Space Geodesy: A Revolution in Crustal Deformation Measurements of Tectonic Processes

Downloaded from specialpapers.gsapubs.
org on March 13, 2016
The Geological Society of America 18 8 8 2 013

0
Special Paper 500
2013
CELEBRATING ADVANCES IN GEOSCIENCE
Space geodesy: A revolution in

crustal deformation measurements of tectonic processes
Roland Bürgmann*
Department of Earth and Planetary Science, 389 McCone Hall, University of California, Berkeley, California 94720, USA
Wayne Thatcher*
U.S. Geological Survey, 345 Middlefield Road, Menlo Park, California 94025-3561, USA
ABSTRACT
During the last ~100 years, tectonic geodesy has evolved from sparse field-based
measurements of crustal deformation to the use of space geodetic techniques involv-
ing observations of satellites and from satellites orbiting Earth, which reveal a variety
of tectonic processes acting over a wide range of spatial and temporal scales. Early
terrestrial measurements using triangulation and leveling techniques characterized
large displacements associated with great earthquakes and led to the recognition of
the fundamental mechanics of seismic faulting and the earthquake cycle. More pre-
cise measurements using ground-based laser ranging allowed for the characteriza-
tion and modeling of interseismic strain buildup and determination of slip rates on
major faults. Continuous and highly accurate point measurements of strain, tilt, and
fault creep have captured intriguing deformation transients associated with slow slip
events on active faults. The greatly improved precision, spatial and temporal resolu-
tion, global coverage, and relatively low cost of space geodetic measurements led to
a revolution in crustal deformation measurements of a range of tectonic processes.
Very Long Baseline Interferometry, the Global Positioning System, Interferometric
Synthetic Aperture Radar, and space-based image geodesy complement each other to
comprehensively capture tectonics in action at scales ranging from meters to global
and seconds to decades. Space geodetic measurements allow for the precise measure-
ment of global plate motions, the determination of strain rate fields and fault slip
rates in distributed plate-boundary deformation zones, and characterization of subtle
intra-plate deformation. These measurements provide increasingly important con-
straints for earthquake hazard studies. Space geodesy also allows for the recognition
and detailed model exploration of a number of transient deformation processes dur-
ing the post-earthquake deformation phase of the earthquake cycle. Measurements of
postseismic deformation transients provide important insights into the mechanisms,
*burgmann@seismo.berkeley.edu; thatcher@usgs.gov
Bürgmann, R., and Thatcher, W., 2013, Space geodesy: A revolution in crustal deformation measurements of tectonic processes, in Bickford, M.E., ed., The Web of
Geological Sciences: Advances, Impacts, and Interactions: Geological Society of America Special Paper 500, p. 397–430, doi:10.1130/2013.2500(12). For permission
to copy, contact editing@geosociety.org. © 2013 The Geological Society of America. All rights reserved.
397
Downloaded from specialpapers.gsapubs.org on March 13, 2016
398 Bürgmann and Thatcher
rheological properties, and dynamics of crustal deformation. Increasingly, seafloor

geodetic measurements provide information about deformation on the 70% of the
Earth’s surface that were previously inaccessible. Future improvements of modern
geodetic techniques promise to further illuminate details of crustal deformation at all
spatial and temporal scales, leading to an improved understanding of the dynamics
of active tectonics.
INTRODUCTION: THE HISTORY OF have decreased with greater distance from the fault (Fig. 1B),
TECTONIC GEODESY understandable in the pre–plate tectonic era. We now know that
displacement rates far from this fault reach a steady value, the
Tectonic geodesy is a field of geophysics that uses geodetic total offset from west to east indicative of the long-term slip rate
measurements of surface deformation to quantify active plate across the fault system.
tectonic motions and to better understand subsurface processes Geodetic measurements made after the destructive 1923
that lead to earthquakes and related deformation. It is a venerable Kanto (Tokyo) M 7.8 earthquake (Scholz and Kato, 1978;
field, dating from at least the 1890s, when earthquakes in Japan, Thatcher and Rundle, 1979), the 1946 Nankaido (SW Japan)
Indonesia, and India produced measurable surface motions M 8.0 earthquake (Fitch and Scholz, 1971; Okada and Nagata,
clearly related to faulting (Bonafede et al., 1992; Segall, 2010; 1953), and the 1906 San Francisco earthquake (Thatcher, 1975)
Yeats et al., 1997). Classical geodetic methods for measuring also demonstrated that transient aseismic deformation followed
relative positions (and their changes) dating from the early 1800s these large earthquakes for decades or more, indicating post-
had long been the only way of detecting surface strains and tilts earthquake stress adjustments in the crust and underlying upper
at the ground surface. But over the past 30 years space geodetic mantle. These led to rudimentary ideas about an “earthquake
methods have transformed a marginal field of geophysics into a deformation cycle” consisting of pre-shock strain accumulation,
major study area in active tectonics. These geodetic techniques coseismic strain release, and postseismic transient movements.
also have proved invaluable for the study of various other pro- Until the early 1960s and 1970s, these observations were not
cesses producing displacements at the Earth’s surface including united into quantitative kinematic models of earthquake cycle
volcanoes, landslides, soft-sediment deformation, subsurface deformation. Steketee (1958) was perhaps the first to realize that
fluid flow, injection and extraction, glaciers, mining and tunnel- crystal dislocation theory could be applied to slippage on faults
ing, and deformation in response to various changing surface in elastic media to predict surface deformations and constrain the
loads, which are not considered in this review. This chapter pro- magnitude (and distribution) of static fault slip (see also Chin-
vides a brief review of the methods and achievements of classical nery, 1961). These were later incorporated into the now classic
tectonic geodesy while focusing on the undeniably revolutionary “buried dislocation” models of elastic strain accumulation (Sav-
advances during the past 30 years. age and Burford, 1973), widely used to model the earthquake
Geodetic triangulation and leveling measurements (see cycle (Fig. 1C). As geodetic data provided increasingly detailed
below) made before and after the 1891 Nobi (Japan) earthquake descriptions of spatial and temporal patterns of deformation
(Mikumo and Ando, 1976) and the 1892 Tapanuli (Sumatra) associated with the earthquake cycle, increasingly sophisti-
earthquake (Reid, 1913) clearly established the relation between cated modeling approaches have been developed to enhance our
coseismic (coincident with the earthquake) movements and the understanding of the underlying processes and their mechanical
fault motions observed at the surface. Although these relation- properties (Segall, 2010). In particular, as rocks can deform by
ships were not universally appreciated at the time, it was the 1906 viscous flow at the high temperatures and pressures found below
M 7.8 San Francisco, California, earthquake on the San Andreas the seismogenic zone, postseismic deformation was recognized
fault (Lawson, 1908) which cemented the idea into earthquake to include a contribution from the relaxation of coseismic stresses
science that earthquake faulting results from the sudden release by transient rock deformation at depth (Thatcher, 1983).
of elastic strains accumulated and stored in the crust for long In the 1960s, the new technology of ground-based laser
periods of time prior to the earthquake itself (Fig. 1A). Given the ranging began to be applied to detect the subtle interseismic fault
rudimentary nature and quality of the 1906-related geodetic mea- motions around the San Andreas fault in California (Savage and
surements, it is in retrospect rather astonishing that these basic Prescott, 1973). This method enabled strain changes that previ-
principles have stood the test of time. Reid (1910) concluded ously required decades to detect with classical methods (Whitten,
that the 1906 earthquake resulted from the “elastic rebound” of 1948) to be systematically mapped at high accuracy (a few parts
strains stored in the crust over a wide region surrounding the San in 107) in only a few years. But these new methods, although
Andreas fault in the ~100 or more years prior to the earthquake, an order of magnitude more precise than classical triangulation
which were released “near” the fault (within ~10 km) at the time measurements, were relatively costly and labor-intensive and
of the earthquake. Reid and his co-workers imagined that the were carried out only by large government organizations like, in
pre-earthquake (“interseismic”) displacements must somehow the United States, the U.S. Geological Survey.
Reid’s View of Interseismic Strain & Displacement
123° 122°
39° Displacement
Strain
SA
N
AN
DR
FAULT
EA
S
FA
Distance from Fault
U
B
LT
Savage and Burford’s Dislocation Model
Interseismic Deformation
0.5
• -1
v = s/π tan (x/D)
38° 0
Mt. Diablo
Normalized Velocity
SF
-0.5
-10 -8 -6 -4 -2 0 2 4 6 8 10
Normalized Distance from Fault
C
PACIFIC
OCEAN Coseismic Deformation
Space geodesy
0.5
-1
u = s/π tan (D/x)
37°
0
A
Normalized Displacement
1906 Displacements -0.5

