Research sections:

Granulite sections:

Origin of Granulites

An annotated reading list


A problem in dealing with this topic in a concise way is that granulite terrains vary widely in character, and granulite-facies rocks present many different sorts of phenomena to be explained. The basic petrological problem is that typical granulite-facies assemblages require PH2O << Ptotal for their stability, and debate has centred around how this low PH2O is achieved. There are three distinguishable models:

  1. Melting occurs, water from intergranular fluid and hydrous phases is partitioned into the melt, and the melt may be removed, taking the water with it.
  2. CO2-rich vapour from some external source passes through the rocks, driving dehydration reactions and diluting or flushing out the resulting aqueous fluid.
  3. In certain terrains the precursor rocks were already rather dry (e.g. water-deficient magmatic rocks and their dehydrated thermal aureoles) so that their subsequent high-grade metamorphism does not require a special mechanism to generate granulite-facies assemblages.

Other features of granulite terrains which need to be explained are:

  1. Depletion in large-ion lithophile elements (LILE).

  2. This was regarded as a major issue in the 1970's when the focus was on the Lewisian gneisses, and it was ascribed either to melt extraction or to deep-seated fluid metasomatism. It is now realised that the Lewisian is atypical, being exceptionally depleted in LILE, whereas many granulite terrains show rather small depletions.
  3. CO2-rich fluid inclusions.

  4. A classic study by Touret in southern Norway established that a change from water-rich to CO2-rich fluid inclusions coincided with the regional orthopyroxene isograd. This has been found to be generally true, though granulite terrains form a spectrum from those in which fluid inclusions are ubiquitous and abundant to those in which they are rare and small.
  5. "Arrested charnockitisation"

  6. A strange phenomenon first described from south India and Sri Lanka, where in acid to intermediate rocks granulite-facies assemblages (charnockite) are found in vein-like arrays superimposed over amphibolite-facies gneisses. These features are now known, on a less conspicuous scale, from a number of other terrains. They have been attributed to CO2 streaming. Other explanations may be possible, but at the very least this phenomenon draws attention to the fact that the amphibolite-granulite transition cannot depend on changes in T or P alone.

Model 1: Melting and melt extraction

The original reference given for this model is usually Fyfe (1973), who suggested that granulites could be regarded as restites after removal of relatively large amounts of partial melt. The model has been suggested for the Lewisian Complex (see above).

A substantial boost for the model came from the realisation that dehydration melting, i.e. vapour-absent incongruent melting of assemblages containing hydrous phases (Ms, Bt, Hbl) is likely to be an important process in high-grade metamorphism. Theoretical treatments were made by Thompson (1982), Powell (1983) and Grant (1985).

Dehydration melting was applied to natural examples in Broken Hill, Australia (Phillips, 1980), New England (Tracy and Robinson, 1983) and Namaqualand, S. Africa (Waters and Whales, 1984; Waters, 1988). Notice, however, that these more recent accounts don't necessarily envisage the extraction of large volumes of melt.

Model 2: Influx of CO2-rich vapour

The type locality for this model is in south India, at classic exposures in stone quarries, e.g. at Kabbaldurga in the Late Archaean amphibolite-granulite transition zone, and further south in the Kerala "khondalite belt" where the phenomena are probably of Pan-African age (550 Ma). The model has been championed by R.C. Newton at Chicago.

Model 3: Metamorphism of dry precursors

The type area for this model is the Adirondack Highlands of upstate New York, where much of the terrain consisted of pre-orogenic pyroxene-bearing igneous rocks and their high-grade thermal aureoles, subsequently metamorphosed to give Grenville age (1000 Ma) granulite-facies assemblages.

These three models are not mutually exclusive, and some authors have suggested interesting scenarios which allow elements of each model to be combined (e.g. Frost and Frost, 1987).

P-T paths and tectonic setting

Again, a historical view is helpful in understanding the nature of the debate. By the end of the 1970's it was assumed that metamorphic rocks in orogenic belts typically experienced a clockwise P-T-time path (perhaps as a result of a casual reading of England and Richardson, 1977, J. Geol. Soc. London, 134, 201-213, a seminal paper of the time, although this paper actually suggested that high-T metamorphic belts might well not do this). So, when Ellis (1980) described an isobaric cooling history from Enderby Land he regarded it as unusual and exceptional. However, many more examples followed, and a variety of models were suggested to account for this type of retrograde path:

  1. Cooling at depth in the lower parts of thickened crust (Ellis, 1987)
  2. After metamorphism in extensional terrains (Sandiford and Powell, 1986)
  3. In terrains heated by the addition of voluminous magmas (Bohlen, 1987, see below)

In some examples where parts of the prograde history could be detected, anticlockwise P-T paths were documented. In many cases, these were in terrains where magmatic accretion had played an important role in crustal thickening and heat advection. Some examples of models or applications are:

There seems now to be a broader recognition of the variety of P-T paths and tectonic settings to be found in granulite-facies terrains. The characteristic features which shed light on their P-T history are reviewed by

Created October 1994