AN INTRODUCTION TO THE GLOBAL POSITIONING

SYSTEM AND SOME GEOLOGICAL APPLICATIONS

T. H. Dixon
JetPropulsionLaboratory
Divisionof EarthandPlanetarySciences
Pasadena,California

Abstract.Receiversequippedto measuredualfrequency data acquisitionand analysisare havinga significantim-

carrierphasesignalsfrom satellites of theGlobalPosition- pact on studies of near-fault crustal deformation and
ing System(GPS)havebeencapable,underspecialcondi- earthquakeprocesses,until recently the province of con-
tions,of determiningrelative horizontalpositionsamong ventional terrestrialgeodetictechniques.The enhanced
stationsseparatedby oneto a few hundredkilometerswith satelliteconstellation,
improvedmodels,and establishment
a precisionof one to severalmillimeterssincethe early of global tracking networkshave extendedseveral mil-
1980s. The major obstaclesto making this capability limetershorizontalpositioningcapabilityto stationsepara-
routine,extendingit to all partsof theglobe,andextending tionsof 1000km or morein virtuallyall partsof the world.
it to longerstationseparations, havebeenequipmentcost, Thisenablesstudyof new classesof tectonicproblemsthat
limitations in the GPS satellite constellation,arduousdata previouslywere difficult to attackwith any geodetictech-
analysis,uncertainties in satelliteorbits,uncertainties in nique.Examplesincludea completekinematicdescription
propagationdelaysassociatedwith variabletropospheric of ongoingcrustaldeformationin broad, complexconti-
watervapor,and difficultiesin resolvingcarrierphasecy- nentalplate boundaryzones,and measurementof relative
cle ambiguities. Recentimprovements haveoccurredin all plate motionat convergentboundarieswhere global mod-
theseareas.The increasingeaseandreducedcostof GPS elsmaybe poorlyconstrained.

INTRODUCTION extendingthe range and accuracyof GPS measurements

are emphasized.I haveaimed for broadcoverageof most
With the adventin the early 1980sof a satellite-based of therelevanttopicsandan intuitiveratherthancomplete
navigationsystemknownastheGlobalPositioning System or rigorous treatment, giving more derailed references
(GPS) operatedby the U.S. Departmentof Defense it whereappropriate.
becamepossiblefor a user with the properreceiverto
obtain almost instantaneousthree-dimensionalposition
informationaccurateto severalmeters.With the comple- A BRIEF HISTORY OF THE GPS PROGRAM
tion of the satelliteconstellationin the early 1990s this
capabilitywill be extendedto virtually all parts of the The spacesegmentof GPS is a constellationof satellites
globe.This is a remarkableachievement and buildson a in highEarthorbitequippedwithpowerfulradiofrequency
number of technologicaladvancesin the last several transmitters andhighlystableatomicclocks.Easton[1978]
decades. Even more remarkable is the fact that. with careful reviewsthe majordevelopments leadingto thiscapability.
attentiontoexperiment configuration anddat•analysis itis In 1967 an early prototypeof a GPS satelliteknown as
possibleto obtain relative position data 3 orders of Timation 1 was launchedinto low Earth orbit (-900 km
magnitudemore precise than the design level of the altitude) as part of a military test program in satellite
system.This enhancedperformanceallows for measure- navigation. Weighingabout40 kg and consuming only 6
ment of crustalstrainand fault motionratesin just a few W of power,it carrieda UHF transmitterslavedto a stable
)rears. quartz
clock,
witha frequency
driftofseveral
parts
in 10•
This paper reviews fundamentalprinciplesof GPS, per day. Additional proof-of-conceptsatellitesfollowed,
discussessome geologicaland geophysicalapplications culminating10 yearslater in the NavigationTechnology
and their accuracyrequirements,and considersimplica- Satellite (NTS) 2, very similar to subsequentGPS
tions for GPS experimentdesign.Recent developments satellites.NTS-2 was launched into a 20,300-km-altitude

Copyright1991 by theAmericanGeophysical
Union. Reviews
of Geophysics,
29, 2 / May 1991
pages249-276
8755-1209/91/91RG-00152 $05.00 Papernumber91RG00152
ß 249 ß
250 ß Dixon: THE GLOBAL POSITIONING SYSTEM 29, 2/REVIEWS OF GEOPHYSICS

orbit, weighed440 kg, consumed400 W of power, and

oneanotherin theskyprovidecorrelated (redundant) range
transmittedtwo L band (-1.2 and 1.5 GHz) timing and
information,an effect known as geometricdilution of
precision(GDOP). If observationsare limited to four
ranging signalsbased on a sophisticatedcesium clock,
witha frequency
driftof lessthantwopartsin 10•3per satellitesby receiverdesign,thesegeometriceffectscanbe
day. One year after NTS-2, the first "Block 1" GPS minimized by choosingsatelliteswhich maximize the
satellitewas launched,part of the operationaltestphaseof volumeof a tetrahedron, definedby the pointsof intersec-
the GPS program. By 1990 the Block 1 constellation tion on a unit spherecenteredon the user, of vectors
includedsix functioningsatelliteslaunchedbetween1978 betweenthe satelliteandthegroundreceiver.
and 1985. It is alsopossibleto obtaindistanceinformation (strictly
It hasbeenrecognizedfor sometimethathigh-precision speaking,the changein distance)throughphasemeasure-
geodetic measurementscould be made by exploiting mentson the carrier signal itself, keepingtrack of the
signalsfrom artificial satellites[e.g.,Prestonet al., 1972; numberof cycles after signal acquisition.Assuming
MacDoran, 1979; Counselmanand Shapiro, 1979]. The perfectclocks,andignoringpropagation effects,
Block 1 constellationhasprovensatisfactoryfor develop-
ing andrefiningexperimentdesignandanalyticalconcepts
and for initiating a number of high-precisiongeodetic
monitoringprograms.The first Block 2 satellite,with a = (v,/œ) (n+ 0)) (2)
numberof improvementsrelative to its forebears,was
wheren is the numberof integercarrierwavelengths at
launchedin February1989.As of thiswriting,a totalof 10
signalacquisition(initially unknown),t• is the phasein
Block 2 satellitesare in operation.A total of 21 Block 2
satellitesplus threesparesare plannedto be in operation cycles, )• isthewavelength,f isthefrequency, andv, isthe
by the end of 1992. They will orbit at an altitudeof about phase velocity (the importance of distinguishingv, and
20,000 km in six orbital planes with 12-hour periods, group velocity, vg,will become apparent). Sincethe
wavelengthof the carrieris considerably shorterthanthat
enablingsimultaneous observationof four or more GPS
of the lower frequencycode modulations(Table 1), the
satellitesin virtuallyall partsof theglobe.
resulting"distance"measurement,thoughambiguous by
the initial numberof wavelengths,is considerably
more
POINT POSITIONING WITH GPS
,,,-ooioo
tho,, o pseudorangemeasurementand is one of the
keys to high-precision GPS measurements [Bossleret al.,
1980;Counselman and Gourevitch,1981;Remondi,1985].
RangeMeasurement
Carrier phaseis not measureddirectly, as this would
An observeron Earth can uniquelylocatehis position
requirevery high samplingrates.Rather,the signalis
by determiningthe distancebetweenhimself and three
mixed ("heterodyned")with a signal generatedby the
satelliteswhoseorbital positionsare alreadyknown.With
receiver's internal clock (local oscillator) and, after
GPS, distanceinformation is basedon the travel time x of a
band-passor low-pass filtering, the resulting lower
satellite signal, obtained by measuringthe difference
frequency"carrierbeat phase"is sampled.Most current
between
thetransmit
(t•)andreceive
(tr)timesat theGPS generationreceiversaccomplishthis "downconversion"
receiverof a specialrangingcode,describedin the next
section.If we ignore transmissionmedia effects on the
with electronics
that includeextensiveanalogcircuitry.
Some newer generationreceivershave largely digital
speedof light c andany timing(clock)errors,thenthetrue
architecture,
reducingproductioncost, size, and power
rangep between satelliteandreceiveris justC(tr -- rs). consumption and enablingdigital samplingof the carrier
Errorsin receiveror satelliteclocksarepresentin therange
phasesignalwith only minimalpreprocessing[Melbourne,
estimate,which for this reasonis referredto as pseudo-
1990].
rangeR, definedmorepreciselyas
TABLE 1. Summaryof GPS SignalCharacteristics
R = p + c(Atr - Ats+ Atp) (1)
Carriers Code Modulations
C/A
whereAtr isthereceiver
clockoffsetfrom"true"(GPSsys- L1 L2 P (L1 only)
tem) time (we ignore any other receiver-induced
errors),
clockoffset,andAt isthedelayassoci- Frequency
At•is thesatellite (carrier) 1.57542 1.2276 10.23 1.023
ated
withallother
error
sources,
rn•ainly
duetoatmos- or chiprate
(codemodulation)
GHz GHz MHz MHz
phericpropagation
effects.Informationfrom a fourthsatel- Wavelength 19.0cm 24.4cm ~30m ~300m
liteallows
a first-order
clockcorrection
(Atr - Ats),andap-
proachesdiscussed below can be appliedto estimateand
correct for At, enabling meter-levelpositioningunder
idealconditionS. For analysispurposeswe considerthe pseudorangeor
Observation
geometryaffectsthequalityof theresulting phaseparameters in termsof what the receiveractually
three-dimensional
position.Satellitesthat appearcloseto sees(the "observable"),
explicitlyaccountingfor major
29, 2 / REVIEWSOF GEOPHYSICS Dixon'THE GLOBALPOSITIONINGSYSTEMß 251

error sources.For example,in simplifiedform the phase wherem• andm2 are the angularfrequencies (m = 2•rf)
observable, sometimes called integrated Doppler or associated withtheL1andL2carriers andA• andAcare
accumulated deltarange,canbe written(in unitsof cycles) the relative amplitudesof the P and C/A codes. The
for a singlereceiver-satellitepairas amplitudeof the C/A signalis higherto facilitateinitial
signalacquisition.One effect of the PRN code modula-
•(t) =-f'c(t) + v(t) + a (3) tionson the carriersis to spreadthe energyof the P code
signal+10.23 MHz aroundthe carriercenterfrequency;
wherethedelayx(= p/c) is thegeometric delaydueonlyto becauseof this wide bandwidththe GPS signal is often
thesatellite-stationgeometry,ignoringpropagation effects, referred to as "spread spectrum."Figure 1 illustrates
v represents errors("noise")includingpropagation effects, schematicallyhow a carrier signalis biphasemodulated
and a is a constant,representing a combinationof the with a PRN code.A receiverwith knowledgeof the code
carrierphasecycleambiguityand an initial phaseoffset structure and an internal clock can recover an estimate of
between satellite and receiver oscillators.Equation (3) signaltransittime by cogenerating the codesequence and
ignorestime-dependent clockerrorsin both the satellite performinga crosscorrelationbetweenthe receivedsignal
andreceiver;also,the variouseffectsrepresented by v and andits internalcode,determining thetimedelaynecessary
a mustbe considered in moredetail.King et al. [1985]and to matchthe two sequences.
Leick [1990] give a completederivationof the phase P-CODE CHIP
1 2 3 4 5 6 7 8 9 1011121314-..
observable equations. SEQUENCE

C/A CODE
1 1 1 1 1 1 1 1 1 1 0 0 0 0...
SignalStructure VALUE

The followingdiscussion
is abbreviatedfrom Spilker C/A CODE
+1

[1978]. GPS satellites transmit two L-band carrier SCHEMATIC

frequencies,each modulatedby severallower frequency P-CODE 1 0 1 1 0 1 0 0 1 0 0 0 1 1"'

VALUE
signals(Table1). Thecarriers(L1 at 1.57542GHz andL2
at 1.22760 GHz) are coherentmultiplesof a 10.23-MHz P-CODE
SCHEMATIC
atomicclock,a stableoscillatorthat providesa frequency
standardon each satellite(L1 = 154 x 10.23 MHz, L2 = 1 "CHIP"=9.77x 10'8 sec
120 x 10.23 MHz). The clock frequencyis actually set -
--
54 L1 CYCLES
slightlylowerthanthisnominalfrequencyto accountfor i i i i
I I I
_ I I I I I
relativisticeffects [McCaskill and Buisson,1985] so that I I I I I I
MODULATED
an observeron the ground"sees" 10.23 MHz almost CARRIER
(C/A) ,
I I I I I I

exactly.
The L1 carrier has two components.The "inphase"
componentis modulatedby the precision(P) code. A
lower frequencycoarse/acquisition (C/A) code is modu-
lated in quadrature,i.e., on the samecarrier frequency
I SHh=Tr ß = P-CODE = C/ACODE
shifted90ø (Figure1). TheL2 carrieris normally
modu- TRANSITION TRANSITION
lated only with the P code. All threecarriersalso are
modulatedwith a low bit rate (50 Hz) data streamtrans-
Figure 1. Schematicillustration of biphase and quadrature
mittingsatellitehealth,ephemeris, andotherhousekeepingmodulationof the L1 carrier signal (1575.42 MHz) by the P
information.These codes,P(t), C(t), and D(t), can be (10.23 MHz) and C/A (1.023 MHz) pseudorandom noise(PRN)
consideredas squarewaves with valuesof +1 and are codes. Small circles on modulated carriers denote phase
termedpseudorandom noise (PRN) codesbecausethey inversion points.Note90ø phaseshiftbetween P andC/A
havesufficientlylongrepeattimesthattheyappearrandom modulations.In this illustrationthereis only one carriercycleper
to a user without knowledge of code structure.For P codechip.In actualitythereare 154L1 cyclesperP codechip.
example,theP coderepeattimeis 37 weeks.Eachsatellite
is assigned a unique1-weekportionof thecode,and,since Receiverdesignaffectsthe type of observablethat can
the number of active satellites will not exceed 24, each be extractedfrom the spectrumof GPS signals.There are
satellite can be uniquely identified by a PRN number currentlytwo basicreceiverarchitectures in commonuse
corresponding to thecodeportiontransmitted. for high-precision
geodesy, code correlating (as described
We canrepresent L 1 andL2 signalsas above) and codeless,where only the carrier phase observ-
able is available. Code-correlatingreceiverscan recover
Sl(t) =A•,t P(t)D(t) cos(tot t) ("reconstruct")the phaseobservableas a by-productof the
(4) correlationprocess.Recoveryof carrierphasewithoutcode
+ AcC(t)D(t) sin (tot t) knowledgerequiresa nonlineardetectionschemesuchas
signalsquaring.In effect,the signalis multipliedby itself,
S2(t) = A•,2P(t)D(t) cos(0•2t) (5) making the original phaseinversions(equivalentto an
252 ß Dixon: THE GLOBAL POSITIONING SYSTEM 29, 2/REVIEWS OF GEOPHYSICS