-10 -8 -6 -4 -2 0 2 4 6 8 10
5m
Normalized Distance from Fault
Figure 1. Measurement and interpretation of deformation associated with the 1906 M 7.8 San Francisco (SF) earthquake mark the beginning of tectonic geodesy and earthquake-cycle
science. (A) Surface displacements from the 1906 Great San Francisco earthquake determined from repeated triangulation measurements in the San Francisco Bay Area (Hayford
and Baldwin, 1908). In this determination, three stations to the east (solid circles) were assumed stationary. (B) Schematic model of interseismic deformation across the San Andreas
fault, from Reid (1910), illustrating the concept of elastic rebound. While Reid correctly recognized the distributed nature of interseismic strain accumulation, lacking knowledge of
plate tectonics, the far field motions in this model go to zero, rather than reflecting steady plate motions. (C) Predicted displacement fields for interseismic and coseismic deformation
from the two-dimensional dislocation model of Savage and Burford (1973). Here, the distances from the fault, x, are normalized by the depth, D, to the base of the coseismic rupture,
or locking depth, below which steady interseismic slip at long-term rates is assumed. The interseismic velocities, v, and coseismic displacements, u, are normalized by the interseismic
slip rates, s , and coseismic slip, s, respectively.
399
Measuring plate tectonic far-field motions was beyond the future problems and anticipated technological developments in
reach of ground-based geodetic techniques, and our knowledge tectonic geodesy.
of the rates of plate tectonic motions was based solely on other
evidence (DeMets et al., 1990). The rates of deformation aver- GEODETIC MEASUREMENT METHODS
aged over several million years could be estimated from the mag-
netic signature of past magnetic pole reversals that are frozen into Geodesy is the science of the shape, deformation, and grav-
oceanic crust along mid-ocean spreading ridges or from offsets ity field of the Earth (Lambeck, 1988). A range of techniques
of geologic features across faults. Space geodetic methods rely exists to measure such quantities. Here we concentrate on mea-
on extremely precise measurements of signals from extraterres- surements of crustal deformation associated with plate motions
trial objects and earth-orbiting satellites, or on observations made and active faults, which is the focus of tectonic geodesy. We
from orbiting spacecraft, and thus do not require line of sight describe terrestrial, space-based, and seafloor geodetic tech-
between observation points. Wegener, who had proposed an early niques and their application to tectonic geodesy with a focus on
“continental drift” model of global plate motions, had presciently the enormous advances made in this field in the last few decades
stated in 1929, “This must be left to the geodesists. I have no using space-geodetic technologies.
doubt that in the not too distant future we will be successful in
making a precise measurement of the drift of North America rela- Terrestrial Geodesy
tive to Europe” (Wegener, 1929). It took until the 1980s, relying
on space geodetic measurements using large radio telescopes and Until the 1980s, tectonic geodesy relied on ground-based
Very Long Baseline Interferometry (VLBI) measurements, to optical or mechanical methods, of which triangulation, trilat-
fulfill this promise (Herring et al., 1986). eration, and leveling were the most common (Lambeck, 1988).
It was the deployment of dedicated satellite constellations, Triangulation is based on the measurement of horizontal angles
in particular the Global Positioning System (GPS) in the 1980s and their changes with time between monuments spaced up to
and Interferometric Synthetic Aperture Radar (InSAR) methods ~100 km apart. Trilateration utilizes the phase delay of a modu-
in the early 1990s, that truly revolutionized tectonic geodesy. lated electromagnetic signal that is transmitted to a reflector cen-
Many ground benchmarks could be occupied economically with tered on a neighboring benchmark to compute the distance to that
GPS methods, first periodically (every few months to years) in neighboring station with high precision. For both triangulation
survey mode, and more recently continuously by networks of and trilateration it is important to survey multiple baselines to and
permanent stations distributed across tectonically active regions. from stations in the network to be able to compute relative posi-
Under favorable ground conditions InSAR mappings provide tions and changes in these positions from deformation. Leveling
semi-continuous coverage of changes in range between points on relies on measuring relative elevations between station pairs a
Earth’s surface and orbiting satellites at precisions of 5–10 mm. few tens of meters apart using a perfectly leveled optical appara-
About 70% of the Earth’s surface is covered by water, includ- tus and two calibrated height rods.
ing almost all divergent plate boundaries, oceanic transform There are a number of other ground-based methods to
faults, and the most actively deforming portions of the world’s study tectonic deformation, which we will not describe in fur-
subduction zones. To study active deformation in offshore areas, ther detail. These include periodically occupied alinement arrays
seafloor geodetic techniques are required. While having been in across creeping fault traces (Galehouse and Lienkaemper, 2003)
development for some years, the technological challenges and and mechanical creepmeters to observe aseismic fault slip at the
high cost continue to make marine geodesy a frontier area of surface by measuring oblique distance changes across a fault
crustal deformation science. trace using rods or wires (Bilham et al., 2004). Creepmeters can
This paper is divided into four subsequent sections. We first capture steady and episodic fault creep at sub-millimeter preci-
describe the basic principles behind ground, space, and sea- sion, and they have proved valuable in studies of slow slip events,
floor geodetic methods, giving the scale, precision, and usage afterslip transients, and triggered slip. In addition to measuring
to which the measurements are applied. We then illustrate how linear strains with laser ranging over kilometer-scale distances, it
GPS and InSAR are applied to map surface deformation over is also possible to precisely measure linear and volume strain at a
large intra-continental regions, using western North America as point using extensometers in boreholes or laser ranging over short
an example and showing how these methods can be used to esti- distances (hundreds of meters scale) in long-baseline strainmeter
mate fault-slip rates, identify stable blocks, and map strain rate installations (Agnew, 1986). Strainmeters provide the most sensi-
fields sometimes found between major faults. Modern ideas tive observations of deformation events in the Earth, at a tempo-
of earthquake cycle deformation are treated next, illustrating ral resolution of a fraction of a second, suitable for study of small
how space-based methods are used to uncover a panoply of deformation transients. Tiltmeters, including borehole and long-
processes and, especially, to link postseismic deformation with baseline systems, provide similarly precise point measurements
the rheological properties (rules that relate the strain response of rotations about horizontal axes (Agnew, 1986). Tide gauges
of rocks to applied stress) of the crust and upper mantle. We measure sea level changes along coastlines. Once corrections for
complete the chapter with our own assessment of outstanding global eustatic sea level changes, local oceanographic effects,
Space geodesy 401
and tidal and seasonal variations have been made, changes in external to the triangulation network to provide a reference frame
sea level reflect local vertical motions. With several decades of leads to an ambiguous displacement field. Displacements can be
continuous data being available in many coastal regions, sub- tied to a frame by making additional assumptions (e.g., zero dis-
millimeter-per-year precision of such uplift-rate measurements placement of several stations, no areal strain, minimal network
can be obtained (Burgette et al., 2009). Precise measurements of rotations, and translations) about the measured deformation. For
relative or absolute changes in gravity from vertical motions and/ measurements of small tectonic motions, it is best to rely directly
or redistribution of subsurface mass can be used to study active on the angle-change measurements and estimate relative station
deformation (Chandrasekhar et al., 2004; Tanaka et al., 2001). displacements or shear strains within one or more station trian-
Tectonic displacements can also be obtained from analysis of gles (Frank, 1966).
repeated ground-based, aerial, or satellite photographic images The triangulation measurement of deformation associated
(Leprince et al., 2008; Michel et al., 1999) or from differencing with the great 1906 San Francisco earthquake continues to repre-
highly precise digital elevation models (DEMs) obtained from sent one of the most important contributions of tectonic geodesy
airborne light detection and ranging (LIDAR) topographic sur- to earthquake science. Californian surveyors soon recognized that
veys (Oskin et al., 2012). ground displacements of several meters during the earthquake
must have significantly distorted the regional triangulation net-
Triangulation—The Beginning of Tectonic Geodesy work in Northern California. Re-measurement of angles between
Triangulation is probably the first geodetic positioning tech- the monuments undertaken to “repair” the network showed that
nique used for surveying applications, going back to the seven- the land to the west of the San Andreas fault moved northwest-
teenth century (Dracup, 1995). Triangulation involves determin- ward by 2–6 m relative to stations well to the east of the fault,
ing the location of a point by measuring angles to it from two based on the difference of coordinates derived from a 1906–1907
or more other points with known positions. These measurements survey and earlier measurements in the mid- to late nineteenth
involve the use of a theodolite, which consists of a rotatable tele- century (Hayford and Baldwin, 1908). These results were of
scope to obtain precise estimates of the horizontal angles formed fundamental importance in the recognition of elastic rebound
between the survey point and two or more distant stations of a and the earthquake cycle (Reid, 1910). The estimates of 1906
network forming the vertices of triangles. The basic technique earthquake displacements shown in Figure 1 were obtained by
had been in use for several centuries for surveying, navigation, assuming that three stations along the eastern edge of the network
and mapmaking. did not move (Hayford and Baldwin, 1908). These displacement
Geodetic-grade triangulation of use for crustal deforma- estimates, computed using the station positions obtained in sur-
tion studies involves repeated, highly precise (of order 0.3 arc s) veys before and after the earthquake, are affected by assumptions
measurements of angular separations between intervisible points made regarding fixed stations and use of the relatively impre-
monumented with permanent survey benchmarks. Moving the cise baseline-length measurement. Segall and Lisowski (1990)
theodolite through a widespread network of stations provides the obtained station displacements by reanalyzing the data using only
angles of many such inter-station triangles (Fig. 1). In addition repeated angle and azimuth measurements and an approximate,
to inter-station angles, it is possible to include sightings to astro- expected displacement field, forgoing the less accurate baseline
nomic objects, which provide constraints on rigid-body rotations measurement and the errors it introduces into the predicted dis-
by orienting the geodetic network in inertial space. Measurement placements. In contrast to the results of Hayford and Baldwin,
of vertical angles allows for the determination of vertical posi- the recalculated displacements are entirely consistent with right-
tions, but this capability was not generally used in tectonic defor- lateral strike slip in the 1906 earthquake. Thatcher et al. (1997)
mation studies. used the angle-change data from along the nearly 500-km-long
By only measuring angles, triangulation provides no infor- rupture zone to directly invert for a model of the coseismic slip
mation about the scale of the network, requiring direct measure- distribution along the San Andreas fault, finding slip values rang-
ment of the length of a short baseline in the network to calculate ing from ~2 m to >8 m.
positions from the angles. By measuring this baseline distance Starting in the 1920s, triangulation networks in California
between two stations in a network using tapes or rods, the rel- were reoccupied at decadal intervals in an effort to measure the
ative position of each station in the network can be computed accumulation of deformation predicted by elastic rebound. These
from the angular information and the measured length. However, measurements eventually documented 30–50 mm/a of right-
the baseline measurements tend to be insufficiently precise for lateral motion across the San Andreas fault system in central Cal-
many applications, and errors of positions increase rapidly with ifornia, thus documenting the plate tectonic motions and inter-
distance from the baseline (Lambeck, 1988). Triangulation mea- seismic strain accumulation along the transform San Andreas
surements of station positions can be precise to within 1–2 parts fault (Savage and Burford, 1970; Whitten, 1948).
in 105 or a few meters for networks spanning several tens of kilo- Triangulation networks ultimately spanned substantial por-
meters (Lambeck, 1988). tions of the globe, including many tectonically active areas.
Changes of positions with time can be used to estimate These data provide valuable survey measurements that have
relative station displacements. The absence of any observations enabled more recent studies of important historic earthquakes
and interseismic deformation (e.g., Bilham and England, 2001; be recovered in subsequent earthquakes, consistent with Reid’s
Song et al., 2008; Thatcher, 1975, 1983; Thatcher et al., 1997; elastic rebound theory. Measurements across the Mojave Des-
Yu and Segall, 1996). Because more than 100 years have passed ert show that the Eastern California shear zone accommodates
since the first historic triangulation measurements, re-observing nearly a quarter of the total North America–Pacific relative plate
these old networks with modern methods can provide valuable motion (Sauber et al., 1986; Savage et al., 1990). Despite the
information on deformation spanning many decades. advances in space geodetic measurements, these freely avail-
able data (http://earthquake.usgs.gov/monitoring/deformation/
Trilateration—High Precision, High Cost geodolite/) continue to be of value for active deformation studies
By the mid-1970s, triangulation measurements of angles (Shen et al., 2011).
were complemented and eventually replaced by electronic dis- Two-color laser ranging provides particularly precise obser-
tance measurements or trilateration as the standard surveying vations relying on the dispersive property of light to reduce the
tool for horizontal position measurements (Savage and Burford, effects of atmospheric signal delay (Langbein et al., 1990).
1973). Trilateration estimates station positions from distance Measuring line lengths with both red and blue lasers reduces the
measurements between control points, relying on the temporal uncertainty associated with changes in atmospheric water vapor
delay of a reflected electromagnetic signal. Precise electronic content. This technique allows for a precision of ~0.5 mm +
distance meters (EDM) were introduced in the 1950s, employ- 0.18 mm-per-km for distances up to ~12 km. Data from a per-
ing microwave, infrared, and later optical wavelengths. The manently installed system at Parkfield, California, provided geo-
modulated carrier signal is reflected off a mirror situated over detic evidence for a transient increase in slip rate on the San
a second benchmark, and the wavelength of the modulation is Andreas fault in the mid-1990s, which accompanied several M
varied to resolve the phase ambiguity of the returned signal. The 4–5 earthquakes on this fault segment (Langbein et al., 1999).
measured phase shift between the emitted and reflected signal
provides precise constraints of the signal traveltime between Leveling—Vertical Motions Only
the two sites. The distance between two points is calculated by Triangulation and trilateration are mostly used for observing
multiplying the measured traveltime by the velocity of light for horizontal positions and deformation. To obtain precise measure-
the local pressure-temperature-humidity conditions. Inter-station ments of relative elevations and vertical displacements, geodetic
distances can be as large as tens of kilometers, but atmospheric leveling has been used since the late nineteenth century (Vanicek
signal delays can produce substantial errors. The measurement et al., 1980). Leveling involves measurement of the elevation dif-
precision can be improved by measuring pressure, temperature, ference between two points using an optical level and two rods
and humidity at, and temperature and humidity by airplane in placed over temporary control points separated by tens of meters.
between the measurement points. Such a system can obtain mea- To measure height differences over large distances, this process
surements accurate to ~3 mm plus ~0.2 ppm (0.2 mm-per-km) is repeated by successively moving the level and rods along a line
baseline length (Savage and Prescott, 1973). By measuring dis- of points between permanent benchmarks that may stretch for
tances between benchmarks forming the vertices of a series of tens to hundreds of kilometers. Most leveling is conducted along
touching or overlapping triangles, horizontal positions relative to roads or railroads, and survey lines may or may not form closed
a reference point and changes in positions can be determined. circuits or networks of lines.
By the late 1980s, nearly 20 years of these labor intensive As the horizontal plane is established as the tangential plane
and costly measurements, conducted in large part by the U.S. to the gravitational equipotential surface, leveling determines
Geological Survey, revealed details of the broad zone of elastic height differences and height changes with respect to the geoid.
strain buildup across the San Andreas fault system (Lisowski et Random errors accumulate as a function of distance from a refer-
al., 1991). These measurements provided a California-wide hori- ence point proportional to the square root of distance, but system-
zontal velocity field even before GPS, but required additional atic errors may further impact the measurement. Highest quality
adjustments owing to limited line-of-sight baselines (Lisowski measurements relying on short baselines, well-calibrated instru-
et al., 1991). The velocity field inferred from repeated surveys of ments and rods, and careful consideration of refraction errors can
trilateration networks spanning much of the fault system revealed achieve precision of as little as 1 mm/km1/2, or 10 mm over a dis-
a fault-parallel shear flow parallel to the local strike of the fault, tance of 100 km (Vanicek et al., 1980). Even today, leveling may
with little fault-normal motions. About 35 mm/a of relative plate provide most precise measurements of vertical deformation over
motion are captured within the span of the trilateration networks, distances of up to a hundred kilometers, but, like other classical
about the same as geologic slip-rate estimates for major faults. methods, it is very time intensive and costly.
The data allowed for estimates of slip rates of individual faults, As with the other geodetic methods relying on line-of-sight
such as the San Andreas, Hayward, and Calaveras faults, using orientations or distances, an important source of error in level-
dislocation models (see section below on the earthquake defor- ing comes from atmospheric refraction. In particular, strong tem-
mation cycle), thus contributing to estimates of seismic hazard perature gradients near the Earth’s surface can lead to systematic,
from these faults. The agreement between these two estimates local slope–correlated errors in the estimated elevation differ-
showed that most of the measured deformation is elastic and will ences. Significant bias of leveling data can also come from poorly
Space geodesy 403
calibrated leveling rods, differential settlement of the instrument These laser ranging systems require clear skies, and this lim-
and the rods during the measurement, variable illumination and its the utility of the method. The French Doppler Orbitography
solar heating of the instrument and rods, and magnetic effects, and Radio-positioning Integrated by Satellite (DORIS) system
which can all produce measurement biases that correlate with involves ground stations sending a radio transmission (at two
topography and/or azimuth (Lambeck, 1988). frequencies to mitigate atmospheric refraction) to two or more
In addition to their analysis of the horizontal deformation receiving satellites in the system (Auriol and Tourain, 2010). The
associated with the 1906 San Francisco earthquake, Hayford orbital motion of the satellites causes a Doppler frequency shift
and Baldwin (1907) also examined leveling data to show that of the signal, from which more precise satellite orbits and ground
the deformation during the earthquake was almost completely positions can be determined. In addition, large-scale changes in
horizontal, thus documenting the strike-slip nature of this event. Earth’s gravity field associated with ~M 9 earthquakes, or glacial
Systematic, elevation-dependent errors played an important isostatic rebound since the removal of continental ice sheets in
role in studies of the “Palmdale Bulge,” a period of apparently the late Pleistocene, can be captured using time series of data
rapid uplift of up to 400 mm in the Transverse Ranges of South- from the GRACE (Gravity Recovery and Climate Experiment)
ern California as deduced from leveling measurements along satellite mission launched in 2002 (Tapley et al., 2004). Precise
10,000-km-long routes, during 1959–1974 (Castle et al., 1976), repeat measurements of the baseline between two identical satel-
and its apparent subsequent collapse (Castle and Bernknopf, lites separated by ~220 km in 450-km-high orbits track changes
1996). A number of re-analyses of these data suggest significant in the Earth’s gravity field at about monthly intervals and with
systematic errors from refraction, rod calibration, and differential ~350-km spatial resolution. Co- and postseismic gravity changes
settling of level and rods (e.g., Jackson et al., 1980). associated with the 2004 M 9.2 Sumatra-Andaman earthquake,
the 2010 M 8.8 Maule, Chile, earthquake, and the 2011 M 9.0
Space Geodesy Tohoku-Oki earthquakes were captured by GRACE (Han et al.,
2010; Panet et al., 2010; Wang et al., 2012b). By utilizing obser-
While many of the terrestrial geodetic techniques described vations of objects in space or observations from space, space
above provide highly precise measurements of crustal defor- geodesy has revolutionized our understanding of tectonic defor-
mation, they have a number of limitations. None of the tech- mation processes at all scales.
niques provide precise measurements of three-dimensional dis-
placements. Terrestrial measurements also require line-of-sight VLBI—Measuring Plate Tectonics in Action
between benchmarks, thus limiting the spatial coverage and pre- While GPS and InSAR currently dominate the space geo-
cision of measurements over longer distances. Generally, tradi- detic arena, Very Long Baseline Interferometry (VLBI) provided
tional field measurements tend to be labor-intensive, complex, the first measurements of current global plate motions. VLBI
and costly and thus were limited in large part to surveying efforts relies on large radio telescopes that simultaneously observe radio
by government agencies. Space geodesy addresses each of these signals from extragalactic quasars billions of light years away
limitations. Unlike terrestrial methods, space geodetic methods from Earth. The development of VLBI began in the late 1960s
do not require line-of-sight between stations in a network, they as a tool for studying compact extragalactic radio sources. Its
do not rely on good weather and daylight, and errors in the mea- potential for geodetic positioning was recognized in the 1970s
surements accumulate much more slowly with increasing dis- (Hinteregger et al., 1972; Shapiro et al., 1974).
tances between stations. Furthermore, once the relevant satellites VLBI utilizes faint signals emitted from stellar sources at
are launched, the measurements are accurate, relatively easy to the edges of the universe that are received at radio telescopes dis-
make, and are often publicly available, either free of charge or tributed on the Earth’s surface. Data from an array of up to ~30
at low cost. Thus, these techniques have rapidly replaced most radio telescopes are combined to create the equivalent of a single
of the traditional geodetic methods. A number of space-based large coherent antenna. While the primary goal is to improve
techniques have been developed to measure crustal deformation the resolution of observations of objects in space by emulating
of which Very Long Baseline Interferometry (VLBI), the Global a telescope with a size equal to the separation between the tele-
Positioning System (GPS) and more recently developed follow- scopes, this requires precise determination of the inter-telescope
up systems, and interferometric synthetic aperture radar (InSAR) baselines from the time delay of the quasar signals. Using com-
have turned out to be the most important. It is interesting to note plex modeling and processing procedures, the data from the radio
that none of these systems were originally developed for the pur- telescopes are analyzed for the time delays between the reception
pose of measuring deformation. Each of these techniques will be of signals from distant radio sources to each telescope. The tele-
discussed in sections that follow. scopes function as interferometers that accurately measure the
Other space geodetic positioning methods introduced in the wavelengths of light and baseline distances. The estimated time
1980s include satellite laser ranging (SLR) and lunar laser rang- delay depends on the angle between the source and the baseline
ing (LLR), which rely on precisely measured distances to satel- between the two stations and on the baseline length. Additional
lites and sites on the moon from the two-way traveltime of laser delays are due to atmospheric refraction. By using observations
beams reflected back to ground stations (Vermaat et al., 1998). at two frequency bands, the dispersive delay from the Earth’s
ionosphere can be eliminated as a major error source in the esti- of the data and the utilization of the carrier phase, as opposed
mation of the distances between the radio telescopes. By utiliz- to coded signals, transmitted in two frequencies, the precision
ing signals from several sources at multiple stations, it is pos- of relative station positions can be improved to the level of few
sible to compute the three-dimensional position of each station millimeters. This precision is about five orders of magnitude bet-
with centimeter-level precision. Repeating these measurements ter than the system was originally designed for. Currently, GPS
every few months enables the detection of changes in the relative tracks the relative three-dimensional positions of thousands of
station positions owing to tectonic motions. The most powerful campaign-mode and continuously operating stations with sub-
aspect of this method lies in its ability to measure these precise centimeter precision. Crustal deformation research with GPS
locations over distances of thousands of kilometers. is being carried out along plate boundaries all over the world,
Integration of the precise baseline information from a net- revealing the complex and variable patterns of the shifting plates
work of radio telescopes allowed for determination of changes in and the complex deformation at their boundaries (Allmendinger
their positions from tectonic motions. Following a few years of et al., 2009; Segall and Davis, 1997; Thatcher, 2003, 2009).
observations, VLBI measurements allowed us for the first time GPS geodesy is based on measuring the distance to several
to directly measure the shifting of the Earth’s plates below us, GPS satellites from the time it takes for a signal to be transmit-
thus fulfilling Wegener’s (1929) dream of measuring active plate ted from the satellites at ~20,000 km altitude to a receiver. If the
motions. Inter-station displacements computed from four years of signal from four or more satellites can be measured at the same
VLBI baseline measurements between radio telescopes in North time, and if we know the position of each satellite, we can com-
America and Europe showed that the two plates are moving apart pute the three-dimensional station position. Each GPS satellite
at ~20 mm/a (Herring et al., 1986). Measurements at stations on transmits two microwave carrier signals with 19 cm and 24 cm
islands on the largely oceanic Pacific plate showed that current wavelength. (GPS satellites launched since 2010 generate a third
plate motions relative to the stable interior of North America are carrier signal with 25.5 cm wavelength.) Most GPS receivers
consistent with rates averaged over a few million years (Argus rely on the code modulated onto the carrier phases, which pro-
and Gordon, 1990; Ward, 1990). For several years, a few smaller, vides information about satellite orbit positions, satellite clock
relatively mobile VLBI systems were deployed in the western corrections, and satellite-specific ranging codes. The code data
USA to provide somewhat more detailed measurements of how are used to determine the distance between the satellites and the
deformation across the plate boundary zone is distributed across antenna feeding the GPS receiver by determining the time off-
the Basin and Range Province and the San Andreas fault system set between the received and the receiver-generated code signal,
(Clark et al., 1987). and multiplying the time difference by the speed of light. To esti-
While initial VLBI measurement precision was on the order mate the true range to the satellites, a number of biases and error
of a few meters (Hinteregger et al., 1972), continued improve- sources must be taken into consideration, namely clock errors
ments of technology and processing approaches achieved cen- in the receiver and satellite, ionospheric and tropospheric signal
timeter level precision in horizontal and vertical components delays, multipath from signal reflections from objects near the
(Herring et al., 1981). Plans for future improvements to VLBI GPS antenna, and receiver noise. Range measurements to four
aim to achieve 1-mm measurement accuracy of global baselines or more satellites, together with information about satellite posi-
(Schlüter and Behrend, 2007). Currently, VLBI continues to tions, allow for measurement of relative GPS station positions
play an important role in the maintenance of global (terrestrial with ~1-m accuracy. A far more precise observable than the code
and celestial) geodetic reference frames and in monitoring the information is the phase of the underlying received carrier sig-
Earth’s angular velocity and the orientation of the rotation axis nal itself. The received carrier is changing in frequency owing
(Altamimi et al., 2012; Schlüter and Behrend, 2007). VLBI is the to the Doppler shift of the moving satellite. The total number of
only space-geodetic method that constrains position in an inertial carrier wavelengths between the satellite and the receiver can be
frame, while all others involve objects rotating about the Earth. estimated by careful evaluation of the satellite phase with time,
However, GPS has proven to be the more suitable, inexpensive, by simultaneously considering phase data from multiple satel-
and equally precise tool to measure tectonic deformation at thou- lites and by utilizing the initial code positioning. As the phase can
sands of points around the globe. be determined at a small fraction of the wavelength, GPS phase
positioning can achieve accuracy of a few millimeters. As with
GPS—Space Geodesy for All the code measurements, atmospheric refraction, orbit uncertain-
Undoubtedly, it was the development of GPS that led to a ties, and multipath can impact the phase measurements and need
revolution in crustal deformation studies in the late 1980s and to be corrected to achieve such precisions even for stations that
early 1990s (Dixon, 1991; Hager et al., 1991). GPS was devel- are thousands of kilometers apart.
oped to provide position, velocity, and timing information glob- The precision of the vertical component is worse than the
ally and at all times. Receivers equipped to measure signals from horizontal by about a factor of three owing to the geometry of the
the GPS satellites operated by the U.S. Department of Defense satellite constellation above the receiver and tropospheric delay
allow the determination of three-dimensional positions accurate errors (Segall and Davis, 1997). It is also important to realize
to several meters in a few seconds. Through careful treatment that GPS and VLBI measure vertical positions with respect to an
Space geodesy 405
idealized ellipsoidal reference frame, as opposed to the Earth’s tion measurements from remote sensing data without requir-
geoid, that provides the datum for leveling surveys. Thus, when ing any presence in the study area. InSAR and GPS are highly
trying to compare or difference vertical positions obtained from complementary in that GPS provides long-term stability, three-
leveling and GPS, knowledge of the local gravity field is required dimensional vector displacements, and better temporal cover-
(Mossop and Segall, 1997). age compared to the global coverage and high spatial resolution
A total of 24 GPS satellites (plus some spares) assure that provided by InSAR. A number of review papers provide more
a sufficient number of satellites are in view at any time any- detailed introductions into the data analysis, characteristics, and
where in the world. To achieve the precision necessary for capabilities of InSAR (Bürgmann et al., 2000; Massonnet and
most crustal deformation studies, it is necessary to mitigate Feigl, 1998; Rosen et al., 2000; Simons and Rosen, 2007). Here,
the effects of signal propagation delays in the ionosphere and we focus on providing a basic overview of InSAR as applied to
the troposphere, to cancel out common error sources by dif- the study of tectonic deformation.
ferencing data from at least two receivers, and to average data InSAR is capable of measuring changes in the line-of-sight
over many hours or days. For baselines longer than ~100 km (LOS) distance between the radar antenna and the surface of the
it is also necessary to compute a more precise position of the Earth between radar flyovers to obtain ~100-km-wide swaths
GPS satellites. For this purpose, data from a worldwide network of surface deformation measurements at a resolution of tens of
of continuously running GPS sites are used. By knowing the meters every few weeks. Radar (RAdio Detection And Rang-
positions of these global stations to within a few centimeters, we ing) imaging involves the illumination of a target with electro-
can compute improved orbit tracks for the GPS satellites. The magnetic microwaves. SAR (Synthetic Aperture Radar) com-
U.S. NAVSTAR GPS constellation was the first of what is now bines signal-focusing techniques with satellite orbit information
a number of similar established or developing global naviga- to produce a high-resolution radar image, with information of
tion satellite systems (GNSS), including the completed Russian both the amplitude and phase of the returned signal from each
GLONASS constellation and the growing constellations of the image pixel. Initially, SAR satellite missions were aimed at pro-
European Galileo and Chinese BeiDou-Compass navigation sys- viding high resolution, all-weather, day-night amplitude images
tems (Hofmann-Wellenhof et al., 2008). of the Earth’s surface, but soon the value of the phase obser-
While early generation GPS receivers were expensive and vations became apparent. The phase difference of two images
bulky, modern GPS instruments are affordable, highly portable, and taken from the same antenna position, but at different times, can
operate under essentially all atmospheric conditions. Increasingly, be measured precisely to reveal shifts of a small fraction of the
rather than only temporarily surveying benchmarks with campaign radar wavelength used. Phase data from two precisely aligned
GPS systems every few months or years, networks of permanently SAR images can be differenced to produce an interferogram,
operating stations that continuously collect and transmit data are which contains information on minute surface displacements
being established. A globally distributed network of >100 such toward or away from the radar between the times of the two
stations, the International GNSS Service for Geodynamics (IGS), image acquisitions. As SAR satellites repeat their orbits only
provides valuable constraints on GPS satellite orbits, Earth rotation within several hundred meters, corrections have to be made to
parameters, satellite clock information, the terrestrial reference separate out the phase change from orbit differences and topog-
frame, and of course global plate motions. Regional continuous raphy, usually using an independently determined digital eleva-
GPS networks now exist in many plate boundary zones including tion model (DEM). Phase unwrapping is a final step of InSAR
~1000 stations in the western USA (see below) and in Japan. An processing that involves converting the phase cycle difference
increasing number of stations within these networks provide high- information (i.e., modulo 2π fringes in a wrapped interfero-
rate (up to 100 Hz), continuous data streams that can provide infor- gram) into LOS displacements (i.e., range-change contours in
mation about deformation events with centimeter-level precision an unwrapped interferogram). Following various processing
within seconds to minutes, rather than days to months. This real- steps, InSAR can sometimes measure relative motions in the
time capability opens up a new range of scientific opportunities for LOS direction within an interferogram with millimeter-level
GPS researchers, including studies of deformation from transient precision. More commonly, obtainable precision is ~5–10 mm
aseismic slip events, rapid characterization of earthquakes, and the owing to error sources we now discuss.
development of enhanced tsunami and earthquake early warning A large number of error sources challenge the production
systems (Hammond et al., 2011b). of such precise and extensive deformation maps with InSAR
(Bürgmann et al., 2000). Any temporal change of the radar return
InSAR—Global and Dense Remote Sensing of Deformation within the target scene (such as from vegetation growth or move-
Since the first successful and spectacular image of surface ment, erosion, construction, cultivation, etc.) may cause decor-
deformation of the 1992 Landers earthquake in the Mojave Des- relation between the two imaging passes used in an interfero-
ert was published two decades ago (Massonnet et al., 1993), gram and thus will lead to a loss of range-change information.
InSAR has matured as a tool that has revolutionized a wide Greater separation between the repeated orbit tracks of the SAR
range of Earth science fields, including tectonic geodesy. InSAR satellite also lead to reduced image coherence. Longer wave-
is unique in its ability to obtain highly precise surface deforma- length radar can penetrate through vegetation, and thus L-band
(~24 cm wavelength) systems have much improved phase cor- ther the quality of InSAR-derived data sets (Ferretti et al., 2011;
relation compared to the shorter wavelength C- (~6 cm) and Hetland et al., 2012; Hooper, 2008; Shirzaei, 2013).
X-band (~3 cm) systems. Longer wavelength systems also allow InSAR studies of tectonic deformation depend on availability
for a wider range of orbit separation baselines without losing of data from frequent and enduring acquisitions by satellites with
image coherence. Thus, for a given satellite system, only image suitable temporal and orbit separation baselines. The first civil-
acquisitions within a range of temporal and orbit baselines will ian SAR mission, the 1978 L-band Seasat satellite, obtained data
be suitable for producing an interferogram. As the velocity of only for three months but provided the crucial first demonstration
wave propagation between the radar and the surface of the Earth of the capability for measuring centimeter-level surface displace-
is affected by the refractive index of the atmosphere, heteroge- ments over agricultural fields in southern California (Gabriel et
neity of the signal delay in the ionosphere and troposphere can al., 1989). The European Space Agency’s 1991–2000 ERS-1 and
produce substantial errors in the surface deformation measure- 1995–2011 ERS-2 missions collected vast amounts of C-band
ment. Atmospheric effects can be reduced by averaging a large SAR data along identical orbits over many actively deforming
number of interferograms, or through the determination of a regions of the world at 35-day repeat intervals. Data collected
so-called atmospheric phase screen during time series process- by the ERS satellites proved crucial for scientific exploration of
ing, to cancel out temporally uncorrelated atmospheric delays. fault and volcano deformation throughout the world, as well as
In some cases it is possible to produce and subtract a model of deformation from various non-tectonic or anthropogenic sources
atmospheric delays from first-order correlation with topography (Bürgmann et al., 2000; Massonnet and Feigl, 1998). While
(e.g., Shirzaei and Bürgmann, 2012) or by using complementary SAR data from the follow-up 2002–2012, C-band Envisat satel-
observations of atmospheric variability made from GPS net- lite cannot be easily used to produce interferograms with ERS
works or other ground-based or satellite sensors (Li et al., 2005). images, it is possible to integrate ERS and Envisat interferograms
InSAR measures a change in distance along the look direc- to produce extensive deformation time series spanning nearly
tion but is not capable of determining the full three-dimensional 20 years (e.g., Shirzaei and Bürgmann, 2013). SAR data useful
displacement vector. Thus, given typical signal incidence angles for InSAR applications also come from the Japanese 1992–1998
of 15°–45° from vertical, InSAR has good sensitivity to verti- JERS-1 and 2006–2011 ALOS L-band satellites, the Canadian
cal displacements but is less sensitive to horizontal motions C-band RADARSAT-1 (launched in 1995) and RADARSAT-2
perpendicular to, and cannot resolve motions parallel to, the (2007) satellites, the German TerraSAR-X (2007) and TanDEM-
approximately north-south–directed satellite track. To measure X (2010) X-band satellites, and the Italian constellation of now
an additional component of the displacement vector, we need to four X-band COSMO Skymed satellites (first launch in 2007).
combine information from interferograms from both ascending Availability of data from these missions is variable, based on
(moving north) and descending (moving south) orbit tracks or geographic constraints, onboard power limitations, and the com-
consider data from multiple satellites. mercial nature of some of the systems. It is hoped that future
Early InSAR studies relied on a single or small number of missions—including the European constellation of C-band Sen-
standard SAR interferograms. Increasingly, large data sets of tinel satellites, the Japanese ALOS-2 mission, and the first scien-
up to hundreds of interferograms are integrated to obtain more tific U.S. SAR-satellite L-band mission—will provide continu-
precise maps of surface motions and time series spanning time- ity and improvements of SAR coverage over actively deforming
varying deformation events, such as transient fault-slip events regions with data being made freely available for scientific stud-
and postseismic afterslip (Hooper et al., 2012). Time series anal- ies. For example, SAR data will be acquired by the Sentinel satel-
ysis also allows for the mitigation of atmospheric artifacts, taking lites for almost every point on Earth at least once every six days,
advantage of the temporally uncorrelated nature of atmospheric opening up the possibility of using InSAR for near real time
delay patterns. The utilization of large sets of interferograms monitoring (Salvi et al., 2012).
also helps to mitigate image coherence problems. By identify-
ing stable points in a series of radar images that maintain their Seafloor Geodesy
coherence over time, it is possible to create a network of phase Traditional terrestrial and space geodetic techniques are lim-
measurements over time and space. Time series analysis of such ited to sites on land, leaving out the ~70% of the Earth’s surface
permanent or persistent scatterers (Ferretti et al., 2001; Hooper et that is covered by water. Methods used to make horizontal or
al., 2004) allows for deformation studies even in quite vegetated vertical displacement measurements on land are generally not
areas, as long as distributed radar-bright and phase-stable scatter- suitable for use in the oceans, as water does not transmit elec-
ers, such as buildings or rock outcrops, exist across an image. An tromagnetic waves well. Instead, many seafloor geodetic meth-
alternative approach lies in determining a time series using only ods rely on the transmission of acoustic waves or measurements
interferograms with very short temporal and orbit-separation of changes in water pressure. Just as most of the terrestrial and
baselines, thus minimizing decorrelation in the images used space geodetic methods we discussed are impacted by refraction
(Berardino et al., 2002; Schmidt and Bürgmann, 2003). New and delay of electromagnetic waves through the heterogeneous
“flavors,” combinations, and improvements of these processing atmosphere, most seafloor geodetic techniques are affected by
techniques continue to be developed and promise to improve fur- the complex refraction of acoustic waves through water.
Space geodesy 407
A number of seafloor geodetic techniques have been devel- below the ship. Acoustic signal traveltimes from the ship to the
oped for studies of oceanic deformation zones. Similar to trilater- seafloor transponders and back are converted to geometric range
ation, acoustic extensometers can determine changes in distance by ray-tracing through a sound speed profile obtained from
between pairs of instruments separated by a few hundred meters conductivity, temperature, and pressure measurements using
to a kilometer from measurements of round-trip traveltime of a device being lowered to the seafloor during the collection of
acoustic pulses (Chadwick et al., 1999). To span larger distances, the GPS and acoustic data. Three or four precision transponders
chains of the acoustic transducers need to be aligned across the are deployed on the seafloor to form an equilateral triangle or
targeted deformation zone. These extensometers have been suc- square inscribed in a circle with the radius of the nominal water
cessfully deployed to study deformation associated with a rift- depth. The ship is continuously positioned near the center of the
zone eruption along the Juan de Fuca Ridge several hundred kilo- array so that the acoustic raypaths will all have nearly the same
meters offshore of the Pacific Northwest (Chadwick et al., 1999). launch angles from the ship, and the azimuths from the ship to
Vertical deformation of the seafloor can be observed by the transponders are uniformly spaced. Thus upper ocean sound-
measuring hydrostatic pressure at seafloor monuments, convert- speed variability will appear to move the seafloor array verti-
ing the pressures to depths, and monitoring the change in those cally, but it will not bias the horizontal position estimate. With
depths over time (Ito et al., 2011; Phillips et al., 2008). Instrument this approach, ~80 h of continuous GPS and acoustic data can
drift and tilt, ocean tides, and natural variations in atmospheric determine the horizontal position of the seafloor array with 4–
and ocean pressure contribute to measurement uncertainties, but 6 mm repeatability in the global reference frame (Chadwell and
precision at the several-centimeter level can be achieved (Phillips Spiess, 2008; Gagnon et al., 2005).
et al., 2008). Phillips et al. (2008) used an array of nine pressure Seafloor geodetic measurements using GPS-A have been
sensors to characterize the vertical deformation field of Kilauea used to measure motions of the oceanic Juan de Fuca microplate,
Volcano’s mobile south flank offshore of the Big Island of Hawaii, to study offshore deformation and coupling across subduction
showing up to 90 mm/a of uplift in this large slump structure zones in Peru and Japan, and to obtain co- and postseismic dis-
~15 km SE of the coastline. Six continuously operating ocean- placements directly in the hanging wall of subduction zone earth-
bottom pressure gauges captured the coseismic deformation of quakes in Japan. GPS-A measurements collected at two stations
the 2011 Tohoku-Oki earthquake, providing solid evidence for offshore of the Cascadia subduction zone indicate motions of the
uplift of >5 m near the trench, which produced the subsequent Juan de Fuca plate consistent with plate rotation models con-
catastrophic tsunami (Iinuma et al., 2012; Ito et al., 2011). The strained by magnetic lineation data (Chadwell and Spiess, 2008;
pressure data also provided detailed information about deforma- Spiess et al., 1998). A third station, closer to the trench, is moving
tion from an M 7.3 foreshock and its rapid afterslip, which pre- in the direction of the predicted plate motion, but at a velocity
ceded the Tohoku-Oki earthquake by two days (Ohta et al., 2012). 20 mm/a less than predicted (Fig. 2; C.D. Chadwell, 2012, per-
Borehole tiltmeters that are routinely used on land to moni- sonal commun.). This suggests elastic and/or permanent defor-
tor volcano deformation can also be deployed offshore. A new mation of the Juan de Fuca plate. Gagnon et al. (2005) obtained
experiment by scientists at the Woods Hole Oceanographic Insti- velocities at three stations on the accretionary complex overlying
tution (WHOI) will deploy a tiltmeter in an existing borehole, the subducting oceanic Nazca plate in Peru, South America. The
together with a seafloor geodetic benchmark and pressure sensor observed velocities are consistent with a fully locked subduction
array, to be installed ~100 km offshore of Vancouver Island in thrust from 2 to 40 km depth. The strong coupling at shallow
the Cascadia subduction zone and which can stream its data in depths suggests that earthquakes rupturing the megathrust with
real time via an existing submarine cable system (Jeff McGuire, high coseismic slip up to near the trench in this region are partic-
2012, personal commun., http://www.neptunecanada.ca/news/ ularly prone to generating large tsunamis (Gagnon et al., 2005).
news-details.dot?id = 31191). The 2011 M 9 Tohoku-Oki earthquake turned out to be such an
The GPS-Acoustic (GPS-A) approach extends precise GPS event, with horizontal displacements as large as 31 m at GPS-A
positioning for crustal motion studies to the seafloor. It com- stations close to the trench, and producing a devastating tsunami
bines GPS and acoustic ranging to measure the centimeter-level caused by the resulting seafloor uplift near the trench (Kido et
position of an array of seafloor transponders in the same global al., 2011; Sato et al., 2011). The data collected at seven GPS-A
reference frame as land-based GPS sites (Spiess, 1985). The sea- stations before and after the earthquake constrain extremely large
floor array can be hundreds of kilometers from shore, allowing (>50 m) coseismic slip near the Japan Trench (Iinuma et al.,
geodetic measurements of plate motion between widely sepa- 2012). There are now ~30 offshore GPS-A stations along the east
rated seafloor points or of plate boundary deformation processes coast of Japan to monitor interseismic coupling and earthquake
across subduction zone margins and oceanic spreading zones. A potential of the Nankai and Japan subduction zones. Twenty
ship or buoy, with three GPS antennas forming a triangle, pro- more GPS-A stations are to be installed along the Japan Trench
vides the interface between GPS and acoustic ranging to the sea- in 2012, and new stations are also planned for the Nankai subduc-
floor. High-rate GPS carrier phase data are sampled at the ship tion zone (M. Kido, 2012, personal commun.).
and on shore to provide the second-by-second positions of the Just as it is possible to measure surface deformation from
shipboard GPS antennas and of a hydrophone that is mounted comparison of land-surface images or digital elevation models,
large submarine deformations can be captured with repeated

