Hofmann on Recycling - outline

Historical Perspective
Continental recycling: Armstrong (1968), DePaolo (1983):
Early Plate tectonic framework:
Steady state creation and destruction of continental crust by arc magmatism and sediment subduction.
Isotope evolution models explain Sr-Nd-206Pb/204Pb evolution, but inconsistent with 207Pb/204Pb-206Pb/204Pb data of oceanic basalts (modeled slopes always too steep).

Ocean crust recycling: Hofmann & White (1980,1982), Chase (1981): Specifically designed to explain OIB magmatism in general. Difficulties when highly depleted “N-MORBs” are recycled (maybe insufficient number of depleted OIBs).

Altered oceanic crust recycling: Staudigel et al. (1995): Rule out severely altered crust, because alteration effects would generate bizarre isotope patterns after 1 Ga recycling time.

Sediment recycling to explain high-⁸⁷Sr/⁸⁶Sr in some OIBs (now called EM-2): Hawkesworth et al. (1979); White & Hofmann (1982).

Recycling of delaminated subcontinental lithosphere: McKenzie & O’Nions, (1983): Intended to explain all hotspot magmatism. No longer relevant even in McKenzie’s thinking.

Recycling of delaminated subcontinental lithosphere to explain EM-1 OIBs: Possible explanation for low e(Nd) e(Hf) e(Sr) ²⁰⁶Pb/²⁰⁴Pb of EM-1s.

Recycling of pelagic sediments to explain EM-1:
OIB: Weaver (1991), Chauvel et al. (1992)
Indian Ocean MORB: Rehkämper & Hofmann (1997)

Recycling of terrigenous sediments to explain EM-2:
(see above)

Recycling of oceanic gabbros: Hofmann & Jochum (1996) to explain high Sr, low Th-U concentrations in Mauna Loa basalts.
Sobolev et al. (2000): Confirmation by much clearer gabbroic fingerprint in rare melt inclusions from Mauna Loa.

Mantle Metasomatism (1): Popular at least since Hanson (1977, “veined mantle”) to explain trace element enrichments in alkalic OIBs. Originally conceived as mechanism for enriching the lithosphere and extracting OIBs from lithosphere. Now mostly out of fashion because of apparent inconsistency with isotope data and popularity of plume model.

Mantle Metasomatism (2): Recycling of metasomatized lithosphere: Sun & McDonough (1989) low-degree melts from asthenosphere infiltrate lithosphere, causing primary enrichment. Recycling of enriched lithosphere to account for “isotopic aging”, specifically to explain HIMU basalts.
Workman & Hart (2003): Same mechanism to explain EM-2 (Samoa).
Have all conceivable reservoirs been exhausted? Probably!

Effect of convective stirring

Early modeling indicated rapid mixing. Therefore, thin oceanic crust would not survive as recycled reservoir:

Biggest objection to crustal recycling models
(and one of the rationales for lithospheric delamination model).

Apparently no longer a problem: Why?

Segregation model of Christensen & Hofmann, (1996), uses density and viscosity differences in lower boundary layer to store dense material.

New 3-D stirring models preserve some unstirred volumes

Degree of homogeneity depends on scale of sampling (“inspection”). Solid state diffusion so slow (centimeter scale) that heterogeneities become physically smaller, but remain chemically heterogeneous:

Detailed sampling of single volcanoes (e.g. HSDP hole on Mauna Kea) and
Melt inclusion studies reveal small-scale heterogeneities.

64k question: How small are the heterogeneities??? > 10 cm, < 5 km.
Melts capable of carrying samples of these heterogeneities into the crust.
Apparent homogeneity of magmas caused by magma chamber mixing.

No real homogenization due to stirring:

1. Basalts may sample EVERYTHING that survives subduction. Homogeneity result of magma mixing.
   
2. (Almost) all oceanic basalts (MORB & OIB) come from “processed” mantle (Sr-Nd-Hf isotopes; almost universally radiogenic Pb isotopes (lead paradox).
   
3. (Almost) all known differentiation processes occur within the upper 100 km of the mantle
   

1. Ridge melting
   
2. Sedimentation
   
3. Subduction-related dehydration & melting
   
4. Continental differentiation processes
   

1. Probably “see” all recycled rock types in OIBs or in MORBs, but may not always recognize them, because of subduction modification
   
2. Why so few EM-type OIBs with sediment or other continental signatures??
   

1. Why so little evidence of severely altered recycled material? Subduction must “clean it out” (see Terry)
   
2. Any primitive mantle in OIBs? Not much, if any. Except for noble gases. Almost all oceanic basalts come from processed mantle (required by Pb, Sr, Nd, Hf isotopes and trace elements). Primordial noble gases either survived in otherwise processed mantle (see Tackley), or they are stored in a separate reservoir, which may be somewhere in the lower mantle, D’’, or the core. Solubility data for noble gases in silicates and metals needed. I (still) favor separate reservoir.
   
3. Alternatives to recycling crust-lithosphere?
   

1. Melt inclusions show chemically and isotopically much more heterogeneous compositions than bulk samples. Full meaning of this not yet understood. Probably indicates sampling of very small-scale heterogeneities that survived convective stirring. “Small” may mean scales between 10 cm and 5 km. Need better understanding of melt extraction and transport processes.
   
2. Recognizing the specific rock types of recycled material useful for understanding (a) overall mantle evolution, (b) understanding convective stirring in the mantle, (c) history of specific mantle regions (e.g. differences between Indian Ocean and Pacific). These objectives cannot be achieved without strong input from high resolution mantle tomography, 3-D mantle dynamics, and much fuller understanding of core-mantle relationships.