Page 121 - Global Tectonics

108 CHAPTER 5

scale problems it is necessary to be able to measure ends of baselines are determined from the signals
across very large distances to very great accuracy. Ter- received at the instruments from several satellites.
restrial methods are extremely time consuming on land, The simultaneous observation of multiple satellites
and impossible to use across major oceans. Since 1980, makes extremely accurate measurements possible
however, the measurement of very long baselines using with small portable receivers. This is now the most

extraterrestrial methods has become possible via the efficient and accurate method of establishing geodetic
application of space technology. control on both local and regional surveys (e.g. Sec-
Three independent methods of extraterrestrial sur- tions 8.5.2, 10.4.3).
veying are available. These are very long baseline inter- Gordon & Stein (1992) summarized the early deter-
ferometry, satellite laser ranging, and satellite radio minations of relative plate motions by these methods.
positioning. The most common and best known Generally, plate velocities averaged over a few years
example of the latter method is the Global Positioning of observation agree remarkably well with those aver-
System (GPS). aged over millions of years. The methods were fi rst
The technique of very long baseline interferometry applied to the measurement of the rate of movement
(VLBI) makes use of the radio signals from extraga- across the San Andreas Fault in California. Smith et
lactic radio sources or quasars (Niell et al., 1979; al. (1985), using SLR, reported that a 900 km baseline
Carter & Robertson, 1986; Clark et al., 1987). The that crossed the fault at an angle of 25° had been
−1
signal from a particular quasar is recorded simultane- shortened at an average rate of 30 mm a . Lyzenga
ously by two or more radio telescopes at the ends of et al. (1986) have used VLBI to measure the length
baselines which may be up to 10,000 km long. Because of several baselines in the southwestern USA and have
of their different locations on the Earth’s surface, the found that over a period of 4 years movement on
−1
signals received at the telescopes are delayed by differ- the fault was 25 ± 4 mm a . These direct measure-
ent times, the magnitude of the delays between two ments of the rate of displacement across the San
−1
stations being proportional to the distance between Andreas Fault are lower than the 48–50 mm a pre-
them and the direction from which the signals are dicted from global models of plate movements (DeMets
coming. Typically, during a 24-hour experiment, 10–15 et al., 1990). However, during the period of observa-
quasars are each observed 5–15 times. This scheme tion, no major earthquakes occurred. Over longer
provides estimates of baseline length that are accurate time intervals, the discrete jumps in fault movement
to about 20 mm (Lyzenga et al., 1986). The usefulness associated with the elastic rebound mechanism of
of this system has been greatly enhanced by the devel- large earthquakes (Section 2.1.5) would contribute to
opment of mobile radio telescopes that frees the tech- the total displacement and provide a somewhat higher

nique from the necessity of using fi xed observatory figure for the average rate of movement. Alternatively,

installations. motion between the Pacific and North American plates
The technique of satellite laser ranging (SLR) calcu- may be occurring along other major faults located

lates the distance to an orbiting artificial satellite or a adjacent to the San Andreas Fault (Fig. 8.1, Section

reflector on the Moon by measuring the two-way travel 8.5.2).

time of a pulse of laser light reflected from the satellite Tapley et al. (1985), using SLR, measured changes in
(Cohen & Smith, 1985). The travel time is subsequently length of four baselines between Australia and the

converted to range using the speed of light. If two laser North American and Pacific plates, and found that the
−1
systems at different sites simultaneously track the same rates differ by no more than 3 mm a from average rates
satellite, the relative location of the sites can be com- over the last 2 Ma. Similarly Christodoulidis et al. (1985)
puted by using a dynamic model of satellite motion, and and Carter & Robertson (1986) measured the relative
repeated measurements provide an accuracy of about motion between pairs of plates and found a strong
80 mm. Periodic repetition of the observations can then correlation with the kinematic plate model of Minster
be used to observe relative plate motions (Christodou- & Jordan (1978). Herring et al. (1986) made VLBI
lidis et al., 1985). measurements between various telescopes in the USA
The technique of satellite radio positioning makes and Europe and determined that the present rate of
use of radio interferometry from the GPS satellites movement across the Atlantic Ocean is 19 ± 10 mm a .
−1
−1
(Dixon, 1991). It is a three-dimensional method by This agrees well with the rate of 23 mm a averaged
which the relative positions of instruments at the over the past 1 Ma.