Mid-ocean Ridge Melting Research Articles

Runaway climate cooling of ocean planets in the habitable zone: a consequence of seafloor weathering enhanced by melting of high-pressure ice

ABSTRACT Terrestrial planets covered globally with thick oceans (termed ocean planets) in the habitable zone were previously inferred to have extremely hot climates in most cases. This is because H2O high-pressure (HP) ice on the seafloor prevents chemical weathering and, thus, removal of atmospheric CO2. Previous studies, however, ignored melting of the HP ice and horizontal variation in heat flux from oceanic crusts. Here, we examine whether high heat fluxes near the mid-ocean ridge melt the HP ice and thereby remove atmospheric CO2. We develop integrated climate models of an Earth-size ocean planet with plate tectonics for different ocean masses, which include the effects of HP ice melting, seafloor weathering, and the carbonate–silicate geochemical carbon cycle. We find that the heat flux near the mid-ocean ridge is high enough to melt the ice, enabling seafloor weathering. In contrast to the previous theoretical prediction, we show that climates of terrestrial planets with massive oceans lapse into extremely cold ones (or snowball states) with CO2-poor atmospheres. Such extremely cold climates are achieved mainly because the HP ice melting fixes seafloor temperature at the melting temperature, thereby keeping a high weathering flux regardless of surface temperature. We estimate that ocean planets with oceans several tens of the Earth’s ocean mass no longer maintain temperate climates. These results suggest that terrestrial planets with extremely cold climates exist even in the habitable zone beyond the Solar system, given the frequency of water-rich planets predicted by planet formation theories.

A joint geochemical–geophysical record of time-dependent mantle convection south of Iceland

The North Atlantic V-Shaped Ridges (VSRs) provide a spatially extensive and clear record of unsteady mantle convective circulation over >40 My. VSRs are diachronous ridges of thick crust formed with a periodicity of ∼5 My along the Mid Atlantic Ridge, south of Iceland. We present data from a set of dredged basalt samples that shows chemical variation associated with two complete VSR crustal thickness cycles where they intersect the Mid Atlantic Ridge. The new dataset also records chemical variation associated with a VSR crustal thickness cycle along a plate spreading flow-line. Inverse correlations between crustal thickness and both incompatible trace element concentrations and incompatible element ratios such as Nb/Y and La/Sm are observed. Geochemical and crustal thickness observations can be matched using a time-dependent mid-ocean ridge melting model with a basal boundary condition of sinusoidally varying potential temperature. Our observations and models suggest that VSRs are generated when hot patches are carried up the plume stem beneath SE Iceland and spread radially outward within the asthenosphere. These patches are then drawn upward into the melting region when passing beneath the Mid Atlantic Ridge. The geometry of the VSRs and the size of the dynamically supported swell suggest that the Iceland Plume is the strongest plume in the Earth at present, with a volume flux of 49±14 km3yr−1.

Renewed melting at the abandoned Húnafloí Rift, northern Iceland, caused by plume pulsing

The abandoned Húnafloí Rift (HR), near Skagi in northern Iceland, provides an opportunity to investigate how relocation of an oceanic spreading ridge influences melt generation. Melting recommenced beneath HR several million years after spreading ceased at 7–4 Ma. Both the older (pre-HR abandonment) and younger (post-abandonment) lavas are tholeiitic basalts with high Fe–Ti and enriched incompatible trace element compositions. The HR lavas are compositionally similar to modern lavas from central Iceland (above the Iceland Plume conduit) and the Eastern Volcanic Zone (a propagating rift). Although older and younger HR lavas are compositionally similar, geological and geophysical observations indicate different mantle melting conditions. For the older lavas, a plume migration reconstruction suggests that the combination of incompatible trace element enrichment and crustal thickness of ∼25 km can be explained by plume-driven upwelling, a scenario similar to central Iceland at present. The smaller volume of the younger HR lavas (<1 km melt thickness) and their greater distance from the plume centre argues against plume-driven upwelling to explain their enrichment. Rare earth element inversion modelling indicates that the younger lavas come from reactivation of the deepest part of the melting region beneath HR. We suggest that the base of the HR melting region was reactivated when a pulse of unusually hot asthenosphere, known to be producing unusually thick crust ∼700 km away from the plume centre today, passed laterally beneath the HR at ∼3 Ma. A thermal and kinematic ridge–plume interaction model indicates that this scenario can explain both the volume and composition of the younger Skagi lavas, provided that a small amount of decompression occurs. The decompression may result either from extension or from buoyancy-driven upwelling of hot mantle beneath the relatively thin lithosphere of the abandoned HR. The hypothesis that a hot plume pulse caused the younger HR lavas is significant not only as an example of interaction between mid-ocean ridge melting and time-dependent mantle convection, but also because this same hot pulse may have subsequently helped promote intensification of the Northern Hemisphere Glaciation when it uplifted the oceanic gateways either side of Iceland.

