Melissa Escudero
Subduction zones are key components in the process of plate tectonics. They form at convergent boundaries where the earth’s crustal plates are colliding. The convergent boundaries on Earth can be split into three different categories based on the kinds of crust involved in the convergence. The three types of convergence are ocean-ocean, which involves two plates of oceanic crust converging and one subducting beneath the other; ocean-continent, where ocean crust is subducting beneath continental crust; and continent-continent, in which two continents collide but neither one subducts. We are therefore only interested in the first two types because they are where the subduction zones form.
The phrase "subduction zone" is an all-encompassing term used to describe the destroyed and over-riding crustal plates and their associated features. The trench, arc-trench gap, and the volcanic arc are the visible representations of a much larger zone that can extend 700+ kilometers into the Earth’s mantle. The angle and depth to which a plate is "eaten" or subducted by the mantle is a function of many variables including mantle viscosity, crust age and temperature, mineralogy, and the stresses applied from the opposing divergent boundary and fore- and back-arc basins. The angles and relative rates of subduction can be observed experimentally by collecting and analyzing earthquake data from different convergent boundaries.
Earthquake data is recorded continuously by seismographs around the world and then collected in central repositories like the US Geological Survey for research and education purposes. The storage bank of earthquake data can be found on the web at www.usgs.gov. This site allows for detailed compiling of tremor events based on given parameters such as longitude and latitude. I was able to collect and plot data for two specific convergent boundaries, one representing each of the two types that result in subduction zones. The ocean-ocean boundary I have investigated is the Fiji-Tonga/Kermadec subduction zone located in the far South Pacific Ocean off the northwestern coast of New Zealand and the ocean-continent boundary is the Cascadia/Juan de Fuca subduction zone off the northwestern coast of the United States.
The Fiji-Tonga subduction zone is characterized by the Pacific plate on the eastern side being subducted by the Indo-Australian plate to the west. This subduction zone is very seismically active, meaning it produces a large number of earthquakes every year. In the earthquake record from 1976 through October 2002, the USGS recorded ~35,000 earthquakes along this zone, many with a depth of up to 700+ kilometers. (neic.usgs.gov) This is a very deep and complete earthquake record, and it can easily be used to better understand the mechanics of the subduction zone. The number of quakes also tells us that the zone is actively subducting the crust, at an average rate of 6-10 centimeters per year. (Tarbuck and Lutgens, 1999)
Deep earthquakes like the ones at Fiji-Tonga provide evidence that the subducted portions of the Pacific plate stay somewhat dense and hard, and that it takes longer to melt into the surrounding viscous mantle (Stein and Stein, 1996). The earthquakes’ close horizontal location relative to the trench’s position at the surface, also gives us an indication that once the Pacific plate has subducted below the Indo-Australian plate, it doesn’t go very far horizontally, instead focusing its descent into the mantle in a vertical direction. According to research models, this actual earthquake data fits the theoretical data for a subduction zone having similar crustal development to this one. (Stein and Stein, 1996) A trend towards a vertical subduction regime (60-65°) is characteristic of oceanic crust that is old and cold, with a small H2O ratio (Ruff and Tichelaar, 1996; Strahler, 1998).
The mineralogy of a rock is responsible for its H2O content and many other definite chemical and physical properties that are related to subduction. However, it is not possible to completely discuss all of the properties at this time, so we must limit our notation to the key fact that differences in rock densities are directly related to their mineral compositions. For example, oceanic crust is composed of dark iron rich minerals, which make the crust dense and hard, allowing it to sink beneath the lighter more silica rich continental crust during ocean-continent convergence. The difference in rock compositions also leads to variations in subduction angles within and across rock types, so that some plates sink at steeper angles than others while being subducted. The proof of variable subduction angles lies in the earthquake data itself (Bebout, 1996).
Scientists have proven that older, colder and less buoyant crust subducts at a steeper angle than does younger, warmer oceanic crust, like that found in the Cascadia subduction zone. The Cascadia subduction zone, unlike the Fiji-Tonga zone, is distinguished by continental crust on one side and oceanic crust on the other rather than two flanks of the same crust. The continental crust on the eastern side is much more buoyant, containing less dense minerals than the oceanic crust to the west. Even though the oceanic crust in the zone is relatively young and warm compared to that of Fiji-Tonga, it is still much denser than the continental crust and gets subducted. Since both crustal types are young and H2O rich, the oceanic crust doesn’t sink into the mantle as easily here, so it forms a very shallow (30- 40°) subduction angle with respect to the continental crust (Peacock, 1996; Bebout, 1996; Strahler, 1998).