active-source, multibeam bathymetric surveys. Fujiwara et al.
(2011) document this capability and show horizontal seafloor
displacements of ~50 m and 7–10 m uplift across the surface
trace of the 2011 Tohoku-Oki megathrust rupture at the trench,
with uncertainties of several meters. Interferometric sonar pro-
vides enhanced precision of bathymetric measurements. Interfer-
ometric swath bathymetry employs two or more vertically spaced
receiving elements and uses the relative phase return delays to
determine the direction from which the signal was received in
addition to the range distance (Geen, 1998). Interferometric sonar
is mostly applied in relatively shallow, <~200 m water depths.
However, by using a precisely navigated underwater vehicle
moving several tens of meters over the seafloor, it appears to be
possible to achieve seafloor bathymetric surveys at centimeter-
level resolution by combining synthetic aperture interferometry
and multibeam techniques (Asada and Ura, 2010). This promises
to provide precise measurements of small-scale deformation pro-
cesses on the seafloor.
CRUSTAL DEFORMATION AND ACTIVE TECTONICS
The Earth’s surface is composed of a mosaic of a dozen or

so large, mostly rigid plates that rotate about instantaneous rota-
tion axes (Euler poles) (e.g., Cox and Hart 1986). However, at the
continental boundaries of these major plates, particularly in the
backarcs of current or recently active subduction zones (Hynd-
man et al., 2005), deformation across the plate boundary zone
is diffuse, spread over hundreds to several thousand kilometers,
depending on the setting (e.g., western USA; Alaska; western
South America; Central Asia; Japan; New Zealand; Aegean).
Prior to the advent and maturation of space geodetic methods it
was difficult to accurately quantify this active, widespread con-
tinental deformation (Gordon, 1998; Thatcher, 2003). However,
during the past decade, GPS point measurements and InSAR
spatial mapping of surface deformation have overcome this defi-
Figure 2. Seafloor geodetic measurements offshore of the Cascadia
subduction zone. Results from three GPS-A stations (orange arrows) ciency, providing detailed pictures of the patterns of the current
show the motion of the Juan de Fuca plate with respect to stable North velocity field over wide regions (e.g., Fig. 3).
America. The station closest to the coast moves in the direction of
plate convergence, but ~20 mm/a more slowly than the geologically Mapping Current Continental Deformation
determined rate (black arrows), suggesting internal deformation of the
subducting plate (GPS-A results from Spiess et al., 1998; Chadwell
and Spiess, 2008; and David Chadwell, 2012, written commun.). On- Space geodetic mapping of continental deformation clearly
land GPS velocities relative to the plate interior (blue arrows, Rob shows that the majority of interplate displacements are concen-
McCaffrey, 2012, personal commun.) reveal the interseismic elastic trated at major subduction zone boundaries, including the Aegean
strain accumulation across the megathrust and regional microplate tec- (Floyd et al., 2010), New Zealand (Wallace et al., 2004), and
tonics. Red dots are earthquake locations from 1990 to 2012 (from Japan (Loveless and Meade, 2010), or at large strike-slip faults
ANSS catalogue), green lines are spreading ridges, and orange lines
are oceanic transform faults. Active volcano-tectonic deformation of such as the San Andreas fault system, western USA (McCaf-
the Axial Seamount on the Juan de Fuca Ridge has been studied with frey et al., 2007), and the North Anatolian fault, Turkey (Reil-
seafloor extensometer and pressure sensor measurements, showing re- inger et al., 2006). However, measurable deformation, as much
peating inflation-eruption cycles (Chadwick et al., 1999; Chadwick et as ~30% of the relative plate motion, and slip rates as large as
al., 2012). A new experiment offshore of Vancouver Island will include ~10 mm/a, are observed on other faults not part of the boundary
a borehole tiltmeter (yellow triangle) together with a seafloor geodetic
benchmark and a pressure sensor array (white circles), with data being zones (Thatcher, 2003).
transmitted in real time via a submarine cable system (red line, Jeff The distributed nature of active continental deformation is
McGuire, 2012, personal commun.). immediately clear from examination of Figure 3, which shows
12
A
2W East
ern C
alifor
nia Shea
r Zon
e
Sierra Nevada - Great Valley Block
12
4W
San Andreas Fault System
N
40
0 100 20 mm/yr
12
N
12
N
11
N
N
2 0 8
W W W
32
38
34
36
Space geodesy
Figure 3. Surface deformation across California, measured with GPS and InSAR, respectively. The maps are shown in oblique Mercator projection about a pole of Pacific plate to North
America plate rotation, determined from GPS measurements (Wdowinski et al., 2007). (A) Horizontal GPS velocities across California with respect to a stable North American reference
frame (R. McCaffrey, 2012, personal commun.). The data reveal the broadly distributed deformation across the Eastern California shear zone and distribution of strain associated with the
San Andreas fault system. Red lines are mapped late Quaternary faults in California (Jennings, 1975). (B) Map of relative range-change rates in the line of sight (LOS) from 2006.5 to
2010 measurements of the Japanese ALOS SAR satellite across California (Tong et al., 2013). Positive velocities (red) show the ground moving away from the satellite, flying along an
ascending orbit and viewing the Earth along an ~38° incidence angle. GPS data are used to constrain long-wavelength deformation, and InSAR recovers the short-wavelength deformation.
The integrated GPS and InSAR measurements reflect the broadly distributed deformation about locked sections of the San Andreas fault system, localized creep along the aseismically
slipping central San Andreas, and vertical deformation associated with groundwater level changes, oil field extraction, sediment compaction, and geothermal fields.
409
GPS and InSAR maps of current movements along the San as a candidate model. Since velocity fields like Figure 3 include
Andreas fault system and the Eastern California shear zone in no direct coseismic effects of earthquakes, and postseismic tran-
California. Figure 3A, a map of the GPS velocity field relative to sient effects from recent events are eliminated or minimized, the
stable (not currently deforming) North America, shows (at this field represents steady-state “interseismic” velocities (see next
scale) a sharp velocity gradient at the San Andreas fault, with section for details on the earthquake cycle). In block modeling,
velocities near 50 mm/a west of this system, ~12 mm/a directly this steady-state field is assumed to be composed principally of
to the east, and ~3 mm/a in the central Nevada Basin and Range two deformation sources: (1) elastic strain accumulation from
Province. Because of large long-wavelength uncertainties in the faults that are locked (non-slipping) from the surface to ~10–
InSAR data from orbit and atmospheric errors, the GPS data are 30 km depths (Savage, 1983; Savage and Burford, 1970), and
used to constrain the regional LOS velocity field. Thus, GPS (2) slip below at rates consistent with the rotation of rigid blocks
and InSAR are highly complementary methods constraining that follow the rules of plate tectonics—that is, the blocks rotate
regional and local deformation patterns, respectively. Figure 3B, as rigid spherical caps about an Euler pole. The interseismic
an InSAR map of LOS velocity between the Japanese L-band model is commonly implemented as the sum of the long-term
radar satellite ALOS and ground points in California (Tong et rigid block rotation and shallow back slip (i.e., deformation from
al., 2013), also shows the high-velocity gradient across the San shallow dislocation slip at the long-term rate, but in the oppo-
Andreas fault system and details of the velocity field not eas- site direction). In some block models, internal straining of the
ily seen in the GPS data (e.g., the creeping segment of the San blocks, presumed to be effects of minor faults not included in the
Andreas fault in central California). An advantage of InSAR over block geometry, is an additional deformation source that needs
GPS is that under favorable conditions a nearly complete spatial to be accounted for (McCaffrey, 2005). In some cases, it is also
mapping of continental deformation is obtained, albeit at reduced important to allow for partial coupling on some block-bounding
precision and subject to vexing errors from atmospheric con- faults by also solving for distributed back slip, such as the sub-
tamination of the range measurement (see discussion of previous duction zones in Japan (Loveless and Meade, 2010). Once the
section). Data such as these constrain kinematic models of the block geometry is specified and the observed velocity field is cor-
present-day deformation, including estimates of fault slip rates rected for fault locking (or included as an adjustable parameter in
and rates of straining between the faults (see below) and models the modeling), the data are then used to estimate the Euler vector
of the earthquake deformation cycle (next section). (latitude, longitude, and rotation rate) for each block. The relative
motion between adjacent blocks then provides an estimate of slip
Modeling Continental Deformation rate and sense of slip at every point on the block-bounding faults.
Misfits of data to the model may then be decreased by trial-and-
There are two commonly used ways in which assumed error adjustment of block geometry, either by modifying bound-
steady-state (i.e., interseismic or long-term average with earth- ary locations, adjusting the locking depth on a fault segment, or
quakes ignored) GPS and InSAR velocity fields are modeled to by adding new blocks.
determine fault slip rates and/or rates of straining. The block- There are both advantages and limitations to the block
modeling approach considers the elastic strain field across shal- modeling approach. The method is conceptually simple, and
lowly locked faults (Savage and Burford, 1973) together with present-day fault slip rates are determined in the analysis and
relative motions of fault-bounded blocks to model continental can be compared with late Quaternary and Holocene slip rates
deformation. The alternative way to interpreting surface velocity estimated by geologic methods. The two independent methods
fields involves calculating maps of distributed horizontal strain often agree well, validating the utility and accuracy of both
rates and modeling continental deformation as flow of a con- methods (see further discussion below). However, slip rates on
tinuous viscous medium. Both of these approaches can be used closely spaced faults cannot be individually determined from
to make assessments of earthquake hazard. Figure 4 illustrates GPS data because of the smooth spatial gradient of the veloc-
this, showing a block model for the western USA (Fig. 4A; R. ity field resulting from fault locking. Furthermore, the effects of
McCaffrey, 2012, personal commun.) and a strain rate map of minor faults not included in the model can often be cumulatively
an overlapping, somewhat smaller region (Fig. 4B; Kreemer et significant, contaminating the block model solution in ways that
al. (2012). are difficult to predict or quantify. Such effects are often lumped
into estimates of internal block strain as three additional hori-
Block Models zontal tensor strain rate parameters included in the inversion for
Block models are analogues of the global plate model applied the Euler vectors. At best these are averages of the true spatially
to continental deformation, and they consider both the rigid rota- varying intra-block strain. At worst they are biased fictions of
tion of fault-bounded blocks on a sphere and the elastic strain the deformation field that influence slip-rate estimates for the
fields about the shallowly locked faults (e.g., McCaffrey, 2002; major faults in unknown ways.
Meade and Hager, 2005). The analysis proceeds in several steps. A major limitation of traditional block modeling is the ab
First, using major active faults as a guide, a mosaic of elastically initio subjectivity in the choice of block geometry, although new
deformable blocks like those shown in Figure 4A is constructed methods are beginning to overcome this deficiency. Specific
A B
42°
Eastern Basin 41°

and Range Province
40°
40°
39°
SNGVB
38°
37°
35°
36°
Pacific Plate 35°

Space geodesy
North American
Plate 50 mm/yr
34°
km 33° 2nd Invariant of strain rate tensor (10 -9/yr)

0 125 250
0 2 6 14 30 72 136 264 5201032 3600
30° 32°
235° 240° 245° 250° -124° -122° -120° -118° -116° -114° -112°
Figure 4. (A) Block model of the western USA (R. McCaffrey, 2012, personal commun.), with block boundaries shown by thin, somewhat linear line segments, Quaternary ac-
tive faults (thin wavy black lines) and GPS sites (red triangles). (B) Map of second invariant of the strain rate tensor for southwestern USA determined by Kreemer et al. (2012).
SNGVB—Sierra Nevada–Great Valley block.
411
case studies have shown that locally varying block geometry or knowledge of either the GPS station locations or the distribu-
can change estimated fault slip rates, sometimes greatly, and tion of active faults. As Simpson et al. (2012) show, cluster analy-
choosing between alternative solutions is usually problematic sis applied to GPS data from the San Francisco Bay Area identi-
(e.g., d’Alessio et al., 2005, their fig. 5; McCaffrey, 2005, his fig. fies four distinct blocks that are bounded by major strands of the
11). Furthermore, there is no easy way to incorporate the often San Andreas fault system (Fig. 5B). Results are similar to those
numerous alternative viable block geometries into the commonly found previously by d’Alessio et al. (2005) using a conventional
used block modeling algorithms, introducing unknown biases block modeling approach and assuming and adjusting an initial
into fault slip-rate estimates. Finally, the block models can be candidate block geometry. A limitation of cluster analysis is that
no more than instantaneous descriptions of the kinematics of a finer block structure with more blocks is not precluded: in fact,
deformation, because the irregular shapes of the blocks will, after randomly distributed velocities can also be clustered. Tests must
many earthquake cycles, accumulate large stresses leading to be applied to verify that the clusters of interest have statistical
additional faulting and changes in block geometry. significance, although the spatial coherence in map view of the
However, a new approach to analyzing GPS velocity fields clustered velocities can also provide a strong visual confirmation
using cluster analysis (Kaufman and Rousseeuw, 1990) shows of significance.
considerable promise in removing the subjectivity in choice of
block geometry (Simpson et al., 2012). Briefly put, this method Strain Rate Maps and Continuum Deformation Models
uses the similarity of groups of GPS velocity vectors to identify An alternative or complementary approach to modeling
groups or clusters in station-velocity space (Fig. 5A). Not sur- space geodetic surface deformation data is to fit it to a spatially
prisingly, the observations forming a velocity cluster often form varying strain rate field. In this method the three components
spatially coherent groupings in map view also (Fig. 5B), thus of the horizontal strain rate tensor are estimated on a regular
defining blocks with no prior assumptions about block geometry grid using the geodetic data and applying some constraints that
SNGVB
Bay Block (BB)