amplitudechangeof +1) unity,givingthesecondharmonic positionat the time of signal transmission, and carrier
of the carrier with no code modulations at double the phasecycle ambiguities. The remainderof this paperis
originalfrequency,or half the wavelength.The advantage devotedto relative(as opposedto poin0 positioning and
of the codelessapproachis that the high-precision part of associated conceptsthat enable high-accuracygeodetic
the signal (the carrier phase) can be utilized without measurements with GPS. Relative positioninginvolves
knowledgeof the classifiedP code, which may not be simultaneous observationof a group of satellitesby a
availablein the future.Disadvantages includereductionof network of groundreceivers.Three-dimensionalvectors,
signal-to-noise ratio, possiblyimportantunder marginal termed baselines, are def'med between all stations in the
observingconditionssuchas periodsof high ionospheric network, relative to one or more fixed stations whose
activity,andreducedeffectivewavelength,makingcarrier positionsare known a priori. The combinationof simul-
phasecycle ambiguityresolutionmore difficult. Also, P taneous networkobservations andtheanalyticaltechniques
code data, which are otherwise useful for clock designed to accommodate suchdataenablesusto eliminate
synchronization, data editing, and, dependingon data or greatlyreducethe errorslistedabove,resultingin the
quality,resolutionof carrierphasecycle ambiguities,are millimeter-to centimeter-level positiondatawe requirefor
obviouslyunavailable.Some civilian receiversemploy a mostgeologicalandgeophysical applications.
hybridapproach,with code-correlating capabilityon L1 to
recoverthe nonclassified C/A codefor clock synchroniza- Frequency Standards, Time,and ReferenceFrames
tion and navigationinformation,anda codeless channelto One key to high-precisiongeodesy with GPS is
recover the second harmonic of L2. simultaneous satelliteobservations by a numberof ground
Block 2 satelliteshave additionalsecurityfeaturesthat receivers.Simultaneityin this case is defined quite
will affect civilian users. "Selective availability" (SA) stringently;in 1 millisecond(ms) a stationat mid-latitudes
reducespoint positioningaccuracyto about 100 m by moves more than 30 cm to the east as a result of Earth
reducingthe accuracyof thebroadcast ephemeris, altering rotation,a GPS satellite,orbitingat about3 km/s,moves3
the clock epoch,and ditheringthe clock frequency,thus m, and a pseudorange signal propagates300 kin. Ul-
affecting both code-correlatingand codelessreceivers. timately,we will relateobservations at widely separated
"Antispoofing"(AS) will be activatedperiodicallyfor test groundstationsto better than a microsecond(gs), al-
purposesand consistsof encryptionof the P code; this though,as we shallsee,physicalclocksynchronization at
would not affect codelessreceivers.Most aspectsof SA anythingnear this level is unnecessary. We nevertheless
and AS will not causeseriousimpacton high-precision require a precise time definition and measurement
geodeticapplications.However, if activated,AS would capability,a methodfor eliminatingclock errors,and the
limit high-precisiondynamic applications,and one abilitytorelatewithgreatprecision thepositions of ground
important aspect of SA (clock dithering) is discussed receiversanywhereon the Earth to satellitesin orbit. The
below. following discussionis summarizedfrom King et al.
[1985],Lambeck[1980, 1988], andLeick [1990].
BothGPS satellitesandreceivershaveprecise"clocks,"
HIGH-PRECISION GEODESY: RELATIVE i.e., high-frequency,highly stableoscillators.A receiver
POSITIONING might employa quartzoscillator,a mechanicalresonator
that exploits the frequency-selective propertiesof the
Uncertaintiesin a GPS point positionmay be several piezoelectric effect, with a fractionalfrequencystability
meters to several tens of meters, althoughMalys and Af/fof about 1 partin 10•øperday.MostGPSsatellites
Jensen [1990] recently reported point position uncer- employ higher-qualityrubidium or cesium frequency
tainties of about 1 m using data from a speciallycon- standards, where atomicresonance phenomena basedon
figuredglobalexperiment.One sourceof uncertaintyin a the energy difference between two states def'me the
GPS point positionis the inherentimprecisionof the P "clock."Forexample, thecesiumclock[Essen andParry,
code group delay measurement,meter level for most 1956] is basedon the splittingof the groundelectronic
receivers,althoughat least one recentmodelachievesa stateof cesium133,depending on whetherthe spinof the
severalcentimeterprecisionwith just severalminutesof unpairedvalenceelectronis parallelor antiparallelto the
averaging[Melbourne, 1990]. For this reasonwe make nuclearspin.The transition betweenthesetwo hyperfme
phase measurementson the carrier itself [Bossleret al., levelshasa frequencyof 9,192,631,770Hz andis thebasis
1980;Remondi,1985], with an inherentprecisionof a few for the currentlyaccepted(SI) definitionof the second
millimeters or better. Major remaining error sources (moreon thisbelow).
include clock biases (in both the satelliteand ground Onemeasure of clockstabilityis theAllan(twosample)
receiver, althoughgroundreceiversare likely to have variance[Allan,1966].FollowingThompson et al. [1986],
larger biases),the atmosphere,includingthe frequency- if thefractional average frequency deviationjr(t) fromthe
dispersiveionosphereand the nondispersive troposphere, nominal frequency f0 overa timeintervalx is
bothof whichaffect signalvelocityand thusour estimate
of satellite-receiverdistance, uncertaintiesin the satellite [(t) = [{(t + x) - d•(t)]/2gfox (6)
29, 2/REVIEWS OF GEOPHYSICS Dixon: THE GLOBAL POSITIONING SYSTEMß 253

then
theAllan
variance
O'2A
is mentedby leap seconds.Thusthe differencebetweenGPS
time andUTC increases by integralseconds,roughlyone
O'2n
(1;)
= <[f'(t+1;)- f'(t)]2
>/2 (7) per yearsince1980whentheyweresetequal.
We will need to relate the orbital positionsof the GPS
where the angular brackets denote the infinite time satellites, computed in a celestial (--inertial) reference
average. TheAllanstandard deviations (N/-•a
2) ofseveralframe,to the locationsof the groundstations,definedin an
commonclocksare plottedin Figure2. Note the long-term Earth-fixed(terrestrial)referenceframe. This requires
stability of cesium clocks, making them attractivefor preciseknowledgeof the Earth's orientationin inertial
satelliteapplications,the short-termstability of quartz space. Earth orientation exhibits a rich spectrum of
oscillators,making them adequatefor groundreceivers, temporal variation. Aside from intrinsic geophysical
which can be periodically synchronizedwith satellite interest, we neeA to understand,measure, and correct for
signals,and the exceptionalstabilityof hydrogenmasers, these effects in order to relate inertial and terrestrial
makingthemthebestchoicefor groundreferencestations. referenceframes.In termsof causeswe can distinguish
externaltorquesactingon theequatorialbulgeof theEarth
(forcedmotion),associatedwith relative orbital motionof
the Earth and moon about the Sun, and free motion,
associated with the responseof the Earth to internalmass
redistribution and correspondingangular momentum
exchanges in the Earth system,includingthe hydrosphere
and atmosphere.Effects include changesin spin axis
directionwith respectto inertial space(precessionand
nutation),spin axis changeswith respectto the Earth's
crest(polarmotion),and changesin spinrate. Following
Soversand Border [1990], it is convenient to consider the
transformation from a terrestrial to inertial reference frame
as a series of rotation matricescorrespondingto these
majoreffects.
-16
0 1 2 3 4 5 6 7 Precession
(P) is a large-amplitude
(-23.5ø) slow
Log• (sec)
circularmotionof the pole with a periodof 26,000 years,
causedby lunar-solarattractionon the equatorialbulge.
Figure 2. Frequencystability of "best commercial"clocks, Accurate models are describedby Kaplan [1981] and
measured by thesquare rootof theAllanvariance(oa,equation Melbourneet al. [1983, 1985]. Nutations(N) are smaller
(7)), asa functionof timeintervalx. Notetheshort-term
stability (maximumamplitude9 s of arc), more rapid oscillations
of quartzcrystaloscillatorsandthelong-termstabilityof Cs and
superimposed on thisprecession,withperiodsfrom9 days
H Maser clocks. Rb, Cs, and H Maser data from Allan et al.
to 18.6 years.They are excitedby externaltorques,but
[1989].QuartzcrystaldatafromHellwig [1979].
their responseincludesa free component.Nutation series
are available from the 1980 International Astronomical
Now that we can measure time, let's define it more Union (IAU) theoryof nutafions[Kaplan, 1981; Seidel-
precisely. Sidereal "time," the kind many people are man, 1982], basedon the theoryof Wahr [1981]. Addi-
familiar with, is basedon the Earth's irregularrotation tional annual and shorter-periodterms are now known
about its axis and is no longer usedas a time standard. from very long baselineinterferometry(VLBI) observa-
Ephemeristime (ET) is more regular and exploitsthe tions [Herring et al., 1986]. Correctionsto the standard
periodicnatureof the orbitalmotionsof the Earth, moon, precession-nutation model can be incorporatedin a
and Sun. Although no longer in use, ET is related to perturbationmatrix, f•. Polar motionrepresentsboth free
currenttime standardsbecausethe atomic (SI) secondis and forced responses.In addition to contributionsfrom
defined equal to the ET second,in mm defined as a externaltorques,possibleinternalexcitationmechanisms
fractionof the tropicalyear 1900.Universaltime (UT1) is include large earthquakes [Slade and Yoder, 1989],
a form of siderealtime, correctedfor someirregularities aseismicdeformation,motionsof the atmosphereand
[Aokiet al., 1982]. Coordinateduniversaltime (UTC) is an hydrosphere,and core-mantlecoupling. The dominant
internationallyacceptedstandardbasedon atomic time, periodis 14 months(theChandlerwobble),with additional
roughly synchronizedto UT1 by adding"leap seconds" annualand shorter-period motions,causingthe pole to
whenrequired.Over the lasttwo decadesthisadditionhas wander in a circle less than a few tens of meters in
averagedaboutone secondper year,reflectingthe change diameter.Pole positionis describedby two coordinates,
x
of the Earth's rotationrate aboutits axis with respectto and y (correspondingtranslationmatrices X and Y),
gravitationalorbitalmotion(thebasisof theET, andhence representingrotationsabouttwo orthogonalaxeslying in
atomic,second).GPS satellitesbroadcast time signalsthat theequatorialplane(1 cmof polarmotioncorresponds to a
are synchronizedwithin 1 gs of UTC but are not incre- rotationof about 0.3 millarc seconds(mas). The time-
254 ß Dixon: THE GLOBAL POSITIONING SYSTEM 29, 2/REVIEWS OF GEOPHYSICS

varyingrotationrate of the Earth, U, and corresponding centerof masswith a precisionof about5 cm or better
changesin length of day (LOD), whose annual and [Smithet al., 1985;Tapleyet al., 1985].ColocatedVLBI
shorter-period variationscouldonly be measuredafter the and SLR sites define the center of mass correction for a
invention of atomic clocks in the mid-1950s, can be VLBI reference framewithcomparable precision.
representedby the differencebetweenatomictime (UTC) The completetransformation from a positionvex:torin a
and UT1. Time seriesof LOD or UT1-UTC showstrong terrestrial referenceframerr toaninertialreference frame
(several millisecondsamplitude) annual and semiannual r• isgivenby
periodsassociatedwith seasonalchangesin wind circula-
tion. Shorter-periodfluctuations(2 week and lunar rI = •PNUXY(co¾+ rcm+ A•o•+ Aoc+ Apd+ Aatm) (8)
monthly) are due to tidal effects, while longer-period
"decadefluctuations"may reflectcore-mantleinteractions whereot is an empiricalcoordinate framescalingfactor.
(seereviewsby Wahr [1988], Lainbeck[1988], and Hide Note thatfor localGPS networks,tidal and atmospheric
and Dickey [1991]). Values for both polar motion termscan be ignored,and the precisionrequirements for
(precision~1 mas)andUT1-UTC (precision~0.1 mswhen the otherparameters may be lessstringentthanfor large
averagedover a day) come from VLBI observationsof apertureor globalnetworks.
quasars[Carter et al., 1985]. Satellitelaserranging(SLR)
[Cohenand Smith, 1985] providescorresponding preci- Eliminating ClockErrors:Singleand Double
sionsof approximately 2-4 masand0.2 msaveragedover Differencing
severaldays[Smithet al., 1985;Tapleyet al., 1985].Lunar A GPSpointpositionis limitedto meter-level accuracy
laser ranging (LLR) providesroughly comparabledata by a combinationof uncertaintiesin clocks, orbits, and
[Dickey et al., 1985]. Sensitivityanalysesindicatethat atmospheric effects.However,for a relativepositioning
systematicerrors in GPS orbit and baselineestimates experiment involvinga networkof groundstationswithina
causedby theselevels of uncertaintyin Earth orientation specificregion,theseerrorsarecommonmodeor nearlyso
are negligible[Lichtenand Border, 1987; Dixon et al., and can be eliminatedor greatlyreducedin subsequent
1991c]. data processing[Bossler et al., 1980]. Consider two
In summary,spin axis changeswith respectto inertial receiverssimultaneouslyobservingtwo satellites.There
space(precession and nutafion)are largelydue to external are four geometricrangeequationsand a maximumof 16
torques(forcedmotion),canbe largein amplitude,andcan observation equations, sinceeachrangecouldbe estimated
be modeledquitewell. Polarmotionis a spinaxischange with botha pseudorange andcarderphaseequation,each
with respectto the Earth's crest, consistsof both forced at two frequencies. For a single-frequency pseudorange
and free motions,is much smallerin amplitudeand less measurement, R (equation(1)), denotingsatellites(id) by
easily modeled.UT1-UTC likewise cannotbe modeledat superscripts andreceivers(1,2) by subscripts, we havethe
presentwith sufficientaccuracy.Both polar motionand followingfor satellitei:
UT1-UTC are measuredadequatelyfor our purposesby
VLBI and SLR. R1 =p• +c(Aq -At / +A •) (9)
Tide-relatedelastic deformationsmay be important,
i i
particularlywhere high accuracyis desiredfor widely R2=P2+c(At2 - At/ +Att;2) (10)
separatedstations(the effectslargely cancelfor stations
closeto eachother).Therearethreemajoreffects:the solid Subtraction givesthe singledifferenceobservableR '.'
Earth tide Asd (maximumamplitude-50 cm); ocean
loading effectsin coastal
areas,Aoc(maximum
amplitude Ri' =pi'+c(Atl- At2+At}'
) (11)
~5cm);andthepoletideApd,
representing
elastic
response
of the Earthto changesin spinaxisorientation(maximum wheretheprimedvariables denotethedifferential
pseudo-
amplitude ~1 cm).Differential atmosphericloadingAatm rangeor true rangeand differentialpropagation delay,
[Rabbeland Schuh,1986] may causeeffectsat the several respectively. Satellite
clockerror(At•) is eliminated,and
millimeter level for large stationseparationsbut is not differential
propagation delayis considerablysmallerthan
believed to be significant for local or regional GPS theoriginaldelay.Similarly,for satellitej,
networks.This effect may become importantwith the
adventof high-precision globalGPSnetworks. Rj'=p/'+c(Atl-At2+
An Earth-fixed reference frame based on VLBI
observations
doesnot preciselydefinethe locationof the Subtractionof equations(11) and (12) gives the double
Earth's center of mass. However, GPS-determined difference observable R ":
positionsare intimately related to GPS satelliteorbits,
which are sensitive to the center of mass. To account for
R"= p"+ c (6t;,") (13)
offsets between the VLBI-based reference frame and the
actualcenterof mass,we definea vector,rc,•, to be whicheliminatesreceiverclockerror and leavesonly
incorporatedin the final transformation.
SLR definesthe differentialpropagationdelays as a significanterror
29, 2/REVIEWS OF GEOPHYSICS Dixon: THE GLOBAL POSITIONING SYSTEM ß 255

