1. Chapter 14: Ocean Intraplate Volcanism
Ocean islands and seamounts
More enigmatic processes and less voluminous than activity at plate margins
No obvious mechanisms that we can tie to the plate tectonic paradigm
As with MORB, the dominant magma type for oceanic intraplate volcanism is basalt, which is commonly called ocean island basalt or OIB
41 well-established hot spots Estimates range from 16 to 122
Figure Map of relatively well-established hotspots and selected hotspot trails (island chains or aseismic ridges). Hotspots and trails from Crough (1983) with selected more recent hotspots from Anderson and Schramm (2005). Also shown are the geoid anomaly contours of Crough and Jurdy (1980, in meters). Note the preponderance of hotspots in the two major geoid highs (superswells).

2. Ocean islands and seamounts Commonly associated with hotspots

3. Plume
Figure 14.2  Photograph of a laboratory thermal plume of heated dyed fluid rising buoyantly through a colorless fluid. Note the enlarged plume head, narrow plume tail, and vortex containing entrained colorless fluid of the surroundings. After Campbell (1998) and Griffiths and Campbell (1990).

4. Two principal magma series
Types of OIB Magmas
Two principal magma series
Tholeiitic (dominant type)
Parent: ocean island tholeiitic basalt (OIT)
Similar to MORB, but some distinct chemical and mineralogical differences
Alkaline series (subordinate)
Parent: ocean island alkaline basalt (OIA)
Two principal alkaline sub-series
Silica undersaturated
Slightly silica oversaturated (less common)
Modern volcanic activity of some islands is dominantly tholeiitic (for example Hawaii and Réunion), while other islands are more alkaline in character (for example Tahiti in the Pacific and a concentration of islands in the Atlantic, including the Canary Islands, the Azores, Ascension, Tristan da Cunha, and Gough)

5. Cyclic, pattern to the eruptive history
Hawaiian Scenario
Cyclic, pattern to the eruptive history
1. Pre-shield-building stage (variable)
2. Shield-building stage begins with tremendous outpourings of tholeiitic basalts
3. Waning activity more alkaline, episodic, diverse, and violent (Mauna Kea, Hualalai, and Kohala).
4. A long period of dormancy, followed by a late, post-erosional stage. Characterized by highly alkaline and silica-undersaturated magmas, including alkali basalts, nephelinites, melilite basalts, and basanites
1) Early, pre-shield-building stage that is more alkaline and variable, but quickly covered by the massive tholeiitic shields
Recent studies of the Loihi Seamount encountered a surprising assortment of lava types from tholeiite to highly alkaline basanites.
2) Shield-building: Kilauea and Mauna Loa (the two nearest the hot spot in the southern and southeastern part of the island) -> 98-99% of the lava in Hawaii
3) Lavas are also more diverse, with a larger proportion of differentiated liquids
The two late alkaline stages represent 1-2% of the total lava output
Note all three OIB series are represented in Hawaii
Is this representative of all islands? Probably not

6. Evolution in the Series
Tholeiitic, alkaline, and highly alkaline
Figure After Wilson (1989) Igneous Petrogenesis. Kluwer.

7. Alkalinity is highly variable
Alkalis are incompatible elements, unaffected by less than 50% shallow fractional crystallization, this again argues for distinct mantle sources or generating mechanisms
The variation in Na/K among the suites makes the possibility much more likely, and leads us to suspect that the mantle is more heterogeneous than we had previously thought

8. Thus all appear to have distinctive sources
Trace Elements
The LIL trace elements (K, Rb, Cs, Ba, Pb2+ and Sr) are incompatible and are all enriched in OIB magmas with respect to MORBs
The ratios of incompatible elements have been employed to distinguish between source reservoirs
N-MORB: the K/Ba ratio is high (usually > 100)
E-MORB: the K/Ba ratio is in the mid 30’s
OITs range from 25-40, and OIAs in the upper 20’s
Thus all appear to have distinctive sources

9. Trace Elements
HFS elements (Th, U, Ce, Zr, Hf, Nb, Ta, and Ti) are also incompatible, and are enriched in OIBs > MORBs
Ratios of these elements are also used to distinguish mantle sources
The Zr/Nb ratio
N-MORBs are generally quite high (>30)
OIBs are low (<10)

10. Trace Elements: REEs
Note that ocean island tholeiites (represented by the Kilauea and Mauna Loa samples) overlap with MORB and are not unlike E-MORB
The alkaline basalts have steeper slopes and greater LREE enrichment, although some fall within the upper MORB field
Note also that the heavy REEs are also fractionated in the OIB samples (as compared to the flat HREE patterns in N- and E-MORB). This indicates that garnet was a residual phase
These melts must have segregated from the mantle at depths > 60 km
Figure After Wilson (1989) Igneous Petrogenesis. Kluwer.

