This may explain why eruption rates are so high a significant eruption occurs once every years , and why it is the only place on Earth where a mid-ocean ridge is exposed above sea level. The relative travel times of seismic waves beneath Iceland seismic waves travel more slowly through hot or partially molten materials have been used to map a narrow anomaly extending to at least km depth, which many believe to be evidence of an Icelandic mantle plume.
However, it is important to keep in mind that km depth is still only the uppermost part of the asthenosphere, nowhere near the 2, km depth of the core-mantle boundary which is generally believed to be the origin of hot mantle plumes. Some scientists believe that the plume does extend into the deep mantle, but cannot be imaged at depth because it is so narrow that it cannot be resolved using seismic imaging techniques. Skip to main content. Adiabatic melting of the mantle, also known as decompression melting.
During decompression melting, upwelling of the mantle results in a rapid decrease in pressure. We have also performed computations to analyse how the plume flow is affected by the position of the root above it. We tested three scenarios:. For the first two cases an upstream flow was observed.
However, the pattern of the plume flow was significantly different for the third case. In this case plume flowed in front of the leading edge of the lithospheric root, that is, in the direction of plate motion. The previously discussed models included the dynamical effect of chemical buoyancy of the depleted residue produced during partial melting of plume material. In the wide root case, we compare the model incorporating chemical buoyancy with one that ignores this effect to isolate its influence on melt generation.
The total melting rate peaks at similar values in both cases after approximately 2 Myr Fig. Subsequently, the melting rate decreases strongly for the case without chemical buoyancy whereas it remains high for another 3—4 Myr in the presence of chemical buoyancy. The distribution of depleted residue at 5. In absence of chemical buoyancy the main zone of strong melting is confined at the plume axis.
This indicates the promotion of small-scale instabilities of material in the thermal boundary layer below the lithosphere by the density deficit of the residue. The associated vertical stirring of the hot plume head significantly enhances pressure-release melting.
Comparison of distribution of residue for the case with a wide root a without and b with depletion buoyancy at 5. In the previous models we assumed an error function geotherm for both the oceanic and the continental lithosphere, which is probably unrealistic at least in the latter case.
The error function profile overestimates the temperature at mid-depth in the lithosphere and implies a thermal boundary layer that is too thick. Aside from the differences in the geotherm, pre-existing small-scale convection may also interfere with plume ascent and melting. In this section, we investigate the influence of these effects in the case of a wide cratonic root.
We use our 3-D model in a first step to derive a modified continental geotherm. In the second step we let the oceanic lithosphere grow in the presence of small-scale temperature perturbations to the same nominal thickness as before.
Finally we insert the plume and follow its interaction with the lithosphere. The parametrization of Doin et al. Running this model for Myr in the presence of small-scale convection we observed only very minor growth of the plate thickness, to the equilibrium thickness Z L equal to km.
The mean geotherm resulting from our numerical calculation is shown in Fig. The oceanic lithosphere is initially rather thin 22 km and defined by an error function profile. Again weak lateral temperature pertubations are added on both the oceanic and continental side and we let the model evolve until the oceanic plate has reached a thickness of 48 km.
In this step we use our standard values for the rheological parameters Table 1. Hence the modelling is not entirely self-consistent, but should in any case result in a more realistic continental geotherm.
Small-scale convection is rather weak and the final oceanic geotherm did not differ significantly from an error-function geotherm, while the thickness of the continental lithosphere grew very slightly.
We used the final state of this run, with the plume-related temperature perturbation added, as initial condition for our plume model. Compared to the wide-root case with error function geotherms and no pre-existing small-scale convection, the melt production rate is slightly lower during the initial stage of high melt generation and is virtually indistinguishable during the later stage Fig.
Thus, melt production is intermediate, in terms of absolute rates and timing, between the case of a root with an error function geotherm and the case without a root, although much closer to the former. This can be understood in terms of a slightly smaller effective thickness of the cratonic lithosphere with a linear geotherm compared to the plate with an error-function geotherm Fig.
However, the difference between the two cases are small and we conclude that details of the thermal structure of the cratonic root do not play a large role for its interaction with a rising plume.
