When a deep earthquake happens, what happens? At its focus or “source,” how long does it take? And over how big a region in space does it happen? Does it happen smoothly, or in a series of jerks? Does it radiate seismic waves of different strength in different directions? If so, what is the pattern? These are some obvious questions about the nature of the deep earthquake source. This chapter focuses on the observed properties of the deep earthquake source. Then, the next chapter will review the underlying physical processes that may be responsible for deep earthquakes.

Only a few features of an earthquake's source are directly observable from individual seismograms. These include the source duration, the scalar moment, the radiated energy (Fig. 6.1), and sometimes the relative arrival times of various subevents within the rupture process. The uncertainties in these observations can be large; for individual earthquakes different investigators often report values that differ by a factor of two or more. All other source parameters – static stress drop, moment tensor elements, etc. – are derived. Determining these requires the introduction of models that often introduce systematic errors that produce even greater uncertainty. As we evaluate the literature on the deep earthquake source it is important for the reader to keep in mind which source parameters are observed and which are derived.

If you asked most seismologists for a brief summary of what they know about intermediate- and deep-focus earthquakes, they might write something like:

Unfortunately, most of the assertions in the above statement aren't strictly true; i.e., some are never true, some are true only sometimes, and some we aren't yet sure about. For example, in global earthquake catalogs about 25% of all earthquakes have reported focal depths exceeding 60 km (Table 1.1); the 75% that are shallower are hardly “nearly all.” Moreover, even for very large earthquakes with MW or mB larger than 8.0, those with focal depths of 60 km and more make up about a quarter of all events. Moreover, we shall see that not all deep earthquakes are associated with oceanic trenches; many important deep quakes occur as isolated events and not within planar groups. Whether deep quakes have higher stress drops and more impulsive source time functions is subject to debate.

Why write a book on deep earthquakes?

On 21 January 1906 an earthquake with MW of 7.4 occurred at a depth of 340km beneath Japan. Fusakichi Omori, one of the world's finest seismologists of that time, located the event using (S-P) intervals, and concluded that at several Japanese seismograph stations the P waves were transverse and the S waves were longitudinal. This puzzled Omori, so he published a paper about this a year later (see Fig. 0.1 and Chapter 3). So began the story of deep earthquakes – Omori's (1907) paper is among the earliest publications in my collection of references on deep earthquakes.

Of course, for many reasons, Omori wouldn't have told the story of the beginning quite this way. In 1907 the magnitude scale hadn't yet been invented, and Omori's paper never uses the terms “P”, “S”, or “(S-P) interval”; rather, it mentions the “1st displacement,” the “2nd displacement,” and the “duration of 1st preliminary tremor.” Omori also didn't know that the earthquake was deep; indeed, he didn't even know that “normal” earthquakes occurred at depths of 40 km or less. Our perspective on earthquakes has changed considerably since 1907; my collection of references on deep earthquakes now has more than 2000 entries. We have learned a great deal about earthquakes over the past 100 years. For example, one thing we have learned is that deep earthquakes are quite common – nearly a third of all earthquakes located by the International Seismological Centre have focal depths exceeding 60 km; of these, about one fifth exceed 300 km.

Solid earth geophysics in 1900

Nowadays, we often forget how recently we came to our present understandings about the structure and dynamics of the Earth (e.g., see Brush, 1980). In the year 1900, Emil Wiechert had just proposed that the Earth had a core, which he said was solid and made of iron (Wiechert, 1897); Clarence Dutton had just coined the word “isostasy” (Dutton, 1889; reprinted 1925); one of the first international geophysics journals – Gerlands Beiträge zur Geophysik – had just published its first issue in 1887; and nobody had yet rejected Alfred Wegener's theory of continental drift, because it hadn't yet been proposed (Wegener, 1915).

In seismology, there was much confusion about the meaning of the various seismic phases observed on the few crude seismographs available at the time (e.g., Dewey and Byerly, 1969). While Poisson had established the existence of P and S waves earlier in the nineteenth century, it wasn't clear to all that they would be observable. Knott (1888), for example, performed a theoretical analysis of the reflection and refraction of waves at discontinuities. He concluded that P and S would be mixed up by conversions and reflections on passing through heterogeneous materials in the earth's crust, so it would be futile to look for a separation of the two kinds of waves, or to infer anything about the origin of the shock from the nature of the surface motion.