11. Trace Elements: REEs
La/Yb (REE slope) correlates with the degree of silica undersaturation in OIBs
Highly undersaturated magmas: La/Yb > 30
OIA: closer to 12
OIT: ~ 4
(+) slopes  E-MORB and all OIBs  N-MORB (-) slope and appear to originate in the lower enriched mantle
In Chapter 10, I argued that MORB tholeiites probably originated in the depleted upper mantle, and alkali basalts in the enriched lower mantle reservoir. Now it appears that E-MORB and ocean island tholeiites also have a source in the deeper reservoir (high % PM can still  tholeiite from an enriched source)

12. MORB-normalized Spider Diagrams
OIBs are enriched in incompatible elements over MORB (values > one)
Broad central hump in which both the LIL (Sr-Ba) and HFS (Yb-Th) element enrichments increase with increasing incompatibility (inward toward Ba and Th)
The pattern we would expect in a sample that was enriched by some single-stage process (such as partial melting of a four-phase lherzolite) that preferentially concentrated incompatible elements.
The pattern is regarded as typical of melts generated from non-depleted mantle in intraplate settings
Figure Winter (2001) An Introduction to Igneous and Metamorphic Petrology. Prentice Hall. Data from Sun and McDonough (1989).

13. Nb/U vs. Nb
Nb/U shouldn’t vary with PM or FX (both highly incompatible)
This diagram suggests most OIB & MORB are derived from a source with non-chondritic Nb/U (higher than chondrite)
If so, very little primitive mantle source left!!
Figure Nb/U ratios vs. Nb concentration in fresh glasses of both MORBs and OIBs. The Nb/U ratio is impressively constant over a range of Nb concentrations spanning over three orders of magnitude (increasing enrichment should correlate with higher Nb). From Hofmann (2003). Chondrite and continental crust values from Hofmann et al. (1986).

14. Isotope Geochemistry
Isotopes do not fractionate during partial melting of fractional melting processes, so will reflect the characteristics of the source
OIBs, which sample a great expanse of oceanic mantle in places where crustal contamination is minimal, provide incomparable evidence as to the nature of the mantle

15. Simple Mixing Models Ternary Binary
All analyses fall between two reservoirs as magmas mix
Ternary
All analyses fall within triangle determined by three reservoirs
Figure Winter (2001) An Introduction to Igneous and Metamorphic Petrology. Prentice Hall.

16. Sr - Nd Isotopes
High values of 87Sr/86Sr and low values of 144Nd/143Nd are associated with mantle enrichment
OIBs show considerably more variation than MORB as seen on next frame
Figure Data from Ito et al. (1987) Chemical Geology, 62, ; and LeRoex et al. (1983) J. Petrol., 24,

17. m
Figure After Zindler and Hart (1986), Staudigel et al. (1984), Hamelin et al. (1986) and Wilson (1989).

18. Mantle Reservoirs 1. DM (Depleted Mantle) = N-MORB source
Shows low values of 87Sr/86Sr and high values of 144Nd/143Nd as well as depleted trace element characteristics
Figure After Zindler and Hart (1986), Staudigel et al. (1984), Hamelin et al. (1986) and Wilson (1989).

19. 2. BSE (Bulk Silicate Earth) or the Primary Uniform Reservoir
Reflects the isotopic signature of the primitive mantle as it would evolve to the present without any subsequent fractionation i.e. neither depleted nor enriched…just plain old mantle
Several oceanic basalts have this isotopic signature, but there are no compelling data that require this reservoir (it is not a mixing end-member), but falls within the space defined by other reservoirs
Figure After Zindler and Hart (1986), Staudigel et al. (1984), Hamelin et al. (1986) and Wilson (1989).