source,which can be calibrated,or solvedfor, independ- bending.There are two main regions to consider:the
ently(seebelow).Note thatfor stationsreasonably closeto frequency-dispersive ionosphere(--50-500 km altitude)
one another(<<satellitealtitude)the effect of orbit error is and the nondispersive neutralatmosphere, especiallythe
also largely common mode. Similar equationscan be troposphere (4}-10 km altitude).Thompsonet al. [1986]
formedfor thecarrierphaseor beatphaseobservable[King give a comprehensive review.
et al., 1985], which is actually more important for Free electronsexist in the ionospherebecausesolar
high-precisionmeasurements. ultravioletradiationis absorbedby gaseousmoleculesin
If there are more than two receiversin the ground the upperatmosphere, liberatingoutershellelectrons.The
network, each pseudorangeor phase measurementcan electrons interactstronglywith any electromagnetic signal
contribute to more than one differenced observable without in the frequencyrange of GPS. Ionosphericeffects are
adding new information;the observables are correlated, proportional to the integratedelectroncontentalongthe
complicating error calculation.While thereare methodsto signalpathandthusdependon solaractivity,the elevation
deal with thesecorrelations[Beutler et al., 1986; Goad and angleof theobservation, time of day,andlatitude.
Mueller, 1988], it may be desirableto operatedirectlyon The traveltime• for a groupdelaymeasurement canbe
undifferenceddata (sometimescalled one-way range or represented by
phase),considering clocktermsasnuisance parameters to be
eliminatedby otherapproaches [e.g.,Goad,1985]. •=p+A B
The simplepictureof clock error cancellationthrough
• • +f• +... (14)
simultaneous observations is complicatedby $A and re- where thetermA/f represents mostof theionospheric
ceiverclockepochand samplingoffsets.King et al. [1985] delayexperiencedat frequencyf, A is constant,and B is
providea completetreatment, but we canillustratetheprob- proportional to the averagemagneticfield strength.If we
lemby considering observation in thepresence of $A, where ignore the third- and higher-orderterms, the differential
the effectsare particularlyimportant. $A affectssatellite groupdelayA• betweenobservations at f2 (the L2
clocksin two ways.First, the satelliteclockcorrectionrela- frequency) andfl (theL1 frequency) is [Spilker,1978]:
tive to "true"GPS time,normallybroadcast in thedatames-
sage,is corrupted. Fortunately, thishaslittleimpacton non-
real time relative positioningapplications.Second,clock
"dithering"changesthebasicP codefrequency, resultingin
commensurate changesin thederivedL1 andL2 signals.Ir-
regularvariations of severalhertzaroundthenominalcarrier
centerfrequencyover periodsof a few minuteshavebeen (15)
observedwith the Block2 satellites.Sincegroundmeasure-
mentsare taggedby receivetime,but SA effectsare com- = '1;1[(fl If2 )2 _ 1]
mononly at the sametransmittime,doubledifferencing or
analogous schemes to eliminateclockerrorscanresultin in-
completeerrorcancellation. The maximumnonsimultaneity where• is thegroupdelayatL1. Similarequations canbe
due to $A for receivers at different locations on the Earth is constructedfor the carrier phase observable.Dual fre-
about 20 ms, which, at currentlevels of $A, inducesmilli- quencyobservations thereforeallow eliminationof major
meter-levelerrorsonly on the longestbaselines. The effects ionospheric effects(but not third-andhigher-order terms)
of nonsimultaneity dueto receiverclocksare usuallynegli- for bothpseudoronge andcarrierphasemeasurements.
gible with the double-differencing approach,sincethese The magnitudeof the ionospheric effectcan be large.
clockscanbe periodicallysynchronized to GPStime,for ex- Recallingthatn = c/v,
ample,with the C/A code. More seriousnonsimultaneity
(severaltenthsof seconds)canresultsimplyfrom different nion= 1- Nq•2/2e0 m•o)2 (16)
receivertypessamplingthe signalat arbitrary(anddifferen0
times.While thereare techniques for dealingwith thisnon- wherenio n is theionospheric indexof refraction appropri-
simultaneity[Feigl et al., 1990a],thesimplestapproachis to ateto thephasevelocityof a particularfrequency, N is the
ensurethatall receiversare in a networksampleat common number density ofelectrons (-10• - 10•2/m3), q•andm•
GPS time. Recent resultssuggestthat with this simple aretheelectron charge andmass, respectively,ande0isthe
precautionthe impactof $A on high-precision geodesyis permittivity. Equivalently, the index of refraction at a
negligible[RockenandMeertens,1990;Feigl et al., 1990a; given frequencycan be estimatedfrom the plasma
Rosenblattet al., 1990]. Additional discussioncan be found frequency
f•,oftheionosphere:
in theworksby Melbourne[1990]andWu et al. [1990].
nion
--1- fv2/2f
2 (17)
The Ionosphere where
The GPS signalis affectedby propagationthroughthe
atmosphere,through changesin velocity, and by ray .fv
2=Nq•2
[•m• (18)
256 ß Dixon:THE GLOBALPOSITIONINGSYSTEM 29, 2/REVIEWS OF GEOPHYSICS

The plasma frequency is the characteristicvibration It shouldbe clearfrom the precedingdiscussionthat

frequencyfor interactionbetweena plasma (the iono- dual frequencyobservations are critical to obtaining
sphere)and an electromagnetic
wave;typicalionospheric high-precisiongeodeticdatawith GPS.Single-frequency
valuesare 10-20 MHz. Signalattenuation
is smallforf>> receivers are an economical alternative for certain
f•,,explaining
thefrequency
selection
forGPScarriers. applications,
but if highprecision
is required,theycan
We are more interestedin the integratedeffect of the onlybe usedon theveryshortbaselines mentioned
above,
index of refractionalong the ray path and requirethe wheredifferentialionospheric
effectsaresmall.
integrated
electron
density
N/(electrons/m2): The impact of ignoring higher-orderterms in the
ionospheric
delaydepends
on theelectrondensityin the
N•=l•th
Ndz (19)ionosphereand the dot productof the ray pathandthe
Earth'smagnetic field.Severalmillimeters
of rangeerror
A usefulrule of thumb is that the delay at zenith in
at low elevationcanresult.Thiscanbe ignoredfor most
equivalent
lengthunits,
L, isgiven
byL -- 10-17N/.For applicationsbutmaybecomemoreimportantin thefuture
zenith observations
the magnitudeof the time delay for improved vertical component estimates and
associated
with theP codeobservablemaybe 10 nsor less troposphericstudies,both of which benefit from low-
atnight
atmid-latitudes
(N/~1017
electrons/m
2)and100 elevation observations.Dual frequencyobservations
ns or morefor daytimeobservations
nearthe geomagnetic enableestimationof theintegrated
electron
densityandin
equator
andpolarregions
(N/ ~ 1028electrons/m•),
principlecouldbe usedto derivemodelsof ionospheric
equivalentto path length variationsof severalmetersto gradients.
severaltensof meters.The delaycanalsovaryas a result
of magneticstormsandis sensitive to the 11-yearsunspot TheNeutralAtmosphere
cycle. In thissectionwe consider
nondispersiveatmospheric
Sincetheionosphere is frequency-dispersive,
thegroup effectson the GPS signal.These effectsare due to the
velocity
vg (associated
with theP codepseudorange troposphere,
tropopause, and mesosphere,
but roughly
observable)
differsfromthephasevelocityv• (associatedthreequarters
of thetotalnondispersive
atmospheric
delay
with thecarrierphaseobservable);theeffectsareequalin and most of the variability is associatedwith the
magnitude butoppositein signandarereferredto asgroup troposphere.
delayand phaseadvance.This is suggested by equations A GPSsignalis bentandslowedin itspassage
through
(16) and(17), wherenion is lessthan1 for a givenfre- theloweratmosphere.
Thedelayxat
mis thedifference
in
quency, implyingthatthephaseof a wavepassing through travel time betweenactualsignalpropagation and the
theionosphere is advancedrelativeto a wavetravelingin theoreticaltransittimein vacuum. It is usuallyexpressed
vacuum.Fundamental principlesare not violated,because asequivalent pathlengthby multiplying by thespeedof
it is vg(thevelocity
of thecodemodulation) thatdeter- lightand can be definedby the difference of two path
minesthe speedat whichinformationis carried. integrals[Daviset al., 1985]:
For many applicationsit is deskableto considerlinear
combinationsof data, to isolate or reduce errors, or to
reducecomputation timeby compressing observations to a X,tm
=l,tm
n(s)ds
- l•,½
ds (21)
singledatatype.A goodexampleis the ionosphere-free
phase
observable
Lcgivenby [e.g.,Blewitt,1989]: wheren(s)is theindexof refraction
at thepoints alongthe
pathandn,a½ = 1 is omittedfromthe second integral.
Evaluation
of the first integralrequiresan atmospheric

If•2L1
-f22L2)
(20) model.All components of theatmosphere
delay,but it is convenient
separately.
to consider
contributeto the
two components
The "dry" delay is associated with molecular
whereL1 andL2 hererepresent carrierphaseranges at the constituents of theatmosphere in hydrostatic
equilibrium
two frequencies for a particular
receiver-satellite
pair (i.e., (including H20), andthe"wet"delayis associated with
L1 = -cO)l/f1 andsimilarly forL2),Lc is theionosphere-watervapornotin hydrostatic equilibrium.Thedrydelay
corrected phase, andtheeffective wavelength is•c = c/(f• is typically200-230 cm at zenith(elevation angle,0 =
+f2) -- 10.7cm(5.4cmforcompletely codelessreceivers). 90ø)ataltitudes nearsealevel,whilethezenith wetdelay
The shortereffectivewavelength of •c impliesthat mightrangefrom3 to 30 cm.Thedelayat otherelevation
resolution of carrierphasecycleambiguities will be more anglesis larger,increasing approximately as 1/sin(0),but
complicated.For this reason, and also becausenoise othereffectsareincorporated in a "mapping function" for
(particularly multipath) is amplified by L•, it maybe greateraccuracy,includingthe finite height of the
deskableto usesingle-frequency measurements for very atmosphere, theverticaldistribution
of components, Earth
short(a few kilometers or less,depending on ionosphericcurvature,andmy bending[e.g.,Blackand Eisner,1984;
activity)baselineswhereionospheric effectsare smallor Lanyi,1984;Daviset al., 1985].Thetotalpathdelaycan
process thetwofrequencies separately. be written as
29, 2/REVIEWS OF GEOPHYSICS Dixon: THE GLOBAL POSITIONING SYSTEMß 257

t,(0)= + (0) (22) have a small bias that dependson site, season,or data
reductionalgorithm.
where p0refers to thepathdelayat zenith (hereafterthe Unfortunately,most WVRs are big and expensive.
zenith delay), subscriptsd and w refer to dry and wet Fortunately, comparison of GPS baseline estimates
components, respectively,andM(0) is a mappingfunction, involving WVR calibration and stochasticestimation
assumedto be azimuthallysymmetric. generally indicates no significant differences between
The dry zenithdelay is determinedby measurement of the two approaches(for example, Figures 3 and 4),
surfacepressure.If suchmeasurements are not available, implying that at current precision levels WVR calibra-
standardatmosphericconditionscan be assumed,and an tion of the wet delay is not required. In fact, it is even
initial estimatefor the dry zenithdelayapproximated from possibleto lump the wet and dry delays togetherand
the standardpressureP (in bars), basedon the surface estimatethem jointly, avoidingall neutral atmosphere
elevationof the station,h, anda scaleheightH, typically7 calibration(includingsurfacepressure)entirely [Tralli
km [Tralli et al., 1988]: and Lichten, 1990]. This works at current levels of GPS
precisionbecausethe mappingfunctionsfor wet anddry
P = 1.013e-4'm (23) delaysare very similarabove10ø or 15ø, the typical
cutoff angle usedwith GPS to avoid groundmultipath
Subsequentanalysisof the GPS data can improve the (see below).
initial estimate and reduce the effect of errors in this In spite of the success of stochastic estimation
parameteron thebaselineestimates. techniques,for GPS experimentsin regionsof high wet
The wet zenithdelay can be estimatedin at leastthree path delay and high variability (for example, some
ways:by measurement of surfacetemperature andrelative conditionsin the tropics), troposphericcalibration is
humiditycoupledwith a simpleatmospheric model [e.g., probablythe dominanterror sourcefor baselinesin the
Chao, 1974]; with a water vapor radiometer(WVR), an geologicallyimportantlength range of several tens to
instrumentthat measuresatmospheric blackbodyradiation several hundred kilometers. Orbital effects dominate at
in the microwaveregion,which is affectedby a rotational longer baseline lengths, and receiver effects and other
moleculartransitionof water vapor near 22.2 GHz [e.g., errors dominate at shorter baseline lengths. This is
Janssen, 1985; Robinson, 1988]; and by stochastic discussed in more detail in the sectionon precisionand
estimation techniques without a priori calibration, accuracy.Troposphericcalibrationdoesnot appearto be
exploiting the data strength of GPS and the known a significanterror sourcein some other regions(for
elevationangledependence of the wet delay [Tralli et al., example, typical conditionsin the southwesternU.S.)
1988; Dixon and Kornreich Wolf, 1990; Dixon et al., when the wet path delay is low and not particularly
1991a]. Stochasticmodels make use of the fact that the variable.Future improvementsin the accuracyof GPS,
wet path delay is likely, in a statisticalsense,to vary particularlyfor the vertical component,will dependin
within a limitedrangeover a shorttime interval.For SM, part on improvedtroposphericcalibrationand estimation
WVR, or other a priori calibrations,residuals(wet delay techniques. By analogywith VLBI, direct line of sight
corrections) can be estimated along with geodetic calibrationwith a WVR or othertechniquemay improve
parametersof interest.This improvesthe final resultbut GPS baseline estimateswhen calibration accuracy
implies errorsin the calibration.Studieshave shownthat exceeds~5-10 mm [Herring, 1986;Elgeredet al., 1991;
SM calibrationby itself can lead to unreliableGPS results Dixon and KornreichWolf, 1990]. It hasbeendifficult to
[Tralli et al., 1988;Dixon et al., 1991a],confirmingearlier deployhigh-accuracy WVRs to a majorityof sitesin a
studiesindicatingpoorcorrelationbetweensurfacerelative GPS networkbecauseof the excessive cost,weight,and
humidityandthe wet pathdelay[Reberand Swope,1972; power consumptionof these sophisticatedinstruments.
Elgered,1982].K. Hurstet al. (Estimationof GPSbaseline Recentdevelopments in the areaof monolithicmicrowave
errorsdueto imperfectretrievalof wet atmospheric delays integratedcircuit(MMIC) technology havethepotentialto
using surface meteorologymeasurements,submittedto improvethe accuracyand reducethe cost,weight,and
Journalof Geophysical Research,1991)compareSM and powerconsumption of WVRs, makingextensivedeploy-
WVR estimatesof wet zenith delay and concludethat ment of these instrumentsviable. Improved stochastic
SM-basedestimates wouldhavean rmserrorproportional modelsare anotherpromisingarea of research.Current
to the magnitudeof the delay, given by 1.2 cm + 0.08 modelsassumeazimuthalsymmetryin the wet delay.
(SMT), where "SMT" is the SM-basedzenith delay in However,azimuthalasymmetryhasbeenobserved[Dixon
centimeters.Tralli et al. [1988] and Tralli and Lichten and KornreichWolf, 1990;Rockenet al., 1991]. Estimat-
[1990] suggested that the errorin SM calibrationcan also ing two stochastic,orthogonalspatialgradientsin addition
show considerabletime variation. In contrast,corrections to thestochasticwetzenithdelayshouldimproveprecision
to WVR calibrationare generallysmall(1-2 cm in zenith in the presenceof asymmetryand shouldbe feasiblewith
wet delay) and constantor nearlyso over severalhours, sufficientdata strength,presumablyavailable with the
implyingthattheseinstruments cangive a goodindication enhanced Block 2 constellation and receiverscapableof
of temporalvariabilityin the wet zenith delay but may trackingmorethanfour satellites.
258 ß Dixon: THE GLOBAL POSITIONING SYSTEM 29, 2/REVIEWS OF GEOPHYSICS

_ a4o/o LIMON - LIBERIA(270 km) _

,,.:.:.:.:.:.:.:.:.:.:.:.:.:.:,

:::::::::::::::::::::::::::::::::::::::::
• STOCHASTIC ESTIMATION --
::::::::::::::::::::::::::::: •

- iiiiii!iii!i!ili!iiiiiiiiiiil
85%86% 87% 87%
86%

86% 87%
• 86%

..
•
iii.dliiill
• ilili
86% 88%
86%
87%
88ø/* --
::::::::.,..__
::i::i:::: STOCHASTIC
RESIDUAL

0.0 0.48 0.80 1.0 1.2 1.8 2.4 0.48 0.80 1.0 1.2 1.8 2.4

O• (cm/x/h-• O• (cm/•--•
WVR CALIBRATION STOCHASTIC ESTIMATION,
NO PRIOR CALIBRATION

Figure 3. Short-termrepeatabilityfor a 270-km baselinein CostaRica with varioustreatmentsof

wet path delay (see also Figure 4). Percentsymbolsaboveeachbar indicatenumberof biases
resolved.With WVR calibration(left side), estimationof a residualdelay, either as constantor
witha stochastic
modelspecified
by therandom
walkparameter,
a (cm/•), improves
the
calibration,but only to the level reachedin the casewhereno calibrationis used(right side) and
theentiredelayis estimatedstochastically [fromDixonandKornreichWolf, 1990].

