In a stimulating paper, Collins (2002, p. 535) asserted that “many granulite terrains were too hot to have formed during continental collision” in spite of evidence to the contrary (e.g., Harley, 1989; O'Brien and Rötzler, 2003). Based on this assertion, Collins (2002, p. 535) postulated that “high-grade metamorphic terrains . . . represent thickened hot orogens that . . . formed in another tectonic setting.” The present SW Pacific margin was suggested as a model, and exemplified by the evolution of the Lachlan orogen. In the model, slab rollback drives extension of the overriding plate, backarc basin formation, and adiabatic decompression melting; it is postulated that underplating by basalt causes granulite facies metamorphism. Extension may be interrupted by arrival of a bathymetric high at the trench, which leads to flat subduction and an episode of shortening; as the orogen thickens, it will cool as it becomes isolated from the asthenosphere. A change from extension and heating to shortening and cooling will generate a counterclockwise pressure-temperature (P-T) path. Such a path conflicts with those generated by Sandiford and Powell (1986, their Fig. 2) during simulated extension, who did not propose that “lithospheric extension is . . . a necessity for granulite facies metamorphism” (Collins, 2002, p. 536).
Estimating temperature and pressure in granulites is difficult, due to resetting of Fe-Mg exchange thermometers during cooling, and a simple compilation of data from the literature of the past thirty years is inappropriate. Pattison et al. (2003) have recalculated the P-T conditions of common granulites using garnet-orthopyroxene thermobarometry corrected for retrograde exchange. Based on their compilation, the T-P range of common granulites is 800–900 °C and 0.4–1.0 GPa, respectively. During the past decade, it has become apparent that some granulites record evidence of extreme thermal conditions (ultrahigh-temperature [UHT] granulite metamorphism [e.g., Harley, 1998]), whereas other granulites record evidence of deep burial (high-pressure [HP] granulite metamorphism [e.g., O'Brien and Rötzler, 2003]). The P-T conditions of the former are T >900 °C, with maximum values of perhaps >1150 °C (e.g., Moraes et al., 2002), at P of 0.5–1.5 GPa, and of the latter are T of 750–1050 °C at P of 1.0–2.0 GPa. Under granulite conditions, prograde reaction commonly is complete and melting is ubiquitous, which obliterates evidence of the prograde evolution, with the exception of mineral and fluid inclusions in porphyroblasts.
Most granulites followed a gross clockwise P-T path, although the prograde path is rarely well constrained and the retrograde path may involve decompression with cooling or heating, or a combination of near-isothermal decompression with near-isobaric cooling segments. This type of P-T evolution is characteristic of convergent margin orogens involving collision (e.g., Brown and Dallmeyer, 1996; Brown, 2001), although at least one margin must have involved subduction. Examples of terrains that record evidence of gross counterclockwise P-T paths are more rare, but include the Pikwitonei (e.g., Mezger et al., 1990), the Acadian (e.g., Schumacher et al., 1990), and the Eastern Ghats belt (e.g., Sengupta et al., 1990). Collins (2002) suggested that a return to rollback leads to a repeat of the orogenic cycle outboard. Although narrow belts of orogenic activity younging outboard is a feature of the Lachlan orogen, this is not the pattern of the Variscan (Paleozoic), Brasilia (Neoproterozoic), or Grenville (Mesoproterozoic) orogens that record UHT and HP granulite facies conditions and that are interpreted as having a collisional terminal phase.
Modeling the thermal evolution of collisional orogens has proven challenging, and published models involve simplification of the possible processes (which include ridge-trench interactions, thickening due to magmatic additions, thickening due to collision, internal advection of heat by magmatism, and collapse due to slab breakoff or delamination). It is the combination of processes in these orogens that makes modeling their thermal evolution challenging, and the failure of conventional models to achieve the peak T recorded is no surprise.
The model proposed by Collins (2002) is inconsistent with the bulk of the data; it does not explain the range of P-T conditions associated with UHT and HP granulite metamorphism, or the clockwise P-T evolution that is characteristic of most granulites. Without undermining the innovative nature of the model proposed by Collins (2002) for the evolution of the Lachlan orogen, or its possible application to other orogens with broadly similar features, such as the Acadian orogen, broader application of the model to explain most granulites is inappropriate, and the model fails to explain even the first-order features of many granulite terrains.