20. 3. EMI = enriched mantle type I has lower 87Sr/86Sr (near primordial)
4. EMII = enriched mantle type II has higher 87Sr/86Sr
(> 0.720), well above any reasonable mantle sources
Since the Nd-Sr data for OIBs extends beyond the primitive values to truly enriched ratios, there must exist an ENRICHED mantle reservoir
Both EM reservoirs have similar enriched (low) Nd ratios (< )
Figure After Zindler and Hart (1986), Staudigel et al. (1984), Hamelin et al. (1986) and Wilson (1989).

21. 5. PREMA (PREvalent MAntle)
Also not a mixing end-member
PREMA represents another restricted isotopic range that is very common in ocean volcanic rocks
Although it lies on the mantle array, and could result from mixing of melts from DM and BSE sources, the promiscuity of melts with the PRIMA signature suggests that it may be a distinct mantle source
Figure After Zindler and Hart (1986), Staudigel et al. (1984), Hamelin et al. (1986) and Wilson (1989).

22. Note that all of the Nd-Sr data can be reconciled with mixing of three reservoirs: DM EMI and EMII since the data are confined to a triangle with apices corresponding to these three components. So, what is the nature of EMI and EMII, and why is there yet a 6th reservoir (HIMU) that seems little different than the mantle array?
Figure After Zindler and Hart (1986), Staudigel et al. (1984), Hamelin et al. (1986) and Wilson (1989).

23. Pb Isotopes Pb produced by radioactive decay of U & Th
Eq U  234U  206Pb
Eq U  207Pb
Eq Th  208Pb
All three elements are LIL
Fractionate into the melt (or a fluid) phase (if available) in the mantle, and migrate upward where they will become incorporated in the oceanic or continental crust

24. Pb is quite scarce in the mantle
Low-Pb mantle-derived melts susceptible to Pb contamination
U, Pb, and Th are concentrated in continental crust (high radiogenic daughter Pb isotopes)
204Pb non-radiogenic: 208Pb/204Pb, 207Pb/204Pb, and 206Pb/204Pb increase as U and Th decay
Oceanic crust also has elevated U and Th content (compared to the mantle)
Sediments derived from oceanic and continental crust
Pb is a sensitive measure of crustal (including sediment) components in mantle isotopic systems
93.7% of natural U is 238U, so 206Pb/204Pb will be most sensitive to a crustal-enriched component
Mantle-derived melts are susceptible to contamination from U-Th-Pb-rich reservoirs which can add a significant proportion to the total Pb
U, Pb, and Th are concentrated in sialic reservoirs, such as the continental crust, which develop high concentrations of the radiogenic daughter Pb isotopes
204Pb is non-radiogenic, so 208Pb/204Pb, 207Pb/204Pb, and 206Pb/204Pb increase as U and Th decay
Oceanic crust has elevated U and Th content (compared to the mantle) as will sediments derived from oceanic and continental crust
Pb is perhaps the most sensitive measure of crustal (including sediment) components in mantle isotopic systems
Since 93.7% of natural U is 238U, equation 9-20 will dominate over equation 9-21, and thus 206Pb/204Pb will be most sensitive to a crustal-enriched component
U  234U  206Pb
U  207Pb
Th  208Pb

25. Figure 14-7. After Wilson (1989) Igneous Petrogenesis. Kluwer.
207Pb/204Pb vs. 206Pb/204Pb data for Atlantic and Pacific ocean basalts
Geochron = simultaneous evolution of 206Pb and 207Pb in a rock/reservoir
= line on which all modern single-stage (not disturbed or reset) Pb isotopic systems, such as BSE, should plot.
~ NONE of the oceanic volcanics fall on the geochron. Nor do they fall within the EMI-EMII-DM triangle, as they appear to do in the Nd-Sr systems
The remaining mantle reservoir: HIMU (high mu) proposed to account for this great radiogenic Pb enrichment pattern

26. m = 238U/204Pb (evaluate uranium enrichment)
HIMU reservoir: very high 206Pb/204Pb ratio
Source with high U, yet not enriched in Rb (modest 87Sr/86Sr)
Old enough (> 1 Ga) to ® observed isotopic ratios
HIMU models:
Subducted and recycled oceanic crust (± seawater)
Localized mantle lead loss to the core
Pb-Rb removal by those dependable (but difficult to document) metasomatic fluids
The similarity of the rocks from St. Helena Island to the HIMU reservoir has led some workers to call this reservoir the “St. Helena component”

