directly in the foreland, but in the Bengal-Ganges-Brahmaputra and Indus-Pakistan sedimentary complexes. In the Ganges-Brahmaputra Delta, for example, sedimentation started to grow rapidly ca. 40 Ma with an increase in sediment flux and development of prograding clastic depositional sequences (Lindsay et al., 1991).
Lastly, we note that in the Lhasa terrane the Miocene high-K calc-alkaline lavas with compositions that show a significant contribution by crustal materials (Miller et al., 1999) are geochemically distinct from coeval leucogranites produced by Greater Himalayan crustal anatexis. This distinction supports the argument by Kohn and Parkinson (2002) that the leucogranites are not the extruded equivalent of modern partial melts in southern Tibet. The latter, if they do exist so extensively (Nelson et al., 1996), highlight a fundamental question of why such melts have rarely extruded along the north-south–striking normal faults widespread in southern Tibet where no magmatism has been identified since ca. 10 Ma.