Tapping of magmas from ubiquitous mantle heterogeneities: An alternative to mantle plumes?

Abstract

Long‐lived heterogeneities in the mantle are indicated by Nd‐Sr‐Pb isotopic systematics. Explanations for these variations have invoked either distinct layers in the earth that are preferentially tapped by some source regions (such as plumes rising from the base of the mantle) or small‐scale heterogeneities which are ubiquitously distributed through the mantle and tapped differentially in various source regions. Consideration of the thermal and mechanical aspects of the latter hypothesis provides an explanation for the differences between mid‐oceanic ridge basalts and off‐axis volcanism. The melting behavior in a region with heterogeneities is modified by lateral conduction of heat, so that the first melted regions are tapped preferentially at off‐axis volcanoes where small degrees of melting occur. A small (less than a few kilometers), easily melted heterogeneity draws heat from its surroundings as it melts during ascent. An increase of melting by a factor of 2 over adiabatic ascent is probable in the most easily melted regions if the melted and unmelted regions have comparable volume. The increase is larger in the earliest stages if melting is confined to small isolated heterogeneities. The depleted iso‐topic ratios of mid‐oceanic ridge basalts can be explained as follows. Heterogeneities with enriched components melt before those with depleted components as material ascends beneath the ridge. Then the enriched melts are removed from the source region before mid‐oceanic ridge basalts are tapped. The first enriched components to melt enter fractures that carry them upward and outward away from the source region of basalts at the ridge axis. This occurs because the direction of shear strain in the upwelling is such that the axis of tension plunges at 45° away from the ridge axis and dikes of melt propagate up perpendicular to the axis of tension. In contrast with the ridge, volcanoes away from the ridge axis preferentially tap the earlier melted components. At the greater source depths of these volcanoes only non‐mid‐oceanic ridge basalt (MORB) components are significantly molten. Another enriched off‐axis source region is the layer formed by the first enriched magma which moved laterally and avoided the source region of basalts at the ridge axis. For small‐scale ubiquitous heterogeneities to explain the differences between mid‐oceanic ridge basalts and off‐axis volcanism, the MORB component must produce a greater volume of melt than the other components. If isotopic mass balances preclude the MORB component from being this abundant in the whole mantle, then the lower mantle is probably a separate reservoir from the upper mantle. In that case, it is conceivable that passive ubiquitously distributed blobs derived from the lower mantle at earlier times act as small‐scale heterogeneities. Both this blob model and the small‐scale heterogeneity model are possible alternatives to active mantle plumes.