27. EMI and EMII High 87Sr/86Sr require high Rb & long time to ® 87Sr
Correlates with continental crust (or sediments derived from it)
Oceanic crust and sediment are other likely candidates

28. Figure 14. 10 After Wilson (1989) Igneous Petrogenesis. Kluwer
Figure After Wilson (1989) Igneous Petrogenesis. Kluwer. Data from Hamelin and Allègre (1985), Hart (1984), Vidal et al. (1984).
Note that HIMU is also 208Pb enriched, which tells us that this reservoir is enriched in Th as well as U
This highly enriched EM component has been called the DUPAL component, named for Dupré and Allègre (1983), who first described it
207Pb/204Pb data (especially from the N hemisphere) ® ~linear mixing line between DM and HIMU, a line called the Northern Hemisphere Reference Line (NHRL)
Data from the southern hemisphere (particularly Indian Ocean) departs from this line, and appears to include a larger EM component (probably EMII)

29. Other isotopic systems that contribute to our understanding of mantle reservoirs and dynamics
He Isotopes
Noble gases are inert and volatile
4He is an alpha particle, produced principally by a-decay of U and Th, enriching primordial 4He
3He is largely primordial (constant)
The mantle is continually degassing and He lost (cannot recycle back)
4He enrichment expressed as R = (3He/4He)
unusual among isotopes in that radiogenic is the denominator
Common reference is RA (air) = 1.39 x 10-6

30. He Isotopes
N-MORB is fairly uniform at 8±1 RA suggesting an extensive depleted (degassed) DM- type N-MORB source
Figure 14.12  3He/4He isotope ratios in ocean island basalts and their relation to He concentration. Concentrations of 3He are in cm3 at 1 atm and 298K.After Sarda and Graham (1990) and Farley and Neroda (1998).

31. He Isotopes
OIB 3He/4He values extend to both higher and lower values than N- MORBs, but are typically higher (low 4He).
Simplest explanations:
High R/RA is deeper mantle with more primordial signature
OIB 3He/4He values extend to both higher and lower values than N-MORBs, but are typically higher.
Fresh glasses (yellow) are generally required to avoid contamination.
Some mineral inclusions, either fluid or melt/glass (other symbols), may also be good.
Low R/RA has higher 4He due to recycled (EM-type?) U and Th.
Figure 14.12  3He/4He isotope ratios in ocean island basalts and their relation to He concentration. Concentrations of 3He are in cm3 at 1 atm and 298K.After Sarda and Graham (1990) and Farley and Neroda (1998).

32. He Isotopes
PHEM (primitive helium mantle) is a hi-3He/4He mantle end-member reservoir with near-primitive Sr-Nd-Pb characteristics.
Figure 14.13  3He/4He vs. a. 87Sr/86Sr and b. 206Pb/204Pb for several OIB localities and MORB. The spread in the diagrams are most simply explained by mixing between four mantle components: DM, EMII, HIMU, and PHEM. After Farley et al. (1992).

33. He Isotopes
Summary:
Shallow mantle MORB source is relatively homogeneous and depleted in He
Deeper mantle has more primordial (high) 3He/4He, but still degassed and less than primordial ( RA) values
PHEM may be that more primitive reservoir
Low 3He/4He may be due to recycled crustal U and Th

34. Re/Os system and Os Isotopes
Both are platinum group elements (PGEs) and highly siderophile (→ core or sulfides)
Mantle values of (187Os/188Os) are near chondritic (~0.13)
Os is compatible during mantle partial melting (→ trace sulfides), but Re is moderately incompatible (→ melts and silicates)
The mantle is thus enriched in Os relative to crustal rocks and crustal rocks (higher Re and lower Os) develop a high (187Os/188Os) which should show up if crustal rocks are recycled back into the mantle.

35. Re/Os system and Os Isotopes
All of the basalt provinces are enriched in 187Os over the values in mantle peridotites and require more than one 187Os-enriched reservoir to explain the distribution.
Volcanics are high (with little overlap to peridotites).
High Re indicates a crustal component.
Peridotites are low (low Re)
Figure 14.13  187Os/188Os vs. 206Pb/204Pb for mantle peridotites and several oceanic basalt provinces. Os values for the various mantle isotopic reservoirs are estimates. After Hauri (2002) and van Keken et al. (2002b).