EBB
BB
NC
AL
East Bay PB
Block (EBB)
HA
SA
Sierra Nevada
Great Valley Block
(SNGVB)
A B
Figure 5. San Francisco Bay Area GPS data, analyzed with four clusters. (A) Colors are used to distinguish the clusters. The vectors pointing
to the centers of the large ellipses are the average velocities for each of the clusters. (B) Colors again indicate distinct clusters. Cluster analysis
of the GPS vectors identified four blocks: Pacific, Bay, East Bay, and Sierra Nevada–Great Valley (SNGV). The three fault families separat-
ing these blocks are: SA—San Andreas; HAY—Southern Calaveras, Hayward, Rodgers Creek, Maacama; NCAL—Northern Calaveras,
Concord, Green Valley, Berryessa, Bartlett Springs. Note the tight clustering of velocities in the SNGV block to the east of NCAL. Note also
the consistency of velocity azimuth for the three blocks to the west of NCAL, compared to the velocity of the SNGV block relative to fixed
North America.
Space geodesy 413
interpolate through un-sampled regions. Figure 4B shows one or both rate estimates. Worldwide comparisons (e.g., Thatcher,
example from Kreemer et al. (2012) for part of the western 2009, his fig. 8) show general agreement within data uncertainties,
USA. This particular model assumes some focusing of strain reassuringly validating the two independent approaches in most
near major faults as well as a smoothing constraint on the pre- cases. Nonetheless, a number of apparently well-documented
dicted strain rate field (see Kreemer et al., 2012, for details). differences occur in several regions (Oskin et al., 2008; Peltzer
Results show high strain rates near the San Andreas fault system et al., 2001), raising troubling questions about the uniform-rate
and Eastern California shear zone, very low rates on the Sierra assumption and/or the accuracy of the methods used in obtaining
Nevada–Great Valley microplate, and diffuse but generally low the geologic and geodetic results.
rates in the Basin and Range Province. How can we best evaluate differences between geodetic and
The resulting maps provide a useful synoptic summary of geologic slip rates? We suggest four criteria:
the velocity field that shows where strain is concentrated and 1. Even-handed assessment of random and systematic
where it is more diffuse. Since there are many free parameters in errors. Geodetic block models often produce fault slip-
the model (3N, where N is the number of grid points), the data are rate estimates with small formal errors, but as described
generally fit well. Where active deformation occurs on closely above, uncertainties in block geometry can produce
spaced faults or is widely distributed over many faults, this unknown systematic error. Modelers commonly assume
method can identify such regions and may well provide a better subjective random uncertainties that are larger (say
description of the deformation than block models can supply. ±3 mm/a, Thatcher, 2009), but the true error is often not
Continuum mechanical models that approximate the litho- well constrained. Changes in deformation rates through
sphere as a thin viscous sheet imply a continuous strain rate field the earthquake cycle (see next section) may bias slip-rate
and broadly account for both the kinematics of the deformation estimates, depending on how late in the cycle a fault seg-
and the forces that drive it (England and Molnar, 1997; Flesch et ment is (Chuang and Johnson, 2011; Hearn et al., 2013;
al., 2000; Floyd et al., 2010; Parsons and Thatcher, 2011). The Johnson et al., 2007; Segall, 2002). Contrasting elastic or
continuous strain rate field is, in this case, an imposed feature of viscoelastic properties across block boundaries can also
the model, and this field is consistent with the forces that drive affect slip-rate estimates (Fay and Humphreys, 2005;
the deformation. Schmalzle et al., 2006). Where isolated fault segments
Strain rate maps also have several limitations. Since the spa- are well mapped geodetically, such as along the Carrizo
tial distribution of GPS sites is invariably irregular and the true Plain segment of the San Andreas fault (Segall, 2002), the
strain rate field is unknown, data sampling is spatially aliased far-field offset rate across the fault and its uncertainty can
(e.g., Fig. 4A), and interpolation across data gaps introduces be directly estimated from the GPS or InSAR data, and
unknown uncertainties in strain rate estimation. Indeed, it can be block modeling results agree with these estimates. Geo-
shown that methods using different assumptions about interpola- logic rate estimates are subject to their own systematic
tion and smoothing of the model strain rate field produce radi- errors. For example, Cowgill (2007) showed that implicit
cally differing results, even in southern California (Hearn et al., and explicit assumptions about the timing of fluvial ter-
2010), where the GPS station coverage is relatively dense (see race incision can affect precisely timing initiation of their
Fig. 4A). Finally, because both strain rate maps and thin viscous fault offsets, often leading to significant differences in
sheet models implicitly or explicitly smooth the deformation fault slip rates of several faults on the Tibetan Plateau and
field, these models cannot provide useful information on geodeti- elsewhere. Geologic estimates are often made in trenches
cally constrained fault slip rates. cut only ~10 m across active faults, and more widespread
off-fault deformation has been documented from both
Relation of Tectonic Geodesy to Tectonic Geology coseismic offset results (Lazarte et al., 1994; Thatcher
and Lisowski, 1987) and geologic mapping across active
One of the chief goals of active fault geology is to map faults faults (Oskin et al., 2008).
and determine late Quaternary (<125 ka) and Holocene (<10 ka) 2. Are slip rate estimates made by more than a single geo-
slip rates across them. If we provisionally assume that rates of detic or geologic method? Single-site slip rate estimates
deformation are constant on these time scales (which may not could be correct, but multiple determinations at several
universally apply), then geodetic model-based measures of fault different sites that agree provide more confident rate esti-
slip can be directly calibrated against those estimated from tec- mates. For example, GPS slip rate estimates for the Altyn
tonic geology. Geodetic estimates of fault slip rates from inter- Tagh fault (Bendick et al., 2000; Hilley et al., 2009) agree
seismic deformation rates depend on assumptions made when with InSAR based estimates from a different site on the
modeling the elastic strain field about shallowly locked faults, same fault (Jolivet et al., 2008) and with block model-
while geologic estimates depend on assumptions on the age of ing results (Thatcher, 2007), all yielding slip rates of
formation and development of offset geologic features. Where ~9 mm/a. These geodetic rates are much smaller than
slip rate differences occur they may indicate temporal changes earlier estimates of 26–31 mm/a based on geological
or significant but unknown random or systematic errors in either methods (Mériaux et al., 2004), which were questioned
and revised by Cowgill and colleagues (Cowgill, 2007; specified time spans. The method then stands or falls on the basis
Cowgill et al., 2009) to rates of 9–14 mm/a, consistent of the accuracy of the original strain rate maps, whose strengths
with the geodetic estimates. and shortcomings have been described above.
3. Is there a mechanically viable quantitative model pro-
posed to explain rate changes? Several studies have THE EARTHQUAKE DEFORMATION CYCLE
proposed significant rate changes on active faults during
Holocene time (Dolan et al., 2007; Oskin et al., 2008; The first conceptual model of the earthquake cycle informed
Peltzer et al., 2001), but none of these investigators have by geodetic data, Reid’s (1910) elastic rebound theory, envi-
presented quantitatively testable models that account for sioned long-lasting periods of steady elastic strain accumulation
these changes. Rate switching over Holocene and late punctuated by sudden release by an earthquake rupture. Dis-
Quaternary time has been postulated for the San Jacinto location theory (Steketee, 1958) provides a simple mechanical
fault and the San Andreas fault in California (Bennett model for such a system in which steady aseismic slip on the
et al., 2004), but more recent geologic studies have put downdip extension of a locked rupture zone (Savage and Bur-
the proposed slip-rate changes in question (Blisniuk et ford, 1973) loads the earthquake source region to eventual fail-
al., 2010). ure and associated elastic deformation of the surrounding crust
4. Is a proposed rate change mechanism consistent with (Chinnery, 1961). Interseismic slip to infinite depth in this model
examples of changes in style and sense of deformation represents a simplified kinematic equivalent of the actual load-
preserved in the geologic record? For example, ca. 1 Ma ing mechanism and produces the observed asymptotic approach
shifts in activity from one active strand of the San of the fault slip rate at large distances from the fault. Over a
Andreas fault system to another nearby strand have been full earthquake cycle, the elastic deformation fields from steady
documented in southern California (Powell and Weldon, deep interseismic and episodic shallow slip (Fig. 1C) combine
1992). Inversions from normal to thrust faulting over the to produce block offset without cumulative deformation of the
past ca. 5 Ma have been observed for backarc faults in adjoining blocks, consistent with the long-term fault slip rate.
both New Zealand and Japan (Sibson, 2009). This approach underlies the block-modeling method to deter-
mine geodetic fault slip rates described in the previous section.
Relation to Seismic Hazard Assessment The recognition of accelerated deformation in the decades after
the great 1906 San Andreas fault earthquake rupture suggests
Because, other factors being equal, fault slip rate and strain that in the years after a large event, the deep-seated shear is
rate are directly related to earthquake occurrence, inferences accelerated owing to the sudden stress increase from the overly-
of slip and strain rate from space geodetic data can be impor- ing earthquake rupture (Thatcher, 1975).
tant inputs to seismic hazard assessment. In fault-slip-rate- High temperatures in the lower crust and upper mantle allow
based hazard assessment (e.g., Working Group on California for viscous behavior of rocks, and thus an alternative view of the
Earthquake Probabilities, 2008), these slip rates are combined deformation below earthquake faults involves a viscous substrate
with a model that specifies the long-term rate of all possible participating in the earthquake cycle (Nur and Mavko, 1974). The
damaging earthquakes in the region of interest (e.g., based simplest viscoelastic earthquake cycle model includes periodic
on Gutenberg-Richter scaling or the characteristic earthquake earthquake slip on a dislocation in a shallow elastic layer repre-
model; Schwartz and Coppersmith, 1984), finally giving the senting the seismogenic zone, overlying a half-space with a visco-
probability that each earthquake in the rate model will occur elastic Maxwell material, which features a combination of linear
during a specified time span (say 10, 20, or 30 years). This elastic and viscous responses to stress (Savage and Prescott, 1978).
method is most useful when there is a large inventory of reli- Interpretation of geodetic data is being challenged by the recog-
able geologic and geodetic slip rate estimates (e.g., California nition that dislocation and viscoelastic models of the earthquake
and the western USA, New Zealand, and Japan), but it is still cycle can be parameterized to produce similar surface deforma-
subject to the biases in the slip rate estimates described above tion patterns, especially in the two-dimensional, anti-plane strain
and the validity of the earthquake rate model. case relevant for strike-slip faults (Savage, 1990). Even today,
In hazard assessments that depend on strain rate maps, these contrasting first-order views involving deep afterslip on the
the method of Kostrov (1974) is used to convert strain rates to fault surface versus viscous relaxation (Thatcher, 1983) form the
moment accumulation rates (or potency rate, which does not starting point of explorations of the earthquake cycle informed
depend on assumed crustal shear modulus). This is accom- by space geodetic data. For simplicity and ease of calculation,
plished by averaging strain rates over a specified area, assuming these early earthquake models involved two-dimensional (cross-
a thickness to the seismogenic crust, and using an assumed shear sectional) implementations, but finite three-dimensional para-
modulus (for moment rate) or not (for potency rate). Then the meterizations are relatively straightforward (Okada, 1985; Pollitz,
hazard assessment follows, using algorithms analogous to those 1992). As observational constraints on fault-related deformation
that start with fault slip rate, assuming an earthquake recurrence improve, increasingly sophisticated mathematical models of the
model and calculating probabilities of event recurrence within earthquake cycle can be pursued (Segall, 2010).
Space geodesy 415
Improved space geodetic methods reveal the detailed spa- (Hearn and Bürgmann, 2005), lateral elastic heterogeneity (Mas-
tiotemporal pattern of crustal deformation. These data lead to terlark, 2003), or the spherical shape of the Earth (Banerjee et al.,
improved insights into the processes involved in earthquake- 2005). Coseismic slip models provide insights into the mechanics
cycle deformation and the rheology of the crust and mantle. In of earthquake slip, but they also represent the starting points of
this section we explore how space geodetic data have (1) allowed investigations of subsequent postseismic deformation transients.
for increasingly detailed kinematic models of coseismic earth-
quake ruptures, (2) revealed a plethora of postseismic relaxation Aseismic Fault Slip Transients
processes, and (3) informed postseismic modeling studies aimed
at resolving the in situ constitutive properties of rocks in the Not all fault slip occurs by dynamic rupture that produces
lower crust and upper mantle. seismic waves, but instead some involves slow fault creep. The
improved monitoring of plate boundary zones with spatially and
Kinematics of Earthquake Ruptures temporally dense geodetic measurements has revealed a broad
spectrum of aseismic slip transients in the shallow seismogenic
The development of GPS and deformation remote sensing zone, as well as deep in the transition region to fully aseismic and
with InSAR rapidly increased the number, precision, and resolu- ductile deformation. While some slow slip transients represent-
tion of geodetic observations of earthquake deformation fields ing afterslip caused by stress increases from a nearby seismic
around the world, and the number of earthquakes, small and event (described in the next section), many appear to be spon-
large, captured by space geodesy probably exceeds one hundred taneous. The size, rate, and duration of such slow slip transients
at this time. Several dozen earthquakes have been captured with span a wide range, and more study is needed to fully explore the
InSAR (Weston et al., 2011), and the growing number of con- kinematics and underlying frictional mechanics of aseismic fault
tinuous GPS networks along many active plate boundary zones is slip transients.
further contributing to the database of coseismic ruptures. Relatively small slow slip transients were first recognized
The coseismic surface deformation field can be inverted for with surface creepmeters, borehole strainmeters, and two-color
detailed models of the geometry and slip distribution of an earth- EDM on sections of the creeping segment of the San Andreas
quake rupture, using numerical optimization techniques aimed fault (Langbein et al., 1999; Linde et al., 1996). Some of these
at finding dislocation parameters that best fit the observations events take place in a matter of a few days (Linde et al., 1996),
(Arnadottir and Segall, 1994). Most kinematic rupture models but more enduring aseismic slip accelerations can continue for
involve slip on multiple rectangular or triangular dislocation ele- several years (Gao et al., 2000; Langbein et al., 1999; Mur-
ments in a uniform elastic half-space, with the detail of the slip ray and Segall, 2005). While most of the slip in these shallow
solution depending on the precision and spatial sampling of the transients is by aseismic creep, they are often accompanied by
geodetic measurements. To avoid mechanically implausible and/ small earthquakes, including identically repeating events that
or overly rough model slip distributions, constraints on the slip- represent seismic failure of small asperities embedded in the
vector orientation and spatial smoothing are commonly applied otherwise aseismically creeping fault surface (Nadeau and
in the inversion. Resolution tests are important to determine to McEvilly, 1999). Using InSAR, Lohman and McGuire (2007)
what level of detail slip on a rupture can be constrained and how showed that an intense swarm of small earthquakes in southern
much smoothing should be applied to a slip solution. In general, California accompanied a shallow slow-slip event, suggesting
the resolution of slip decreases with depth and distance from the that often earthquake swarms more generally reflect a transient
surface deformation measurements. By complementing the geo- slip episode.
detic data with constraints from seismic waveforms, it is possible Improved GPS and seismic monitoring has revealed the
to further improve the model result and to constrain additional ubiquity of slow slip transients on subduction-zone megath-
information about the nucleation and slip history of the coseismic rusts. Particularly interesting has been the recognition of exten-
rupture (Kim and Dreger, 2008; Wei et al., 2011b). sive slow-slip transients downdip of the seismogenic zone of a
The most recent major earthquake in the San Andreas fault number of subduction zones (Beroza and Ide, 2011; Peng and
system was the 2010 M 7.2 El Mayor–Cucapah earthquake of Gomberg, 2010; Schwartz and Rokosky, 2007). These deep slip
Baja California in Mexico. The event was captured by the con- transients sometimes recur at regular intervals and are accompa-
tinuous GPS network to the north of the U.S.-Mexico border, but nied by a new class of seismic signal, so-called “non-volcanic
GPS coverage in Baja California was limited. A combination of tremors,” which help to further illuminate their propagation in
InSAR, SAR, and satellite image pixel offset, together with seis- space and time (Obara, 2004; Rogers and Dragert, 2003). While
mic waveform data, allowed for determination of a detailed dis- fault offsets in these deep-seated slow-slip transients are only on
tributed slip model of this complex rupture along multiple fault the order of a few centimeters, they can involve very large fault
segments (Wei et al., 2011b; Fig. 6). areas of several thousand square kilometers. The characteriza-
In some cases the high quality of data and/or size of an earth- tion of numerous slow slip events of various sizes showed that
quake allow for the consideration of higher order model features, slow earthquakes have scaling characteristics that are different
such as increasing rock rigidity with depth in the lithosphere from those of ordinary earthquakes. Whereas earthquakes grow
416
−116°00' −115°30' −115°00' −114°30'

A HNPS
B SE NW
ACT SAR Azimuth Offset Data SPOT Offset Data
5
33°30' 4
DHLG
3
LMS GLRS
2
CRRS GMPK
USGC 1
33°00'
along strike offset (m)

0
-10 0 10 20 30 40 50 60 70 80 90 100 110 120
F4 NW Distance (km)
32°30'
F2 0
32
radar look 10 22 32
32 22
38°
Depth (km)
F1
22
22
32
F4
Bürgmann and Thatcher
F3 F2
32°00' cm
LOS motion
ion
(meters) F3 C 10 0 100 200 300 400 500 600
F1
ALOS PALSAR paths A211-212
Figure 6. Coseismic deformation and slip of the 2010 M 7.2 El Mayor–Cucapah earthquake (modified from Wei et al., 2011b). (A) Example of unwrapped ascending orbit ALOS
interferograms spanning the earthquake, showing displacements in the LOS direction with an average incidence angle of 38° (Eric Fielding, 2012, personal commun.). Black arrows
show horizontal coseismic displacements of continuously operating stations of the PBO network. Rectangles show the surface projection of the model rupture segments, and the star
is the epicenter. (B) Coseismic surface offsets along the rupture trace, estimated from pixel offsets of ALOS SAR amplitude (blue) and SPOT optical (red) images. (C) Preferred
coseismic slip model of Wei et al. (2011b), inverted from GPS, ALOS, and Envisat interferograms, SAR azimuth offsets, SPOT optical satellite imagery pixel offsets, and seismic
waveform data. Color contours and arrows show cumulative slip amplitude and slip direction, respectively, and contour lines (seconds) are isochrons illustrating the propagation of
the seismic rupture front relative to the onset of slip at the hypocenter obtained from seismic data.
Space geodesy 417
in seismic moment with the cube of their duration, slow earth- larly suitable tool to observe shallow afterslip, triggered slip, and
quakes seem to exhibit a linear relationship between their seis- fault creep transients independent of prior earthquakes (Johanson
mic moment and duration (Ide et al., 2007). Slow slip and tremor et al., 2006; Wei et al., 2009). These data can be used to constrain
are quite easily triggered or modulated by very small changes the rate-state frictional parameters of aseismic fault slip, which
in stress (a few kilopascals or less) owing to teleseismic waves, are important for improved understanding of earthquake nucle-
tides, atmospheric pressure changes, seasonal hydrological loads, ation and slow slip transients (Hsu et al., 2006; Johnson et al.,
and static stress changes from tectonic or volcanic deformation 2006; Perfettini et al., 2005).
(Peng and Gomberg, 2010). This susceptibility to small changes Afterslip deep in the aseismic parts of the crust has been
in loads suggests a weak fault rheology, possibly due to near- inferred from GPS and InSAR measurements of some earth-
lithostatic fluid pressures in the fault zone. quakes, but it can be challenging to separate the contributions from
Scientific interest in slow slip transients is also great, because deep afterslip and viscous flow (Freed et al., 2006b; Hearn et al.,
such events may trigger or represent the final nucleation process 2002; Ryder et al., 2007). To diagnose deep afterslip, it is impor-
of destructive earthquakes. For example, seafloor pressure sen- tant to have precise and well-distributed measurements of three-
sors, GPS data, and seismicity patterns revealed that an ~3-week- dimensional deformation surrounding a finite earthquake rupture
long episode of small and large foreshocks and aseismic slip near (Hearn, 2003). It can also be useful to consider if observed post-
the hypocenter of the 2011 M 9 Tohoku earthquake propagated seismic transients can be matched by physical models of stress-
toward the initial rupture point and eventually prompted the driven deep afterslip, rather than just by kinematic, optimized
unstable dynamic rupture of the mainshock (Kato et al., 2012; distributed-slip models (Hearn et al., 2009; Johnson et al., 2009).
Miyazaki et al., 2011; Ohta et al., 2012). It is not clear how com- Thanks to comprehensive postseismic GPS measurements, it has
mon such precursory slip transients are, or how often similar been possible to confirm contributions of deep afterslip early after
events occur without being followed by large earthquakes. the 1999 Izmit, Turkey, earthquake (Hearn et al., 2009), whereas
substantial deep afterslip following the 1992 and 1999 Landers
The Spectrum of Postseismic Relaxation Processes and Hector Mine, California, earthquakes can probably be ruled
out (Freed et al., 2007; Pollitz et al., 2001). Pollitz et al. (2001)
Earthquake ruptures change the stress field in the surround- utilized GPS and InSAR data spanning the first nine months fol-
ing lithosphere. A number of postseismic processes act to relieve lowing the M 7.1 1999 Hector Mine earthquake to show that the
the imposed loads, leading to transient surface deformation. Space observed quadrant uplift pattern surrounding the rupture is oppo-
geodetic observations of post-earthquake deformation transients site to that expected for localized shear beneath the earthquake
reveal an increasingly complex image of the earthquake cycle, rupture, but is consistent with viscous relaxation in the lower crust
involving multiple processes that go well beyond the end-member and upper mantle (Fig. 7B). The lower crust was found to have an
models considered in early investigations of postseismic deforma- ~30 times higher viscosity than the upper mantle (Fig. 7C).
tion (Thatcher, 1983). Models of these relaxation processes can The discrimination of discrete afterslip versus distributed
reveal valuable information about the rheology of, and fluid flow viscous flow is easier for earthquakes with a significant dip-slip
in, rocks and fault zones. As a number of processes contribute component, as the spatial patterns of the predicted deformation
to postseismic deformation, including aseismic afterslip, viscous fields differ substantially. For example, broad postseismic uplift
flow of rocks, aftershocks, poroelastic rebound, and fault zone following normal-faulting earthquakes in the Basin and Range
dilatancy, it is essential to find diagnostic patterns in the surface Province rule out significant deep afterslip on or below the rup-
deformation that allow for separating the contributions of the dif- ture (Gourmelen and Amelung, 2005; Hammond et al., 2007;
ferent processes. Successful space geodetic explorations of post- Nishimura and Thatcher, 2003). In contrast, the first 15 months
seismic deformation transients rely on detailed knowledge of of deformation following the 1999 Chi-Chi, Taiwan, thrust earth-
(1) the pre-earthquake “background” deformation field, (2) the quake appear to have been dominated by afterslip on a detach-
coseismic rupture geometry and slip distribution, and (3) the com- ment downdip of the coseismic rupture (Hsu et al., 2007).
plete spatial and temporal distribution of three-dimensional defor- Viscoelastic relaxation occurs in the lower crust and upper
mation resulting from the postseismic relaxation processes. mantle, where hot rocks respond to coseismic stress changes by
Afterslip occurs on the same fault or faults that ruptured dur- viscous, crystal-plastic flow that relaxes the applied stress and
ing an earthquake, usually immediately surrounding and deep thus reflects the rheology of rocks at high temperatures and
below a rupture. Afterslip in the upper crust suggests that some pressures. Model exploration of postseismic transients puts con-
faults are velocity-strengthening: that is, they are unlikely to fail straints on the depth extent and rheology (equations describing
dynamically and generally slip aseismically at rates depending the relationship between applied stress and resulting strain) of
on the surrounding stresses (Marone et al., 1991). Triggered the relaxing regions. For a linear viscous Maxwell rheology,
slip also may occur on neighboring faults, sometimes even at the relaxation time is defined as the ratio of twice the viscos-
large distances from the coseismic rupture, and is often induced ity over the shear modulus of the relaxing material. Thus, higher
by seismic shaking rather than static stress changes (Wei et al., viscosities and lower rigidities of rocks at depth result in slower
2011a). The dense spatial mapping of InSAR makes it a particu- postseismic deformation transients. If relaxation times are long
relative to the earthquake recurrence times, surface deforma- quakes produce the most enduring and far-reaching deformation
tion will be relatively steady with time throughout much of the transients from viscous relaxation in the upper mantle. Trench-
earthquake cycle (Savage and Prescott, 1978; Segall, 2002). ward motions of GPS stations in the continental interiors indicate
Postseismic studies across the plate boundary zones of western that even four decades after the great Mw 9.3 1964 Alaska and Mw
North America generally find that viscous relaxation dominantly 9.5 1960 Chile megathrust ruptures, deformation rates are strongly
occurs below ~40–60 km in the upper mantle (see references in perturbed by the continued flow of rocks in the underlying mantle
recent reviews by Bürgmann and Dresen, 2008; Hammond et al., (Khazaradze and Klotz, 2003; Zweck et al., 2002). Continuous
2007; Thatcher and Pollitz, 2008). Here, and in many other for- GPS network and GRACE gravity measurements following more
mer or present backarc regions, the lower crust is found to have recent M≈9 megathrust ruptures in Sumatra, Chile, and Japan are
a substantially higher viscosity than the asthenospheric mantle. capturing the early phases of postseismic relaxation and are provid-
This suggests substantial weakening and/or removal of the litho- ing further information about the rheological properties of the man-
spheric mantle lid and/or localized brittle deformation and duc- tle asthenosphere (Hu and Wang, 2012; Panet et al., 2010; Pollitz
tile flow on the downward extension of major faults in the lower et al., 2008a). Models parameterized based on these observations
crust (Shelly and Hardebeck, 2010). suggest the possibility that even >300 years since the 1700 M≈9
The duration and spatial extent of postseismic deformation Cascadia subduction earthquake, surface velocities across much
depend on event magnitude, and therefore great subduction earth- of the western USA are perturbed by viscous relaxation transients
A C
Figure 7. Early postseismic observations and models of deformation, following the M 7.1 1999 Hector Mine earthquake, by Pollitz et al. (2001).
(A) Map showing historic earthquakes in the Mojave Desert from 1992 to 2000. (B) The left and right panels show predictions of line-of-sight
(LOS) displacement from viscoelastic and afterslip models, respectively, compared to the wrapped ERS-2 interferogram in the middle, spanning
20 October 1999–21 June 2000. One color cycle represents 28 mm of LOS ground displacement. (C) Panels in upper right evaluate misfit of in-
dependent InSAR (left) and GPS (right) data sets as a function of mantle viscosity (ηm) and crust-to-mantle viscosity ratio (ηc/ηm). The viscosity
combination used to generate the viscoelastic forward model in bottom left (B) is shown with a black circle.
Space geodesy 419
(Pollitz et al., 2008b, their fig. 19). Figure 8 shows current veloc- restraining and releasing steps, respectively (Peltzer et al., 1996).
ity fields in the upper plate regions of the 2011 Tohoku, Japan; Evidence for substantial poroelastic rebound has been found with
1960 Valdivia, Chile; and 1700 Cascadia earthquakes, illustrating InSAR following some events (Jacobs et al., 2002; Jónsson et al.,
the first-order deformation patterns associated with early, interme- 2003) but ruled out after some others (Barbot et al., 2008), sug-
diate, and late stages of the subduction earthquake cycle. Linear gesting that hydrological conditions vary widely among tectoni-
viscous-strength estimates for the upper mantle in the hanging cally active regions.
wall of these and other subduction thrust earthquakes average Fault-zone dilatancy recovery is another relaxation pro-
~1019 Pa s at depths >60 km, similar to the values found in the cess that was captured by InSAR mapping of near-fault-surface
western USA (Wang et al., 2012a). These model explorations of deformation following the 2003 Bam, Iran, earthquake (Fielding
space geodetic measurements show that the mechanical litho- et al., 2009). A narrow zone of subsidence along the fault sec-
sphere in plate boundary zones is greatly thinned compared to tion with the greatest coseismic slip at depth can be explained
values >100 km for continental interiors found in geodetic studies by the recovery of damage and fault dilatation produced by the
of isostatic rebound from glacial unloading (Milne et al., 2001; earthquake. Time series analysis of more than a hundred SAR
Sella et al., 2007; Tamisiea et al., 2007). In the next section, we interferograms showed that fault-zone compaction in the upper
describe further insights about rheological properties in the lower kilometer of the fault zone evolved logarithmically over the
crust and upper mantle gained from model studies of the tempo- 3-year measurement period.
ral and spatial evolution of postseismic deformation captured by
space geodesy. Probing Rock and Fault Rheology with Earthquake
Poroelastic rebound is due to movement of pore fluids Cycle Deformation
through rocks to re-equilibrate pore-pressure changes caused by
an earthquake. This process leads to postseismic deformation Modeling studies of postseismic deformation represent rock-
that is measurable at the Earth’s surface. It was first discovered mechanics experiments of lithospheric dimensions. Increasingly
in the InSAR-measured deformation field of the 1992 Landers, detailed characterization of the spatial and temporal patterns of
California, earthquake, where postseismic displacements in fault postseismic deformation with space geodetic observations helps
stepovers partly recovered the coseismic uplift and subsidence in put further constraints on the rheological properties or constitutive
A NE Japan B Chile C Cascadia