We obtain two significant results when we incorporate a moving lithospheric root into a model of plume with a diapiric head rising to the bottom of the lithosphere.
First, the melting rate increases strongly, compared to the case without root, in the early stages when the plume head flattens below the lithosphere. The temporal variation of melt production is similar to that observed for LIPs and the hotspot trails associated with them White A second interesting observation is the occurrence of volcanism of different ages at a given location.
This implies that lithospheric roots have significant influence on the spatio-temporal distribution of volcanism. In the present model, we get a sharp peak in the melting rate about 0. These values imply production of a total melt volume of 4—5 million km 3 , which is comparable with the estimates of basalts in various flood basalt provinces Richards et al.
After about 5 Myr the melt production rate drops sharply to about 0. The initial peak is absent when the lithosphere is of uniform thickness, and there is only a moderate variation in melt production with time. We explain the enhanced melting in the plume head by the lateral confinement caused by the topographic step in the lower boundary of the lithosphere. It restricts lateral spreading to a smaller area and reduces diffusive cooling of the plume head.
Distribution of LIPs support the association of flood basalt volcanism with continental lithosphere of heterogeneous thickness. In this scenario, our model can explain large melting rates during the initial phase of plume—lithosphere interaction. The second aspect of occurrence of volcanism of different ages at a single location is also interesting. Our model indicates that most of the initial-stage melting takes place in a region up to km upstream of the plume conduit.
Subsequent partial melting, after the dispersal of the plume head and movement of the root away from the plume location, is confined to the region near the plume axis. If we assume that all the melt produced at a given time is rapidly implaced on the surface of the lithosphere above the location where it was produced, it implies that the upstream region which had experienced a previous episode of volcanism once again experiences a fresh episode of volcanism as the lithsophere moves to the right in the box and passes over the plume.
This is clearly seen in Figs 6 a and b but is largely missing in Fig. These are shown in Fig. The figure indicates the presence of relatively older age volcanics north of DVP Locations 6, 7 and in the western part of DVP Location 1 but younger age volcanics in the north Deccan Locations 2—4, 8. Courtillot et al. They related the secondary event to the opening of the Cambay graben in the northwest part of India. Although these dates do not clearly support the presence of volcanics of different ages at a single location, these certainly point towards such a possibility.
The secondary peak shown by Courtillot et al. Locations are shown in Fig. An example of lateral flow of plume material from beneath the thick cratonic root to the surrounding thin lithosphere is seen in the Cenozoic magmatism of eastern Australia.
In this region, the oldest volcanism is seen in the northern part of eastern Australia. It becomes younger towards south as a result of northward movement of the plate.
Tychkov et al. They performed 2-D numerical modelling of plume flow beneath a heterogeneous plate to explain this pattern of volcanism. Thereafter, the plate moved over the plume bringing thick lithosphere above the plume.
This caused plume material to flow towards the eastern edge of the lithospheric root, resulting in a shift in the location of volcanism. Our numerical models also show that the plume material will flow towards the edge of the root and that the flow can take place in the direction opposite to that of plate motion.
Sleep et al. There are some obvious limitations to our model. One such limitation is that it suggests presence of volcanism on the normal thickness lithosphere rather than above the root.
This is contrary to the presence of flood basalts on continental lithosphere. In our model, the thickness of the root inhibits any partial melting beneath it. If we identify the thin lithosphere in our model as an oceanic plate and the root as a continental lithosphere, then the pattern of melting suggests presence of vast regions of basalt volcanism offshore rather than on the continental shield region. The presence of basalts related to the Deccan Volcanism in drilled oil wells has been reported from the western offshore of India Ramanathan Previous numerical models, initially developed to explain the mature-stage Hawaiian volcanism and then extended to model the flood basalt volcanism, also suffer from this limitation because in these models the rigid lithospheric lid is uniformly thin, more suited for an oceanic lithosphere.
Our present study is a step forward, including complexities of lithospheric structure. Continental lithosphere is highly heterogeneous, consisting of thick cratonic roots surrounded by relatively thin lithosphere of mobile belts.