Do deep earthquakes occur randomly in time and in space? Or does their occurrence exhibit regular patterns – perhaps correlated with other phenomena? If you consider the big picture and don't focus too hard on the details the answers are simple: deep earthquakes are almost random with respect to time; and they are very nonrandom in space. But what about those details? In this chapter I use a descriptive approach to characterize the most obvious features of deep earthquake clustering. Then I apply some simple statistical methods to evaluate various more subtle features.

Spatial clustering

Earthquakes and subduction zones

Geographically, the world's deep earthquakes are definitely not random; the great majority lie near deep-ocean trenches beneath well-defined subduction zones (see Figs. 2.1 and 2.2). Moreover, of those with depths exceeding 300 km, about two-thirds occur in the Tonga–Kermadec region (see Fig. 1.14). When plotted in cross sections oriented normal to the trench axis, intermediate- and deep-focus quakes commonly form approximately planar groups called Wadati–Benioff zones (Figs. 5.1 and 5.2).

What is the thickness of these planar zones? For intermediate-depth seismicity there are a few places where travel times from regional network data are available and there have been careful relative relocations of deep seismicity. At a depth of about 100–125 km beneath western Argentina, Smalley and Isacks (1987) determined that all quakes occurred within a zone approximately 20 km thick, with 90% confined to a 12 km thick region (Fig. 5.3).

When a deep earthquake happens, what happens mechanically? Is the rock failure process similar for deep and shallow earthquakes? Or are deep earthquakes an entirely different phenomenon? Since seismic rupture involves the rupture of a crystalline material, what happens at the molecular level?

For the most part scientists who tackle these questions work in rock mechanics laboratories, although seismological observations do place some constraints on the answers. Moreover, at present scientists don't all agree about the answers. This chapter will approach these questions from a historical perspective, focusing on the early proponents for various important ideas, and then summarizing our current state of knowledge.

In general, seismological evidence indicates that temperature and pressure are critical factors affecting where deep earthquakes do and don't occur. Deep earthquakes are almost nonexistent except beneath subduction zones, where cold lithosphere convects downward into the mantle. Moreover, deep-focus earthquakes occur in zones where the subducted lithosphere is coldest, i.e., where the thermal parameter Φ, the product of lithospheric age and vertical descent rate, is highest (see Section 2.4 and Figs. 2.8 and 4.7). Thus, another important question we should ask is: how do temperature and pressure affect the mechanical processes that allow rock to fracture or deform?

Distributions of parameters describing individuals within a population generally contain information about the processes that affect the individuals; e.g., the age distribution of an animal species will depend on whether individuals most often succumb to episodes of bad weather, to random predation of adults, or to “old age.” In this chapter, the individuals are earthquakes, the populations reside in earthquake catalogs, and the parameters of interest are focal depth and “size,” i.e., magnitude or moment. What can the distributions of focal depth and size tell us about the physical processes that control deep seismicity?

What is focal depth?

Although Wadati originally established that deep earthquakes existed by evaluating (S-P) isochrons and felt intensity data, nowadays most catalogs report depths determined using one of three types of data. These are: (1) P arrival times; (2) (pP–P) intervals; and (3) body and surface waveforms.

Depths from P arrival times

The most common method to locate earthquakes – the method used to construct the catalog of the International Seismological Centre (ISC) – is to compare observed P arrival times with arrivals predicted by a reference earth model. The location is then the latitude, longitude, depth, and origin time that produces the best fit between model and observations; in principle this is unique if there are P times recorded at four or more stations. The advantage of this method is that it is widely applicable since P arrivals from numerous stations are often available for earthquakes with magnitudes as small as 4 or less and as large as 8 or more.

Are deep earthquakes useful? That is, do they provide us with information that helps solve problems of practical importance to humankind? Or, are deep earthquakes just another abstruse phenomenon, justifying attention from scientists only “because they are there”?