Multipath range observables with systematic, time-dependent

Antennas for almost all GPS receivers are omnidirec- sinusoidal signals associatedwith variable receiver-
tional, enabling signals from several satellitesto be satellitegeometryovera pass.The magnitudeof multipath
receivedsimultaneously. Dependingon the immediate is roughlyproportionalto wavelength,and the effectsare
environment and the gainpatternat low-elevationangles, considerablylarger for P code pseudorangerelative to
suchantennas can be susceptible
to interference ("multi- carrier phase.Low-angle observationstend to be most
path")frommultiplearrivalsof thesamesignalbecause of affected,
andfor thisreason
a cut-offangleof 10ø-20
ø
reflections
fromnearbyobjects.Reflectionsat thesatellite above the horizon is usually employed.This has the
alsooccur[Younget al., 1985]butcanbe ignoredfor most fortunateeffect of minimizingerrorsdue to third- and
applications.
Multipathcorruptsthephaseand/orpseudo- higher-ordertermsin the ionospheric
delay and minimiz-
LIMON - LIBERIA (270 km)

I I I I i i

--•'--- WVRCALIBRATION • [ i ..,.T

' ' Figure 4. Baseline estimates
-O- STOCHASTIC ESTIMATION with and without WVR cali-
brationof the wet path delay
for thedatashownin Figure3.
Estimates with WVR calibra-
tion include a stochastic resi-
dualmodel(a = 1.2cm/x/-•).
Estimates without WVR cali-
w
i
brationemploya similar sto-
>
chastic model for the entire
wet delay. Error bars and el-
lipses are one sigma [from
Dixon and Kornreich Wolf,
-4
19901.
-4
-4 -2 0 2 -4 -2 0 2
EAST (cm) LENGTH (cm)
29, 2/REVIEWS OF GEOPHYSICS Dixon: THE GLOBAL POSITIONING SYSTEMß 259

ing tropospheric calibrationerrors,includingthosedue to natureof the perturbingforces on the satellitesand the

incorrectmappingfunctions.However,lack of low-angle requirement for high-accuracy meansthatthismodelis not
observationsis one of the conlributingfactors to poor adequatefor our purposes.Rizos and Stolz [1985] sum-
resolution of the vertical component with GPS, so marize major accelerationson the GPS satellites.Atmos-
multipathcontrol,as well as improvements in the other phericdragis negligibleat thesealtitudes,and the Earth's
areasjust mentioned, is highlydesirable. Multipathcanbe nonsymmetricgravity field, while significant, is ade-
minimizedthroughuseof antennabackplanesor RF (radio quatelydescribedwith currentmodels;the more poorly
frequency)absorbentmaterial around the base of the determinedshortwavelengthcomponents, relatedmainly
antenna, mounting antennasclose to the ground (to to lithosphericstructuresin poorly surveyedpartsof the
minimizetheeffectsof groundreflection),andcarefulsite Earth,havelittle effect on the high-altitudeGPS satellites.
selection,choosingsiteswell away from planar-reflecting Additionallargeperturbations are the gravitationaleffects
surfacessuchasbuildingsor vehicles. of the moonandSun,whichcanbe accuratelydetermined,
Since multipathis related mainly to the geometryof andsolarradiationpressure.
nearbyobjects,andsincethisgeometryis usuallyconstant One approachto the orbit problemis to begin with an
over several days or longer, the temporalsignatureof estimateof the six orbitalcomponents at someinitial time
multipathtendsto repeatfrom day to day, retardedby an and numericallyintegratethe equationsof motionusing
amountequalto theoffsetrisetimeof the satellites(4 min accurate modelsfor variouspertubingforcesandresulting
earlier each day). The repetitivenature of multipathin accelerations,predictinglocation and velocity of the
principlecan be usedto estimateand correctits major satellitesat later epochs;groundtrackingdatacanbe used
effects,thoughfew such studieshave been reported.If in a least squaresadjustmentto improve the initial
uncorrected,multipath can be an important sourceof position/velocity estimates,the modelparameters,and the
systematicerror, dependingon environment,antenna/ subsequent position/velocityestimates.The lengthof the
backplanedesign,and lengthof observingsession.Carrier orbitarc over whichthe equations are integratedmay be
phasemultipathtendsto haveperiodsshorterthan 10-20 only a few hours ("short arc") in which case the force
min, thus observationsover severalhoursor longer will modelscanbe relativelysimple,or may extendto several
averageoutmostof theeffects. weeks("multiday"or "long arc"), requitingsophisticated
forcemodels.Solarradiationpressure produces a largeand
Orbits somewhatunpredictableperturbingforce becauseof the
GPS satellites orbit about three Earth radii above the large cross-sectionarea of the solar panels, complex
surfacein six orbitalplanes.For high-precision geodesyit satellitegeometry,and variablesatellitealbedo,resulting
is necessaryto know precisely the positions of the inaccelerations oforder 10-7m/s 2andperturbing theorbits
satellitesat the time of observation, with acceptableerrora manymetersin just a few hours.Althougha solarradiation
functionof desiredbaselineprecisionand stationsepara- pressuremodel for the GPS satellitesaccountsfor albedo
tion; for millimeter-levelperformanceon baselineslonger andgeometry,largeresidualshavebeenobserved[Fleigel
than about 100 km, meter-levelprecisionin the orbit et al., 1985]. It is therefore common to solve for at least
estimatesis required. one additionalaccelerationparameterin the estimation
Satellite orbits can be describedby six parametersat process,representing departuresfrom the ideal model.In a
someinitial epochanda forcemodelto definesubsequent spacecraft-centered coordinatesystemwherez pointsto the
time evolution. The epoch state parametersare three centerof the Earthandy is alongthe solarpanelsupport
componentsof position (x,y,z) and three velocities in beam and normal to the spacecraft-Sundirection, un-
Cartesian coordinates,or equivalently six Keplerian modeledaccelerations areoftenobserved in they direction
elements:the semimajoraxis (a) and eccentricity(e), ("y-bias"), perhaps related to thermal radiation or
describingthe size and shapeof the elliptical orbit; a misalignedsolarpanels[Fleigelet al., 1985; Schutzet al.,
parameterdescribingthe positionof the satellitewithin 1990]. Residualstend to be largestduring the 2-month
that orbit (for example,M, the "meananomaly"or f, the eclipseseason,when the spacecraftperiodicallyenter the
"true anomaly"); and the inclination (i), argument of Earth's shadow [Schutzet al., 1990]. Lichten and Border
perigee(co),and fight ascensionof ascendingnode (f•), [1987] adjustedsolar pressurecoefficientsin all three
describingthe positionand orientationof the orbitalplane. components for data arcs up to 1 week, usingconstant
In the absenceof additionaldisturbingforces,only the corrections to the nominalmodel.Lichtenand Bertiger
anomalyis timedependent. A simpleKepleriandescription [1989] usedthreecomponentstochasticcorrectionsto the
of an orbit in a plane can be obtainedby integrating nominalmodelfor dataarcslongerthan 1 week.
Newton'slaws of motion,giving Informationon satellitetrajectoriesduringtheperiodof
theexperiment maycomefromthebroadcast ephemeris in
r = a(1 - e2)/[1+ e cos(f)] the GPS signal, or additional GPS data taken simul-
taneouslyfrom siteswhosepositionsare well knownfrom
where r is radial distance from the center of the Earth independentmeasurementssuch as VLBI or SLR. The
[Kaula, 1966]. For near-circularorbitslike GPS, e--0. The resulting tracking data allow generation of accurate
260 ß Dixon- THE GLOBAL POSITIONING SYSTEM 29, 2/REVIEWS OF GEOPHYSICS

ephemerides (satellitelocationsas a functionof time) and vantageous to employa multidayarc analysis,insteadof

define the reference frame for the experiment. The the singleday, shortarc approach.The advantages of
subnetworkof known groundstationsis calleda fiducial multidayarc analysisare described by LichtenandBorder
network. Uncertainties in fiducial site location, or in the [1987]. With observations
over more than one revolution,
groundsurveysconnectingthe phasecentersof the large orbital periodsare more accuratelydeterminedand the
VLBI antennasto the ground mark used by the GPS
antenna,or low-qualityGPS data at one or moreof these
12
sites during an experiment,are a major sourceof sys-
tematicerror in GPS geodesy.
VERYICALI O'
10 o
Thatpartof theerrorin a baseline
estimate
•t,t dueto
orbiterror•oa,canbeestimated
fromtheruleof thumb
-- •oa,L/h, whereL is baselinelength,h is satellite I o

altitude,-20,000km, and • is a geometricfactor(4).2)

[Lichten, 1990a]. Since the error in the broadcast Siope•
ephemeris canbe 50-100 m with selectiveavailability,we 0 Ilb•)Slope
=1x10 -9'
•o o
w ,,.

requirea robustfiducialnetworkfor any experimentwith o_

w
baselineslonger than about 50-100 km if we desire II o
centimeterlevel or better accuracy.The nature of the
requirednetworkdependson the experiment.In general,at • I I . I t
leastthreestationsare required,with adequatenorth-south
and east-west extent. Stations too close to each other ß I ß

provide redundantinformationand may not contribute NO'RTH'

I
U.S. only
greatly to network strength.On the other hand, mutual Global network
satellite visibility is deskablefor cancellationof clock
errors,one or two redundantstationsprotectagainstdata
outages,and some pmximal stationscan facilitate cycle
ambiguityresolution.The optimumnumberand geometry 4
of stations may be difficult to predict. A covariance
analysis,which estimatesuncertainties in baselinevectors
from certainassumptions aboutthe dataand its errors,can 2
be useful in testingtrial fiducial networks[Dixon et al.,
1985;Freymuellerand Golombek,1988].
In some cases,covarianceanalysesmay suggestthe o
need for fiducial sites in areas where VLBI or SLR sites
are not available,or where the necessarygroundtie data
EAST I I .... 0 , 0ß
are not available.It may still be advantageousto deploya
receiverto sucha locationto provideadditionalgeometric
strengthto the trackingdata.KornreichWolfet al. [1990]
used data from the "CASA" experimentto show that
. a$\•/-
addingtwo trackingsitesin the southwest Pacificand two
in Europeto a fiducialnetworkconsisting of threestations .• -

in the U.S. improvedthe repeatabilityof the horizontal

componentsof long (>400 kin) baseline estimatesin o o o 1
CentralAmerica and northernSouthAmerica(Figure 5), / '
even thoughthe actuallocationof the new stationswas
uncertainat the 10-20 cm level. The positionof such
tracking sites is estimatedin the data analysisto avoid 200 400 600 800 1000
introducingsystematicerror.
In spiteof our besteffortsto establisha robustfiducial BASELINE LENGTH (km)
networkwe may find that logisticallimitationsor data
outageslimit the qualityof the resultingobservations. An Figure5. Improvement in short-term (3-8 days)repeatability
(equation(30)) for seventeenbaselinesin Central and northern
importantclue to the causeof low-quality data can be
SouthAmericaasa function of fiducialnetwork
configuration.
found by plotting repeatability(day-to-dayscatter)of
Single-day orbitalarcsare used."U.S. only"indicates three
baseline
estimates
asafunction
ofbaseline
length;
ahighVLBIsimsinthe
U.S.
"Global"
indicates
this
network
supple-
dependence
onlength
(>2-3parts
inl0sforbaselines
mented
bytwoEuropean
and
two
southwest
Pacific
stations.
Best
longer
than100-200
kin)suggests
thepossible
influence
fitstraight
lines
areshown
(modified
fromKornreich
Wolf
etal.,
of orbit errors (Figures5 and 6). It may then be at- [1990].
29, 2/REVIEWS OF GEOPHYSICS Dixon: THE GLOBAL POSITIONING SYSTEM ß 261

positionsof the orbital nodesbetterdefined,with cor- six-stationGPS global trackingnetwork, expectedto be

respondingimprovementin the ability to resolvecarder operationalin 1991,for trackingthe U.S.-FrenchTOPEX-
phasecycle ambiguitiesthroughoutthe networkand the Poseidonsatellite,a low Earth orbiterfor oceanographic
precisionandaccuracyof all components of the baseline research to be launchedin 1992.Neilan et al. [1990]givea
vector.Figure6 showstherepeatability
of a setof baseline currentsummaryof internationaltrackingnetworks.
vectors from the northern Caribbean before and after a
multiday arc analysis [Dixon et al., 1991a]. Geodetic Reducing
the Data
estimateswere improvedby factorsof 2 or morewith the Considera hypotheticalGPS experimentinvolving5
longer arcs for baselineslonger than about 500 kin. daysof observations at 10 groundstations,with the posi-
Single-dayarcsandthe existingthree-fourstationfiducial tionsof threefiducialsiteswell known.Eachstationmight
network were apparently adequate to minimize the observesix satellitesfor an averageof 5 hourseachday
contribution of orbit error for shorter baselines. (the total time spanof observations might be 8 hoursor
more, but not all satellitesare visible simultaneously).
BIASES RESOLVED EAST Pseudorange and/orphasedataarecollectedcontinuously,
E3 SINGLE DAY ARCS
5 but let's considerphasedataaveragedat 1-min intervals.
[]MULTI DAY ARCS
4
The totalnumberof "datapoints"collectedin 1 day is then
10 stationsx 6 satellitesx 5 hoursx 60 pointsper hour=
3
18,000. From this large data set we might estimatethree
positioncomponentsfor each unknown ground station
(21), six components (threeposition,threevelocityat the
initial epoch)for eachsatellite(36), a y bias term for each
$ • o satellite(6), perhapsone troposphere parameterfor each
NORTH groundstation(10) (additionalterms are requiredfor a
stochasticmodel), and clock terms every minute for all
satellites(1800) and ninegroundstations(2700), sinceone
stationactsas a referenceclock,giving a total of 4573 un-
knowns.The systemis overdetermined, with data points
greatlyexceedingunknowns.Sincethe dataare noisy,we
cannotformulatesimultaneous equationsto solvedirectly
for the desiredparameters.The challengeis to find, in a
least squaressense,the best estimateof the parameters
VERTICAL from availabledata,usingwherepossibleadditionalinfor-
mation suchas troposphericcalibrationand orbit models.
Therearenumerous approaches to GPSdataanalysis,andI
havetriedto generalizethefollowingdiscussion. However,
6
somedetailsare specificto analyticalapproaches familiar
to me, andI apologizefor thisobviousbias.
The timeevolutionof pseudorange or phaseobservables
(e.g., equation(3)) is largely a function of observation
0 - ,
200
,
300
,
400 51•0
•
9;0 1100 1300 geometry.Given prior knowledgeof the approximate
BASELINE LENGTH (km) groundstationpositionsin a terrestrialreferenceframe,
satellite orbits (from the broadcastephemeris),Earth
Figure 6. Histograms of short-termrepeatability(equation(30)) orientation, andinitialestimates of atmospheric delay,it is
for 15 baselinesin the northernCaribbeanfor singleday and possible to generate fairly accurate models for the
multiday orbital arc analysis.Carrier phasecycle ambiguities observables. With thesemodels,andknowledgeof how the
("biases")areresolvedfor eachtechniqueto the extentpossible. observableob would changegiven a small changein a
Improvementin the longer(>500 km) baselineswith multiday parameter,x, (given by the partial derivative,•5(ob)/fx),
arcsresultsin part from the increasedabilityto resolvethebiases
leastsquares or Kalmanfilter techniques canbe appliedto
[from Dixon et al., 1991a].
theobservables to improveestimates of x. FollowingKing
et al. [1985], givenan observablemodelc, an observation
The burdenof establishing adequatefiducialnetworks l, andan observation errorv (all functionsof time),
for regionalGPS experiments will be considerably reduced
in the nextfew yearsas variousnationalandinternational c= l + v (25)
agenciesestablishglobaltrackingnetworks.The Coopera-
tive International
GPSNetworkhasbeenin operationwith Thiscanbe linearizedby expandingtheleft-handsidein a
six or more stations since 1989, administered in the U.S. Taylorseriesanddroppingsecond-andhigher-order
terms,
by the National GeodeticSurvey.NASA will sponsora expressingc as a vectorof initial estimatesof the observ-
262 ß Dixon: THE GLOBAL POSITIONING SYSTEM 29, 2/REVIEWS OF GEOPHYSICS

