CIG Research Projects

Water, mid-ocean ridges and dynamic geochemical layering of the Earth's mantle

Victoria E. Lee • email
Donald J. DePaolo • email

Dept of Earth and Planetary Science, University of California
Berkeley CA 94720-4767

From AGU Fall 2004 poster U41A-0716

This research is supported by the National Science Foundation through the Petrology and Geochemistry Program, Grant EAR0408521

Figure 1

There are key elements of the solid Earth cycle where chemistry is done. Most of what we infer about the mantle depends on how well we understand the processes. Two of the most important are the subduction factory and the mid-ocean ridge factory.

Figure 2

Figure 3

Introduction

Seismological observations and numerical simulations both suggest that there is significant material exchange between the upper mantle (UM), defined as above the 650 km seismic discontinuity, and the lower mantle (LM). Geochemical observations indicate that the mantle that melts at mid-ocean ridges (MOR) is different from that melted at intraplate hot spots (OIB), suggesting that the LM is chemically different from the UM. An important additional observation is that typical MOR mantle is “depleted,” but only in incompatible trace elements, indicating that it has been affected by removal of only a small degree of partial melt (ca. 1%). [Figure 1]

One observation is that OIB reservoirs are less depleted and more heterogeneous than MORB. This diagram from Kellogg et al. (2002) shows the difference in terms of Sr and Nd isotopes. For Nd the peaks in the distributions are different by 4 epsilon units. One question is how to maintain a difference in a mantle that is convecting. We argue that this difference could be a result of differentiation at mid-ocean ridges coupled with segregation during deep subduction. [Figure 2]

The other important observation is that depleted mantle is not depleted very much. To generate it from primitive mantle requires removal of a small melt fraction, about 1%. Therefore it must be generated where small melt fractions are common. This is not generally in the main melting region at mid-ocean ridges, plumes or subduction zones. In all of these areas the melt fractions for the bulk of the magma are much larger, in the range 1 - 20%. Only in regions of incipient melting in the presence of water is it likely that this type of incipiently depleted material is formed. The bulk of the volume of the mantle that undergoes melting is in fact melted only incipiently due to the effects of water. [Figure 3]

Mid-Ocean Ridge Factory

A key component of the mantle differentiation cycle is the effect of water on melting under MOR’s, and possibly also in island arcs and plumes. The presence of ca. 200-500 ppm H₂O causes melting to begin at greater depth (ca. 100 – 150 km). The resulting oceanic lithologic column is composed of three major sections (rather than two): basaltic crust (BC; 4-9 km), depleted harzburgite/lherzolite (DHL; ca. 40 - 70 km thick), and incipiently melted lherzolite (IML; 30 to 70 km thick).

Aging of the oceanic lithosphere for 70 My results in the BC-DHL package cooling to an average temperature of 400°C, whereas the IML section remains near ambient mantle potential temperature. If the BC-DHL package is preferentially carried to LM depths due to its lower temperature, negative buoyancy, and high viscosity, and the IML is preferentially retained in the UM, the result will be a dynamic geochemical layering in terms of trace elements, isotopes, and heterogeneity The efficiency of this separation does not need to be high to be effective at generating dynamic layering of the mantle. [Figure 4]

Figure 4

Figure 5

Box Model to Describe the Upper Mantle–Lower Mantle Isotopic Contrast

The final equation for the two-reservoir box model in Figure 5 (in terms of Sm-Nd isotopes) is:

where:

The continental crust is ignored in this model, but can be assumed to exist, and cause the bulk system ε_Nd to be displaced from zero.

The steady state value (which may not apply but is a limiting value) is:

K_Nd and K_Sm are derived from the concentrations in the "deeply subducted" (IML) part of the oceanic lithosphere section divided by the starting mantle concentrations. The mass ratios and the UM residence time are the adjustable parameters.

Application of a simple box model to the Sm-Nd system gives a steady-state UM-LM difference in ε_Nd of 4 units (the difference between mean MORB and mean OIB), for mass fluxes between UM and LM corresponding to about 50% of the current rate of generation of oceanic lithosphere. Hence this model works if slab penetration into the LM occurs, but is about 50% efficient. The LM would also tend to have more variable ε_Nd, by about a factor of 2. The model generates subtle major element compositional layering, with average LM being enriched in Ca and Al by a factor of about 1.05 relative to UM. The results are not sensitive to the relative size of UM and LM, but the overall Nd mass balance for the Earth requires that UM be small. The critical issue is the fate of the three components in the oceanic lithosphere during deep subduction.

The figure above summarizes results of one box model calculation (DHL = 45 km) and shows the relationships between the mass ratio of upper and lower mantle and the exchange flux of material from the upper to the lower mantle. To achieve 4 epsilon units difference and maintain a relatively small upper mantle mass means the acceptable solution must be in the box shown. For M_U/M_L ≈ 0.3 and m_UL/M_U ≈ 0.3, the mass flux of material going to the lower mantle is about 30 x 10²⁵ g/Gyr (90 km³/yr), which is less than half of the modern production rate of DHL.

Over 2 billion years a quasi-steady-state UM will develop that is more “depleted” and relatively homogeneous, and the LM will be less-depleted-to-enriched, and more heterogeneous. Melting at hot spots can also contribute to the production of this layering, because in the presence of water, melting typically produces a large volume of IML and smaller volumes of DHL, and the latter is more likely to be part of the cold lithosphere. Subduction of continent-derived sediiment with BC enhances the isotopic contrast between UM and LM.

Back to CIG Research