September 2017 Letter from the Director

Living on an Active Planet
September was marked by a number of devastating natural events that provided a not so subtle reminder of why a better understanding of the workings of our home planet is worth pursuing. As they went about their job pumping heat from the warm Atlantic into the upper atmosphere and from there into space, hurricanes Harvey, Irma, and Maria impacted many of our colleagues, friends, and families. Discussions in the "Handling Hazards" workshop hosted by Carnegie last year made clear that the basic physics of hurricanes are reasonably well understood, and the ability to predict their path once they form has advanced greatly over just the last decade. When a storm is the size of the whole Florida peninsula, or when it lingers long enough to dump many tens of inches of rain on flat lying ground, human impacts are unavoidable. Recent experience demonstrates that knowing that a hurricane is coming, and knowing what to expect when it arrives, can greatly mitigate, though not totally erase, its impact on humans in its path. Our best wishes for a quick recovery go out to all of those affected.
Flooding from Hurricane Harvey in a residential area of Southwest Texas, August 31, 2017. (Credit: U.S. Department of Defense)
Bending Plates
The "Handling Hazards" workshop also considered the case of earthquakes. Presently, only probabilistic estimates can be made of the likelihood of a damaging earthquake in any given area or time period. The September 7 magnitude 8.1 earthquake that occurred at 70 kilometers depth off the western Mexican coast happened in an area that was expected to experience large events, at least on timescales of decades to hundreds of years. Earthquakes in this setting are typically caused by slip along the plate boundary between the Cocos oceanic plate and the overriding North American plate that are moving towards each other at a rate of 7.1 cm (2.8 inches) per year. In 1985, such an event resulted in thousands of deaths in Mexico City, along with widespread damage to the city's buildings and infrastructure. More surprising was the September 19 magnitude 7.1 quake that occurred at 50 km depth, well inland from the previous quake. This earthquake too is associated with the downgoing Cocos plate, but it occurred so far inland because, in this area, the Cocos plate is moving nearly horizontally beneath Mexico. The damage from this earthquake was magnified by the amplification of the ground movement by the soft lake sediments on top of which Mexico City is built.

Seismic image of the downgoing Cocos oceanic plate beneath Mexico.  The plate is defined both by the occurrence of earthquakes (the black diamonds) and by various types of seismic imaging (the red and blue diamonds).  The large red circle shows the location of the Sept. 19 earthquake, at the hinge point where the Cocos plate begins its steep descent into Earth’s interior.  (Credit: U.S. Geological Survey)

Both the September 7 and 19 earthquakes had unusual faulting mechanisms, at least for subduction zone earthquakes; they were caused by tensional, not compressional, stresses in the plate. A possible explanation notes that both earthquakes occurred close to where the horizontally moving Cocos plate bends, first to begin its subduction beneath Mexico, and then again under central Mexico where it begins its descent into the deep mantle. This bending leads to extensional stresses in the top of the Cocos plate as it "cracks" at the hinge point of the bend.

Flat Slab Subduction

This type of so-called "flat slab subduction" has been a long time research target for DTM seismologists. Flat slab subduction was first recognized in South America in the early 1970s by DTM staff scientists David James and Selwyn Sacks and a competing group from Cornell. The fascination with the behavior of subducting plates continues with the work of DTM staff scientist Lara Wagner. The September 19 Mexican earthquake demonstrates one of many reasons this topic is worth pursuing. Normally, when an oceanic plate sinks beneath a continental plate, the oceanic plate descends at a steep angle to great depths in the mantle. As the majority of subduction zone related earthquakes occur near the top, colder and hence more rigid, portion of the downgoing plate, steep subduction limits the more damaging shallow quakes to near the surface intersection of the two plates, usually marked by an oceanic trench. The biggest earthquakes on Earth are produced in this tectonic setting. Most such earthquakes thus occur well offshore, so damage from ground shaking is reduced compared to if they occurred closer to on-land population centers. Unfortunately, this damage mitigating feature of subduction zone earthquakes is more than compensated for by the fact that large, quick, movements of plates beneath the ocean create tsunamis, which often cause more damage than the ground shaking induced by the earthquake.

