IBMs Second Arc Jim Gill UCSC Sub Fac

IBM’s Second Arc Jim Gill, UCSC • Sub. Fac has transformative potential. . . • But beware of consensus: Some of what Rummie said he knew wasn’t true.

Logic of IBM Focus Site: Compare and Contrast I vs M • Now: – Izu shallower slab, Marianas steeper – Marianas backarc spreading; Izu rifting – Izu subducts more loess; Mariana more OIB-type vclastic sediment • Before: – SPVB: more melting in Izu backarc – WPB: north has old arcs; south was oceanic crust

Thanks in part to MARGINS… • Lots of data are in databases * Georoc: CIP 1017 entries; Izu Arc 3367 * Petdb: Mariana Trough 1463 * IFREE/Ganseki: ~400 • ~10% are ‘high quality comprehensive data’ for single samples (Majors, ICPMS TE, Sr-Nd. Pb±Hf, Useries) * Izu Arc/Backarc: ~300 * Mariana CIP+NSP: ~100 • Glass/Melt inclusion data including volatiles * Through space (Kent, Shaw, Kelley) * Through time (Straub)

Pre-MARGINS Transformative Consensus A. Three components contribute to magmas • Mantle wedge: I-DMM – Best seen in HREE, HFSE • AOC ‘Fluid’ (Pearce ‘shallow component’) – Best seen in As, Sb, B, U, Ba, Pb • Sediment ‘Melt’ (Pearce ‘deep component’) – Best seen in 10 Be, Th, LREE B. Fluid and Melt are added separately (ternary mixing is the norm)

Why Transformative? • Technical breakthrough: HQCD show elementisotope systematics that were interpreted as implicating sediment melt. • If sediment melts, so does wet-AOC. • Slab melt is everywhere. • If slab melt is everywhere, the old numerical models are wrong. • Informs testable models (e. g. , Kimura, Baker) • The time scale of slab dehydration, flux melting, melt ascent, and differentiation is human.

However, remember Rummie. • The mantle is not uniformly D-IMM. The VF vs RA difference is not just more sediment in RA. • Ba tracks melt as well as fluid; Th is fluid mobile; much Pb comes from the mantle in the RA; accessory minerals matter. • U-Th disequilibria require fast processes but may not require two stages. • Therefore, we still don’t yet understand the steady state forcing functions. • Non-steady-state processes matter and we understand them even less. • How does all this affect crustal evolution? • Sub. Fac needs post-MARGINS transformation.

IBM Vocabulary • Mariana (Along strike) – CIP – NSP • Izu (Across strike) – Volcanic Front – BAK (extensional zone) – WS(reararc seamount chains) WS BAK VF

How well does the current 3 -component consensus explain an iconic figure? F M DMM

Mariana CIP F M DMM+M DMM

Izu VF: Higher Ba/La but lower Ba/Yb; lower La/Yb in Second Arc; no slab melt

Izu WS: 3 -12 Ma; 1 -2% H 2 O in MI (Kent)

Izu BAK: <2 Ma; 0. 5 -1. 5% H 2 O in MI; mostly decompression melts - M DM ed S Izu Little overlap of I and M

Elliott et al 1997’s Tranformative Science: or, Pearce’s talk plus isotopes Analytically challenging, especially between labs. However, also works with Ce/Ce*.

CIP: Elliott Plus Woodhead

Izu VF: flat trend for isotopically depleted mantle: Th not from sed melt; maybe AOC fluid?

Izu WS: another flat trend; not just sediment melt; Isotopically enriched mantle?

Izu BAK: purest DMM ± AOC melts

Analytical improvements; I≠M now; Izu VF≠BAK≠WS; RA more ‘Indian’ going north; CIP between sediment and AOC

Lots of mantle Pb in RA (from sulfide? ) 50% mantle Pb

See Tollstrup poster: I≠M; Izu BAK is least “Indian” because of AOC melt? ; BAK most like First Arc y al r Te r tri s e ra Ar

Rummie summary questions • • • Is there initial along-strike variability in the mantle as well as in recent slab components? (Can we distinguish a ‘deus ex sedimentus’ from enriched mantle without 10 Be? ) How much of this mantle variation is inherited by the crust? How much of the across-strike variability of arc and backarc magmas is due to the percent of flux melting, the nature of the flux, or the nature of the mantle? What aspects of that ‘nature of the flux’ reflects its solute content (fluid vs melt) vs the composition of its source (AOC, sediment) vs the refractory mineralogy of the source (rutile, zircon, monazite, phengite)? How does this relate to variations in P-T? Why are the rear arc melts so similar throughout Mariana, Izu, NEJ when volcanic front melts are so different? What caused the differences between Izu and Mariana to increase after first backarc spreading? Can intra-crustal differentiation create “bulk continental crust” without recycling significant continent-derived sediment?

More consensus: There are consistent spatial and temporal variations in components • DMM mantle is variably depleted – More depleted in VF than RA (trenchward advection even w/o spreading) – More depleted in Izu after SPV backarc spreading • Slab fluid is ubiquitous – Fluid effect decreases from VF to RA because of less %fluid, distillation of fluid source, less depleted mantle, more sediment in RA • Sediment-rich (few%) slab melt is more localized – More in Marianas: least in central CIP (Maug to Guguan) – Less in Izu: absent from VF ± BAK – Little slab sediment melting in Izu since 3 Ma • Quaternary volcanoes track along-strike variations in sediment 206 Pb/204 Pb, Th/La

All IBM

Add Kasuga RA (~NSP)