Mid-Ocean Ridge Melting: Constraints from Lithospheric Xenoliths at Oahu, Hawaii

by melting a lherzolitic source. Instead, they reflect mixing between One plagioclase–spinel lherzolite and four spinel lherzolite xenoliths two components: (1) the pooled melt derived from lherzolitic source from Oahu, Hawaii, contain clinopyroxene grains that show homoin a passive melting regime with an Fmax of 30% (~80% of the geneous rare earth element (REE) abundances and smooth REE mixture) and (2) melt derived from garnet pyroxenite by 40% patterns with systematic depletion of light REE (LREE). These fractional melting (~20% of the mixture). This model produces five xenoliths are mid-ocean ridge basalt (MORB) magma-depleted 7·1 km of crust. residues with compositions that were not modified by later metasomatism. Trace element systematics of these xenoliths were used to investigate the melt production rate (dF/dP) within a 90-my-old residual mantle column (RMC). Such rates were calculated as

Mid-Ocean Ridge Melting: Constraints from Lithospheric Xenoliths at Oahu, Hawaii

One plagioclase–spinel lherzolite and four spinel lherzolite xenoliths from Oahu, Hawaii, contain clinopyroxene grains that show homogeneous rare earth element (REE) abundances and smooth REE patterns with systematic depletion of light REE (LREE). These five xenoliths are mid-ocean ridge basalt (MORB) magma-depleted residues with compositions that were not modified by later metasomatism. Trace element systematics of these xenoliths were used to investigate the melt production rate (dF/dP) within a 90-my-old residual mantle column (RMC). Such rates were calculated as ratios of the difference in extents of depletion (dF) to the difference in equilibrium pressures (dP) between two xenolith samples. The extents of melting were modeled from REE, Sr and Zr abundances in clinopyroxene; and equilibrium pressures were inferred from the two-pyroxene geobarometer of Mercier et al. (1984, Contributions to Mineralogy and Petrology, 85, 391–403). Equilibrium pressures range from 21 to 7 kbar and extents of melting vary from 2 to 8%. Together, these data constrain the maximum extent of melting in the garnet lherzolite stability field to be <2% and a dF/dP of 0.43%/kbar within the stability field of spinel lherzolite. Uncertainties in the estimates of equilibrium pressure and extent of depletion lead to a slightly broader range of dF/dP values (0.26-0.78%/kbar). These values are significantly lower than that of ∼1.2%/kbar suggested by most previous studies. With the best estimated mean dF/dP of 0.43%/kbar, only 3.9 km of crust could have been generated by melting lherzolite in the pressure range of 26–7 kbar. The thickness and composition of the crust that overlies the 90 Ma Oahu RMC require a higher extent of melting. Based on the REE abundances, most samples from the 90 Ma crust and East Pacific Rise can be explained as pooled melts derived from the lherzolitic source in a passive melting regime with a maximum extent of melting (Fmax) of 30%. This model produces a dF/dP of 4.4%/kbar in the pressure range of 7–2 kbar and generates an additional 1.8 km of crust. In detail, the high [Sm/Yb]DM ratios in most East Pacific Rise samples cannot be explained by melting a lherzolitic source. Instead, they reflect mixing between two components: (1) the pooled melt derived from lherzolitic source in a passive melting regime with an Fmax of 30% (∼80% of the mixture) and (2) melt derived from garnet pyroxenite by 40% fractional melting (∼20% of the mixture). This model produces 7.1 km of crust.

The generation of a compositional lithosphere by mid-ocean ridge melting and its effect on subsequent off-axis hotspot upwelling and melting

The generation of a ∼50–500 times more viscous and 1–2% less dense depleted restite layer by melting and melt extraction beneath a mid-ocean ridge can be a significant barrier to subsequent off-axis upwelling and melting. This compositional lithosphere is formed by melting at a mid-ocean ridge which results in the “dewatering” and increase in Mg# of olivine crystals in the restite residue from melt-extraction; effects which can produce up to a 1000-fold increase in the viscosity of the restite relative to upwelling asthenosphere [1,2]. Compositional lithosphere may be the principle mechanical barrier to plume ascent and melting beneath young seafloor. The formation of compositional lithosphere at the ridge axis would also lead to uniform horizontal motion beneath the thickening thermal boundary layer — this may be the physical reason underlying the excellent bathymetric fits of age-plate cooling models which assume this simple pattern of flow. The presence of a ∼ 1021 Pa-s restite layer within and above the primary melting generation region of a mid-ocean ridge may also provide a mechanism for viscous pressure gradients to effectively focus migrating melt to a narrow region of axial volcanism and, simultaneously, to restrict ridge upwelling rates to the “passive” (plate separation) upwelling speeds implied by recent measurements of UTh disequilibria in MORB. These last two effects critically depend on whether or not the presence of migrating melt will locally weaken the restite layer beneath the ridge-axis to an “asthenosphere” as opposed to a “compositional lithosphere” viscosity. Finally, melting at off-axis plumes that ceases at the base of a relatively uniform thickness compositional/thermal lithosphere would modulate different temperature plume inputs into a uniform temperature asthenosphere that reflects the solidus temperature near the base of the lithosphere. This may be the origin of the relatively uniform asthenosphere temperature that is sampled by melting and crustal production at the global mid-ocean ridge system.