36. O Isotopes (stable)
Sufficiently light to mass fractionate during several geologic processes
Fractionation during melting, crystallization, and gas exsolution is minor, so most strictly silicate systems cluster at d18O ≈ 5.5 ± 0.2‰.
d18O of MORBs reach 6‰ and OIBs up to 7‰ or more.
The change is small, but higher values correlate with trace element and Sr-Nd-Pb- Os values indicative of enriched sources.
d18O of near-surface waters (and sediments equilibrated with such waters) range from 8 to 25‰.
High d18O in mantle systems is most readily explained by contamination by material affected by surface waters.

37. Other Mantle Reservoirs
FOZO (focal zone): another “convergence” reservoir toward which many trends approach. Thus perhaps a common mixing end-member
FOZO
Figure After Hart et al., 1992).

38. Mantle Reservoirs
FOZO

39. EMI, EMII, and HIMU: too enriched for any known mantle process
EMI, EMII, and HIMU: too enriched for any known mantle process...must correspond to crustal rocks and/or sediments
EMI
Slightly enriched
Deeper continental crust or oceanic crust
EMII
More enriched
Specially in 87Sr (Rb parent) and Pb (U/Th parents)
Upper continental crust or ocean-island crust
If the EM and HIMU = continental crust (or older oceanic crust and sediments), only ® deeper mantle by subduction and recycling
To remain isotopically distinct: could not have rehomogenized or re-equilibrated with rest of mantle

40. The Nature of the Mantle
N-MORBs involve shallow melting of passively rising upper mantle
→ a significant volume of depleted upper mantle (lost lithophile elements and considerable He and other noble gases).
OIBs typically originate from deeper levels.
Major- and trace-element data → the deep source of OIB magmas (both tholeiitic and alkaline) is distinct from that of N-MORB.
Trace element and isotopic data reinforce this notion and further indicate that the deeper mantle is relatively heterogeneous and complex, consisting of several domains of contrasting composition and origin. In addition to the depleted MORB mantle, there are at least four enriched components, including one or more containing recycled crustal and/or sedimentary material reintroduced into the mantle by subduction, and at least one (FOZO, PHEM, or C) that retains much of its primordial noble gases.
MORBs are not as homogenous as originally thought, and exhibit most of the compositional variability of OIBs, although the variation is expressed in far more subordinate proportions. This implies that the shallow depleted mantle also contains some enriched components.

41. The Nature of the Mantle
So is the mantle layered (shallow depleted and deeper non-depleted and even enriched)?
Or are the enriched components stirred into the entire mantle (like fudge ripple ice cream)?
How effective is the 660-km transition at impeding convective stirring??
This depends on the Clapeyron slope of the phase transformation at the boundary!

42. No Effect Retards Penetration Enhances Penetration
→ 2-Layer Mantle Model more likely
→ Whole-Mantle mixing more likely
Figure Effectiveness of the 660-km transition in preventing penetration of a subducting slab or a rising plume

43. Mantle dynamics water and cold surface
Figure Schematic diagram of a 2-layer dynamic mantle model in which the 660 km transition is a sufficient density barrier to separate lower mantle convection (arrows represent flow patterns) from upper mantle flow, largely a response to plate separation. The only significant things that can penetrate this barrier are vigorous rising hotspot plumes and subducted lithosphere (which sink to become incorporated in the D" layer where they may be heated by the core and return as plumes). After Silver et al. (1988).
Continental Gradient higher than Oceanic Gradient
Range for both
Highest at Surface
water and cold surface
In the future we will often use average values rather than the ranges

44. Mantle dynamics water and cold surface
Continental Gradient higher than Oceanic Gradient
Range for both
Highest at Surface
water and cold surface
In the future we will often use average values rather than the ranges
Figure Whole-mantle convection model with geochemical heterogeneity preserved as blobs of fertile mantle in a host of depleted mantle. Higher density of the blobs results in their concentration in the lower mantle where they may be tapped by deep-seated plumes, probably rising from a discontinuous D" layer of dense “dregs” at the base of the mantle. After Davies (1984) .

