Science Writing Competition June 05 Runner Up
Tectonics over time: Steady state or Variable system?
by Caroline Burberry
Set a pan full of water and on the stove and add some frozen peas. Bring it to a rolling boil and watch the peas follow the convection cells in the water. Convection is a manner of transferring heat between regions; in this system, heat is transferred from the pan to the water and some of the water is heated up and rises. Cold water flows in to take its place and is heated up in turn. The hot water, meanwhile, cools down by transferring its heat to the air and sinks back to the base of the pan. So your peas, although they will never know it, are displaying an example of Bénard convection where the heat from the stove is the driving force.
Now consider a soft-boiled egg. This isn’t an accurate analogy on the finer points, but it helps to visualise the principles of Earth structure. A soft-boiled egg has three distinct, almost concentric parts: a shell, the almost solid white and a fluid or partially solid yolk; these correspond to the Earth’s lithosphere, asthenosphere and core respectively. If you hit the egg on a hard surface in the first stage of removing the shell, a network of cracks develops. These cracked pieces remain firmly set on the white.
Combining Frozen Peas and Boiled Eggs
Although frozen peas and boiled eggs are not a normal culinary combination, the next thought experiment needs to combine the two ideas. Calculations based on the size and physical properties of the Earth indicate that the asthenosphere, like the peas, will convect relatively vigorously at the high internal heat and pressure of the Earth, over geological timescales. But where the peas convected as the water molecules moved apart, cooled and sank together again, the asthenosphere convects when individual atoms move about in the structure of the mantle rock.
Unlike our egg analogy, the fragments of the lithosphere (called plates) are able to move about on the asthenosphere and become the top boundary layer of a convecting system. Within the asthenosphere, there are additional concentric layers, where the high internal pressure has caused the rock structure to change. The key discontinuity or boundary layer is 670km deep in the Earth.
Calculations indicate that asthenospheric convection will take the form of thin, narrow, hot anomalies (thought to be plumes rising from the D’’ layer) and equally narrow, sinking colder anomalies, rather like the convection cells shown by the peas (fig 1). In the scientific literature, there is debate in about whether the asthenosphere convects in two layers, one above and one below the 670km discontinuity or whether there is convection that extends throughout the whole Earth structure. In the latter case, the upper boundary layer is the lithosphere and the lower boundary layer is the D’’ layer.
The lithosphere as a boundary layer
Whether the system convects as a single or double layer, the lithosphere is an upper boundary layer. The thermal convection cells are coupled to the base of the plates and lithosphere is carried along by drag forces at the lithosphere-asthenosphere boundary. There are three principal interactions of plates at their margins.
Two plates move away from each other at a ridge and molten asthenosphere wells up to fill the gaps, solidifying to form lithopshere. Where two plates move together, either there is crumpling to create a mountain chain, or one is forced to descend (subduct) into the asthenosphere, creating the downwelling cold anomalies of the convection system and driving the plate motion. Lastly, plates can slide past each other on a major fault, with the friction generating earthquakes.
The driving force
Convection of the asthenosphere is somehow coupled to the convection and heat flow inside the core, probably via the D’’ layer. New tomographic images and probes from other seismic studies indicate that the D’’ layer is far more unstable and complex than previously thought. The layer develops instabilities which propagate to form the narrow columns of hot material known as plumes.
The heat produced from the core may be a remnant heat from the formation of the Earth, or continuously generated via radioactive decay of elements in the core. Convection inside the core may be driven by internal differences such as differences in composition or the changing magnetic field in the core. A more complex cause of heat and instabilities in the D’’ layer is the force across the CMB from the rotation of the Earth. As the Earth rotates on its axis about the sun, the different layers of the Earth rotate at varying speeds, and this exerts a force on the D’’ layer.
Whatever the actual mechanism, it is believed that the core layers convect and heat is conducted to the base of the asthenosphere at the boundary of the outer core, thus heat flow from the base of the asthenosphere drives the convection cells, the so-called bottom-up drive.
Single or Double?
To solve the single or double layer dispute, centring on whether plumes and slabs pass through the 670km discontinuity, the interior of the Earth is probed by earthquake waves (an art known as tomography). If the wave signal arrives earlier than expected, it may have travelled through a colder (thus more rigid) part of the Earth’s interior. If the wave signal is delayed, the wave is likely to have travelled through a region of hotter material. In this way, a large portion of the asthenosphere can be imaged.
Images of upwelling hot anomalies (plumes) have been produced, where some plumes seem to originate at the 670km discontinuity but others can be traced almost from the D’’ layer. Other discontinuities in the lower mantle complicate the picture. Tomographic images of subducting slabs (the downwelling part of the convection) have also been seen, where the slab appears to pass through the mantle discontinuity at 670km. Other images show slabs being caught up and delayed, or spread out horizontally at the 670km discontinuity. The convection system evidently has some component of both the single and double layered systems.