ables,
c.,based ofx,x•,plusincrementalconstant parameters (y), a vector of process noise
onapriorivalues
corrections,[?ffob)/15x]dx, by thedataand parameters(p) and v, definedhere as a zero meanwhite
to be constrained
its errors.The linearizedobservationequationin matrix noise vector:
form is
Z= AxX +At, P+AyY+v (28)
c. + A dx= l + v (26)
where z is the difference between an observation and its
where A (the design matrix) is the matrix of partial modelvalue.Parametersmodeledas processnoiseinclude
derivatives, dxis a vectorof smallcorrections tox., andv clocks(modeledas whitenoise)andwet tropospheric path
is now the vector of postfit residuals.The quantity delays (modeledas random walks). Measurementsare
minimizedis theweightedsumof squared residuals,given groupedinto finite time intervalsor batches(typicallyone
byvrWv,where W, theweight matfix,istheinverse of P, to severalminutes)and processedsequentially,leadingto
the covariance matfix. The solution is updatedparameterestimatesand covariancesafter each
batch.After filteringall dataa smoothingalgorithmcanbe
• = x. + dx (27) applied that works backward in time to update the
estimatesand covariances,for example, allowing inves-
where • isthebestestimate ofx, dx= N-1ArW(/- c.), tigationof the time-varyingbehaviorof the troposphere
andN isthenormal equation matrix given byArWA.This [e.g.,Dixon andKornreichWolf, 1990].
approachmay be appliedto bothdifferenced and undif- We sawfrom equations(2) and(3) that the initial phase
ferenceddata.It is alsopossibleto transformundifferenced measurementupon acquisitionof the carder signal is
data to derived parameterscontainingno clock terms biasedby an unknown number of cycles. Assuminga
through orthogonalizationmethods without forming receivermaintainslock on the signal,the range change
normalequations[Lawsonand Hanson,1974]. Additional between the receiver and satellite can be determined, and
discussion of GPS dataanalysisapproaches canbe found the initial range (the cycle ambiguity) can be estimated
in worksby Goad [1985],Bocket al., [1985a,b], Lindler along with the geodeticparametersof interest.However,
and Wells [1985], Bock et al. [1986], Beutler et al. [1986, this degradesthe accuracyof horizontalbaselinecompo-
1987],SchaffrinandGraffarend[1986],andLeick[1990]. nentsrelative to the case where the cycle ambiguity is
Most of the datadiscussed in thisreportwereprocessed knownor can be fixed to the correctvalue.Techniquesto
with the GPS Inferred Positioning System (GIPSY) resolvethe ambiguityrely on the fact that, given enough
software developedat the Jet PropulsionLaboratory. data,the rangebias can be estimatedto betterthan half a
GIPSY processesundifferenceddata using a modified carderwavelength(cycle),after which the bias is fixed to
Kalman filter, describedby Lichten [1990b]. Briefly, a the nearest integer value. Orbits, ionosphericeffects,
conventionalKalman filter updatesmeasurements from tropospheric effects(Figure 7), multipath,and othererror
one observationepochto the next on the basisof the sourcescan corruptGPS signalssuchthat errorsbecome
covariance matrix P. Factorization of P into upper significantrelative to one half wavelength,affectingour
triangular (U) anddiagonal (D) matrices (P = UDUr) abilityto fix ambiguitiesto thecorrectvalue.It is common
improvesaccuracyandcomputational efficiency[Thornton practiceto resolvethe ambiguitiesfirst on shorter(<100
and Bierman, 1980] and is the basis for the parameter kin) baselineswhere the effect of most of theseerrors is
estimation.The formulationof the linearizedleastsquares reduced[Bocket al., 1985b; Abbot and Counselman,1987;
observation equationdiffersslightlyfrom(26), partitioning Dong andBock,1989;Blewitt, 1989].The methodusedfor
parametersinto a satellitestate vector(x), a vectorof mostof the data presentedhere also exploitsthe fact that

lOO

L!.I
X
,'7' 95
Figure 7. Percentage of carder phase cycle
ambiguities(biases)resolvedas a functionof the
stochastic(randomwalk) modelusedto characterize
C9 90
the wet zenithpathdelayfor a GPS networkin the
northernCaribbean.Optimum model has random
C.) 85
walkparameter
ct= 1.0cmP•-•(compare
toFigure
3) [from Dixon et al., 1991a].

80
0

RANDOMWALKPARAMETER
(a..cm/,J'•'r)
29, 2/REVIEWSOF GEOPHYSICS Dixon'THE GLOBALPOSITIONINGSYSTEMß 263

theionospheric groupdelayof theP codemodulation is et al. [1991a, b], Blewitt [1990], and Beyis [1991]. From
the samemagnitude(thoughoppositesign)as the phase ourdiscussion of theprecision of carderphasemeasure-
delay,providedthecorrectnumberof cyclesis associated ments and ionosphericeffects it should be clear that
with thephasemeasurement [Melbourne,1985;Wubbena, phase-tracking, dual-frequencyreceiversare critical. P
1985].Blewitt[ 1989]givesa completediscussion. codepseudorange capabilityis usefulbut not criticalfor
Another"bootstrap"techniquethat assistsbias resolu- editingcycle slips,high-precision clock synchronization,
tion involvesthe "wide-lane"ambiguity.A linearcombi- andcarderphasecycleambiguityresolution.If P codeis
nationof observables,
thewide-lanephase L,•,definedas not present,C/A code pseudorangecan provide clock
synchronization to the microsecondlevel or better. The
Lw =- (•)1- 1•2)•Lw ability to tracksix or moresatellitesis helpful.If possible,
(29) commonreceiversandantennasshouldbe usedthroughout
= (f 1L1 -- f2 L2 )/fl -- f2 ) the network to ensure common time tagging in the
presenceof SA and reducethe effect of antennaphase
isuseful
because
thewide-lane
wavelength,
• = c/(f1-f2) centervariation [Rockenand Meertens, 1989].
-• 86 cm, is more easilyresolvedthan the "narrow-lane" In addition to adequatelysamplingthe geophysical
(•1, L2)ambiguities
(compare (20)and(29)and phenomenonof interest,two other aspectsof network
equations
definitions
of • and•). Subsequent canbe designare important.First, if baselineslongerthanabout
techniques
ambiguities 50-100 km are to be determinedwith high accuracy,an
appliediterativelyto resolvethenarrow-lane
(X•-- 10.7cm)asdiscussed above. WithP codereceiversadequate(three or more station) fiducial network is
the wide-laneambiguitymay be resolvableevenif the P important.A covarianceanalysiscan be performedto
codedataarerelativelynoisy(P codeerroraveragedover determine optimum configuration.Variations in the
several hours justneeds tobelessthan•/2, i.e.,<43cm). number,geometry,andqualityof fiducialstationsbetween
Codelessreceivershave an effectivewide-lanewavelength experimentscan lead to large, spuriousvariations in
of only 43 cm, thus the wide-laneambiguityis more geodeticpositionestimates[Larson, 1990a]. Second,the
difficult to resolve. regionof interestmustbe coveredwith a sufficientlydense
Most high-precision relative positioningexperiments networkof stationsto ensureresolutionof carrierphase
with GPS are "static,"in the sensethat a given receiver cycleambiguities. If a largeregionmustbe coveredwith a
observes in onepositionfor manyhoursor daysto obtain sparsenetwork,at least 2-3 baselinesshouldbe shorter
datastrength adequate to resolvethecycleambiguities and than 100 km. If necessary,the networkcan be coveredin
reduce other errors. However, over short (<30 km) two or more observingsessionswith a few common
baselines,it is also possibleto operatein a "kinematic" stations[e.g.,Kellogget al., 1989].
mode, with a receivermaintaininglock on the satellite With the satelliteconstellationavailableup until 1990,
signalsand movingfrom stationto station,continuously high-precisionresults have been achieved in several
recordingdata. As long as lock is maintained,only one regionswith 3-5 daysor moreof observation, assumingat
cycle ambiguityper receiver-satellite pair needsto be least 7-8 hoursof observationper day [e.g., Tralli and
resolved,regardless of the numberof stationscoveredby Dixon, 1988; Dong and Bock, 1989; Dixon et al., 1990;
therovingreceiver.The technique hasobviousadvantages Kornreich Wolf et al., 1990; Kellogg et al., 1990].
for rapidsurveyof largenumbers of points.A variationof Observingfor two or more days allows a check of data
this approach,"rapidstatic"surveying, doesnot require qualityby treatingeachday independently in the subse-
continuous signallock,exploitinghigh-quality P codedata quentanalysisandcomparingresultsandreducestheeffect
or priordatato resolvetheambiguities directly[Blewittet of randomerror by averaging.Another advantageof
al., 1989]. severaldays of observationis that multiday orbital arc
The assumption of linearity(equations(26) and (28)) analysisis possible.
means that the initial model for the observables must be The durationof observations in a givenday dependson
sufficientlyaccuratethat second-and higher-orderterms experiment locationandsatelliteconstellationgeometry,as
canbe neglected. Accuracycanbe estimated afterthe fact well as receiverdesign,multipath environment, and
by inspecting the solutionoffsetfromtheassigned a priori logisticalconsiderations. Experiencewith the Block 1
value. The solutions should be iterated unless the satellite constellationusingmainly four-channelreceiversallows
epochstatesare accurateto about1 km, andinitialground somegeneralizations. King andBlewitt[1990] suggestthat
positionsaccurateto severalhundredmeters[Lichten, observing for 6.5 hours or less per day may degrade
1990b]. results.Freymuellerand Golombek[1988] compared7-
and9-hourobservations for a GPS experimentin northern
Experiment Design SouthAmericawith covarianceanalysesand suggested
On the basis of the preceding discussionwe can improvementwith the longerobservingtime. Davis et al.
summarizethe importantaspectsof a successfulhigh- [1989] reportedthat carrierphasecycleambiguitiescould
precision GPSgeodetic experiment. Additionaldiscussionsnot be resolved reliably even on short (30-50 km)
canbe foundin worksby King andBlewitt [1990],Dixon baselineswith short(4.5 houror less)observingsessions.
264 ß Dixon:THEGLOBAL
POSITIONING
SYSTEM 29, 2/REVIEWS
OF GEOPHYSICS

Optimumobservingtimesmay decrease
oncetheBlock2 address
mainly"bestachievable"
accuracies;
in somecases
constellationapproaches operationalcapability,assumingan application mayhavelessstringent accuracyrequire-
receiversare capableof trackingmorethanfour satellites ments,andconsiderable savings in fieldoperation
anddata
anddepending on SA effects.High-quality P codedata,if analysiscostsmay be possible by accepting slightly
available,
couldalsoshorten observingtimesbyproviding reducedaccuracy.
constraints
oncarrierphasecycleambiguities. Bothprecisionandaccuracy of GPSbaselineestimates
Finally,siteselection and siteintegrityare important. dependstronglyonwhether or notthecarrier phase cycle
Sitesshouldhaveadequate sky visibilityandbe located ambiguitieshavebeenresolved [Blewitt,1989,Dangand
away from buildingsor other reflectingsurfacesthat Back, 1989].Bias fixing improvesthe precision and
introducemultipath.Instabilityof the groundmonument accuracy of GPShorizontal baseline componentestimates
maybe a substantial sourceof noisein a long-term GPS by factorsof 2-3; the verticalcomponent is usually
experiment,introducing spuriouspositionshiftsunrelated unaffected(Figure8). Theabilityto resolvethebiasesis a
to tectonicsignals,and is an importantand oftenover- good indicationthat certainerrorshave been reducedto
lookedaspectof GPSexperiment design.Monuments not somethreshold value;in general,it is difficultto resolve
emplaced in bedrock aremostsusceptible
andmayrequire thebiases
withlowerquality
data.Forconsistency,
mostof
specialdesign[Wyatt,1989]. the examplesdiscussedin this section will therefore
involve bias-fixed data.
As with mostphysicalmeasurements, it is usefulto
PRECISION AND ACCURACY distinguishrandom errors, the effects of which can be
reduced by averaging moredata,andsystematic errors.
Theprecisionandaccuracyachievable in a givenGPS The majorrandomerror sourcein GPS is the noiseof the
experimenthaveimprovedgreatlyin the lastdecadewith phase observable.
Systematic errorsareof twotypes.The
betterexperiment
design,
receiver
hardware, andanalytical firstisduetoerrorsintheinput(nonestimated)parameters,
techniques,
and furtherimprovements are likely. I will for example, fiducialsitelocation,
Earthorientation,
and

6
6
SINGLEDAYARCS EAST
5
[] BIAS ESTIMATED
[] BIAS-FIXED
4
4

3
3

1
1

0 - i
, 0 Figure8. Histograms
of short-term
repeatability
(equation(30)) for 15 baselinesin the northern
NORTH
4
Caribbeanbefore and after resolutionof carrier
phasecycle ambiguities (bias fixing). Note
improvement
in horizontal
components forbaselines
shorterthan about500 kin. Biasescouldnot be
resolved
on thelonger
baselines
withsingle-day
orbitalarcs(compare
to Figure6) [fromDixonet
al., 199la].

VERTICAL

i i ß

200 300
11 oo 1300

BASELINE LENGTH (km)

29, 2/REVIEWS OF GEOPHYSICS Dixon: THE GLOBAL POSITIONING SYSTEMß 265

geocentercoordinates.Sensitivityanalysessuggestthat repeatability,the scatterof a seriesof GPS measurements