35˚S 50˚N
GPS GPS
100 km 40_
+4 cm 20_
+2 mm/yr
40˚N m/yr
66 m
m/yr
36m
40˚S
100 km
45˚N
100 km
03/11/2011 Mw9
45˚S GPS
35˚N 72
mm 20 mm/yr
/yr
18 mm/yr
140˚E 145˚E 75˚W 70˚W 125˚W 120˚W
Figure 8. The subduction megathrust earthquake cycle. Panels show contemporary geodetic observations of motions relative to the plate in
the hanging wall measured with GPS. (A) Following the 11 March 2011 Tohoku, Japan, earthquake, where all sites moved seaward during
the first half year after the earthquake (GPS data provided by Geospatial Information Authority, GSI, of Japan). (B) Forty years after the
Mw 9.5 Valdiya, Chile, earthquake, with coastal and inland sites moving in the direction of plate convergence to the north, but in opposite
directions in the hanging wall of the rupture (Wang et al., 2007). (C) Three hundred years after the 1700 Mw 9.0 Cascadia earthquake, where
all sites move landward (McCaffrey et al., 2007). Stars are epicenters. Contours show coseismic slip distributions estimated for the three
events (Yan Hu, 2012, personal commun.; Wang et al. (2012a).
equations of rocks at depth. The nature and magnitude of the Large earthquakes often occur where the lithospheric geol-
postseismic response depends on the distribution of tempera- ogy, rheology, and temperature vary strongly across the fault sys-
ture, rock type, stress, and fluids in the lithosphere, and is thus tem, suggesting that postseismic transients probe laterally varying
highly variable depending on the particular tectonic environment, rheology. For example, subduction zones separate oceanic litho-
lithospheric structure, and geologic history of the study region. sphere that is about to be subducted from upper-plate regions of
Precise geodetic measurements and modeling of the postseismic variable makeup. Continental fault zones also commonly bound
deformation resulting from the viscous relaxation of earthquake lithospheric blocks of different composition, thickness, and ther-
stresses thus allow us to probe the rheological properties of rocks mal conditions. Well-distributed and precise geodetic measure-
deep in the Earth and to determine appropriate mathematical for- ments help to map out such first order variations in rheology. Hu
mulations of in situ, real-time stress-strain relations. Postseismic et al. (2004) found that vertical deformations measured with tide
afterslip and spontaneous or triggered slow slip events similarly gauges, following the 1960 M 9.5 Chile earthquake, are better
reveal important information about the rheology of faults, often matched if the viscosity below the incoming oceanic plate is an
modeled in the framework of rate-and-state friction. order-of-magnitude higher than in the continental upper mantle.
GPS or InSAR time series of deformation following several Other studies have explored three-dimensional models of sub-
large recent earthquakes suggest that effective model viscosities duction zones (Hu and Wang, 2012; Masterlark et al., 2001; Pol-
of rocks in the lower crust and/or upper mantle increase with time litz et al., 2008a), but it is difficult to obtain diagnostic data con-
to match the rapidly decaying surface deformation rates (Pollitz, straints on the varying mechanical properties of the oceanic plate
2003, 2005; Ryder et al., 2007). A bi-viscous model material (the from onshore data. Continued monitoring of seafloor GPS-A
so-called Burgers body), which includes a component exhibiting stations offshore of the 2011 M 9 Tohoku earthquake rupture,
transient low viscosities evolving to a higher, steady-state flow in combination with the dense on-land GPS network in Japan,
strength, matches this pattern (Hetland and Hager, 2005; Pollitz, promises to provide much improved information on differences
2003, 2005). Rock-mechanics experiments at high temperatures in rheological properties across the subduction plate boundary in
and pressures suggest that rocks in the lower crust and upper man- northeastern Japan.
tle deform by dislocation creep, for which strain rate is related to In central Asia, Ryder et al. (2011) found lower postseismic
stress to a power n, where n is found to range between 2 and 5 deformation rates to the north of the 2001 M 7.8 Kokoxili earth-
depending on composition (Bürgmann and Dresen, 2008). Post- quake rupture, suggesting that the strike-slip rupture occurred in
seismic relaxation models employing power-law relaxation param- a zone separating the Tibetan Plateau, with its thicker, weaker
eters determined experimentally for steady-state flow of mantle crust, from a stronger Qaidam Basin block to the north. Pollitz et
olivine match the postseismic deformation time series following al. (2012) used GPS time series of the postseismic deformation
several recent earthquakes (Freed and Bürgmann, 2004; Freed et transients from the 2010 M 7.2 El Mayor–Cucapah earthquake
al., 2006a; Hearn et al., 2009). In these models, background stress in Baja California to document lower viscosities of the upper
levels have to be very low for the coseismic stress increase to lead mantle below the Salton Trough, California, Province, bounded
to the strongly diminished effective flow strength inferred from the by stronger lithospheric blocks on either side of the southernmost
early deformation transients (Freed et al., 2010; Hearn et al., 2009). San Andreas fault system. Consistent changes in seismic veloci-
Freed et al. (2012) find that the fit to the deformation tran- ties in the upper mantle, and heat flow across the plate boundary,
sients captured at far-field (>50 km from rupture, Fig. 9A) con- lend further support to this finding (Pollitz et al., 2012). Precise
tinuous GPS stations following the 1999 Hector Mine earthquake measurements of surface deformation across sections of the San
can be further improved by combining steady-state power-law Andreas fault system that are currently in the late, interseismic
flow and transient relaxation into a single constitutive relation phase of the earthquake cycle can also reveal signatures of lateral
(Fig. 9B). Their model results suggest a transient creep phase that variations in elastic and visco-elastic properties (Fay and Hum-
lasts for about a year and has an effective viscosity an order of phreys, 2005; Malservisi et al., 2001; Schmalzle et al., 2006).
magnitude lower than subsequent steady-state flow (red line in As the active deformation mechanisms and their rheological
Fig. 9B), consistent with laboratory observations (Freed et al., properties depend on a large number of parameters and condi-
2012). Linear diffusion creep is ruled out, as the small grain size tions (including composition, temperature, stress, grain size, and
required to match the rapid decay of surface velocities (orange water content), it is a challenging but worthwhile effort to explore
line in Fig. 9B) conflicts with petrological information. Mod- space-geodetic and complementary geologic and seismologic
els assuming nonlinear dislocation creep of olivine in the upper data sets to explore the in situ rheology of the lower crust and
mantle match the surface motions, assuming pre-earthquake upper mantle.
strain rates close to interseismic observations and a geothermal The spatio-temporal evolution of aseismic slip also provides
gradient consistent with surface heat-flow constraints (Fig. 9B). valuable information about the rheological properties of faults.
If postseismic relaxation in the mantle is by power-law creep, the The description of frictional behavior of faults has been devel-
increasing stress following the earthquake will lead to a transient oped empirically from laboratory rock-mechanics experiments,
reduction of viscosity that slowly recovers as stresses decay to which find that the frictional fault strength varies with the rate
background values (Fig. 9C). of fault slip and the evolving state of a fault (Dieterich, 2007;
Space geodesy Coseismic stress change 421

Yucca Mountain
A C
.05
Shear stress (MPa)

GPS 7-year .03
10 mm postseismic SMYC .01
displacements
Calculated based -.01
on relaxation in Ne -.03
Ca v
the upper mantle lifo ada
rn -.05
ia
-.10
7-yr Postseismic stress change
.010
.005
Shear stress (MPa)