Such a scenario also exists for the DVP which is surrounded by mobile belts in its north and west. The western part is bounded by a long linear fault system extending all along the west coast of India. Therefore, our results can be used to infer that the pattern of volcanism can be significantly different than anticipated previously based on melting beneath a normal lithosphere.
Another limitation of our model is in the choice of initial plume thermal anomaly. We assume that a large plume head rises above the km phase boundary and starts flattening. We anticipate that a relatively small plume head, as obtained in self-consistent numerical models of plume generation, can possibly reduce the amount of melting but would maintain the pattern of melting as obtained in present computations.
AM is grateful to M. Albers, N. We thank Norman Sleep, Cindy Ebinger and an anonymous reviewer for helpful comments. Albers M. A local mesh refinement multigrid method for 3-D convection problems with strongly variable viscosity , J. Google Scholar.
Christensen U. Channeling of plume flow beneath mod-ocean ridges , Earth planet. Basu A. Renne P. Das Gupta D. Teichmann F. Poreda R. Early and late alkali igneous pulses and a high- 3 He plume origin for the Deccan flood basalts , Science , , — Campbell I. Griffith R. Implications of mantle plume structure for the evolution of flood basalts , Earth planet.
Coffin M. Eldholm O. Large igneous provinces: crustal structure, dimensions and external consequences , Rev. Cordery M. Davies G. Genesis of flood basalts from eclogite-bearing mantle plumes , J. Courtillot V. Feraud G. Maluski H. Vandamme D. Moreau M. Besse J. Doin M. Fleitout L. Mantle convection and stability of depleted and undepleted continental lithosphere , J.
Duncan R. Pyle D. Durrheim R. Mooney W. Evolution of Precambrian lithosphere: Seismological and geochemical constraints , J. Ebinger C. Sleep N. Cenozoic magmatism throughout east Africa resulting from impact of a single plume , Nature , , — Farnetani C.
Richards M. Numerical investigations of the mantle plume initiation model for flood basalt events , J. Thermal entrainment and melting in mantle plumes , Earth planet. Griffiths R. Stirring and structure in mantle starting plumes , Earth planet. Hirth G. Kohlstedt D. Water in the oceanic upper mantle: Implications for rheology, melt extraction and the evolution of the lithosphere , Earth planet.
Ito G. Shen Y. Wolfe C. Mantle flow, melting, and dehydration of the Iceland mantle plume , Earth planet. Jordan T. Virtually all of the igneous rocks that we see on Earth are derived from magmas that formed from partial melting of existing rock, either in the upper mantle or the crust. Partial melting is what happens when only some parts of a rock melt; it takes place because rocks are not pure materials. Most rocks are made up of several minerals, each of which has a different melting temperature.
The wax in a candle is a pure material. If instead you took a mixture of wax, plastic, aluminum, and glass and put it into the same warm oven, the wax would soon start to melt, but the plastic, aluminum, and glass would not melt Figure 3. Again this is partial melting. As you can see from Figure 3. It is most likely that this is a very fine-grained mixture of solid wax and solid plastic, but it could also be some other substance that has formed from the combination of the two. In this example, we partially melted some pretend rock to create some pretend magma.
We then separated the magma from the source and allowed it to cool to make a new pretend rock with a composition quite different from the original material it lacks glass and aluminum. The main differences are that rocks are much more complex than the four-component system we used, and the mineral components of most rocks have more similar melting temperatures, so two or more minerals are likely to melt at the same time to varying degrees.
Another important difference is that when rocks melt, the process takes thousands to millions of years, not the 90 minutes it took in the pretend-rock example. Contrary to what one might expect, and contrary to what we did to make our pretend rock, most partial melting of real rock does not involve heating the rock up.
The two main mechanisms through which rocks melt are decompression melting and flux melting. Decompression melting takes place within Earth when a body of rock is held at approximately the same temperature but the pressure is reduced. This happens because the rock is being moved toward the surface, either at a mantle plume a. If a rock that is hot enough to be close to its melting point is moved toward the surface, the pressure is reduced, and the rock can pass to the liquid side of its melting curve.
At this point, partial melting starts to take place.
0コメント