45. Mantle dynamics water and cold surface
Continental Gradient higher than Oceanic Gradient
Range for both
Highest at Surface
water and cold surface
In the future we will often use average values rather than the ranges
Figure layer mantle model with a dense layer in the lower mantle with less depletion in lithophile elements and noble gases. The top of the layer varies in depth from ~ 1600 km to near the core-mantle boundary. After Kellogg et al. (1999).

46. Various mantle convection models.
After Tackley (2000). Mantle Convection and Plate Tectonics: Toward
an Integrated Physical and Chemical Theory. Science, 288,

47. A Model for Oceanic Magmatism
The upper mantle is (variably) depleted (DM), and is the source of N-type MORB.
The lower mantle is a major source for E-MORB and OIB magmas.
Both regions are heterogeneous, containing stirred (stretched and folded) disrupted remnants of “marble-cake” crust, lithospheric mantle, and sediment recycled by subduction (as illustrated in the magnified insert for the upper mantle).
The lower mantle contains more of these (denser) constituents: enriched mantle (EM), high-m mantle (HIMU), high-3He/4He (FOZO), and perhaps primitive mantle (BSE) and “prevalent” mantle (PREMA) reservoirs as well.
The 660-km seismic discontinuity is tentatively retained as the boundary layer between upper and lower mantle, but is left gradational to allow for other options, such as a deeper boundary near 1700 km or whole-mantle convection with progressively less depleted and more dense deeper mantle containing more recycled material.
Figure Schematic model for oceanic volcanism. Nomenclature from Zindler and Hart (1986) and Hart and Zindler (1989).

48. Can map the geographic distribution of the isotopic data, suggesting some very large-scale distribution regimes (in addition to the small-scale “marble-cake” regimes)
Contours are for “D8/4” which is a quantitative way of estimating the distance that an isotopic data set for 208Pb and 204Pb plots above the NHRL
Why this enriched anomaly extends as a band across the southern hemisphere at about 30o S is still a mystery
Figure From Hart (1984) Nature, 309,

49. Partial Melting in a Plume
Figure Diagrammatic plume-tail melting model. Rising plume material (heavy arrows are flow lines) is hotter toward the axis. Fluid-present melting of mantle lherzolite may begin at depths of about 350 km (stippled area), but the extent of such melts depends on the amount of fluid present and is probably minor. Melting of recycled crustal pods and stringers (red to green streaks) may also begin near this depth and such melting may be more extensive locally. Major lherzolite melting occurs at depths near 100 km. The melt fraction is greatest near the plume axis, producing picrites and tholeiites. The extent of plume asymmetry depends on plume flux and plate velocity. Plume-head melting is much more extensive (Chapter 15). Based on Wyllie (1988b).

50. Odd:
Tholeiites exhibit enriched isotopic characteristics and alkalic is more depleted (opposite to usual mantle trends for OIA-OIT).
Probably due to more extensive partial melting in the plume axial area (→ tholeiites) where the deep enriched plume source is concentrated
Less extensive partial melting (→ OIA) in the margins where more depleted upper mantle is entrained
Figure Nd/144Nd vs. 87Sr/86Sr for Maui and Oahu Hawaiian early tholeiitic shield-building, and later alkaline lavas. From Wilson (1989). Copyright © by permission Kluwer Academic Publishers.

51. A possible explanation for the late-stage Hawaiian melt rejuvination
Figure 14.22  Melt production in a numerical model of the Hawaiian plume assuming homogenous peridotitic material. Note that melting begins at about 160 km and melt flux is greatest within km of the plume axis and deeper that 120 km. Of particular interest is the second melting event 300 km downstream of the primary melt zone, a result of the reascension of plume material that previously advected slightly downward beneath the lithosphere. Heavy black lines are mantle flow streamlines. After Ribe and Christensen (1999).

52. Figure 14.23  A schematic cross-section through the Earth showing the three types of plumes/hotspots proposed by Courtillot et al. (2003). “Primary” plumes, such as Hawaii, Afar, Reunion, and Louisville are deep-seated, rising from the D" layer at the core-mantle boundary to the surface. “Superplumes” or “superswells” are broader and less concentrated, and stall at the 660-km transition zone where the spawn a series of “secondary” plumes. “Tertiary” hotspots have a superficial origin. From Courtillot et al. (2003).