This chapter's main thesis is that a disproportionate fraction of what we know about the structure and dynamics of the Earth's interior comes specifically from observations of intermediate- and deep-focus earthquakes. Thus, this chapter briefly reviews several topics where deep earthquakes provide the essential information that constrains our present knowledge.

Structure of the Earth

Gross earth models and the structure of the core

What do we know about the structure of the Earth, and from whom did we learn it? Between 1930 and 1945 three individuals, Harold Jeffreys, Beno Gutenberg, and Kiyoo Wadati were responsible for many of the seismological observations used to construct travel time tables and develop models of gross earth structure (see Frohlich, 1987). One thing is very clear if you reread their papers published in journals such as Gerlands Beiträge zur Geophysik, Geophysical Magazine, and Monthly Notices of the Royal Astronomical Society, Geophysical Supplement – records of deep earthquakes were fundamentally important in the construction of these models.

What are deep-focus and intermediate-focus earthquakes?

When we call an earthquake “deep,” what do we mean? As we shall see, there is surprisingly little agreement about this among seismologists. However, in this book I use these definitions:

1. • A shallow or shallow-focus earthquake is any earthquake with a believable reported focal depth less than 60 km;

2. • An intermediate-focus, intermediate-depth, or intermediate earthquake is any earthquake with a believable reported focal depth equal to or greater than 60 km but less than 300 km;

3. • A deep-focus earthquake is any earthquake with a believable reported focal depth of 300 km or more;

4. • A deep earthquake is any deep-focus or intermediate-focus earthquake, i.e., a quake with a believable reported focal depth of 60 km or more.

Although it is confusing for “deep-focus” and “deep” to specify distinctly different depth categories, it is often desirable (as with the title of this book) to specify intermediate- and deep-focus earthquakes together. Since about 1970 a few scientists have called them “mantle” earthquakes instead of “deep” earthquakes (e.g., Isacks and Molnar, 1971). This is misleading since not all shallow-focus earthquakes occur within the crust; e.g., nearly all oceanic intraplate earthquakes are shallow but occur in the mantle. Finally, there are now already hundreds of published papers which used “deep earthquake” to specify both deep-focus and intermediate-focus events.

Realistically, what should you learn from this book? Almost no one reads a scientific monograph from cover to cover; and, even when they do, there can be so much focus on details that it's easy to lose track of what's important. Thus, in this chapter I review what a century of research has taught scientists about deep earthquakes, and then discuss some cultural issues that surround the business that is deep earthquake research.

What we know and what we don't know

What we have learned in the last 100 years

What do we know today about deep earthquakes that we didn't know when Fusakichi Omori was pondering the peculiarities of the 21 January 1906 Japanese earthquake (see the Preface)? Omori was then well into his career and, since he died in 1923, didn't survive to see much of the progress that this book chronicles. It is interesting to imagine what Omori and Harold Jeffreys (1891–1989) might have thought about this book. Jeffreys, of course, had a long and distinguished career in geophysics, as he published his first scientific paper in 1915, his last paper in 1987 (Jeffreys and Shimshoni, 1987), and grappled with several of the research questions that this book chronicles. Each chapter in this book presents information that would have intrigued Omori and about which Jeffreys would have had some firm opinions. What have we learned in 100 years?

Introduction

This chapter summarizes the geographic distribution of the world's deep earthquakes by presenting information about known or alleged deep seismicity for each of 27 geographic regions (Fig. 10.1) as well as a catch-all “other” region (region 28) and the Earth's Moon (region 29). For each region it discusses the geometry of the Wadati–Benioff zone, describes the pattern of focal mechanisms, briefly reviews the literature, and provides information about “significant” earthquakes, i.e., unusually large or well-studied events. In most of the 29 regions the occurrence of deep earthquakes is undeniable. However, in a few regions (e.g., region 1 near the Hjort Trench) deep earthquakes have been reported and the regional tectonics suggest their occurrence is possible, but their existence is unconfirmed.

Catalogs

In this chapter I regularly compare locations reported in several different catalogs; to avoid repetitive citations I abbreviate these as follows.