only fiducialsite locationerroris significant[Lichtenand takenover monthsor years,which will showthe effectsof
Border, 1987; Dixon et al., 1991c]. The secondtype of slowly varying errors.For stationsnot separatedby an
systematic erroris associatedwith modelsconstraining the activefault,no motionis assumedon thebaseline,and(30)
estimatedparameters. For example,errorsin modelsof the can be used.For stationsseparatedby an activefault we
Earth'sgravityfield and the solarflux causeerrorsin the canassumesteadymotionandlook at the scatterof points
subsequentorbit and baseline estimates. Sensitivity abouta best fit straightline on a plot of positionversus
analysessuggestthat theseorbit-relatederrorsare negli- time [Davis et al., 1989; Larson, 1990b]. An important
gible. Other systematicerrors of this type result from questionis the relation between short- and long-term
assumptions and modelsconcerningthe atmosphere,for repeatability.
example, ignoring third- and higher-orderterms of the The overall precisionof an individual GPS baseline
ionospheric correction,useof incorrectmappingfunctions estimatewill dependon severalfactors,whosecontribution
(bothdry and we0 especiallyat low elevationangles,and to the total will be the squareroot of the summedsquares
the assumption of azimuthalsymmetryin thisdynamicand (rss) of the individual error terms. Some of these errors
nonhydrostatic medium.Althoughit is difficult at present will be constant,while otherswill dependon baseline
to quantify these systematicerrors, the data presented length.A simpleformalismfirst proposedby Savageand
belowsuggestsomebounds. Prescott [1973] to describethe precisionof Geodolite
One measureof the uncertaintyof a GPS vectorbaseline distancemeasuringequipmentis
estimate is the formal error, based on propagationof
randomdatanoisethroughthe estimationprocess.Formal • = (a2 + b2L2)1/2 (31)
error basedonly on data noise,even if scaledto reflect
postfit residuals,underestimates true error as it does not wherec•is thestandard deviation,
L is baselinelength,and
accountfor all randomerrorsnor most systematicerrors. a and b are the constantand length-dependent sourcesof
However,formal error doesallow us to propagateuncer- error,respectively,
with a = 3 mm andb = 2 x 10-7. For
taintiesand describecorrelations in a statisticallyrigorous comparison,the length componentof a VLBI position
manner. vectorhasa = 5.4 mm andb = 2.4 x 10-9 [Clark et al.,
Assumingthateachday of datais treatedindependently, 1989]. Equation (31) has also been used to describethe
the scatter of daily solutions for a GPS experiment precisionof GPS measurements as a functionof baseline
spanningseveraldays (the "short-term"repeatability)is length[Dixonet al., 1990;Hager et al., 1991]. As we shah
anothermeasureof GPS performanceand is one indicator see,thissimpleerrormodelhasweaknesses whenapplied
of precision.The repeatabilityR of a vector baseline to GPS,but is adequateto addresssomebasicquestions, in
component(east,north,or vertical)is theroot-mean-square particular the relation between short- and long-term
(rms)scatterabouttheweightedmean: repeatability.Best fitting curves (31) through some
short-termrepeatabilitydata for GPS experimentsin the
southwesternU.S. [Dixon et al., 1990] and the northern
(ci
i--1 -O•i Caribbean[Dixonet al., 1991a]give, for horizontal(north
R: • (30) andeast)components,
a = 3-8 mm,andb • 1-2 x 10-g
(Figure9). The eastcomponent is determinedmorepoorly
i=1•/ than the north, becausethe satellitegroundtrackswere
orientedmainly north-south, giving large range-change
where n is the numberof days, c is the estimateof the "signals"in the north-south
direction,andsmallersignals
component
on dayi, (c) is theweighted
average
of the in the east-west direction. The difference is reduced with
component overall days,and•/ is the formal bias-fixeddata.The verticalcomponent
estimate is themostpoorly
error.Short-termrepeatabilityhasbeenusedto assessthe determinedwith GPS, reflectingthe geometriclimitation
relative merits of various experimentaland analytical thatsatellitesare observedin the upperhemisphere only,
approaches[Tralli and Dixon, 1988; Dang and Back, andhighersensitivity(abouta factorof 3 relativehorizonø
1989;Blewitt, 1989;Dixon et al., 1991a].Note that if just tal components)to errors in troposphericcalibration
a few days of data are availablefor a small numberof [Herring, 1986]. Typical valuesfor verticalrepeatability
baselines,thestatisticalsignificance
of theobservedscatter are in the range 20-60 mm, with the higher values
may be limited. Also, short-termrepeatabilitywill not indicativeof highandvariablewettropospheric pathdelay
reveala slowlyvaryingsystematic error,for example,due (Figure9). Verticalrepeatabilityis nota strongfunctionof
to propagationdelay, multipath, or fiducial network baselinelengthandthusis notwell fit by equation(31).
inconsistencies,which might be constantover several Thequestion of short-versuslong-term repeatability
can
days.On the other hand,sincerandomerrorsare reduced be partiallyanswered by comparingthesedatato datafrom
with repeatedmeasurements, repeatabilitymight actually laterexperiments. Figure10 showsresultsfrom sevenGPS
be pessimistic.
With moredatanowavailableit is possible experimentsconductedover a 3-year period between
to use another indicator of precision,the "long-term" Mojave(on the North Americanplate)and Vandenberg
266 ß Dixon: THE GLOBAL POSITIONING SYSTEM 29, 2 / REVIEWSOF GEOPHYSICS

overa 2-yearperiodof 3-8 mm (north),3-6 mm (eas0,

+ Vertical and11-33 mm(vertical).Thusit appears thatfor carefully
[] East configured experiments(for example, robust fiducial
6 ß Nort,h networks,observingsessions of adequatelength,resolved
carderphasecycleambiguities), therepeatability observed
over severalyearsis not significantlydifferentfrom that
4 + observedoverseveraldays.Theseareencouraging results,
+
+ suggesting
thatthemagnitude of slowlyvaryingsystematic
+ -!- errors is similar to or less than the level of random errors
+ + + + •.•
manifested
overseveraldays.
[] + The accuracyof any long baselinemeasurement
is
•.+•,-1'•,•"•11'",
0 'i• ,e ,
.....
I , I ,
difficult to determinerigorously,since completely

15
MOJAVE-VANDENBERG (350 km)
+ + Vert,cal

+ + [] East + 10 --

o•' 6 ' + + ß Nort,'-

4' + +
• . ++ + o
uJ
m _ + []
21 ß ß

o/ 0 200 400
, , ,-, ,-,', ',1
600 800 1000 1200

LENGTH (km)

Figure 9. Short-term(3-8 days)repeatability

(equation(30)) for
experiments conducted in 1986 in southern
California(top) and
the northernCaribbean(bottom).Curvesare bestfit (equation
(31)) throughhorizontalcomponents only. Note the higher
scatterin the vertical componentestimatesfor the northern
Caribbean,in part reflectinghigher levels of tropospheric
-5
calibrationerror [fromDixon et al., 1990, 1991a]. 30

(on the Pacificplate),a baselinewhereVLBI dataare also

20
available.Straight-linefits throughboth data sets show
northwest motionof Vandenberg withrespectto Mojave,at
a rateanddirectionbroadlyconsistent with whatwe expect 10
to seefrom a globalplatemodel[DeMetset al., 1990].The
scatter of the GPS data about the best fit line is 6 mm
(north),10 mm (east),and 34 mm (vertical),not greatly 0
differentfromtheshort-term repeatabilitynotedin Figure9
for a baselineof thislengthandin fact similarto the VLBI
scatter[Dixonet al., 1990].Thissuggests thatat leastin this -10 I i I • I

location,for thisperiod,thecontribution of slowlyvarying 1986 987 1988 1989

YEAR
systematic errorswaslow.Daviset al. [1989]investigated
long-term(~2 year) repeatabilityfor six short(<50 km)
baselines anda singlelonger(223 km) baselineandfound Figure 10. Comparisonof GPS and VLBI results for the
rmsscatters aboutbestf'ltlinesof 2-12 mm (north),2-19 Mohave-Vandenbergbaselinein California between 1986 and
mm (eas0, and 4-49 mm (vertical),where the larger 1988. Solid line is VLBI global solutionGLB223; data from
values were for nonbias-fixed cases. Larson [1990b] individualVLBI experiments omittedfor clarityandhavescatter
presented dataon long-termrepeatability for 10 baselines similarto GPS data(fromLarson[1990a] reprintedin the work
between91 and 464 km in lengthand foundrms scatters by Dixon et al. [1990]).
29, 2/REVIEWS OF GEOPHYSICS Dixon: THE GLOBAL POSITIONING SYSTEMß 267

independenttechniquesof known accuracy are not are independentof baselinelength.Atmospheric(mainly

availablefor comparison.However, if two independent tropospheric) errorsdependon baselinelengthonly up to a
techniquesprovide similar results, we can be more certain distance, after which atmosphericvariation is
confident in each. For baselines <50 km in length, uncorrelated betweentwo sitesandthecorresponding error
conventionalterrestrialgeodetictechniquescan be used. reachesan approximatelyconstantvalue, dependingon
Davis et al. [1989] reportcomparisons
with creepmeters calibrationor estimationaccuracy.The correlationlength
andalignmentarrayson lines~10 km in length,suggesting for troposphere-related errorsat the level of a few mil-
GPS rate accuraciesof 1-2 mm/yr. The same study limeters,i.e., at a level exceedingreceivernoise,will vary
describedGPS comparisons with Geodolitetrilateration depending on site,season,andlocalweatherbut is likely to
data on lines in the range 10-50 km and found no be on the orderof tropospheric thickness,perhapsranging
detectable differences within the standard deviations of the from a few kilometers to a few tens of kilometers. For
measurements. For longer station separations,VLBI example,Ware et al. [1985] noticedsignificantdifferences
providesimportantdata for comparison. However,VLBI in the wet delay for stationsseparatedby 22 km. Orbit-
and GPS are not completelyindependent,sharing,for related errors are expectedto exhibit an almost linear
example,a similarsensitivityto tropospheric
effects.Also, dependenceon baseline length up to several thousand
GPS relieson Earth orientationdataprovidedby VLBI, as kilometers,
atalevelof1-2parts
in 108ofbaseline
length
well as fiducial site location data. Nevertheless, VLBI-
GPS comparisonsallow the best available standardfor
GPS accuracyassessment for baselineslongerthanabout + Vertical
50 km, and I will use the term accuracyto describethe [] East

results of GPS-VLBI comparisons,recognizing these ß North

-!- -!-

limitations.Figure 11 [Dixon et al., 1990] showsdiffer-

encesbetweenVLBI andGPS baselineestimatesplottedas
a functionof baselinelengthfor oneexperiment,with best
fit curves (equation (31) and Figure 9) summarizing
short-termrepeatabilityin the same experiment.The
comparisonsuggeststhat accuracyis very similarto the
short-termrepeatability,again limiting the magnitudeof
+ + • ß
systematic error.Figure 10 suggests that it is possibleto
achievea similarlevel of performance over longertime
. _' • , ,O ,O • , I ,
spans.Other GPS-VLBI comparisons in Californiahave
200 400 600 800 1000 1200
been reported: the 223-km Palos Verdes-Vandenberg
baseline(six experiments between1986 and 1988) [Davis LENGTH (km)
et al., 1989]; the 245-km Owens Valley-Mojave baseline
(a singleexperimentin 1987) [DongandBock,1989],and
Figure 11. GPS resultsfrom an experimentin Californiain 1986,
the 224-km Palos Verdes-Mojavebaseline (six experi- plottedas differencesfrom VLBI solutions.Curvesare best fit
mentsbetween1986 and 1988) [Larson, 1990b]. All these throughshort-termrepeatabilitydata from the same experiment
comparisons indicateno significantsourceof systematic (Figure 9) and are not based on the GPS-VLBI data. Note
errorof a type that wouldcauseGPS-determined ratesto similarity between "precision"(basedon repeatabilitycurves)
differ from VLBI-determinedrates.It is interestingto note, and"accuracy"(basedon VLBI comparison)[fromDixon et al.,
however,that comparisons involvingGPS and fixed-site 1990].
VLBI oftenexhibitnear-constant offsetswhich fortunately
do not affect the rate estimates.Figure 10 is a good or betterassuminggoodfiducialcontrol.Note that for the
example (see also Larson [1990b, Figure 4]. The best rangeof baselinelengthstypicalof mostGPS applications
explanation for theseoffsetsis thatthesitetiesrelatingthe the projectionof localverticalontothe horizontalbaseline
phasecentersof the large, fixed VLBI antennasto the componentsdue to Earth curvatureis not critical to the
nearby ground marks used by GPS may be in error. error analysis. With these constraintsin mind we can
However,this hypothesis remainsto be verified.Future modify(31) to give a simpleerrormodelmoreapplicable
work on GPS accuracyshould also include GPS-SLR to GPS:
comparisons,since VLBI and GPS have some common
error sources. o= [a2 + b2(1- e-œA) • + c•L•]•/• (32)
One problem with the two-parametererror model
describedby equation(31) is that•dual-frequency GPS wherea is thereceiverandsetup error,b is theasymptotic
measurements haveat leastthreesignificant errorsources, troposphericcontribution to baseline error for station
each with a different dependenceon baseline length. separationsexceeding several troposphericcorrelation
Receivernoiseand setup errorsare typicallyat the level of lengths, ), is the tropospheric correlation length,andc is
1-2 mm, determinedby "zero length"baselinetests,and orbit-related error.An examplemodelis shownin Figure
268 ß Dixon: THE GLOBAL POSITIONING SYSTEM 29, 2/REVIEWS OF GEOPHYSICS

12a.Note thecrossoverpointat about400 km, whereorbit 8

errorsstartto dominate.Tropospheric
errordominatesfor
station
separations
longer
than
about
10kmandshorter Total•
than
400kminthisexample.
Figure
12also
shows
some
long-termrepeatability
datafor a seriesof GPS experi- tr
ments in California for comparisonto the model. The ,m 4 posphere
captionliststhe criteriausedto minimizeinterexperiment
differences,
hopefullyensuringthatdatawith similarerror z
characteristics
havebeencompared.
However,
byusing¸N 2 Receiver

long-termrepeatability,
we have undoubtedly
overa rangeof tropospheric
conditions.
averaged O
-,-
Orbit
Mostof theGPS-VLBI
comparisons
performed
todate 0
arerestrictedto a fairly limitedregionin the southwestern

U.S.,
andanote
of
caution
isin
order
Sensitivity when
studies applying
indicate
that the8•I' ' ' ' ' ' '©'' '
resultingargumentson GPS precisionand accuracyto

other
regions. fiducial
network
are
less
uncertainties
orientation. are
Geodetic
susceptible
tothis
the
largest
networks contributor
to
systematicorbit and baseline error, the exact amount
dependingon experimentlocationandbaselinelengthand
inthesouthwestern
effect
because
ofthe
U.S.
proximal
<• • © [-'] ß North
4

location
ofseveral
high-quality
fiducial
sites,
i.e.,sitesm
n 2E3 [3 East
with a long historyof VLBI occupations
and well- rr • Model
(Total)
established
localtiesbetween
theVLBIantenna
phase 0 , • ., • , • , ' ,
center and the GPS mark. Even so, uncertainties in
VLBI-GPS ties at the 1-2 cm level are common [Tralli
andLichten,
1990;Dixonetal., 199lc]. Theeffectofthese 40

errors
grows
with
distance
from
the
fiducial
network
but
•30• O O O
will not causeincreasedscatterover severaldays,as the
orbitgeometryis essentiallyconstantoverthisperiod.The
effects may even be small over 1-2 year periodsif the
satellite constellationhappensto be similar for several
consecutiveexperiments,and if the same or similar 2o
fiducial networksare used.However,over the 5-10 year
or longer periods likely tocharacterize
manyGPSmm
10 Ve'•'•cal
experiments (see
below),
suchfidelity
is unlikely,
mm Model
implying the need for care in the establishment and 0I • I , i • I • I •
maintenance of fiducial networks. 0 1 O0 200 300 400 500
When we perform a seriesof GPS experiments,we
LENGTH (km)
generallywish to obtain estimatesof strainrate or fault
motion rate. Many geologicproblemsrequire rate ac-
curaciesin the range 1-5 mm/yr [Jordanand Minster,
1988].In orderto planexperiments it is importantto know Figure12. (a) Individualerrorcomponents andtotal(rss)error
how long it will taketo acquirea damsetwith a givenrate fromequation (32)forthehorizontal componentof a hypotheti-
accuracy,or how often an experimentmustbe repeated calGPSexperiment, witha = 2 mm("receiver"),
b = 4 ram,• =
withina fixed period(for example,5-10 years)to achieve 10km(b and• together define"troposphere"),
andc = 1 x 10-8
thataccuracy.
The uncertainty
in a rateestimate
•r is a ("orbit").(b) Long-term repeatability
(weighted
rmsscatterabout
function
of thesinglemeasurement
position
accuracy a bestfit straightline)for GPShorizontalcomponent measure-
the intervalbetweenexperimentsAt (assumedconstantin mentsin central and southernCalifornia between 1986 and 1988
therelationbelow),andthe totaltime spanof observations, compared
to anerrormodel(equation (32)).All dataplottedare
T, and can be estimatedfrom [Coateset al., 1985]: fromTI-4100receivers,
represent
fourormoreexperiments,
and
have all or mostcarrierphasebiasesresolved.Data are from
Daviset al. [1989] andLarson[1990b]."Model"curveis same
as"total"in Figure12a. (c) Similarto Figure12b,for vertical
component. Modelcurvefromequation(32)), with a = 5 ram,b
C•r c•,•
=-•- [(1+T/At)(2
12T/At
+T/At) (33) = 20mm,k= 10km,andc= 1 x 10-8.
29,2/REVIEWS
OFGEOPHYSICS Dixon:
THEGLOBAL
POSITIONING
SYSTEM
ß269

Figure 13 plots the rate uncertainty

(1(•) for single In the limit of shortlengthand time scales(implying
measurement uncertainties
of 1 and2 cm (1(•), for several dense networksand frequentobservations),high-accuracy
valuesof At. While thesinglemeasurement uncertainty is GPSis currentlylimitedmainlyby economic andlogistical
clearlyimportant,the mostefficientway to reducethe constraints,receiver limitations, and data processing
uncertainty
in a rateestimate is to extendthetotaltime efficiency.In the nearfuture,with receiverimprovements
spanof observations.
Comparisonof modernGPSorother and the enhanced constellation, ~5-mm horizontal
geodeticdataandlowerqualityhistorical data accuracy
geodetic shouldbeachievable
overa widerangeof station
taken20 or moreyearsago is a goodexampleof this separations
with kinematicor rapid static techniques,
principle.Significant
compressional
deformationalong enabling densespatialsampling.
Asreceivercostsfall and
partsof thesouthern
SanAndreas in Caldor- the efficiency of data analysisimproves,continuous
faultsystem
nia hasbeendocumented
with thisapproach[Feigl et al., monitoringof densenetworkswith GPS also becomes
1990b; Webb, 1990]. feasible,allowinggreatlyimprovedtemporalsampling.
This capabilitywill complement seismicnetworksand
10
1 •.• I ' ' I ' strainobservatories
(D. M. Tralli, Spectralcomparisonof
q,,, =- 1•m,1 yr' continuousGlobal Positioning System and strainmeter
• %, '.........
O.......
2cm/lyr measurements of crustal deformation, submitted to
• %,, ....
I---1cm/3yr of
Geophysical
Research
phenomena
Letters,1991),enablingmonitoring
at periodsjust beyondthe limit of long-
•O,.....
"%,
'""O..
----13--
"1-1,
2cm/3
yr periodseismometers,for example,creepandstrainevents
(Figure 14) and other possibleeffectsrelatedto stress
migration.Since 1-mm horizontalprecisionin theoryis
achievablewith severalminute samplingof carrierphase
andpreciseP codepseudorange
dataovershort(<10 km)

12
10 STRAIN = 10 plate boundary
0 , I • I • I , I , zone
0 2 4 6 8 10 tectonics
103 yr
10 volcanic post
TIME AFTER FIRST EXPERIMENT (yrs) 10 post-seismic
events rebound, glacial
rebound
coupling

Figure 13. Uncertainty

of a geodetically
determined
rateestimate •'
asa function
of time,T (equation of single ••
(33))fortwovalues 108
SLR
measurement
uncertainty,
o• (1 and2 cm,listedfirstin legend) LU 1 yr
and two valuesof experimentrepeatinterval,At (1 and3 years, <
listedsecond).Note that rate uncertaintyis betterthan3 mm/yr ca VLBI

after10years
forallcases
shown. • 10
GPS-1990

1 day
GEOLOGICAL APPLICATIONS AND ACCURACY lO4 GPS-1995 ?