Refertilization in a Convecting Mantle: Effects on Primary MORB Compositions

The widely held view is that mid-ocean ridge basalts (MORBs) are produced by polybaric fractional melting of the sub-oceanic mantle in which the composite primary MORBs are derived from a relatively fertile mantle by 10-25% total melting (McKenzie, 1985; Johnson et al., 1990). The differences in primary MORB compositions are inferred to result principally from differences in the pressure and extent of melting (Klein and Langrnuir, 1987; Kinzler and Grove, 1992). The compositional variations between abyssal peridorites, in this scenario, reflect different degrees of fractional melting in which the least depleted peridotites represent the residues of ~10% melting and the most depleted peridotites represent the residues of ~25% melting (Dick and Fisher, 1984; Michael and Bonatti, 1985). It has been suggested, however, that the abundances of moderately incompatible elements (e.g., Na) in abyssal peridotites are not compatible with these rocks having been formed as residues of variable degrees of partial melting (Elthon, 1992). It was suggested instead that abyssal peridotites are a mixture of two distinct components. One component is a depleted harzburgite that has either no clinopyroxene (CPX) or that has < 2% CPX with very low abundances of Na20 (0.01 wt. %), Ti (300 ppm), Zr (0.1 ppm), and Ce (0.02 • chondrites). The other component is a basaltic liquid, which is potentially highly variable in composition. If the chemical variations in abyssal peridotites reflect some type of melt addition (or refertilization) process into depleted harzburgite mantle, there are profound implications for how the source regions of MORBs have evolved. Most importantly, if the mantle melts to produce a harzburgite (< 2% CPX) on a routine basis, the differences in primary composite MORBs will reflect most strongly the differences in the local mantle composition as the mantle enters the midocean ridge melting regime. In light of the variable compositions of mantle xenoliths and peridotite massifs as well as the strong evidence for melt addition processes in these samples, it might also be expected that the sub-oceanic mantle is both variable in composition and is influenced by meltaddition (refertilization) processes. It is proposed here that the MORB source region has been through the mid-ocean ridge melting regime at least 1 to 5 times during normal mantle evolution and consists of a highly depleted harzburgite mantle (similar to that found in peridotite massifs) mixed with basaltic/komatiitic material that has refertilized the mantle. In this scenario, the ascending mantle beneath midocean ridges varies significantly in composition, depending on the amount and type of basaltic/ komatiitic mantle added. Within the melting regime, the mantle melts to whatever extent is necessary to deplete the mantle to harzburgite (+ a few percent clinopyroxene) where thermal considerations stop further melting. It is proposed, therefore, that differences in primary MORB compositions are principally due to differences in the amount and composition of basaltic/komatiitic material added into the source region.

Implications of melting for stabilisation of the lithosphere and heat loss in the Archaean

The increased depth and volume of melting induced in a higher temperature Archaean mantle controls the stability of the lithosphere, heat loss rates and the thickness of the oceanic crust. The relationship between density distributions in oceanic lithosphere and the depth of melting at spreading centres is investigated by calculating the mineral proportions and densities of residual mantle depleted by extraction of melt fractions. The density changes related to compositional gradients are comparable to those produced by thermal effects for lithosphere formed from a mantle which is 200°C or more hotter than modern upper mantle. If Archaean continental crust formed initially above oceanic lithosphere, the compositional density gradients may be sufficient to preserve a thick Archaean continental lithosphere within which the Archaean age diamonds are preserved. The amount of heat advected by melts at mid-ocean ridges today is small but heat advected by melting becomes proportionally more important as higher mantle temperatures lead to a greater volume of melt and as the rate of production of oceanic plates increases. Archaean tectonics could have been dominated by spreading rates 2–3 times greater than now and with mantle temperatures between ca. 1600°C and 1800°C at the depth of the solidus. Mid-ocean ridge melting would produce a relatively thick but light refractory lithosphere on which continents could form, protected from copious volcanism and high mantle temperatures.