.003
Hector .001
Mine
-.001
Landers
-.003
-.005
LA San
Andreas -.010
Fault Viscosity 1 month after earthquake
crus 0 km
t
50 km 5e20 Viscosity (Pa s) man
tle 40
3e20
8
B β=10, d=3.5 mm 1e20
80
β=10, εb=0.1 5e19
3e19 120
6
1e19
SMYC
5e18
Observed
4 Long-term background viscosity
litho 0 km
sph
5e20 ere
Viscosity (Pa s)
40
2 3e20
asth
β=10, εb=0.05 1e20
eno
sph
ere 80
β=1, εb = 0.1 5e19
0 3e19 120
0 1 2 3 4 5 6 7 1e19
Years after quake 5e18
Figure 9. Insights on mantle rheology and deformation mechanisms from enduring and far-reaching deformation transients of the
1999 Hector Mine earthquake (Freed et al., 2012). (A) Cumulative postseismic surface displacements (transient component in ex-
cess of pre-earthquake rates) from seven years of GPS measurements following the earthquake. Gray arrows are model predictions,
assuming dislocation creep of wet olivine in the upper mantle below 50 km. LA—Los Angeles. (B) Observed (black) and modeled
time series at station SMYC in southwestern Nevada. Models with a transient viscosity 10 times lower than steady-state (β = 10)
can capture the rapid decay of observed GPS velocities. A linear bi-viscous diffusion creep model (orange line) can match the
displacements but requires very small olivine grain size (d = 3.5 mm). Dislocation creep models match the surface motions if pre-
earthquake background strain rates, εb, are of order 0.1 μstrain/a (red line). Dislocation creep models without transient relaxation
(blue) or lower background strain rates (cyan) do not match the observed time series. (C) Co- and postseismic model shear stress
changes and post-earthquake and long-term viscosity structure inferred from finite element models of the postseismic relaxation are
by Freed et al. (2012). In power-law materials, deforming by dislocation creep, the transiently decreased effective viscosity below
the rupture regains steady-state background levels as the coseismic stress increases decay to long-term values. Results are shown
for the best fitting model based on dislocation creep of wet olivine with a power-law stress exponent of n = 3.40, a steady-state to
transient viscosity ratio of β = 10, and a transient relaxation time of τ = 1 a (red line in B).
Marone, 1998). Detailed geodetic observations of afterslip, epi- California (Shelly and Hardebeck, 2010) argues that this fault
sodic slow slip events, and triggered slip allow for testing these is a narrowly focused shear zone to the base of the California
empirical constitutive relationships on natural faults. Where crust (~25 km). Thus, at least in this region, it seems likely that
the contribution of aseismic fault slip to postseismic deforma- the lower crust behaves largely elastically rather than deform-
tion is well resolved, the spatio-temporal evolution of afterslip ing by bulk ductile flow. If we assume that fault slip rates are
in response to coseismic stress changes can be used to constrain unchanged over geodetic and (say) Holocene time scales, we
fault rheological properties. For example, the evolution of the may use Holocene geologic slip-rate estimates to bound geo-
shallow afterslip following the 2004 M 6 Parkfield earthquake detic model slip rates (e.g., McCaffrey, 2005; Segall, 2002). All
suggests rate-state frictional parameters near the transition from of these seismological and geological data also lead to better
potentially unstable to stable friction, which is consistent with the bounds on much more poorly known earthquake cycle para-
range of experimental values reported for serpentinite that crops meters such as sub-crustal effective viscosity. Further comple-
out along the San Andreas fault, and which may be the cause of mentary information may come from structural-geologic field
its dominantly aseismic behavior in this region (Johnson, 2006). observations and rock-mechanical laboratory data, which pro-
Dynamic models of fault slip at Parkfield through the coseismic, vide insights into the distribution of deformation processes
postseismic, and interseismic phases of the earthquake cycle and their constitutive properties that we try to capture with our
help to map out the varying frictional properties along the San deformation models.
Andreas fault, and promise to lead to a more complete under- An entirely different approach to the non-uniqueness prob-
standing of the physics of active faults (Barbot et al., 2012). Rate- lem is to use Bayesian statistical methods to generate probabil-
state friction models have also been used to explore spontaneous ity density functions (pdfs) for each model parameter. In this
transient-slip events below the locked megathrust zone. To suc- method a large number (e.g., millions) of multi-parameter model
cessfully model the observed pattern of slow slip, fault properties solutions consistent with the data, and loose a priori bounds on
should be near the transition from stable to unstable sliding, and these parameters, are generated. For each model parameter (say,
fluid pressures in the fault zone should be near lithostatic val- coseismic slip, fault slip rate, effective viscosity), a plot of the
ues (Liu and Rice, 2007). Under these conditions it is also found number of times that parameter lies within a given range for
that transients are more easily triggered by rather modest stress all the solutions provides a pdf for each model parameter (e.g.,
perturbations associated with teleseismic waves, tides, and other Johnson et al., 2007; Minson et al., 2013). The pdfs thus provide
external loads. It is also possible that dilatancy during an accel- an easily grasped graphical depiction of the range of acceptable
erating slip event rapidly decreases fluid pressures and thus pro- parameters and their most likely values.
vides an additional mechanism to stabilize fault slip and produce
the observed slow-slip behaviors (Segall et al., 2010). Increas- Block Versus Continuum Active Deformation
ingly detailed geodetic observations of such slow-slip transients
will surely further improve our ability to model and understand Geodetic data provide clear examples of where either block
these intriguing events and their possible role in initiating some or effectively continuum deformation occurs. For example, the
large earthquakes. Tarim Basin in China (Avouac and Tapponnier, 1993), the Sierra
Nevada Great Valley microplate in California (Argus and Gor-
OUTSTANDING PROBLEMS FOR THE FUTURE don, 2001), and the South Aegean region of Greece (Nyst and
Thatcher, 2004; Reilinger et al., 2006) are not currently deform-
Non-Uniqueness of Inferred Deformation Sources at Depth ing measurably except on or near their faulted boundaries. In
contrast, the Ventura and Los Angeles Basins in California and
Inverting surface displacement data to estimate the location, parts of western Anatolia in Turkey (Aktug et al., 2009) consist
geometry, and other parameters of deformation sources at depth of closely spaced faults with comparable (or unknown) slip rates,
is inherently non-unique, making any particular solution subject where block descriptions are not useful and deformation may
to unavoidable uncertainties. How can we then increase our con- best be approximated as continuous. Denser data coverage and
fidence in the geodetic modeling results? independent geological slip-rate constraints may lead to better
Integration of geologic and seismological constraints elim- resolution of the deformation in these regions and perhaps to the
inates some model parameters from the inversion process, lead- definition of smaller blocks. Gradational behavior is also likely,
ing to better bounds on the remaining unknowns. For example, with most deformation taking place on the boundaries of active
if good micro-earthquake locations are available, the maximum regions, but minor seismicity and very low slip-rate faults occur-
hypocentral depths of small events may define the position of ring within the region (e.g., Central Basin and Range Province,
the brittle-ductile transition in the crust (Sibson, 1982, 1984), western USA; Hammond et al., 2011a; Thatcher et al., 1999).
providing a bound on the maximum depth of coseismic slip in Indeed, infrequent intra-plate earthquakes as large as M≈7–7.5
large earthquakes. The depth and spatial focusing of very small occur far from plate boundaries (e.g., Johnston and Schweig,
non-volcanic tremor events between ~15 and ~30 km directly 1996), testifying to the existence of isolated fault zones whose
beneath the surface trace of the San Andreas fault in central behavior is clearly not block-like.
Space geodesy 423
Ultimately, the choice of block versus continuum model fault-slip-rate estimates for major strike-slip faults in California.
approaches is also one of scientific objective. If we are interested Hammond et al. (2007) also have demonstrated that these effects
in describing the kinematics of deformation by faulting in the are important for slip rate estimation of Basin and Range nor-
brittle upper crust, block models provide an efficient description mal faults in Nevada. More broadly, ghost transients may lead to
of contributions by major (block bounding) and minor (intra- erroneous inferences of steady-state strain rates away from major
block strain) faults. If the goal is to understand the forces driving faults and in the interiors of arguably stable blocks.
distributed continental deformation, continuum models of crustal Finally, it is unclear whether laboratory-based experiments
flow driven by plate boundary and buoyancy forces provide a suit- on the deformational behavior of small rock samples can be
able approach to illuminate these relationships (Humphreys and extrapolated to the large-scale motions of the lithosphere. First
Coblentz, 2007; Thatcher, 2009). Advanced geodynamic models of all, many of the results from the lab are semi-empirical rules
of continental deformation that incorporate both frictional fault rather than truly physics-based “laws” (Bürgmann and Dresen,
mechanics and realistic ductile rheologies at depth may provide 2008), so the actual deformational processes are sometimes not
the means to integrate these two end-member approaches. fundamentally understood. Furthermore, spatial variations in
material rock properties and strain-softening behavior may lead
True Rheology of the Lithosphere and to strain localization in the weakest portion of the lithosphere-
Uppermost Asthenosphere scale deforming zone. For example, it is well known that while
intact rock bodies in the seismogenic upper crust can support
While new space geodetic measurements are providing an high (average ~200 MPa) differential stresses, major fault zones
increasing inventory of postseismic transient case histories, we are weaker by at least an order of magnitude (e.g., Townend and
are still unsure about the true rheological stratification of the lith- Zoback, 2000). Correspondingly, the ductile roots of major faults
osphere and the top of the asthenosphere, and the “laws” (really may also deform at stresses considerably lower than surrounding,
empirical rules) that govern the rheology. Furthermore most largely intact country rock in the lower crust and uppermost man-
inferences of ductile rheology are based on observations made tle. Exhumed, now inactive large-scale lower crustal and upper-
during the decade or less following major earthquakes, and it is most mantle shear zones provide evidence for focused ductile
quite unclear whether estimates are representative of the effective deformation (Dijkstra et al., 2004; Hanmer, 1988). Earthquake
viscosity operative during the majority of the earthquake defor- cycle models that explore the implications for surface deforma-
mation cycle, usually hundreds to thousands of years. Finally, the tion of both the laboratory-based flow laws (Freed et al., 2012;
true geometry of the deforming regions is unknown. For example, Hearn et al., 2009) and of focused ductile shear zones (Hearn et
we are unsure whether most of the postseismic poroelastic defor- al., 2013) will shed light on the applicability of these viewpoints
mation involves flow of water concentrated near the surface, and to better understand the processes governing earthquake-related
if ductilely deforming rock flow is generally focused in relatively surface deformation.
narrow shear zones or involves broadly distributed bulk flow.
Observations made near the end of the earthquake cycle, as Inadequacies of Spatial Coverage, Temporal Resolution,
well as in the immediate postseismic interval, are crucial in con- and Data Precision
straining the effective viscosity and hence the duration of post-
seismic transient deformation of the upper mantle throughout the Although GPS and InSAR data distribution is quite dense in
cycle (Hearn et al., 2009). Recent work using the pre- and post- parts of the western USA, New Zealand, Japan, and the eastern
seismic GPS data from the region of the M 7.5 Izmit, Turkey, Mediterranean, many active regions have patchy to nonexistent
earthquake suggests lower bounds for the viscosity of the upper coverage (e.g., western Tibet, Iran, all offshore active zones).
mantle of 1019 to 1020 Pa/s (Hearn et al., 2009). However, the Even in well-monitored areas, improved spatial coverage, tempo-
truly appropriate value or even its upper bound remains poorly ral resolution, and precision of geodetic measurements promise
known. Since effective viscosity estimates tend to increase with to substantially enhance our ability to measure and characterize
the duration of the postseimic transient observation interval, and crustal deformation processes.
the magnitude of the transient becomes smaller with time, it is Campaign, semi-continuous (Blewitt et al., 2009), and
observationally challenging to tightly constrain the appropriate expanding continuous GPS networks will certainly augment on-
upper bound on the effective viscosity from earthquake cycle land coverage. Likewise, InSAR mapping, particularly using a
observations alone. new generation of planned radar satellites dedicated to imaging
This issue is important in determining the degree to which ground deformation (see InSAR section), will fill in large gaps
the steady-state surface velocity field is contaminated by long- worldwide. With the notable successes of seafloor geodesy to
lived “ghost” transient deformation (Gourmelen and Amelung, measure both interseismic strain accumulation at subduction
2005; Hetland and Hager, 2003) and how to appropriately cor- zones (Gagnon et al., 2005) and large coseismic displacements
rect GPS observations for this effect (Hearn et al., 2013). As in the M 9.0, 2011 Tohoku, Japan, earthquake (Kido et al., 2011;
Chuang and Johnson (2011) and Hearn et al. (2013) show, these Sato et al., 2011), expanded networks are being planned in Japan
uncertainties can sometimes significantly influence GPS-based (M. Kido, 2012, personal commun.) and are discussed for the
Cascadia subduction zone and other offshore plate-boundary Most geodetic techniques are limited to measuring compo-
deformation zones. More strategically sited, continuous GPS sta- nents of displacements or strain at or very close to the Earth’s
tions and other measurement systems will better detect and quan- surface. What are the prospects for “deep-Earth geodesy”? If the
tify a range of postseismic relaxation processes, transient slow- precision and spatial resolution of GRACE-type satellite gravity-
slip events, and other currently, poorly understood time-varying change measurements can be improved by an order of magnitude
deformation transients. Innovative new analysis techniques can or two, we will be able to routinely monitor the redistribution
be expected to sharpen the resolution provided by both currently of mass at depth associated with a wide range of tectonic and
available data and that obtained by new networks to be installed volcanic processes. The propagation of seismic waves through
during the next decade. the lithosphere is influenced by active deformation processes.
Despite the challenge of separating deformation-related signal
What Next in Tectonic Geodesy? changes from those caused by various other near-surface and
subsurface processes (e.g., hydrologic processes), there may be
Would a review paper on tectonic geodesy written in the exciting opportunities for studying crustal deformation at depth
1960s have been able to anticipate the new space-geodetic tech- using improved active-source experiments or passive seismic
nologies and the breakthrough discoveries they enabled, which observations (Daley et al., 2007; Taira et al., 2009).
we describe in this contribution? Can we now envision what
new methods and capabilities lie on the horizon for the field of ACKNOWLEDGMENTS
crustal deformation measurements? As described in the previous
section, and as we require increasingly precise measurements of We thank David Chadwell, William Chadwick, Eric Field-
three-dimensional displacements, strain, and gravity changes, we ing, Andy Freed, Yan Hu, Motoyuki Kido, Corné Kreemer, Jeff
aim for temporal resolution down to fractions of a second, we McGuire, Rob McCaffrey, Susan Merle, Fred Pollitz, Manoochehr
seek near-real-time data availability, and we want fully global Shirzaei, and Xiaopeng Tong for providing figure files, data, or
coverage on land and across oceanic regions. What follows is other information for this review. Bob Simpson, Fred Pollitz,
likely to miss the most important advances in future decades, but Chris Johnson, and Xiaopeng Tong provided helpful reviews of
we hope that some of these ideas may prove to become reality. an earlier version of this manuscript. We thank Tom Herring, Jack
It is easiest to envision substantial improvements of exist- Loveless, and Paul Segall for their thoughtful reviews.
ing technologies. Beyond the next generation of SAR satellites,
we envision increasingly large constellations of identical space- REFERENCES CITED
craft, which will allow for about daily repeat-imaging from three
or more LOS directions, to obtain dense time series of three- Agnew, D.C., 1986, Strainmeters and tiltmeters: Reviews of Geophysics, v. 24,
dimensional deformation across vast regions of the Earth. Alter- doi:10.1029/RG1024i1003p00579.
Aktug, B., Nocquet, J.M., Cingöz, A., Parsons, B., Erkan, Y., England, P.,
natively, substantially more powerful radar systems launched to Lenk, O., Gürdal, M.A., Kilicoglu, A., Akdeniz, H., and Tekgül, A., 2009,
high, geosynchronous orbits could provide effectively continu- Deformation of western Turkey from a combination of permanent and
ous InSAR monitoring (Edelstein et al., 2005). Targeted InSAR campaign GPS data: Limits to block-like behavior: Journal of Geophysi-
cal Research, v. 114, B10404, doi:10.1029/2008JB006000.
measurements can also be obtained from airplanes and drones, Allmendinger, R.W., Loveless, J.P., Pritchard, M.E., and Meade, B., 2009,
as demonstrated by the ongoing UAVSAR program by the Jet From decades to epochs: Spanning the gap between geodesy and struc-
Propulsion Laboratory (http://uavsar.jpl.nasa.gov/index.html). tural geology of active mountain belts: Journal of Structural Geology,
v. 31, p. 1409–1422, doi:10.1016/j.jsg.2009.08.008.
A large fleet of unmanned airborne InSAR, LiDAR, and other Altamimi, Z., Métivier, L., and Collilieux, X., 2012, ITRF2008 plate motion model:
geodetic systems would allow for highly precise and timely mea- Journal of Geophysical Research, v. 117, B07402, doi:10.1029/2011JB008930.
surements of active tectonic events. Argus, D., and Gordon, R., 2001, Present tectonic motion across the Coast
Ranges and San Andreas fault system in central California: Geological
The cost and energy demands of geodetic-grade GPS sys- Society of America Bulletin, v. 113, p. 1580–1592, doi:10.1130/0016
tems continues to drop rapidly, and thus it seems reasonable to -7606(2001)113<1580:PTMATC>2.0.CO;2.
expect that future geodetic GNSS networks may collect data Argus, D.F., and Gordon, R.G., 1990, Pacific–North American plate motion
from very long baseline interferometry compared with motion inferred
from many thousands of small sensors. Redundancy in large from magnetic anomalies, transform faults, and earthquake slip vectors:
numbers may obviate the need for the installation of costly stable Journal of Geophysical Research, v. 95, p. 17,315–17,324, doi:10.1029/
geodetic monuments. JB095iB11p17315.
Arnadottir, T., and Segall, P., 1994, The 1989 Loma Prieta earthquake imaged
Similarly, we hope that cheaper and compact seafloor sys- from inversion of geodetic data: Journal of Geophysical Research, v. 99,
tems can be developed that allow for the deployment of large p. 21,835–821,855.
regional geodetic networks across submarine deformation zones. Asada, A., and Ura, T., 2010, Full-swath bathymetric survey system with synthetic
aperture and triangle-arrayed interferometric techniques for autonomous
The full range of seafloor geodetic observables (e.g., distance underwater vehicle, in OCEANS 2010 Proceedings, 20–23 Sept. 2010, p. 1–6.
ranging, pressure sensors, acoustic interferometry) should be fur- Auriol, A., and Tourain, C., 2010, DORIS System: The new age: Advanced
ther developed for improved precision, range, and stability and Space Research, v. 46, doi:10.1016/j.asr.2010.1005.1015.
Avouac, J.P., and Tapponnier, P., 1993, Kinematic model of active deforma-
reduced cost. The use of remotely operated vehicles and buoys, tion in central Asia: Geophysical Research Letters, v. 20, p. 895–898,
instead of costly ship deployments, should aid in this effort. doi:10.1029/93GL00128.
Space geodesy 425
Banerjee, P., Pollitz, F., and Bürgmann, R., 2005, The size and duration of 2001, from height and gravity changes: Geophysical Research Letters,
the Sumatra-Andaman earthquake from far-field static offsets: Science, v. 31, doi:10.1029/2004GL020768.
v. 308, p. 1769–1772, doi:10.1126/science.1113746. Chinnery, M.A., 1961, The deformation of the ground around surface faults:
Barbot, S., Hamiel, Y., and Fialko, Y., 2008, Space geodetic investigation of Bulletin of the Seismological Society of America, v. 51, p. 355–372.
the coseismic and postseismic deformation due to the 2003 Mw7.2 Altai Chuang, R., and Johnson, K., 2011, Reconciling geologic and geodetic model
earthquake: Implications for the local lithospheric rheology: Journal of fault slip-rate discrepancies in Southern California: Consideration of
Geophysical Research, v. 113, B03403, doi:10.1029/2007JB005063. nonsteady mantle flow and lower crustal fault creep: Geology, v. 39,
Barbot, S., Lapusta, N., and Avouac, J.-P., 2012, Under the hood of the earth- doi:10.1130/G32120.32121.
quake machine: Toward predictive modeling of the seismic cycle: Sci- Clark, T.A., Gordon, D., Himwich, W.E., Ma, C., Mallama, A., and Ryan, J.W.,
ence, v. 336, doi:10.1126/science.1218796. 1987, Determination of relative site motions in the Western United States
Bendick, R., Bilham, R., Freymueller, J., Larson, K., and Yin, G., 2000, Geo- using Mark III Very Long Baseline Interferometry: Journal of Geophysi-
detic evidence for a low slip rate in the Altyn Tagh fault system: Nature, cal Research, v. 92, p. 12,741–12,750, doi:10.1029/JB092iB12p12741.
v. 404, doi:10.1038/35003555. Cowgill, E., 2007, Impact of riser reconstructions on estimation of secular
Bennett, R., Friedrich, A., and Furlong, K., 2004, Codependent histories of the variation in rates of strike-slip faulting: Revisiting the Cherchen River
San Andreas and San Jacinto fault zones from inversion of fault displace- site along the Altyn Tagh Fault, NW China: Earth and Planetary Science
ment rates: Geology, v. 32, doi:10.1130/G20806.20801. Letters, v. 254, p. 239–255, doi:10.1016/j.epsl.2006.09.015.
Berardino, P., Fornaro, G., Lanari, R., and Sansosti, E., 2002, A new algorithm Cowgill, E., Gold, R.D., Xuanhua, C., Xiao-Feng, W., Arrowsmith, J.R.N., and
for surface deformation monitoring based on small baseline differential Southon, J., 2009, Low Quaternary slip rate reconciles geodetic and geo-
SAR interferograms: IEEE Transactions on Geoscience and Remote logic rates along the Altyn Tagh fault, northwestern Tibet: Geology, v. 37,
Sensing, v. 40, p. 2375–2383, doi:10.1109/TGRS.2002.803792. p. 647–650, doi:10.1130/G25623A.1.
Beroza, G.C., and Ide, S., 2011, Slow earthquakes and nonvolcanic tremor: Cox, A., Hart, B.H., 1986, Plate Tectonics: How It Works: Palo Alto, California,
Annual Reviews of Earth and Planetary Science, v. 39, doi:10.1146/ Blackwell, 392 p.
annurev-earth-040809-152531. d’Alessio, M. A., Johansen, I. A., Bürgmann, R., Schmidt, D. A., and Murray,
Bilham, R., and England, P., 2001, Plateau ‘pop-up’ in the great 1897 Assam M. H., 2005, Slicing up the San Francisco Bay area: Block kinematics and
earthquake: Nature, v. 410, p. 806–809, doi:10.1038/35071057. fault slip rates from GPS-derived surface velocities: Journal of Geophysi-
Bilham, R., Suszek, N., and Pinkney, S., 2004, California Creepmeters: Seismo- cal Research, v. 110, B06403, doi:10.1029/2004JB003496.
logical Research Letters, v. 75, p. 481–492, doi:10.1785/gssrl.75.4.481. Daley, T.M., Solbau, R.D., Ajo-Franklin, J.B., and Benson, S.M., 2007, Contin-
Blewitt, G., Hammond, W.C., and Kreemer, C., 2009, Geodetic observation uous active-source seismic monitoring of CO2 injection in a brine aquifer:
on contemporary deformation in the northern Walker Lane: 1. Semi- Geophysics, v. 72, p. A57–A61, doi:10.1190/1.2754716.
permanent GPS strategy in Oldow J.S., and Cashman, P.H., eds., Late DeMets, C., Gordon, R.G., Argus, D.F., and Stein, S., 1990, Current plate
Cenozoic Structure and Evolution of the Great Basin-Sierra Nevada motions: Geophysical Journal International, v. 101, p. 425–478,
Transition: Geological Society of America Special Paper 447, p. 1–15, doi:10.1111/j.1365-246X.1990.tb06579.x.
doi:10.1130/2009.2447(01). Dieterich, J., 2007, Applications of rate- and state-dependent friction to mod-
Blisniuk, K., Rockwell, T., Owen, L.A., Oskin, M., Lippincott, C., Caffee, els of fault slip and earthquake occurrence: Treatise on Geophysics, v. 4,
M.W., and Dortch, J., 2010, Late Quaternary slip rate gradient defined p. 107–129, doi:10.1016/B978-044452748-6/00065-1.
using high-resolution topography and 10Be dating of offset landforms on Dijkstra, A.H., Drury, M.R., Vissers, R.L.M., Newman, J., and Van Roermund,
the southern San Jacinto Fault zone, California: Journal of Geophysical H.L.M., 2004, Shear zones in the upper mantle: Evidence from alpine-
Research, v. 115, B08401, doi:10.1029/2009JB006346. and ophiolite-type peridotite massifs: Geological Society [London] Spe-
Bonafede, M., Strehlau, J., and Ritsema, A.R., 1992, Geophysical and structural cial Publication 224, p. 11–24, doi:10.1144/GSL.SP.2004.224.01.02.
aspects of fault mechanics—A brief historical review: Terra Nova, v. 4, Dixon, T.H., 1991, An introduction to the global positioning system and some
p. 458–463, doi:10.1111/j.1365-3121.1992.tb00581.x. geological applications: Reviews of Geophysics, v. 29, p. 249–276,
Burgette, R.J., Weldon, R.J., II, and Schmidt, D.A., 2009, Interseismic uplift doi:10.1029/91RG00152.
rates for western Oregon and along-strike variation in locking on the Cas- Dolan, J.F., Bowman, D.D., and Sammis, C.G., 2007, Long-range and long-term
cadia subduction zone: Journal of Geophysical Research, v. 114, B01408, fault interactions in Southern California: Geology, v. 35, doi:10.1130/
doi:10.1029/2008JB005679. G23789A.23781.
Bürgmann, R., and Dresen, G., 2008, Rheology of the lower crust and upper Dracup, J.F., 1995, Geodetic surveys in the United States: The beginning and the
mantle: Evidence from rock mechanics, geodesy and field observations: next one hundred years 1807–1940, in http://www.ngs.noaa.gov/PUBS_LIB/
Annual Review of Earth and Planetary Sciences, v. 36, p. 531–567, geodetic_survey_1807.html, ed., ACSM/ASPRS Annual Convention & Expo-
doi:10.1146/annurev.earth.36.031207.124326. sition Technical Papers, Volume 1: Bethesda, Maryland, p. 24.
Bürgmann, R., Rosen, P.A., and Fielding, E.J., 2000, Synthetic aperture radar Edelstein, W., Madsen, S., Moussessian, A., and Chen, C., 2005, Concepts and
interferometry to measure Earth’s surface topography and its deforma- Technologies for Synthetic Aperture Radar from MEO and Geosynchro-
tion: Annual Review of Earth and Planetary Sciences, v. 28, p. 169–209, nous Orbits: SPIE Proceedings, v. 5659, doi: 10.1117/1112.578989.
doi:10.1146/annurev.earth.28.1.169. England, P.C., and Molnar, P., 1997, Active deformation of Asia: From
Castle, R.O., and Bernknopf, R.L., 1996, The southern California uplift and kinematics to dynamics: Science, v. 278, p. 647–650, doi:10.1126/
associated earthquakes: Geophysical Research Letters, v. 23, p. 3011– science.278.5338.647.
3014, doi:10.1029/96GL02894. Fay, N.P., and Humphreys, E.D., 2005, Fault slip rates, effects of elastic hetero-
Castle, R.O., Church, J.P., and Elliott, M.R., 1976, Aseismic uplift in Southern Cali- geneity on geodetic data, and the strength of the lower crust in the Salton
fornia: Science, v. 192, p. 251–253, doi:10.1126/science.192.4236.251. Trough region, southern California: Journal of Geophysical Research,
Chadwell, C.D., and Spiess, F.N., 2008, Plate motion at the ridge-transform v. 110, B09401, doi:10.1029/2004JB003548.
boundary of the south Cleft segment of the Juan de Fuca Ridge from Ferretti, A., Prati, C., and Rocca, F., 2001, Permanent scatterers in SAR inter-
GPS-Acoustic data: Journal of Geophysical Research, v. 113, B04415, ferometry: IEEE Transactions on Geoscience and Remote Sensing, v. 39,
doi:10.1029/2007JB004936. p. 8–20, doi:10.1109/36.898661.
Chadwick, W.W., Jr., Embley, R.W., Milburn, H.B., Meinig, C., and Stapp, Ferretti, A., Fumagalli, A., Novali, F., Prati, C., Rocca, F., and Rucci, A., 2011,
M., 1999, Evidence for deformation associated with the 1998 eruption A new algorithm for processing interferometric data-stacks: SqueeSAR:
of Axial Volcano, Juan de Fuca Ridge, from acoustic extensometer mea- IEEE Transactions on Geoscience and Remote Sensing, v. 49, p. 3460–
surements: Geophysical Research Letters, v. 26, p. 3441–3444, doi:10 3470, doi:10.1109/TGRS.2011.2124465.
.1029/1999GL900498. Fielding, E.J., Lundgren, P.R., Bürgmann, R., and Funning, G.J., 2009, Shal-
Chadwick, W.W., Nooner, S.L., Butterfield, D.A., and Lilley, M.D., 2012, Sea- low fault-zone dilatancy recovery after the 2003 Bam, Iran, earthquake:
floor deformation and forecasts of the April 2011 eruption at Axial Sea- Nature, v. 458, p. doi:10.1038/nature07817.
mount: Nature Geoscience, v. 5, p. 474–477, doi:10.1038/NGEO1464. Fitch, T.J., and Scholz, C.H., 1971, Mechanism of underthrusting in southwest
Chandrasekhar, D.V., Mishra, D.C., Singh, B., Vijayakumar, V., and Bürgmann, Japan: A model of convergent plate interactions: Journal of Geophysical
R., 2004, Source parameters of the Bhuj earthquake, India, of January 26, Research, v. 76, p. 7260–7292, doi:10.1029/JB076i029p07260.
Flesch, L.M., Holt, W.E., Haines, A.J., and Shen-Tu, B., 2000, Dynamics of the Hammond, W.C., Brooks, B.A., Bürgmann, R., Heaton, T., Jackson, M., Lowry,
Pacific–North American plate boundary in the western United States: Sci- A.R., and Anandakrishnan, S., 2011b, Scientific value of real-time Global
ence, v. 287, p. 834–836, doi:10.1126/science.287.5454.834. Positioning System data: Eos (Transactions, American Geophysical
Floyd, M.A., Billiris, H., Paradissis, D., Veis, G., Avallone, A., Briole, P., Union), v. 92, p. 125–132.
McClusky, S., Nocquet, J.M., Palamartchouk, K., Parsons, B., and Han, S.C., Sauber, J., and Luthcke, S., 2010, Regional gravity decrease
England, P.C., 2010, A new velocity field for Greece: Implications for after the 2010 Maule (Chile) earthquake indicates large-scale
the kinematics and dynamics of the Aegean: Journal of Geophysical mass redistribution: Geophysical Research Letters, v. 37, L23307,
Research, v. 115, B10403, doi:10.1029/2009JB007040. doi:10.1029/2010GL045449.
Frank, F.C., 1966, Deduction of earth strains from survey data: Bulletin of the Hanmer, S., 1988, Great Slave Lake shear zone, Canadian shield: Recon-
Seismological Society of America, v. 56, p. 35–42. structed vertical profile of a crustal-scale fault zone: Tectonophysics,
Freed, A. M., and Bürgmann, R., 2004, Evidence of powerlaw flow in the Mojave v. 149, p. 245–264, doi:10.1016/0040-1951(88)90176-X.
desert mantle: Nature, v. 430, doi:10.1038/nature02784, p. 548–551. Hayford, J.F., and Baldwin, A.L., 1907, The earth movements in the Califor-
Freed, A.M., Bürgmann, R., Calais, E., and Freymueller, J.T., 2006a, Stress- nia earthquake of 1906: Appendix 3 of U.S. Coast and Geodetic Survey
dependent power-law flow in the upper mantle following the 2002 Report, p. 69–104.
Denali, Alaska, earthquake: Earth and Planetary Science Letters, v. 252, Hayford, J.F., and Baldwin, A.L., 1908, The earth movements in the California
doi:10.1016/j.epsl.2006.1010.1011. earthquake of 1906, in Lawson, A.C., ed., The California Earthquake of
Freed, A.M., Bürgmann, R., Calais, E., Freymueller, J.T., and Hreinsdót- April 18, 1906, Report of the State Earthquake Investigation Commis-
tir, S., 2006b, Implications of deformation following the 2002 Denali, sion, Volume 1: Washington, D.C., Carnegie Institution of Washington,
Alaska, earthquake for postseismic relaxation processes and litho- p. 114–145.
spheric rheology: Journal of Geophysical Research, v. 111, B01401, Hearn, E.H., 2003, What can GPS data tell us about the dynamics of post-
doi:10.1029/2005JB003894. seismic deformation?: Geophysical Journal International, v. 155,
Freed, A.M., Bürgmann, R., and Herring, T.A., 2007, Far-reaching tran- p. 753–777.
sient motions after Mojave earthquakes require broad mantle flow Hearn, E.H., and Bürgmann, R., 2005, The effect of elastic layering on
beneath a strong crust: Geophysical Research Letters, v. 34, L19302, inversions of GPS data for earthquake slip and stress changes: Bul-
doi:10.1029/2007GL030959. letin of the Seismological Society of America, v. 95, p. 1637–1653,
Freed, A.M., Herring, T.A., and Bürgmann, R., 2010, Steady-state laboratory doi:10.1785/0120040158.
flow laws alone fail to explain postseismic observations: Earth and Plan- Hearn, E.H., Bürgmann, R., and Reilinger, R., 2002, Dynamics of Izmit earth-
etary Science Letters, v. 300, doi:10.1016/j.epsl.2010.1010.1005. quake postseismic deformation and loading of the Duzce earthquake
Freed, A.M., Hirth, G., and Behn, M.D., 2012, Using short-term postseis- hypocenter: Bulletin of the Seismological Society of America, v. 92,
mic displacements to infer the ambient deformation conditions of p. 172–193, doi:10.1785/0120000832.
the upper mantle: Journal of Geophysical Research, v. 117, B01409, Hearn, E.H., McClusky, S., Ergintav, S., and Reilinger, R.E., 2009, Izmit
doi:10.1029/2011JB008562. earthquake postseismic deformation and dynamics of the North Ana-
Fujiwara, T., Kodaira, S., No, T., Kaiho, Y., Takahashi, N., and Kaneda, Y., tolian fault zone: Journal of Geophysical Research, v. 114, B08405,
2011, The 2011 Tohoku-Oki Earthquake: Displacement reaching the doi:10.1029/2008JB006026.
trench axis: Science, v. 334, p. 1240, doi:10.1126/science.1211554. Hearn, E., Johnson, K., Sandwell, D., and Thatcher, W., 2010, SCEC
Gabriel, A.K., Goldstein, R.M., and Zebker, H.A., 1989, Mapping small UCERF workshop report: http://www.scec.org/workshops/2010/
elevation changes over large areas—Differential radar interferometry: gps-ucerf3/FinalReport_GPS-UCERF3Workshop.pdf.
Journal of Geophysical Research, v. 94, p. 9183–9191, doi:10.1029/ Hearn, E.H., Pollitz, F.F., Thatcher, W., and Onishi, C.T., 2013, How do “ghost
JB094iB07p09183. transients” from past earthquakes affect GPS slip rate estimates on south-
Gagnon, K.L., Chadwell, C.D., and Norabuena, E., 2005, Measuring the onset ern California faults?: Geochemistry Geophysics Geosystems, v. 14,
of updip locking in the Peru-Chile Trench at 12°S from GPS and acoustic doi:10.1002/ggge.20080.
measurements: Nature, v. 434, p. 205–208, doi:10.1038/nature03412. Herring, T.A., Corey, B.E., Counselman, C.C., III, Shapiro, I.I., Rönnäng,
Galehouse, J.S., and Lienkaemper, J.J., 2003, Inferences drawn from two B.O., Rydbeck, O.E.H., Clark, T.A., Coates, R.J., Ma, C., Ryan, J.W.,
decades of alinement array measurements of creep on faults in the San Vandenberg, N.R., Hinteregger, H.F., Knight, C.A., Rogers, A.E.E., Whit-
Francisco Bay Region: Bulletin of the Seismological Society of America, ney, A.R., Robertson, D.S., and Schupler, B.R., 1981, Geodesy by radio
v. 93, p. 2415–2433, doi:10.1785/0120020226. interferometry: Intercontinental distance determinations with subdeci-
Gao, S.S., Silver, P.G., and Linde, A.T., 2000, Analysis of deformation data at meter precision: Journal of Geophysical Research, v. 86, p. 1647–1651,
Parkfield, California: Detection of a long-term strain transient: Journal of doi:10.1029/JB086iB03p01647.
Geophysical Research, v. 105, p. 2955–2967, doi:10.1029/1999JB900383. Herring, T., Shapiro, I., Clark, T., Ma, C., Ryan, J., Schupler, B., Knight, C.,
Geen, M., Applications of interferometric swath bathymetry, in OCEANS ’98 Lundqvist, G., Shaffer, D., Vandenberg, N., Corey, B., Hinteregger, H.,
Conference Proceedings, 1998 (28 Sept.–1 Oct. 1998): Piscataway, New Rogers, A., Webber, J., Whitney, A., Elgered, G., Ronnang, B., and Davis,
Jersey, IEEE, v. 2, p. 1126–1130. J., 1986, Geodesy by radio interferometry: Evidence for contemporary
Gordon, R.G., 1998, The plate tectonic approximation: Plate nonrigidity, diffuse plate motion: Journal of Geophysical Research, v. 91, p. 8341–8347,
plate boundaries, and global plate reconstructions: Annual Review of Earth and doi:10.1029/JB091iB08p08341.
Planetary Sciences, v. 26, p. 615–642, doi:10.1146/annurev.earth.26.1.615. Hetland, E.A., and Hager, B.H., 2003, Postseismic relaxation across the Cen-