REQUIREMENTS 1 hr

In this sectionwe review severalapplications

of GPS 2
eadhquakes
10
and their corresponding accuracyrequirements. The
examples encompass a widerangeof stationseparations, 10¸ 101 102 103 104
an importantfactor in determining both the accuracy SPATIAL SCALE (km)
achievablewith GPS and critical aspectsof experiment
design.
Figure14 (modifiedfromMinsteret al. [1990]shows Figure 14. Cartoonshowingorder of magnitudeestimatesfor
thetemporalandspatialscales
of a varietyof geologicalspatialandtemporalscalesof somegeologicalphenomena(lower
phenomenawheregeodetic monitoringcanprovide useful caseletters,stipplepattern)comparedto detectioncapabilityof
information.The regionsof applicabilityfor several some geodetic techniques(upper case letters, diagonal line
geodetictechniques arealsoshown.
It is apparentthateven pattern),
assuming
10-7 strainor higherandtypicaloperating
at currentlevelsof performance,
GPScancontribute to the conditions.OB, observatoryinstruments(for example,continu-
studyof a widevarietyof phenomena. This figurealso ously recording strainmeters,tiltmeters); VLBI, very long
baselineinterferometry;SLR, satellitelaserranging;EDM, two
indicatesa roughestimateof GPSperformance 5 years
color laser electronicdistancemeasuringequipment,assuming
hence, based on the expectedimpact of some recent
daily measurements. Modified fromMinsteret al., [ 1990].
developments.
270 ß Dixon: THE GLOBAL POSITIONING SYSTEM 29, 2 / REVIEWSOF GEOPHYSICS

baselines,
it is temptingto speculate
on thedetectability
of zonethrustfaultsaregivenby Savage[1983] andThatcher
seismic precursorsproposedby Lorenzetti and Tullis and Rundle[1984]). The fault-parallelsurfacedisplace-
[1989] with a densearray of continuouslyoperatingGPS mentratev duringan earthquakecyclefor a simpleelastic
stations. half-spacemodelis [SavageandBurford,1973]:
In the limit of long length scales,GPS accuracywill
improveas globaltrackingnetworksevolveto includea v= (Vo/•)tan-• (x/D) (34)
uniform set of well-distributed,high-accuracyground
systemswith strict configurationcontrol,and as force where D is the fault or "locking" depth related to the
models for solar radiation pressure and other orbit brittle-ductiletransitiondepth,x is distanceperpendicular
perturbations are ref'med.This will enablea numberof to thefault,andv0 is the totalsliprateat largedistances
platemotionandregionalplateboundaryzoneproblemsto from the fault or beneaththe locking depth.If T is the
be attacked.Even with currentperformance thereare many earthquakerecurrence interval,thenthecoseismicslipthat
moreapplications of GPS thanwe canaddresshere,but to relieves
accumulated
elasticstrain(andstress)
is justvoT.
give someflavorfor thediversity,I will considertwo types Geodeticdatamay be invertedfor D and v. Constraints
on
of applications,dividedarbitrarilyby lengthscale. D are also availablefrom seismicity(large earthquakes
tend to nucleate near the base of the fault, at the brittle-
Short (0-100 km) Baselines:The Near Field ductiletransition),
andv0maybedetermined
fromfar-field
Geodetic measurements within about 50-100 km of a spacegeodeticmeasurements (VLBI, SLR, or GPS) or, for
fault systemmay provideconstraints on crustand mantle a simpleplateboundary,froma globalplatemotionmodel.
rheology,faultingmechanics, andtheearthquake cycle.This For strike-slipfaultsin continentalcrust,D is typically
lengthscaleis alsoappropriate for monitoring uppercrustal about10-20 kin, on thebasisof inversionof geodeticdata
deformation associated with inflation/deflation of shallow [e.g.,King et al., 1987],seismicdepthdata,andrheologi-
magmachambers.Until recently,mostsuchstudieswere cal properties of crustalmaterialsfrom laboratoryexperi-
performedby conventionalterrestrialgeodetictechniques ments [Sibson, 1983]. Figure 15 plots differencesin
(triangulation,
tri!aterafiondistancemeasurement, andlevel- fault-paralleldisplacementrates for elastic half-space
ing). The accuracy,siteflexibility,easeof dataacquisition, modelswith D = 10, 15, and 20 kin. The largestrate
andincreasingly straightforwarddamanalysisallowsGPS to differences are observed at horizontal distances from the
complement, and in somecases,replacetheseapproaches. faultroughlyequalto fault depth,givingsomeguidanceas
The accurate three-dimensional vector nature of GPS to stationplacement.Distinguishingamongthe models
positioningalso allowsconstraints to be placedon block would require resolutionof fault-paralleldisplacement
rotations,somethingthat is not possiblewith conventional ratesat the 1-3 mm/yrlevel.
techniques alone [e.g.,Davis et al., 1989].Minsteret al.
, [

[1990],Bilham [1991], andHager et al. [1991] review ap- 41 I ' I I

•, 10-15
plicationof geodeticstraindetermination to near-fieldand
earthq•e studies.In thissectionI will focusmainlyon one • 15-20

aspect,namelytheaccuracyrequiredfor GPSdatato distin- • 10-20

guishamongsomeplausiblemodelsof faultingandcrustal
rheology.
A seriesof GPSexperiments allowsdetermination of sur-
face strainand strainrate tensors,reflectingstresses and
rheologicalproperties in thecrustandmantle.Simpleelastic
half-spacemodelsand more complicatedrheologicaldis-
tributionshavebeenusedto explainobserved surfacedefor- -4 , , i •
mation [e.g., Melosh, 1976, 1977; Savageand Prescott, -40 -20 0 20 40

1978;Spenceand Turcotte,1979;Turcotteet al., 1984;Li DISTANCE (km)

and Rice, 1987; Li and Lira, 1988]. The mannerin which
deformationchangesin space(away from a fault) and time
(beforeandaftera largeearthquake) mayrevealinfomarion Figure15.Differencein faultparallelsurfacedisplacement
rate
abouttherheologyof the lowerlayersbut may alsorequire forseveral
modelsassuming a locked,vertical
strike-slip
faultin
anelastic
halfspace(equation (34))forfaultdepths,
D = 10, 15,
largenumbersof accuratemeasurements on any givenfault
and20km.A farfielddisplacement rateof 30 mm/yris assumed.
segment to describe adequately
thespatialandtemporaldis-
tributionof strain.Importantquestions
includewhetherGPS
dataallow discrimination amongthesemodels,whataccu- A more realistic model for deformation near a fault
racyis required,andwhetherwe nee• to incorporateany ex- accountsfor the layeredrheologyof the crustand upper
traordinary featuresintotheGPSexperimentdesign? mantleandthecouplingthatoccursbetweenlayers.Recent
Considerthe surfacestrainpatternfor a locked,vertical modelsassumethat lower layer or layers behave like
strike-slipfault (strainmodelsappropriateto convergence Maxwellviscoelastic solids.Like its mechanical
analogue,
29, 2/REVIEWS OF GEOPHYSICS Dixon: THE GLOBAL POSITIONING SYSTEM ß 271

a springand dashpotin series,a viscoelasticsubstance degreesor betterin azimuth[DeMetset al., 1990]. These
behaveselasticallyover shorttime scalesand viscously models are based on spreadingrates from mid-ocean
over long time scales.Viscousrelaxationin the viscoelas- ridges,determined frommagneticanomalies averagedover
tic "asthenosphere" dampsshort-termperturbations in the severalmillion years, azimuthsof transformfaults that
overlyingelastic"lithosphere."The rate of stressrelaxa- averageover a similartime interval,andearthquake slip
tion is related to a time constant2•l/!.t, where •1 is the vectors representing essentially instantaneouslocal
viscosityof the lower layer and !x is the averagemodulus azimuthinformation.There is no reasonto suspectthat
of rigidity of the two layers.Expressions for the surface plate motionaveragedover severalmillion yearsdiffers
strainspredictedby couplingmodelsare givenby Savage over shorter(for example, geodetic)time scaleswhen
and Prescott [1978], Turcotte et al. [1984], and Li and measured betweenplateinteriors,away from the transient
Rice [1987]. Inspectionof these models suggeststhat effectsof plate boundarydeformation,unlessa plate
differencesin fault-parallelsurfacedisplacement ratesof boundaryis undergoing rapid evolution.This hypothesis
1-2 mm/yr or better are required for discrimination. hasbeenlargelyconfirmedby NASA's CrustalDynamics
However,for mostpartsof the earthquake cycle,predicted Project,which hasmadevelocitydeterminations between
surfacedeformationwith the elastichalf-spacemodeland severalmajorplateswith VLBI [e.g.,Argusand Gordon,
eventhe simplest(two layer) viscoelastic couplingmodel 1990] and SLR [e.g., Harrison and Douglas, 1990].
may be surprisinglysimilar and thus difficult to dis- Nevertheless,there are some plate boundarieswhere
criminatewith GPS or any othergeodetictechniquealone relativeplatemotionis poorlyconsmined,andwhereGPS
[Savage and Prescott, 1978; Thatcher, 1983; Savage, (as well as VLBI and SLR) measurementscan make
1990].On theotherhand,time-varyingsignalsoverpartof uniquecontributions. Plates lacking a spreadingcenter
the earthquakecycle are a potentiallyusefulobservable boundary (for example,thePhilippineplate)or with only
[Thatcher, 1983; Li and Rice, 1987], and we should not small or poorly definedspreadingcenterboundaries (for
rely on long(> 10-15 years)observation timesto achievea example,the Caribbeanplate) will have poorly defined
given rate accuracy.Davis et al. [1989] obtainedrate relativemotionrateswith respectto adjacentplates.Plate
uncertaintiesof 1 mm/yr in horizontalcomponentsand boundaries wheretransform faulttrendsor earthquake slip
3-4 mm/yr in the verticalcomponentfor baselinesup to azimuths may be biased due to the influence of
12 km in lengthin 3 yearsusingGPS data takenseveral heterogeneous continentalcrust(for example,the Pacific-
timesper year. North Americaplate boundarynorthof the southernGulf
GPS is also well suited to define the coseismic strain of California)will havepoorlydetermined relativemotion
field associatedwith earthquakes.These strains may azimuths.Predictedmotionat subduction zonesmay be
exhibit great spatial variation within a few tens of biasedbecauseof systematic biasesin trenchearthquake
kilometersof the fault, particularlyfor complex fault slip vectors[DeMets et al., 1990]. In caseswhere local
geometry,implyingthe needfor densesurveys.It may be data are lackingor biased,relativeplate motion is deter-
useful to separatecoseismicrupture from postseismic minedmainlyfromclosureconditions on otherplates,and
relaxationphenomena,implying a requirementfor rapid GPS or otherspacegeodetictechniquecanprovideuseful
responseand frequentresurveyafter an event.If required complementaryinformation. GPS stands out as the
accuracies are not high (severalcentimetersare sufficient appropriate techniqueif economicor logisticalconsidera-
for most studiesof coseismicoffset),kinematicor rapid tionsprecludethe useof SLR or VLBI, bothof whichhave
static techniquesenabling survey of large numbersof larger, more cumbersomeground systems,and if the
points may be useful. Larsen [1990] used kinematic problemcan be solvedwith stationseparations lessthan
techniquesto measure coseismicoffsets for the 1987 1000-2000km (bothVLBI andSLR havebetterperform-
Superstition Hills earthquake sequence in southern anceat longerstationseparations).
California. Figure16 showsthe uncertainty in relativevelocitiesof
someplatepairswith convergent boundaries. If at leastone
LongBaselines (100 > 1000 km) of theplatesisoceanic, nearbymeasurement sitesspanning
Long baselinemeasurements with GPS are challenging the trenchplateboundary may not exist.One of the goals
becausethe influenceof orbit errorsis directly propor- of the CASA (Centraland SouthAmerica)GPS experi-
tional to baselinelength,and becausecarrierphasecycle ments[Kellogget al., 1989;Kelloggand Dixon, 1990] is
ambiguity resolutionbecomesmore difficult at station to determine the rate and direction of subduction of the
separationsexceeding 100-200 km. Special care is Cocos and Nazca plates beneathCentral and northern
requiredin networkdesign,and covarianceanalysesprior SouthAmerica,amongthe mostpoorlydetermined plate
to the actualexperimentmay be important.The geologic pairsin theNUVEL 1 platemotionmodel[DeMetset al.,
rationalefor pursuingsuch difficult GPS experiments, 1990](Figure16).Thisrequiresmeasurements fromCocos
someexamples,andpreliminaryresultsare givenbelow. Island(the only point of land on the Cocosplate) and
Relativevelocitiesof mostof the majorplatesare now MalpeloIsland(theonlypointof landon theNazcaplate
well known from globalgeologicmodels,with precisions within 500 km of mainlandSouthAmerica)to pointsin
of a few millimetersper year or betterin rate and a few Central and South America, defining numerouslong
272 ß Dixon: THE GLOBAL POSITIONING SYSTEM 29, 2 / REVIEWSOF GEOPHYSICS

INCREASING UNCERTAINTY has at least three major faults that are importantin
! ! accommodating thetangential
components of relativeplate
CO-NA •////•.///•.////////•J motionbetweenthe Pacific and North Americanplates.
! I

NZ-SA //////////•///////////////////• Rotations are predicted theoretically [Ivins, 1989] and