This shallow dip angle can be observed by plotting the zone’s earthquake depths. They are all less than 80 kilometers in depth. The angle can also be seen in the arc-trench gap, the distance from the point of subduction in the trench, to the volcanic arc. The volcanic arc is formed where part of the melting plate is erupted onto the surface of the over-riding plate to relieve the pressure buildup caused by subduction. The presence of an arc denotes that the subducted plate can only be within a certain subsurface distance under the line of the arc. At the Fiji-Tonga trench, the volcanic arc is right next to the trench, but at Cascadia, the arc trench gap distance is over 100 kilometers in some areas. This, in connection with the observed earthquake locations, leads to an important idea that the larger the distance between the trench and the volcanic arc, the shallower the subduction angle between the plates (Ruff and Tichelaar, 1996).
The rate of subduction at the Cascadia zone is much lower than for Fiji-Tonga, subducting the crust at a rate of only 2-5 centimeters per year and producing only ~16,000 earthquakes in the 1976 to 2002 time period (www.usgs.gov). This is partly due to Cascadia’s lack of an active mid-ocean ridge in direct opposition to the subduction geometry. Because of the activity along the San Andreas transform fault system along the west coast of North America, the mid-ocean ridge that drives the Cascade subduction zone has been offset, increasing the effects of the low activity ridge and pushing the crust at some oblique angles rather than straight ahead and thus slowing the rate of crust destruction in the zone (Goldfinger, 1996).
The Cascadia zone’s active volcanic arc produces an excess of shallow focus earthquakes. These magmatically driven quakes interfere with the plotting of the subduction related earthquakes in 3D, producing a blob feature instead of the needed plane. For this reason, a best-fit plane was not possible for this convergent boundary.
When all of the Fiji-Tonga earthquakes are plotted on a graph according to depth, one can see a trend from east to west with increasing depth. A 3D plot of this same data provides an outline of a planar feature extending from the ocean floor into the mantle at an acute angle. By fitting the earthquake data to a quadratic equation by depth, latitude and longitude, a "best-fit plane" is created that mimics the overall shape of the subsurface feature. The computer fits a plane to the data points giving us the best fit for the set as a whole. This gives us the best view of what the slab might look like underground. Physical models produced with clay slabs in the lab have confirmed that the plane fitting is accurate and have provided a view of subduction in various stages (Shemenda, 1994).
My initial goal in gathering this earthquake data was to find out which convergent boundary produced the most earthquakes, which had the deepest earthquakes and to see if I could relate earthquake depth to the angle of plate subduction. I have indeed shown the variability in angles of subduction for each of the two convergent boundaries, shown which margin had a greater frequency of earthquakes and therefore a greater rate of subduction, and which had the deepest earthquakes and why. With the help of computer analysis and modeling I have successfully observed how historical earthquake data can be used to reinforce the current accepted models for subduction and to understand some of the forces that drive plate tectonic activity. All these of course, with a little help from the professionals…
Graphics from the presentation can be found on Melissa's website:
www.public.asu.edu/~beaman/computers/finalmain.html
specifically, the 3D depth plots are at:
www.public.asu.edu/~beaman/computers/fijitopo.jpg
www.public.asu.edu/~beaman/computers/cascadetopo.jpg
References
Bebout, Gray E., Scholl, David W., Kirby, Stephen H., Platt, John P.; Subduction Zones Top to Bottom; American Geophysical Union, 1996
Bebout, Gray E.; "Volatile Transfer and Recycling at Convergent Margins: Mass-Balance and Insights From High-P/T Metamorphic Rocks (Overview)", Geophysical Monograph 96,179-180, 1996
Goldfinger, Chris, Kulm, L.D., Yeats, R.S., Hummon, C., Huftile, G.J., Niem, A.R., Fox, C.G., and McNeill, L.C.; "Oblique Strike-Slip Faulting of the Cascadia Submarine Forearc: The Daisy Bank Fault Zone off Central Oregon", Geophysical Monograph 96, 65-67, 1996
Peacock, Simon M.; "Thermal and Petrologic Structure of Subduction Zones", Geophysical Monograph 96, 119-129, 1996
Ruff, Larry J., and Tichelaar, Bart W.; "What Controls the Siesmogenic Plate Interface in Subduction Zones?", Geophysical Monograph 96, 108-110, 1996
Stein, Seth and Stein, Carol A.; "Thermo-mechanical Evolution of Oceanic Lithosphere: Implications for the Subduction Process and Deep Earthquakes (Overview), Geophysical Monograph 96, 1-9, 1996
-------
Shemenda, Alexander I.; Subduction: Insights from Physical Modeling; Kluwer, 37, 40, 51, 1994
Strahler, Arthur N; Plate Tectonics; Geobooks, 273, 1998
Tarbuck, Edward J. and Lutgens, Frederick K.; Earth, An Introduction to Physical Geology, 6th Ed.; Prentice Hall, 459-463, 484-485, 495, 497, 1999
US Geological Survey: www.usgs.gov and neic.usgs.gov
General Mapping Tools, University of Hawaii