Dancing continents: change with time
A global-scale consideration of the lithospheric processes described above implies that plate motions across the entire surface of the globe are inextricably linked, thus subduction in one part of the Earth must be balanced by crust production at another. Back-tracking the motions and relative motions of the plates in time further implies that the configuration of continents on the Earth has changed (illustrated in animation on www.scotese.com). As previously discussed, the movement of the plates is part of the surface expression of the asthenosphere flow. The complexity of continental positions implies that the convection cells are not as simple as in our pan of peas. The apparent structure and distribution of convection cells keeps changing.
Several times in the past, reconstructions suggest that all the continental crust was close together, forming supercontinents. As the supercontinents broke up and the fragments were carried in a roundabout way to their current positions, oceans and seas opened and closed. A starting analogy is that of a ballroom waltz, where the steps and pathways of the dancers are prescribed. The effect for the viewer is of a slow and choreographed motion, where the dancers move together and apart; slowly progressing around the room.
Further evidence that ridge spreading rates, and hence the convection system, can vary are found in modern observations. We see spreading variation across ocean-ocean scale (there are some ridges that spread very fast (e.g. in the South Indian Ocean) and some, such as the Gakkel Ridge in the Arctic, that spreads apart very slowly. We also see variation along individual ridge lengths. This implies that asthenospheric flow is non-steady, changing with space as well as time, with the potential for abrupt changes.
However, some scientists believe that the rates and the process have remained the same since 180 million years ago, with an average of 3.4 square kilometres of crust created each year and constant crust destruction over space and time.
Has convection changed steadily?
The Earth must have been hotter in its early days. Heat was produced by impacts of space debris onto the proto-Earth and compression of the early planet. Heat flow might easily have been so intense that the entire Earth’s surface would have been molten. As the impacts ceased, the Earth began to cool. Early lithosphere would have been moved around rapidly as the asthenosphere convected heat away from the core. In the Canadian Shield, there is evidence for a mountain chain formed from continents colliding over 500 million years ago. This collision was over in 10 million years, yet similar scale events today take 10’s of millions of years. India has been colliding with the rest of Asia for about the last 50 million years, implying that tectonic processes have slowed down since the creation of the world, as the Earth has cooled.
Pulses of change?
In contrast to those who publicise a steady-state model, many scientists believe that the spreading rate is changeable, or indeed cyclic, and may have pulsed over time. Repeated supercontinent cycles appear to have occurred through geological time, each, presumably, with a major reorganisation of the asthenosphere convection. There are five known supercontinents in Earth history, formed as plates carrying continental fragments collide. These huge landmasses are left in the middle of a ring of subduction zones, giving a large downwelling zone directly under the supercontinent. Over time the supercontinent blankets and insulates the underlying asthenosphere sufficiently that a new upwelling zone is established that eventually rips the continent apart. The Great Rift Valley of eastern Africa is an example of a continental mass where an upwelling heat source has started eastern Africa ripping apart. In this case, however, the blanketing effect has not induced a strong enough upwelling to complete the process.
A model of variable spreading rates gives a possible explanation for cyclic variations in sea levels and can potentially explain the remarkable coincidence of a long period of high sea level, a large amount of volcanism, faster spreading rates and the fragmenting of the supercontinent Pangea in the Cretaceous period (about 70 million years ago, before the death of the dinosaurs). Fragmenting is driven by the upwelling zone produced underneath the supercontinent, sometimes referred to as a superplume. There is ample evidence for plume activity having played a part in separating Africa and South America (opening the Atlantic Ocean) and the increased heat flow may have caused ridge spreading rates to speed up.
A large upwelling zone and high heat flow would cause a doming effect in the crust, particularly in the ocean basins, near the spreading zones. This dome would have the effect of displacing the water in the ocean basin, giving an apparent sea level high. If this isn’t immediately obvious, try placing a breakfast bowl upside down in an empty washing up bowl and then part filling the washing up bowl, so that the water is just covering the breakfast bowl. Notice how the water level falls when the breakfast bowl is taken out.
The observed pulse in the Cretaceous is either triggered by a natural variation within the system, or an anomaly representing a great and huge change in the entire circulation system. “Plumes or Not? Yes, and Plenty!” was the title of the talk given by R Montelli, from Princeton University, at a recent conference. The evidence for plume activity within the Earth indicates that there are components of both single and double layer convection. Perhaps the best scenario that we have yet is from modelling of episodic convection, where sometimes a deep plume breaks the 670km discontinuity, giving a pulse of whole mantle (single layer) convection, superimposed on a more usual regime of double layer convection.