Gourmelen, N., and Amelung, F., 2005, Postseismic mantle relaxation tral Nevada Seismic Belt: Journal of Geophysical Research, v. 108,
in the Central Nevada Seismic Belt: Science, v. 310, doi:10.1126/ doi:10.1029/2002JB002257.
science.1119798. Hetland, E.A., and Hager, B.H., 2005, Postseismic and interseismic displace-
Hager, B.H., King, R.W., and Murray, M.H., 1991, Measurements of crustal ments near a strike-slip fault: A two-dimensional theory for general linear
deformation using the Global Positioning System: Annual Review of viscoelastic rheologies: Journal of Geophysical Research, v. 110, B10401,
Earth and Planetary Sciences, v. 19, p. 351–382, doi:10.1146/annurev doi:10.1029/2005JB003689.
.ea.19.050191.002031. Hetland, E.A., Musè, P., Simons, M., Lin, Y.N., Agram, P.S., and DiCaprio,
Hammond, W.C., Kreemer, C., and Blewitt, G., 2007, Geodetic constraints C.J., 2012, Multiscale InSAR Time Series (MInTS) analysis of sur-
on contemporary deformation in the northern Walker Lane: 3, Central face deformation: Journal of Geophysical Research, v. 117, B02404,
Nevada Seismic Belt postseismic relaxation, in Oldow J.S., and Cashman, doi:10.1029/2011JB008731.
P.H., eds., Late Cenozoic Structure and Evolution of the Great Basin- Hilley, G.E., Johnson, K.M., Wang, M., Shen, Z.-K., and Bürgmann, R., 2009,
Sierra Nevada Transition: Geological Society of America Special Paper Earthquake-cycle deformation and fault slip rates in Northern Tibet:
447, p. 33–54, doi:10.1130/2009.2447(03). Geology, v. 31, doi:10.1130/G25157A.25151.
Hammond, W.C., Blewitt, G., and Kreemer, C., 2011a, Block modeling of Hinteregger, H.F., Shapiro, I.I., Robertson, D.S., Knight, C.A., Ergas, R.A.,
crustal deformation of the northern Walker Lane and Basin and Range Whitney, A.R., Rogers, A.E.E., Moran, J.M., Clark, T.A., and Burke,
from GPS velocities: Journal of Geophysical Research, v. 116, B04402, B.F., 1972, Precision geodesy via radio interferometry: Science, v. 178,
doi:10.1029/2010JB007817. p. 396–398, doi:10.1126/science.178.4059.396.
Space geodesy 427
Hofmann-Wellenhof, B., Lichtenegger, H., and Wasle, E., 2008, GNSS–Global Johnson, K.M., Bürgmann, R., and Freymueller, J.T., 2009, Coupled afterslip
Navigation Satellite Systems: GPS, GLONASS, Galileo, and More: New and viscoelastic flow following the 2002 Denali Fault, Alaska, Earth-
York, Springer, 516 p. quake: Geophysical Journal International, v. 176, doi:10.1111/j.1365
Hooper, A., 2008, A multi-temporal InSAR method incorporating both persis- -1246X.2008.04029.x.
tent scatterer and small baseline approaches: Geophysical Research Let- Johnston, A.C., and Schweig, E.S., 1996, The enigma of the New Madrid earth-
ters, v. 35, L16302, p. doi:10.1029/2008GL03465. quakes of 1811–1812: Annual Review of Earth and Planetary Sciences,
Hooper, A., Zebker, H., Segall, P., and Kampes, B., 2004, A new method for v. 24, p. 339–384, doi:10.1146/annurev.earth.24.1.339.
measuring deformation on volcanoes and other natural terrains using Jolivet, R., Cattin, R., Chamot-Rooke, N., Lasserre, C., and Peltzer, G., 2008,
InSAR persistent scatterers: Geophysical Research Letters, v. 31, Thin-plate modelling of interseismic deformation and asymmetry across
doi:10.1029/2004GL021737. the Altyn Tagh fault zone: Geophysical Research Letters, v. 35, L02309,
Hooper, A., Bekaert, D., Spaans, K., and Arıkan, M., 2012, Recent advances doi:10.1029/2007GL031511.
in SAR interferometry time series analysis for measuring crustal Jónsson, S., Segall, P., Pedersen, R., and Björnsson, G., 2003, Post-earthquake
deformation: Tectonophysics, v. 514–517, p. 1–13, doi:10.1016/j ground movements correlated to pore-pressure transients: Nature, v. 424,
.tecto.2011.10.013. p. 179–183, doi:10.1038/nature01776.
Hsu, Y.J., Simons, M., Avouac, J.P., Galetzka, J., Sieh, K., Chlieh, M., Nata- Kato, A., Obara, K., Igarashi, T., Tsuruoka, H., Nakagawa, S., and Hirata, N.,
widjaja, D., Prawirodirdjo, L., and Bock, Y., 2006, Frictional afterslip 2012, Propagation of slow slip leading up to the 2011 Mw9.0 Tohoku-Oki
following the 2005 Nias-Simeulue earthquake, Sumatra: Science, v. 312, earthquake: Science, v. 335, doi:10.1126/science.1215141.
doi:10.1126/science.1136960. Kaufman, K., and Rousseeuw, P.J., 1990, Finding Groups in Data, an Introduc-
Hsu, Y.-J., Segall, P., Yu, S.B., Kuo, L.-C., and Williams, C.A., 2007, Tempo- tion to Cluster Analysis: New York, Wiley & Sons, 342 p.
ral and spatial variations of post-seismic deformation following the 1999 Khazaradze, G., and Klotz, J., 2003, Short- and long-term effects of GPS mea-
Chi-Chi, Taiwan earthquake: Geophysical Journal International, v. 169, sured crustal deformation rates along the south central Andes: Journal of
p. 367–379, doi:10.1111/j.1365-246X.2006.03310.x. Geophysical Research, v. 108, doi:10.1029/2002JB001879.
Hu, Y., and Wang, K., 2012, Spherical-Earth finite element model of short- Kido, M., Osada, Y., Fujimoto, H., Hino, R., and Ito, Y., 2011, Trench-normal
term postseismic deformation following the 2004 Sumatra earth- variation in observed seafloor displacements associated with the 2011
quake: Journal of Geophysical Research, v. 117, B05404, doi:10 Tohoku-Oki earthquake: Geophysical Research Letters, v. 38, L24303,
.1029/2012JB009153. doi:10.1029/2011GL050057.
Hu, Y., Wang, K., He, J., Klotz, J., and Khazaradze, G., 2004, Three-dimensional Kim, A., and Dreger, D.S., 2008, Rupture process of the 2004 Parkfield earth-
viscoelastic finite element model for postseismic deformation of the great quake from near-fault seismic waveform and geodetic records: Journal
1960 Chile earthquake: Journal of Geophysical Research, v. 109, B12403, of Geophysical Research, v. 113, B07308, doi:10.1029/2007JB005115.
doi:10.1029/2004JB003163. Kostrov, B.V., 1974, Seismic moment and energy of earthquakes and seismic
Humphreys, E.D., and Coblentz, D.D., 2007, North American dynamics flow of rock: Izvestiya, Academy of Sciences, USSR, Physics of the Solid
and western U.S. tectonics: Reviews in Geophysics, v. 45, RG3001, Earth, v. 1, p. 23–44.
doi:10.1029/2005RG000181. Kreemer, C., Hammond, W.C., Blewitt, G., Holland, A.A., and Bennett, R.A.,
Hyndman, R.D., Currie, C.A., and Mazzotti, S.P., 2005, Subduction zone back- 2012, A geodetic strain rate model for the Pacific–North American plate
arcs, mobile belts, and orogenic heat: GSA Today, v. 15, no. 2, p. 4–10. boundary, western United States: Reno, Nevada Bureau of Mines and
Ide, S., Beroza, G.C., Shelly, D.R., and Uchide, T., 2007, A scaling law for slow Geology Map 178, scale 1:1,500,000.
earthquakes: Nature, v. 447, p. 76–79, doi:10.1038/nature05780. Lambeck, K., 1988, Geophysical Geodesy, the Slow Deformations of the Earth:
Iinuma, T., Hino, R., Kido, M., Inazu, D., Osada, Y., Ito, Y., Ohzono, M., Tsu- Oxford, UK, Clarendon Press, 718 p.
shima, H., Suzuki, S., Fujimoto, H., and Miura, S., 2012, Coseismic Langbein, J.O., Burford, R.O., and Slater, L.E., 1990, Variation in fault slip
slip distribution of the 2011 off the Pacific Coast of Tohoku Earthquake and strain accumulation at Parkfield, California: Initial results using two-
(M9.0) refined by means of seafloor geodetic data: Journal of Geophysical color laser geodimeter measurements, 1984–1988: Journal of Geophysi-
Research, v. 117, B07409, doi:10.1029/2012JB009186. cal Research, v. 95, p. 2533–2552, doi:10.1029/JB095iB03p02533.
Ito, Y., Tsuji, T., Osada, Y., Kido, M., Inazu, D., Hayashi, Y., Tsushima, H., Langbein, J., Gwyther, R.L., Hart, R.H.G., and Gladwin, M.T., 1999, Slip-rate
Hino, R., and Fujimoto, H., 2011, Frontal wedge deformation near the increase at Parkfield in 1993 detected by high-precision EDM and bore-
source region of the 2011 Tohoku-Oki earthquake: Geophysical Research hole tensor strainmeters: Geophysical Research Letters, v. 26, p. 2529–
Letters, v. 38, L00G05, doi:10.1029/2011GL048355. 2532, doi:10.1029/1999GL900557.
Jackson, D.D., Lee, W.B., and Liu, C.C., 1980, Aseismic uplift in Southern Lawson, J.J., 1908, The California earthquake of April 18, 1906: Report of the
California: An alternative interpretation: Science, v. 210, p. 534–536, State Earthquake Investigation Commission: Washington, D.C., Carnegie
doi:10.1126/science.210.4469.534. Institution of Washington Publication, v. 87, 451 p.
Jacobs, A., Sandwell, D., Fialko, Y., and Sichoix, L., 2002, The 1999 (Mw 7.1) Lazarte, C.A., Bray, J.D., Johnson, A.M., and Lemmer, R.E., 1994, Surface-
Hector Mine, California, earthquake: Near-field postseismic deformation breakage of the 1992 Landers Earthquake and its effects on structures:
from ERS interferometry: Bulletin of the Seismological Society of Amer- Bulletin of the Seismological Society of America, v. 84, p. 547–561.
ica, v. 92, p. 1433–1442, doi:10.1785/0120000908. Leprince, S., Berthier, E., Ayoub, F., Delacourt, C., and Avouac, J.P., 2008, Mon-
Jennings, C.W., 1975, Fault map of California, with locations of volcanoes, itoring Earth Surface Dynamics with Optical Imagery: Eos (Transactions,
thermal springs, and thermal wells: California Division of Mines and American Geophysical Union), v. 89, p. 1–2, doi:10.1029/2008EO010001.
Geology, Geologic Data Map No. 1, scale 1:750,000. Li, Z., Muller, J., Cross, P., and Fielding, E. J., 2005, Interferometric syn-
Johanson, I.A., Fielding, E.J., Rolandone, F., and Bürgmann, R., 2006, Coseis- thetic aperture radar (InSAR) atmospheric correction: GPS, Moder-
mic and postseismic slip of the 2004 Parkfield earthquake from space- ate Resolution Imaging Spectroradiometer (MODIS), and InSAR
geodetic data: Bulletin of the Seismological Society of America, v. 96, integration: Journal of Geophysical Research, v. 110, B03410, doi:10
p. 269–282, doi:10.1785/0120050818. .1029/2004JB003446.
Johnson, E.A., 2006, Water in nominally anhydrous crustal minerals: Specia- Linde, A.T., Gladwin, M.T., Johnston, M.J.S., Gwyther, R.L., and Bilham,
tion, concentration, and geologic significance, in Keppler Hans, S.J.R., R.G., 1996, A slow earthquake sequence on the San Andreas fault: Nature,
ed., Water in Nominally Anhydrous Minerals: Chantilly, Virginia, Miner- v. 383, p. 65–68, doi:10.1038/383065a0.
alogical Society of America, v. 62, p. 117–154. Lisowski, M., Savage, J.C., and Prescott, W.H., 1991, The velocity field along
Johnson, K.M., Bürgmann, R., and Larson, K., 2006, Frictional properties on the San Andreas fault in central and southern California: Journal of Geo-
the San Andreas fault near Parkfield, California, inferred from models physical Research, v. 96, p. 8369–8389, doi:10.1029/91JB00199.
of afterslip following the 2004 earthquake: Bulletin of the Seismological Liu, Y., and Rice, J.R., 2007, Spontaneous and triggered aseismic deformation
Society of America, v. 96, p. 321–338, doi:10.1785/0120050808. transients in a subduction fault model: Journal of Geophysical Research,
Johnson, K., Hilley, G., and Bürgmann, R., 2007, Influence of lithosphere vis- v. 112, B09404, doi:10.1029/2007JB004930.
cosity structure on estimates of fault slip rate in the Mojave region of Lohman, R.B., and McGuire, J.J., 2007, Earthquake swarms driven by aseismic
the San Andreas fault system: Journal of Geophysical Research, v. 112, creep in the Salton Trough, California: Journal of Geophysical Research–
B07408, doi:10.1029/2006JB004842. Solid Earth, v. 112, B04405, doi:10.1029/2006JB004596.
Loveless, J.P., and Meade, B.J., 2010, Geodetic imaging of plate motions, slip Nadeau, R.M., and McEvilly, T.V., 1999, Fault slip rates at depth from recur-
rates, and partitioning of deformation in Japan: Journal of Geophysical rence intervals of repeating microearthquakes: Science, v. 285, p. 718–
Research, v. 115, B02410, doi:10.1029/2008JB006248. 721, doi:10.1126/science.285.5428.718.
Malservisi, R., Furlong, K.P., and Dixon, T.H., 2001, Influence of the earth- Nishimura, T., and Thatcher, W., 2003, Rheology of the lithosphere inferred
quake cycle and lithospheric rheology on the dynamics of the Eastern from postseismic uplift following the 1959 Hebgen Lake earthquake:
California Shear Zone: Geophysical Research Letters, v. 28, p. 2731– Journal of Geophysical Research, v. 108, doi:10.1029/2002JB002191.
2734, doi:10.1029/2001GL013311. Nur, A., and Mavko, G., 1974, Postseismic viscoelastic rebound: Science,
Marone, C.J., 1998, Laboratory-derived friction laws and their application to v. 183, p. 204–206, doi:10.1126/science.183.4121.204.
seismic faulting: Annual Review of Earth and Planetary Sciences, v. 26, Nyst, M., and Thatcher, W., 2004, New constraints on the active tectonic defor-
p. 643–696, doi:10.1146/annurev.earth.26.1.643. mation of the Aegean: Journal of Geophysical Research, v. 109, B11406,
Marone, C.J., Scholz, C.H., and Bilham, R.G., 1991, On the mechanics of doi:10.1029/ 2003JB002830.
earthquake afterslip: Journal of Geophysical Research, v. 96, p. 8441– Obara, K., 2004, Episodic slow slip events accompanied by non-volcanic trem-
8452, doi:10.1029/91JB00275. ors in southwest Japan subduction zone: Geophysical Research Letters,
Massonnet, D., and Feigl, K.L., 1998, Radar interferometry and its application v. 31, L23602, doi:10.1029/2004GL020848.
to changes in the earth’s surface: Reviews of Geophysics, v. 36, p. 441– Ohta, Y., Hino, R., Inazu, D., Ohzono, M., Ito, Y., Mishina, M., Iinuma, T., Nak-
500, doi:10.1029/97RG03139. ajima, J., Osada, Y., Suzuki, K., Fujimoto, H., Tachibana, K., Demachi,
Massonnet, D., Rossi, M., Carmona, C., Adragna, F., Peltzer, G., Feigl, K., T., and Miura, S., 2012, Geodetic constraints on afterslip characteristics
and Rabaute, T., 1993, The displacement field of the Landers Earth- following the March 9, 2011, Sanriku-oki earthquake, Japan: Geophysical
quake mapped by radar interferometry: Nature, v. 364, p. 138–142, Research Letters, v. 39, L16304, doi:10.1029/2012GL052430.
doi:10.1038/364138a0. Okada, A., and Nagata, T., 1953, Land deformation of the neighborhood of
Masterlark, T., 2003, Finite element model predictions of static deformation Muroto Point after the Nankaido great earthquake in 1946: Bulletin of
from dislocation sources in a subduction zone: Sensitivities to homoge- the Earthquake Research Institute, Tokyo University, v. 31, p. 169–177.
neous, isotropic, Poisson-solid, and half-space assumptions: Journal of Okada, Y., 1985, Surface deformation due to shear and tensile faults in a
Geophysical Research, v. 108, doi:10.1029/2002JB002296. half-space: Bulletin of the Seismological Society of America, v. 75,
Masterlark, T., DeMets, C., Wang, H.F., Sánchez, O., and Stock, J., 2001, p. 1135–1154.
Homogeneous vs heterogeneous subduction zone models: Coseismic and Oskin, M., Perg, L., Shelef, E., Strane, M., Gurney, E., Singer, B., and Zhang,
postseismic deformation: Geophysical Research Letters, v. 28, p. 4047– X., 2008, Elevated shear zone loading rate during an earthquake cluster in
4050, doi:10.1029/2001GL013612. eastern California: Geology, v. 36, doi:10.1130/G24814A.24811.
McCaffrey, R., 2002, Crustal block rotations and plate coupling, in Stein, S., Oskin, M.E., Arrowsmith, J.R., Corona, A.H., Elliott, A.J., Fletcher, J.M.,
and Freymueller, J., eds., Geodynamics Series: Washington, D.C., Ameri- Fielding, E.J., Gold, P.O., Garcia, J.J.G., Hudnut, K.W., Liu-Zeng, J., and
can Geophysical Union, v. 30, p. 101– 112. Teran, O.J., 2012, Near-field deformation from the El Mayor–Cucapah
McCaffrey, R., 2005, Block kinematics of the Pacific–North America plate Earthquake revealed by differential LIDAR: Science, v. 335, p. 702–705,
boundary in the southwestern United States from inversion of GPS, seis- doi:10.1126/science.1213778.
mological, and geologic data: Journal of Geophysical Research, v. 110, Panet, I., Pollitz, F., Mikhailov, V., Diament, M., Banerjee, P., and Grijalva, K.,
B07401, doi:10.1029/2004JB003307. 2010, Upper mantle rheology from GRACE and GPS postseismic defor-
McCaffrey, R., Qamar, A.I., King, R.W., Wells, R., Khazaradze, G., Wil- mation after the 2004 Sumatra-Andaman earthquake: Geochemistry Geo-
liams, C.A., Stevens, C.W., Vollick, J.J., and Zwick, P.C., 2007, Fault physics Geosystems, v. 11, Q06008, doi:10.1029/2009GC002905.
locking, block rotation and crustal deformation in the Pacific North- Parsons, T., and Thatcher, W., 2011, Diffuse Pacific–North American plate
west: Geophysical Journal International, v. 169, doi:10.1111/j.1365 boundary: 1000 km of dextral shear inferred from modeling geodetic
-1246X.2007.03371.x. data: Geology, v. 39, p. 943–946, doi:10.1130/G32176.1.
Meade, B.J., and Hager, B.H., 2005, Block models of crustal motion in south- Peltzer, G., Rosen, P., Rogez, F., and Hudnut, K., 1996, Postseismic rebound in
ern California constrained by GPS measurements: Journal of Geophysical fault step-overs caused by pore fluid flow: Science, v. 273, p. 1202–1204,
Research, v. 110, B03403, doi:10.1029/2004JB003209. doi:10.1126/science.273.5279.1202.
Mériaux, A.-S., Ryerson, F.J., Tapponnier, P., Van der Woerd, J., Finkel, R.C., Peltzer, G., Crampe, E., Hensley, S., and Rosen, P.A., 2001, Transient strain
Xu, X., Xu, Z., and Caffee, M.W., 2004, Rapid slip along the central accumulation and fault interaction in the Eastern California shear zone:
Altyn Tagh fault: Morphochronologic evidence from the Cherchen He Geology, v. 29, p. 975–978, doi:10.1130/0091-7613(2001)029<0975
and Sulama Tagh: Journal of Geophysical Research, v. 109, B06401, :TSAAFI>2.0.CO;2.
doi:10.1029/2003JB002558. Peng, Z., and Gomberg, J., 2010, An integrated perspective of the continuum
Michel, R., Avouac, J.-P., and Taboury, J., 1999, Measuring near field between earthquakes and slow-slip phenomena: Nature Geoscience, v. 3,
coseismic displacements from SAR images: Application to the Land- p. 599–607, doi:10.1038/ngeo940.
ers Earthquake: Geophysical Research Letters, v. 26, p. 3017–3020, Perfettini, H., Avouac, J.-P., and Ruegg, J.-C., 2005, Geodetic displacements and
doi:10.1029/1999GL900524. aftershocks following the 2001 Mw = 8.4 Peru earthquake: Implications
Mikumo, T., and Ando, M., 1976, A search into the faulting mechanism of for the mechanics of the earthquake cycle along subduction zones: Journal
the 1891 great Nobi earthquake: Journal of Physics of the Earth, v. 24, of Geophysical Research, v. 110, B09404, doi:10.1029/2004JB003522.
p. 63–87, doi:10.4294/jpe1952.24.63. Phillips, K.A., Chadwell, C.D., and Hildebrand, J.A., 2008, Vertical deforma-
Milne, G.A., Davis, J.L., Mitrovica, J.X., Scherneck, H.-G., Johansson, J.M., tion measurements on the submerged south flank of Kilauea volcano,
Vermeer, M., and Koivula, H., 2001, Space-geodetic constraints on gla- Hawai’i, reveal seafloor motion associated with volcanic collapse: Journal
cial isostatic adjustment in Fennoscandia: Science, v. 291, p. 2381–2385, of Geophysical Research, v. 113, B05106, doi:10.1029/2007JB005124.
doi:10.1126/science.1057022. Pollitz, F.F., 1992, Postseismic relaxation theory on the spherical earth: Bulletin
Minson, S.E., Simons, M., and Beck, J.L., 2013, Bayesian inversion for finite of the Seismological Society of America, v. 82, p. 422–453.
fault earthquake source models I—Theory and algorithm: Geophysical Pollitz, F., 2003, Transient rheology of the uppermost mantle beneath the
Journal International, doi:10.1093/gji/ggt180 (in press). Mojave Desert, California: Earth and Planetary Science Letters, v. 215,
Miyazaki, S., McGuire, J.J., and Segall, P., 2011, Seismic and aseismic fault p. 89–104, doi:10.1016/S0012-821X(03)00432-1.
slip before and during the 2011 off the Pacific coast of Tohoku earth- Pollitz, F.F., 2005, Transient rheology of the upper mantle beneath cen-
quake: Earth Planets Space, v. 63, doi:10.5047/eps.2011.07.001. tral Alaska inferred from the crustal velocity field following the 2002
Mossop, A., and Segall, P., 1997, Subsidence at The Geysers geothermal Denali earthquake: Journal of Geophysical Research, v. 11, B08407,
field, N. California, from a comparison of GPS and leveling sur- doi:10.1029/2005JB003672.
veys: Geophysical Research Letters, v. 24, p. 1839–1842, doi:10 Pollitz, F.F., Wicks, C., and Thatcher, W., 2001, Mantle flow beneath a continen-
.1029/97GL51792. tal strike-slip fault: Postseismic deformation after the 1999 Hector Mine
Murray, J., and Segall, P., 2005, Spatiotemporal evolution of a transient slip earthquake: Science, v. 293, p. 1814–1818, doi:10.1126/science.1061361.
event on the San Andreas fault near Parkfield, California: Journal of Geo- Pollitz, F., Banerjee, P., Grijalva, K., Nagarajan, B., and Bürgmann, R., 2008a,
physical Research, v. 110, B09407, doi:10.1029/2005JB003651. Effect of 3D viscoelastic structure on postseismic relaxation from the
Space geodesy 429
2004 M = 9.2 Sumatra earthquake: Geophysical Journal International, Schlüter, W., and Behrend, D., 2007, The International VLBI Service for Geod-
v. 173, p. 189–204, doi:10.1111/j.1365-246X.2007.03666.x. esy and Astrometry (IVS): Current capabilities and future prospects: Jour-
Pollitz, F.F., McCrory, P., Svarc, J., and Murray, J., 2008b, Dislocation models nal of Geodesy, v. 81, p. 379–387, doi:10.1007/s00190-006-0131-z.
of interseismic deformation in the western United States: Journal of Geo- Schmalzle, G., Dixon, T., Malservisi, R., and Govers, R., 2006, Strain accu-
physical Research, v. 113, B04413, doi:10.1029/2007JB005174. mulation across the Carrizo segment of the San Andreas Fault, Califor-
Pollitz, F., Bürgmann, R., and Thatcher, W., 2012, Illumination of rheo- nia: Impact of laterally varying crustal properties: Journal of Geophysical
logical mantle heterogeneity by the M 7.2 2010 El Mayor-Cucapah Research, v. 111, B05403, doi:10.1029/2005JB003843.
earthquake: Geochemistry Geophysics Geosystems, v. 13, Q06002, Schmidt, D.A., and Bürgmann, R., 2003, Time dependent land uplift and sub-
doi:10.1029/2012GC004139. sidence in the Santa Clara Valley, California, from a large InSAR data set:
Powell, R.E., and Weldon, R.J., 1992, Evolution of the San Andreas fault: Journal of Geophysical Research, v. 108, doi:10.1029/2002JB002267.
Annual Review of Earth and Planetary Sciences, v. 20, p. 431–468, Scholz, C.H., and Kato, T., 1978, The behavior of a convergent plate boundary:
doi:10.1146/annurev.ea.20.050192.002243. Crustal deformation in the south Kanto district, Japan: Journal of Geo-
Reid, H.F., 1910, Permanent displacements of the ground, in the California physical Research, v. 83, p. 783–797, doi:10.1029/JB083iB02p00783.
earthquake of April 18, 1906, in Report of the State Earthquake Investiga- Schwartz, D.P., and Coppersmith, K.J., 1984, Fault behavior and characteristic
tion Commission, Volume 2: Washington, D.C., Carnegie Institution of earthquakes: Examples from the Wasatch and San Andreas fault zones:
Washington, p. 16–28. Journal of Geophysical Research, v. 89, p. 5681–5698, doi:10.1029/
Reid, H.F., 1913, Sudden earth movements in Sumatra in 1892: Bulletin of the JB089iB07p05681.
Seismological Society of America, v. 3, p. 72–79. Schwartz, S.Y., and Rokosky, J.M., 2007, Slow slip events and seismic tremor at
Reilinger, R., McClusky, S., Vernant, P., Lawrence, S., Ergintav, S., Cakmak, R., circum-Pacific subduction zones: Reviews of Geophysics, v. 45, RG3004,
Ozener, H., Kadirov, F., Guliev, I., Stepanyan, R., Nadariya, M., Hahubia, doi:10.1029/2006RG000208.
G., Mahmoud, S., Sakr, K., ArRajehi, A., Paradissis, D., Al-Aydrus, A., Segall, P., 2002, Integrating geologic and geodetic estimates of slip rate on the
Prilepin, M., Guseva, T., Evren, E., Dmitrotsa, A., Filikov, S.V., Gomez, San Andreas fault system: International Geology Review, v. 44, p. 62–82,
F., Al-Ghazzi, R., and Karam, G., 2006, GPS constraints on continental doi:10.2747/0020-6814.44.1.62.
deformation in the Africa-Arabia-Eurasia continental collision zone and Segall, P., 2010, Earthquake and Volcano Deformation: Princeton, New Jersey,
implications for the dynamics of plate interactions: Journal of Geophysi- Princeton University Press, 458 p.
cal Research, v. 111, B05411, doi:10.1029/2005JB004051. Segall, P., and Davis, J.L., 1997, GPS applications for geodynamics and earth-
Rogers, G., and Dragert, H., 2003, Episodic tremor and slip on the Cascadia quake studies: Annual Review of Earth and Planetary Sciences, v. 25,
subduction zone: The chatter of silent slip: Science, v. 300, p. 1942–1943, p. 301–336, doi:10.1146/annurev.earth.25.1.301.
doi:10.1126/science.1084783. Segall, P., and Lisowski, M., 1990, Surface displacements in the 1906 San Fran-
Rosen, P.A., Hensley, S., and Joughin, I.R., 2000, Synthetic Aperture cisco and 1989 Loma Prieta earthquakes: Science, v. 250, p. 1241–1244,
Radar Interferometry: Proceedings of the IEEE, v. 88, p. 333–382, doi:10.1126/science.250.4985.1241.
doi:10.1109/5.838084. Segall, P., Rubin, A.M., Bradley, A.M., and Rice, J.R., 2010, Dilatant strength-
Ryder, I., Parsons, B.E., Wright, T.J., and Funning, G.J., 2007, Postseismic ening as a mechanism for slow slip events: Journal of Geophysical
motion following the 1997 Manyi (Tibet) earthquake: InSAR observa- Research, v. 115, B12305, doi:10.1029/2010JB007449.
tions and modeling: Geophysical Journal International, v. 169, doi: Sella, G.F., Stein, S., Dixon, T.H., Craymer, M., James, T.S., Mazzotti, S., and
10.1111/j.1365-1246X.2006.03312.x. Dokka, R., 2007, Observation of glacial isostatic adjustment in stable
Ryder, I., Bürgmann, R., and Pollitz, F., 2011, Lower crustal relaxation beneath North America with GPS: Geophysical Research Letters, v. 34, L02306,
the Tibetan Plateau and Qaidam Basin following the 2001 Kokoxili earth- doi:10.1029/2006GL027081.
quake: Geophysical Journal International, v. 187, doi: 10.1111/j.1365- Shapiro, I., Robertson, D., Knight, C., Counselman, C., Rogers, A., Hintereg-
1246X.2011.05179.x. ger, H., Lippincott, S., Whitney, A., Clark, T., Niell, A., and Spitzmesser,
Salvi, S., Stramondo, S., Funning, G.J., Ferretti, A., Sarti, F., and Mouratidis, A., D., 1974, Transcontinental baselines and the rotation of the Earth mea-
2012, The Sentinel-1 mission for the improvement of the scientific under- sured by radio interferometry: Science, v. 186, p. 920–922, doi:10.1126/
standing and the operational monitoring of the seismic cycle: Remote science.186.4167.920.
Sensing of Environment, v. 120, doi:10.1016/j.rse.2011.1009.1029. Shelly, D.R., and Hardebeck, J., 2010, Precise tremor source locations and ampli-
Sato, M., Ishikawa, T., Ujihara, N., Yoshida, S., Fujita, M., Mochizuki, M., and tude variations along the lower-crustal central San Andreas Fault: Geo-
Asada, A., 2011, Displacement above the hypocenter of the 2011 Tohoku- physical Research Letters, v. 37, L14301, doi:10.1029/2010GL043672.
Oki Earthquake: Science, v. 332, 6036, doi:10.1126/science.1207401. Shen, Z.K., King, R.W., Agnew, D.C., Wang, M., Herring, T.A., Dong, D., and
Sauber, J., Thatcher, W., and Solomon, S.C., 1986, Geodetic measurement of Fang, P., 2011, A unified analysis of crustal motion in Southern Califor-
deformation in the central Mojave Desert, California: Journal of Geo- nia, 1970–2004: The SCEC crustal motion map: Journal of Geophysical
physical Research, v. 91, p. 12,683–12,693. Research, v. 116, B11402, doi:10.1029/2011JB008549.
Savage, J.C., 1983, A dislocation model of strain accumulation and release at a Shirzaei, M., 2013, A wavelet based multitemporal DInSAR algorithm for
subduction zone: Journal of Geophysical Research, v. 88, p. 4984–4996, monitoring ground surface motion: Geoscience and Remote Sensing Let-
doi:10.1029/JB088iB06p04984. ters, v. 10, p. 456–460, doi:10.1109/LGRS.2012.2208935.
Savage, J.C., 1990, Equivalent strike-slip earthquake cycles in half-space Shirzaei, M., and Bürgmann, R., 2012, Topography correlated atmospheric
and lithosphere-asthenosphere earth models: Journal of Geophysical delay correction in radar interferometry using wavelet transforms: Geo-
Research, v. 95, p. 4873–4879, doi:10.1029/JB095iB04p04873. physical Research Letters, v. 39, L01305, doi:10.1029/2011GL049971.
Savage, J.C., and Burford, R.O., 1970, Accumulation of tectonic strain in Shirzaei, M., and Bürgmann, R., 2013, Time-dependent model of creep on the
California: Bulletin of the Seismological Society of America, v. 60, Hayward fault from joint inversion of 18 years of InSAR and surface
p. 1877–1896. creep data: Journal of Geophysical Research, doi:10.1002/jgrb.50149
Savage, J.C., and Burford, R.O., 1973, Geodetic determination of relative plate (in press).
motion in central California: Journal of Geophysical Research, v. 78, Sibson, R.H., 1982, Fault zone models, heat flow, and the depth distribution of
p. 832–845, doi:10.1029/JB078i005p00832. earthquakes in the continental crust of the United States: Bulletin of the
Savage, J.C., and Prescott, W.H., 1973, Precision of Geodolite distance mea- Seismological Society of America, v. 72, p. 151–163.
surements for determining fault movements: Journal of Geophysical Sibson, R.H., 1984, Roughness at the base of the seismogenic zone: Contrib-
Research, v. 78, p. 6001–6007, doi:10.1029/JB078i026p06001. uting factors: Journal of Geophysical Research, v. 89, p. 5791–5800,
Savage, J.C., and Prescott, W.H., 1978, Asthenosphere readjustment and the doi:10.1029/JB089iB07p05791.
earthquake cycle: Journal of Geophysical Research, v. 83, p. 3369–3376, Sibson, R.H., 2009, Rupturing in overpressured crust during compressional
doi:10.1029/JB083iB07p03369. inversion—The case from NE Honshu, Japan: Tectonophysics, v. 473,
Savage, J.C., Lisowski, M., and Prescott, W., 1990, An apparent shear zone p. 404–416, doi:10.1016/j.tecto.2009.03.016.
trending north-northwest across the Mojave Desert into Owens Valley, Simons, M., and Rosen, P.A., 2007, Interferometric Synthetic Aperture radar
Eastern California: Geophysical Research Letters, v. 17, p. 2113–2116, geodesy, in Schubert, G., ed., Treatise on Geophysics: New York, Else-
doi:10.1029/GL017i012p02113. vier, v. 3, p. 391–443.
Simpson, R.W., Thatcher, W., and Savage, J.C., 2012, Using cluster analysis Vanicek, P., Castle, R.O., and Balazs, E.I., 1980, Geodetic leveling and its
to organize and explore regional GPS velocities: Geophysical Research applications: Reviews of Geophysics, v. 18, p. 505–524, doi:10.1029/
Letters, 39, L18307, doi:10.1029/2012GL052755. RG018i002p00505.
Song, S.G., Beroza, G.C., and Segall, P., 2008, A unified source model for the Vermaat, E., Degnan, J.J., Dunn, P.J., Noomen, R., and Sinclair, A.T., 1998,
1906 San Francisco earthquake: Bulletin of the Seismological Society of Satellite laser ranging, status and impact for WEGENER: Journal of Geo-
America, v. 98, p. 823–831, doi:10.1785/0120060402. dynamics, v. 25, p. 195–212, doi:10.1016/S0264-3707(97)00036-7.
Spiess, F.N., 1985, Suboceanic geodetic measurements: Geoscience and Wallace, L.M., Beavan, R.J., McCaffrey, R., and Darby, D.J., 2004, Subduc-
Remote Sensing: IEEE Transactions, v. GE-23, p. 502–510. tion zone coupling and tectonic block rotations in the North Island,
Spiess, F.N., Chadwell, C.D., Hildebrand, J.A., Young, L.E., Purcell, J.G.H., New Zealand: Journal of Geophysical Research, v. 109, B12406,
and Dragert, H., 1998, Precise GPS/acoustic positioning of seafloor refer- doi:10.1029/2004JB003241.
ence points for tectonic studies: Physics of the Earth and Planetary Interi- Wang, K., Hu, Y., Bevis, M., Kendrick, E., Smalley, R., Jr., Vargas, R.B., and
ors, v. 108, p. 101–112, doi:10.1016/S0031-9201(98)00089-2. Lauría, E., 2007, Crustal motion in the zone of the 1960 Chile earth-
Steketee, J.A., 1958, Some geophysical applications of the elasticity the- quake: Detangling earthquake-cycle deformation and forearc-sliver
ory of dislocation: Canadian Journal of Physics, v. 36, p. 1168–1197, translation: Geochemistry Geophysics Geosystems, v. 8, Q10010,
doi:10.1139/p58-123. doi:10.1029/2007GC001721.
Taira, T., Silver, P.G., Niu, F., and Nadeau, R.M., 2009, Remote triggering Wang, K., Hu, Y., and He, J., 2012a, Deformation cycles of subduction earth-
of fault-strength changes on the San Andreas fault at Parkfield: Nature, quakes in a viscoelastic Earth: Nature, v. 484, p. 327–332, doi:10.1038/
v. 461, doi:10.1038/nature08395. nature11032.
Tamisiea, M.E., Mitrovica, J.X., and Davis, J.L., 2007, GRACE gravity data Wang, L., Shum, C.K., Simons, F.J., Tapley, B., and Dai, C., 2012b, Coseis-
constrain ancient ice geometries and continental dynamics over Lauren- mic and postseismic deformation of the 2011 Tohoku-Oki earthquake
tia: Science, v. 316, p. 881–883, doi:10.1126/science.1137157. constrained by GRACE gravimetry: Geophysical Research Letters, v. 39,
Tanaka, Y., Okubo, S., Machida, M., Kimura, I., and Kosuge, T., 2001, First L07301, doi:10.1029/2012GL051104.
detection of absolute gravity change caused by earthquake: Geophysical Ward, S.N., 1990, Pacific–North America plate motions: New results from very
Research Letters, v. 28, p. 2979–2981, doi:10.1029/2000GL012590. long baseline interferometry: Journal of Geophysical Research, v. 95,
Tapley, B.D., Bettadpur, S., Watkins, M., and Reigber, C., 2004, The gravity p. 21,965–21,981, doi:10.1029/JB095iB13p21965.
recovery and climate experiment: Mission overview and early results: Geo- Wdowinski, S., Smith-Konter, B., Bock, Y., and Sandwell, D.T., 2007, Spatial
physical Research Letters, v. 31, L09607, doi:10.1029/2004GL019920. characterization of the interseismic velocity field in southern California:
Thatcher, W., 1975, Strain accumulation and release mechanism of the 1906 Geology, v. 35, doi:10.1130/G22938A.22931.
San Francisco earthquake: Journal of Geophysical Research, v. 80, Wegener, A., 1929, The Origin of Continents and Oceans: Dover Earth Science.
p. 4862–4872, doi:10.1029/JB080i035p04862. Wei, M., Sandwell, D., and Fialko, Y., 2009, A silent Mw 4.7 slip event of
Thatcher, W., 1983, Nonlinear strain buildup and the earthquake cycle on the October 2006 on the Superstition Hills fault, southern California: Journal
San Andreas fault: Journal of Geophysical Research, v. 88, p. 5893–5902, of Geophysical Research, v. 114, B07402, doi:10.1029/2008JB006135.
doi:10.1029/JB088iB07p05893. Wei, M., Sandwell, D., Fialko, Y., and Bilham, R., 2011a, Slip on faults in the
Thatcher, W., 2003, GPS constraints on the kinematics of continental defor- Imperial Valley triggered by the 4 April 2010 Mw 7.2 El Mayor–Cuca-
mation: International Geology Review, v. 45, doi:10.2747/0020-6814 pah earthquake revealed by InSAR: Geophysical Research Letters, v. 38,
.2745.2743.2191. L01308, doi:10.1029/2010GL045235.
Thatcher, W., 2007, Microplate model for the present-day deforma- Wei, S., Fielding, E., Leprince, S., Sladen, A., Avouac, J.-P., Helmberger, D.,
tion of Tibet: Journal of Geophysical Research, v. 112, B01401, Hauksson, E., Chu, R., Simons, M., Hudnut, K., Herring, T., and Briggs,
doi:10.1029/2005JB004244. R., 2011b, Superficial simplicity of the 2010 El Mayor–Cucapah earth-
Thatcher, W., 2009, How the continents deform: The evidence from tec- quake of Baja California in Mexico: Nature Geoscience, v. 4, p. 615–618,
tonic geodesy: Annual Reviews of Earth and Planetary Science, v. 37, doi:10.1038/ngeo1213.
doi:10.1146/annurev.earth.031208.100035. Weston, J., Ferreira, A.M.G., and Funning, G.J., 2011, Global compilation of
Thatcher, W., and Lisowski, M., 1987, Long-term seismic potential of the San interferometric synthetic aperture radar earthquake source models: 1.
Andreas fault southeast of San Francisco, California: Journal of Geo- Comparisons with seismic catalogs: Journal of Geophysical Research,
physical Research, v. 92, p. 4771–4784, doi:10.1029/JB092iB06p04771. v. 116, B08408, doi:10.1029/2010JB008131.
Thatcher, W., and Pollitz, F.F., 2008, Temporal evolution of continental litho- Whitten, C.A., 1948, Horizontal earth movement, vicinity of San Francisco,
spheric strength in actively deforming regions: GSA Today, v. 18, no. 4/5, California: Eos (Transactions, American Geophysical Union), v. 29,
doi:10.1130/GSAT01804-01805A.01801. p. 318–323, doi:10.1029/TR029i003p00318.
Thatcher, W., and Rundle, J.B., 1979, A model for the earthquake cycle in Working Group on California Earthquake Probabilities (WGCEP), 2008, The
underthrust zones: Journal of Geophysical Research, v. 84, p. 5540–5556, uniform California earthquake rupture forecast, version 2 (UCERF 2):
doi:10.1029/JB084iB10p05540. U.S. Geological Survey Open-File Report 2007-1437, http://pubs.usgs
Thatcher, W., Marshall, G.A., and Lisowski, M., 1997, Resolution of fault slip .gov/of/2007/1437/.
along the 470-km-long rupture of the great 1906 San Francisco earth- Yeats, R.S., Sieh, K., and Allen, C.R., 1997, The Geology of Earthquakes: New
quake and its implications: Journal of Geophysical Research, v. 102, York, Oxford University Press, 569 p.
p. 5353–5367, doi:10.1029/96JB03486. Yu, E., and Segall, P., 1996, Slip in the 1868 Hayward earthquake from the
Thatcher, W., Foulger, G.R., Julian, B.R., Svarc, J., Quilty, E., and Bawden, analysis of historical triangulation data: Journal of Geophysical Research,
G.W., 1999, Present-day deformation across the Basin and Range prov- v. 101, p. 16,101–16,118.
ince, western United States: Science, v. 283, p. 1714–1718, doi:10.1126/ Zweck, C., Freymueller, J.T., and Cohen, S.C., 2002, The 1964 great Alaska
science.283.5408.1714. earthquake: Present day and cumulative postseismic deformation in the
Tong, X., Sandwell, D., and Smith-Konter, B., 2013, High-resolution interseis- western Kenai Peninsula: Physics of the Earth and Planetary Interiors,
mic velocity data along the San Andreas fault from GPS and InSAR: Jour- v. 132, p. 5–20, doi:10.1016/S0031-9201(02)00041-9.
nal of Geophysical Research, v. 118, doi:10.1029/2012JB009442.
Townend, J., and Zoback, M.D., 2000, How faulting keeps the crust strong:
Geology, v. 28, p. 399–402, doi: 10.1130/0091-7613(2000)28<399
:HFKTCS>2.0.CO;2. MANUSCRIPT ACCEPTED BY THE SOCIETY 15 JANUARY 2013
Printed in the USA

