Gutenberg discontinuity

The Gutenberg (the G) (Revenaugh and Jordan, 1991b) or 8$^{\circ}$-discontinuity (Thybo and Perchuc, 1997) marks the transition from the lid, which is defined as the uppermost mantle beneath the Moho, to the low-velocity-zone (LVZ), which characterizes the transition to the convecting mantle. The G specifies the kinematically defined lower boundary of tectonic plates (Kanamori and Press, 1970; Knopoff, 1983) and the G is identified by an impedance decrease.
The LVZ is part of some oceanic (Revenaugh and Jordan, 1991b) and continental (Thybo and Perchuc, 1997) upper mantle velocity models. Beneath some old continents, the LVZ seems to be absent. Figure 2.3 shows upper mantle models for the Pacific and the Phillipine island regions (PA5, PHB3) and for continental Australia (AU3) (Gaherty et al., 1999). For comparison the global Earth models IASP91 (Kennett and Engdahl, 1991) and PREM (Dziewonski and Anderson, 1981) are added.

Figure 2.3: Models of P-wave velocity. Displayed models are PA5, PHB3, AU5 after Gaherty et al. [1999], shown as black lines. The models are computed for three different corridors beneath oceanic (PA5, PHB3) and continental (AU5) regions. For comparison the global Earth models IASP91 (Kennett and Engdahl, 1991) and PREM (Dziewonski and Anderson, 1981) are added as grey lines.

Figure 2.4: Schematic cross section of oceanic upper mantle, showing the compositional boundary hypothesized to generate the Gutenberg discontinuity (Gaherty et al., 1999). Beneath the ridge, the mantle material melts due to decompression in the melt separation zone (MSZ). Any volatiles (H$_2$O, CO$_2$) in the mantle will enter the melt phase, resulting in a dry layer of depleted peridotite residuum overlying 'damp', normal mantle. This compositional boundary is preserved as the plate ages. The G represents the fossilized base of the MSZ. A geotherm is given to illustrate that the G in this model does not deepen with age, as it would if it were to correspond to a critical isotherm.

The oceanic models show a pronounced decrease in P-wave velocity at depths of 70 km and 90 km for the Pacific and the Phillipines, respectively, indicating the beginning of the LVZ with the G. The LVZ ends at depths of approximately 170 km.
In contrast, the continental model AU3 does not show a LVZ, but rather a constant velocity down to depths of $\sim$250 km. Divergent from this negative result beneath continental Australia, the G has been found beneath continents, as North America (Early-Rise experiment and GNOME nuclear explosion), Europe (FENNOLORA experiment), and Siberia (QUARZ PNE-profile, RIFT experiment) (Thybo and Perchuc, 1997).
The G has been studied using different seismic methods like ScS reverberations (Revenaugh and Jordan, 1991b; Gaherty et al., 1999), travel time studies (Gutenberg, 1948; Thybo and Perchuc, 1997) and P $\rightarrow$ S conversions at the discontinuity (Bock, 1991).
The low-velocities may indicate the presence of (i) volatile-free high-temperature melt; (ii) volatile components that may exist as a separate fluid phase or dissolved in partial melt or (iii) variations in mineralogy. The similar depth of the discontinuity in different tectonic regions found by some studies (Thybo and Perchuc, 1997) is interpreted as an indication for petrological changes. All models can explain the strong reflection coefficient of 4.5% to 7% and the strong attenuation of seismic waves in this layer.
More recently, Gaherty et al. (1999) gave an explanation for the G in oceanic areas based on the theory of a compositional boundary. They found, in agreement with previous studies, indicated by the constant depth of the discontinuity beneath different tectonic settings, that the G is not a thermally controlled transition, but a compositional boundary. The compositional boundary is set by the depth of melting during the production of the oceanic crust at ridges (Figure 2.4). The melting occurs due to decompression melting beneath the ridge in a narrow melt separation zone (MSZ). Any volatile (H$_2$O, CO$_2$) will enter the melt phase. The result is a dry layer depleted of Al, Ca and Fe relative to Mg. The underlying undepleted mantle is water saturated undisturbed mantle. The G represents the compositional boundary between these two layers. The compositional boundary is transported with the plate and the discontinuity is preserved as the plate ages. Far from the ridge, the G marks the contrast of the volatile content and thus represents the fossilized base of the MSZ. In this model, the depth is relatively constant with increasing age of the lithosphere. For comparison, a critical isotherm is given in Figure 2.4. If the G would correspond to such a geotherm the discontinuity would deepen with increasing age of the overlying plate, a behaviour not observed by seismology.
The compositional boundary model cannot explain the detection of the G beneath continental regions, where the G more likely represents the top of a layer containing partial melt (Nielsen et al., 1999).