Wagner's high resolution seismic imaging of the subducting Nazca plate beneath Peru, that required the deployment of nearly 100 seismometers, shows that the subducted plate is severely contorted where the flat portion of the slab meets the steep portion of the slab to the south. To the north, the older portions of the flat slab are falling apart, and new regions of normal steep-dip subduction are beginning to form. The flat part of the subducting plate occurs at the landward extension of a ridge of thickened oceanic crust, known as the Nazca Ridge. The Nazca Ridge is one of many such ridges formed when an oceanic plate moves over a volcanic hot spot. Just to the north of the Nazca Ridge is the Carnegie Ridge, discovered in 1929 by the DTM research vessel Carnegie, that formed when the oceanic plate migrated over the mantle hot spot responsible for the active volcanism of the Galapagos Islands. The thickened crust beneath the Nazca Ridge apparently adds enough buoyancy to the downgoing plate to make it unwilling to subduct compared to the surrounding thinner oceanic crust.

Three dimensional cross section of the crust and mantle beneath Peru. Beneath the crust, the colored region shows the surface of the underlying oceanic Nazca plate descending beneath South America. While most of this plate is subducting at a steep angle into Earth's interior, a small portion just south of Lima is moving horizontally beneath Peru at a depth of about 70-80 km. This "flat slab" reinitiates its plunge into the deep mantle just west of the Peru-Brazil border. Image based on the tomographic results of Antonijevic, Wagner et al. (2015).
Besides the direct earthquake-related consequences of flat-slab subduction, this process dramatically impacts larger scale geologic responses to plate subduction. In normal subduction settings, the descending oceanic plate carries water with it into the mantle. As the subducting plate heats up, it releases this water into the overlying mantle where the water lowers the melting point of the mantle to create the water-rich, and hence explosive, magmas erupted from volcanoes in such settings. DTM staff scientists, Diana Roman and Erik Hauri, are currently studying the composition and eruption characteristics of such volcanoes and how they are influenced by the depth of the underlying descending plate in the Aleutian Islands.

DTM Staff Scientist Diana Roman collecting samples from the Cleveland Volcano, with Tana Volcano on Chuginadak Island in the background in the Aleutian Island chain.  (Credit: Anna Barth, LDEO)

Flat slab subduction interferes with this pattern because it does not allow the water release to occur at sufficient temperature to instigate melting in the mantle. The volcanic chains thus "shut off" in areas of flat subduction. In extreme cases, such as in the western US some 50 million years ago, flat slab subduction moved the line of volcanism as far inland as Wyoming and Colorado where the flat slab reinitiated its descent into the deep mantle. In these cases, the compressive stresses associated with plate collision also are carried inland, providing one explanation for the rise of the Rocky Mountains and for creating the thick crust and high elevations of the Altiplano in South America. The role of flat slab subduction against western North America was the primary target of the High Lava Plains project, a recent DTM study that explored the reasons why volcanism in the Pacific Northwest extends well to the east of the Cascades.

Cartoon model developed from a multidisciplinary project that explored the geology and geophysics of the Pacific Northwest east of the Cascades.  These four cross sections of the Pacific Ocean plates subducting under Oregon and Washington show the transition from a period of flat-slab subduction between about 20-50 million years ago to the current steep subduction.  The transition instigated flow in the mantle that drives the volcanism that covers much of eastern Oregon and western Idaho.  Figure from a paper first-authored by former DTM postdoc Maureen Long.

Basic understanding of the forces driving plate tectonics and plate responses to these forces is unlikely to directly improve our ability to predict earthquakes.  Nevertheless, understanding the fundamental aspects of the inner dynamics of our planet, a key research goal at DTM, can help predict where earthquakes are more or less likely to occur and thus aid in planning for where the most stringent building standards are needed.

Richard Carlson, Director, DTM
Carnegie Institution for Science