Geological Society of America Special Papers

Space geodesy: A revolution in crustal deformation measurements of tectonic
processes
Roland Bürgmann and Wayne Thatcher
Geological Society of America Special Papers 2013;500; 397-430

doi:10.1130/2013.2500(12)
E-mail alerting services click www.gsapubs.org/cgi/alerts to receive free e-mail alerts when new articles cite
this article
Subscribe click www.gsapubs.org/subscriptions to subscribe to Geological Society of America

Special Papers
Permission request click www.geosociety.org/pubs/copyrt.htm#gsa to contact GSA.
Copyright not claimed on content prepared wholly by U.S. government employees within scope of their
employment. Individual scientists are hereby granted permission, without fees or further requests to GSA,
to use a single figure, a single table, and/or a brief paragraph of text in subsequent works and to make
unlimited copies of items in GSA's journals for noncommercial use in classrooms to further education and
science. This file may not be posted to any Web site, but authors may post the abstracts only of their
articles on their own or their organization's Web site providing the posting includes a reference to the
article's full citation. GSA provides this and other forums for the presentation of diverse opinions and
positions by scientists worldwide, regardless of their race, citizenship, gender, religion, or political
viewpoint. Opinions presented in this publication do not reflect official positions of the Society.
Notes
© 2013 Geological Society of America

Space Geodesy: A Revolution in Crustal Deformation Measurements of Tectonic Processes

Uploaded by

Copyright:

Available Formats

Space Geodesy: A Revolution in Crustal Deformation Measurements of Tectonic Processes

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Space Geodesy: A Revolution in Crustal Deformation Measurements of Tectonic Processes

Uploaded by

Copyright:

Available Formats

Downloaded from specialpapers.gsapubs.

org on March 13, 2016

The Geological Society of America 18 8 8 2 013

Space geodesy: A revolution in

398 Bürgmann and Thatcher

rheological properties, and dynamics of crustal deformation. Increasingly, seafloor

1906 Displacements -0.5

400 Bürgmann and Thatcher

Space geodesy 401

402 Bürgmann and Thatcher

Space geodesy 403

404 Bürgmann and Thatcher

Space geodesy 405

406 Bürgmann and Thatcher

Space geodesy 407

408 Bürgmann and Thatcher

large submarine deformations can be captured with repeated

CRUSTAL DEFORMATION AND ACTIVE TECTONICS

The Earth’s surface is composed of a mosaic of a dozen or

Sierra Nevada - Great Valley Block

410 Bürgmann and Thatcher

Eastern Basin 41°

Pacific Plate 35°

km 33° 2nd Invariant of strain rate tensor (10 -9/yr)

412 Bürgmann and Thatcher

Bay Block (BB)

Space geodesy 413

414 Bürgmann and Thatcher

Space geodesy 415

−116°00' −115°30' −115°00' −114°30'

along strike offset (m)

Space geodesy 417

418 Bürgmann and Thatcher

Space geodesy 419

A NE Japan B Chile C Cascadia

140˚E 145˚E 75˚W 70˚W 125˚W 120˚W

420 Bürgmann and Thatcher

Space geodesy Coseismic stress change 421

Shear stress (MPa)

Shear stress (MPa)

422 Bürgmann and Thatcher

Space geodesy 423

424 Bürgmann and Thatcher

Space geodesy 425

426 Bürgmann and Thatcher

Space geodesy 427

428 Bürgmann and Thatcher

Space geodesy 429

430 Bürgmann and Thatcher

Printed in the USA

Geological Society of America Special Papers

Geological Society of America Special Papers 2013;500; 397-430

Subscribe click www.gsapubs.org/subscriptions to subscribe to Geological Society of America

© 2013 Geological Society of America

You might also like