Neogenerotationsin partsof theplateboundaryzonehave
CO-CA /•///////////////////•
! !
beenobservedwith paleomagnetic and otherdata [Carter
CA-SA et al., 1987; Luyendykand Hornafius, 1987; Nur et el.,
0 8 16 24
1989]. Kinematicmodelsof the region [e.g., Bird and
F Rosenstock, 1984;WeldonandHumprheys,1986;Rundle,
FO.05 FO.01 1986]arebeingtestedwith GPS,employinga networkthat
includes islands of the southern California borderland and
station separationsmainly in the range 100-500 km
Figure 16. Uncertaintyin relativeplatemotionvectors(NUVEL [Agnewet al., 1988; Dixon et al., 1990]. Precisionof the
1 model)[DeMetset al., 1990] for plate'pairswherelocalmotion GPS-based rate estimates should be commensurate with
is constrainedmainlyby trenchslip vectors;mostotherplate the models (a few millimeters per year) [Bird and
pairsin themodelarebetterdetermined. Uncertaintyis basedon Rosenstock,1984; Jordan and Minster, 1988], achievable
theF ratiotestfor plateclosure[Gordonet al., 1987]."F.01" and in 5 yearswith yearly GPS measurements
(Figure 13).
"F.05" indicate1% and5% risk levels,respectively
(99% and Preliminary
resultsarereviewedby Hageret al. [1991].
95% confidencelevels).CO, Cocos;NA, North America;NZ, Total relative motion between the Pacific and North
Nazca; SA, South America; CA, Caribbean. Modified from
DeMets et al. [1990].
Americaplatescanbe measured by spanning
therelatively
simpleridge-transformboundaryin the southernGulf of
California,avoidingtectoniccomplexitiesassociatedwith
baselines. It was known in advance that a robust fiducial the TransverseRangesand Basin and Range extension.
networkwouldbe requiredto performsuchmeasurements, The relevantbaselinesare 300-500 km in length.GPS
and covarianceanalyseswere performed to determine measurements were made in 1985 and 1989, allowinga
optimal configurations [Freyrnueller and Golombek, preliminarydeterminationof the relative motion vector
1988]. These analysessuggestedthat a North American (Figure 17). Using a sensitivityanalysisto quantify
fiducial network by itself was inadequate.A global randomand systematicerror,Dixon et al. [1991c] showed
network of tracking stations was instituted for this that the Pacific-North America relative motion vector
experiment,the first civilian globalGPS network[Neilan determinedby GPS in the southernGulf of Californiais
et al., 1989].Preliminaryresultsusingthisnetworktendto consistent, withinlc• totalerror(+_7mm/yrin rate,_+6
ø in
confirm the predictionsof the covarianceanalyses,with azimuth),to the NUVEL 1 globalmodel [DeMetset al.,
significantimprovements in long baselineestimateswhen 1990].
the expandedtrackingnetworkis used[KornreichWolf et
al., 1990;Freyrnuellerand Kellogg,1990;Figure5). The
secondCASA experimentis scheduled for 1991. _•1 I I
Relativemotionbetweenplatesis occasionally accom- 3.0

modated within a continent. Continental crust is both

weaker and older on average than oceanic crust, and
inherited weaknessesof varying orientation may be
reactivatedunder the newer stressregime of the plate
boundary[Ivinset al., 1990], complicatingthe boundary's
geometryand evolution,and oftenleadingto broadzones
of distributeddeformation.Examplesof complexcontinen-
tal transform boundaries include the San Andreas fault
zone in California, the Motogua-Polochicfault zone in
Central America, and the Alpine fault zone in New
Zealand.High-precisionGPS measurements canbe useful I I I I

4.0 3.0 2.0 1.0

in characterizingboth the integratedmotion acrossthe
WEST VELOCITY(cm/yr)
entire deformation zone and details of crustal behavior
near individual faults within the plate boundaryzone,
includingrotationsof quasi-rigidblocksbetweenfaults. Figure 17. Vector diagram summarizingrelative motion of
Knowledgeof individualfault ratesand the totalplaterate Pacificplate(at CaboSanLucas)with respectto North America,
canlead to development of internallyconsistent
kinematic based on GPS measurement,and NUVEL 1 global vector,
block models.In effect, a completemassbalancecan be closure fatting vector (CFV), and best fitting vector (BFV)
obtainedbecauseadequatespatial samplingis possible [DeMets et al., 1990]. Error ellipsesare lo from Dixon et al.
with GPS.For example,the southernSanAndreassystem [1991].
29, 2/REVIEWS OF GEOPHYSICS Dixon: THE GLOBAL POSITIONING SYSTEMß 273

CONCLUSIONS Bilharn,R., Earthquakesand sealevel: Spaceand terres•al met-

rologyona changingplanet,Rev.Geophys.,29(1), 1-29, 1991.
Bird, P. and R. W. Rosenstock, Kinematicsof present•ust and
I have coveredin a generalway the major principles manfie flow in SouthernCalifornia,Geol. Soc. Am. Bull., 95,
behindGPS, particularlythoserelevantto high-precision 946-957, 1984.
geological and geophysicalapplications.GPS is an Black,H. D., andA. Eisner,CorrectingsatelliteDopplerdatafor
impressivenew measurementtool; the rate of improve- tropospheric effects,J. Geophys.Res.,89, 2616-2626, 1984.
mentin the qualityandquantityof geodeticmeasurements Blewitt, G., Carrier-phaseambiguityresolutionfor the Global
Positioning Systemappliedto geo&tic baselinesup to 2000
with GPS over the last decadeis astonishing.We can km, J. Geophys.Res.,94, 10,187-10,203, 1989.
expectto seecontinuedimprovements in receiveraccuracy Blewitt, G., GPS techniquesfor monitoringgeodynamicsat
and portability,reducedreceivercost,decreasedtime and regional scales, paper presented at Second International
costfor dataprocessing, improvedmodelsand calibration Symposium on PrecisePositioningwith Global Positioning
of major error sources,corresponding increasesin the System,Int. Assoc.of Geod.,Ottawa, 1990.
Blewitt, G., U. J. Lindqwister,and K. W. Hudnut,Densification
numberof diversityof scientificapplications,and some of continuously operatingGPS arraysusing a rapid static
veryinterestingscientificresults. surveytechnique(abstract),EosTrans.AGU, 70, 1054, 1989.
Bock,Y., R. I. Abbot,C. C. Counselman Ill, R. W. King,andS.
A. Gourevitch,Three-dimensional geo&tic control by
ACKNOWLEDGMENTS. Comments on a preliminary interferometry with GPS: Processing of GPS phaseobserv-
versionof themanuscript
by Eric Ivins, SteveLichten,andLarry ables,paperpresented at First InternationalSymposium on
PrecisePositioningWith Global PositioningSystem,Int.
Young were very helpful. I am indebtedto Jim Davis and Bob Assoc. of Geod., Rockville, Md., 1985a.
King for thorough reviews, and editor-in-chief Marcia Bock, Y., R. I. Abbot, C. C. CounselmanIll, S. A. Gourevitch,
Neugebauerfor patience and many helpful suggestions.Ken andR. W. King,Establishment of three-dimensional
geodetic
Hurst,ChrisRocken,DaveTralli, andBob King kindlyprovided controlby interferometry with theGlobalPositioning System,
manuscripts in advanceof publication.This work wasconducted J. Geophys.Res.,90, 7689-7703, 1985b.
at the JetPropulsionLaboratory,CaliforniaInstituteof Technol- Bock,Y., S. A. Gourevitch, C. C. Counselman HI, R. W. King,
ogy,undercontractto NASA. and R. I. Abbot, Interferometricanalysisof GPS phase
Marcia Neugebauerwas the editorresponsible for thispaper. observations,Man. Geod., 11,282-288, 1986.
She thanksJamesDavis and RobertKing for their assistance in Bossler,J. D., C. C. Goad,andP. L. Ben&r, Usingthe Global
evaluatingthe technicalcontentandBrian Tinsleyfor servingas Positioning System(GPS) for geodeticpositioning,Bull.
Geod., 54, 553-563, 1980.
a cross-disciplinary
referee.
Carter,J. N., B. P. Luyendyk,andR. Terres,Neogeneclockwise
tectonicrotationof theEasternTransverse Ranges,California,
Geol. Soc.Am. Bull., 98, 199-206, 1987.
Carter,W. E., D. S. Robertson,
andJ. R. Mackay,Geodeticradio
interferometric surveying: Applications and results, J.
REFERENCES Geophys.Res.,90, 4577-4587, 1985.
Chao,C. C., The tropospherecalibrationmodelfor MarinerMars
Abbot, R. I., and C. C. Counselman,Demonstrationof GPS orbit 1971,JPL Tech.Rep.,32-1587, 61-76, 1974.
determinationenhancementby resolutionof carrier phase Clark,T. A., W. G. Melbourne,C. Reigber,L. E. Young,andT.
ambiguity(abstract),EosTrans.AGU, 68, 1238, 1987. P. Yunck,MicrowaveTechniques, in The Interdisciplinary
Agnew, D.C., Y. Bock, R. King, T. Jordan,B. Hager, and D. Roleof SpaceGeodesy, vol. 22, editedby I. I. MuellerandS.
Jackson,GPS measurementsin central and southernCalifor- Zerbini,p. 43-102, Springer-Verlag, New York, 1989.
nia, CSTG Bull., 10, 71-87, 1988. Coates,R. J., H. Prey,G. D. Mead,andJ. M. Bosworth,
Space
Allan, D. W., Statisticsof atomic frequencystandards,Proc. age geodesy:The NASA CrustalDynamicsProject,IEEE
IEEE, 54, 221-230, 1966. Trans. Geosci.RemoteSens.,GE-23, 360-368, 1985.
Allan, D. W., M. A. Weiss,andT. K. Peppler,In searchof the Cohen, S.C., and D. E. Smith, LAGEOS scientific results:
best clock, IEEE Trans. lnstrum. Meas., 38, 624-630, 1989. Introduction, J. Geophys. Res.,90, 9217-9220, 1985.
Aoki, S., B. Guinot,G. H. Kaplan,H. Kinoshita,D. D. McCar- Counselman,C. C., and S. Gourevitch,Miniature interferometer
thy, and P. K. Seidelman,The new def'mitionof universal terminalsfor Earthsurveying: Ambiguityandmultipathwith
time,Astron.Astrophys., 105, 359-361, 1982. the Global PositioningSystem,IEEE Trans. Geosci.Remote
Argus,D. F., and R. G. Gordon,Comparisonof Pacific-North Sens.,GE-19, 244-252, 1981.
America plate motion &termined from very long baseline Counselman, C. C., and I. I. Shapiro,Miniatureinterferometer
interferometry with that&terminedfrom magneticanomalies, terminalsfor earthsurveying, Bull. Geod.,53, 139-163, 1979.
transformfaultsandearthquake slipvectors,J. Geophys.Res., Davis,J. L., T. A. Herring,I. I. Shapiro,A. E. E. Rogers,andG.
95(11), 17,315-17,324, 1990. Elgered, Geodesyby radio interferometry:Effects of
Beutler, G., W. Gurtner, I. Bauersima, and M. Rothacher, atmospheric
modelingerrorson estimates
of baselinelength,
Efficientcomputation of the inverseof the covariance
matrix Radio Sci., 20, 1593-1607, 1985.
of simultaneous GPS carrier phasedifferenceobservations, Davis,J. L., W. H. Prescol•,J. L. Svarc,andK. Wendt,Assessment
Man. Geod., 11, 249-255, 1986. of GlobalPositioning Systemmeasurements for studies
of crustal
Beutler, G., I. Bauersima, W. Gurmer, M. Rothacher, and T. &formation,J. Geophys. Res.,94, 13,635-13,650,1989.
Schildknecht,Evaluation of the 1984 Alaska Global Position- DeMets,C., R. G. Gordon,D. F. Argus,and S. Stein,Current
ing System campaignwith the Bemese GPS software,J. platemotions,Geophys.J. lnt., 101, 425-478, 1990.
Geophys.Res.,92, 1295-1303, 1987. Dickey,J. O., XX Newhall, andJ. G. Williams, Earthorientation
Bevis, M., GPS networks:The practicalside,Eos Trans.AGU, datafromlunarlaserrangingandan erroranalysis of polar
72, 49-56, 1991. motionservices, J. GeophysRes.,90, 9353-9362, 1985.
274 ß Dixon: THE GLOBAL POSITIONING SYSTEM 29, 2/REVIEWS OF GEOPHYSICS

Dixon, T. H., and S. Komreich Wol•, Some tests of wet Hager,B. H., R. W. King, andM. H. Murray,Measurementof
tropospheric
calibration
for theCASAUnoGlobalPositioning crustaldeformationusingGPS,Ann.Rev. Geophys.,
in press,
Systemexperiment,
Geophys. Res.Len.,17, 203-206, 1990. 1991.
Dixon, T. H., M.P. Golombek, and C. L. Thornton,Constraints Harrison,C. G. A., andN. B. Douglas,Satellitelaserrangingand
on Pacific plate kinematics and dynamics with Global geologicconstraints
on plate motion,Tectonics,9, 935-952,
PositioningSystem (GPS) measurements, IEEE Geosci. 1990.
RemoteSens.,GE-23, 491-501, 1985. Hellwig, H., Microwave time and frequencystandards,
Radio
Dixon, T. H., G. Blewitt, K. Larson,D. Agnew, B. Hager, P. Sci., 14, 561-572, 1979.
Kroger,L. Scrumeda,and W. Strange,GlobalPositioning Herring,T. A., Precisionof verticalpositionestimatesfromvery
Systemmeasurements of crustaldeformationin southern long baseline interferometry, J. Geophys. Res., 91,
California: Some constraintson performance,Eos Trans. 9177-9182, 1986.
AGU, 71, 1051-1056, 1990. Herring, T. A., C. R. Gwinn, and I. I. Shapiro,Geodesyby
Dixon, T. H., G. Gonzalez,S. Lichten, and E. Katsigris,First radio interferometry:Studiesof the forced nutationsof the
epochgeodeticmeasurements with the Global Positioning Earth, part 1, Data analysis, J. Geophys. Res., 91,
Systemacross
thenorthern Caribbean
plateboundary
zone,J. 4745-4754, 1986.
Geophys.
Res.,96, 2397-2415,1991a. Hide, R., and J. O. Dickey, The Earth's variable rotation,
Dixon, T. H., D. M. Tralli, G. Blewitt, and J.P. Dauphin, Science,in press,1991.
Geodetic baselinesacrossthe Gulf of California using the Ivins, E. R., T. H. Dixon, and M.P. Golombek, Extensional
GlobalPositioning System,Am.Assoc.Petrol.Geol.,in press, reactivationof an abandoned thrust:A boundon shallowingin
1991b. thebrittleregime,J. Struct.Geol.,12, 303-314, 1990.
Dixon, T. H., G. Gonzalez, S. Lichten, D. Tralli, G. Ness, and J. Ivins, E. R., New aspectsof rotational dynamicswithin the
P. Dauphin, Preliminarydeterminationof Pacific-North North-American-Pacific ductileshearzone,in DeepStructure
America Plate motion in the southern Gulf of California with andPast Kinematicsof AccretedTerranes,Geophys.Monogr.
the Global PositioningSystem,Geophys.Res.Lett., in press, Ser., vol. 50, editedby J. W. Hillhouse,pp. 179-201, AGU,
1991c. Washington,D.C., 1989.
Dong, D., and Y. Bock, GPS network analysiswith phase Janssen,M. A., A new instrumentfor the determinationof radio
ambiguityresolutionappliedto crustaldeformation studiesin path delay due to atmosphericwater vapor, IEEE Trans.
California,J. Geophys.Res.,94, 3949-3966, 1989. Geosci.RemoteSens.,GE23, 485-490, 1985.
Easton,R. L., The NavigationTechnologyProgram,Nay. J. Inst. Jordan,T. H., andJ. B. Minster,Beyondplatetectonics: Looking
Nay.,25, 107-120, 1978. at plate deformationwith spacegeodesy,in The Impact of
Elgered,G., Tropospheric wet pathdelaymeasurements, IEEE VLBI on Astrophysics and Geophysics,editedby M. J. Reid
Trans.Antennas Propag.,AP-30,502-505, 1982. andJ.M. Moran,pp. 341-350, 1988.
Elgered,G., J. L. Davis, T. A. Herring,and I. I. Shapiro, Kaplan,G. H., The IAU resolutionsof astronomical constants,
Geodesyby radiointerferometry: watervaporradiometryfor time scales and the fundamental reference frame, Circ. 163,
estimationof thewet delay,J. Geophys.Res.,in press,1991. U.S.Nay. Observ.,Washington,D.C., 1981.
Essen,L., and J. V. L. Parry,An atomicstandardof frequency Kaula, W. M., Theoryof SatelliteGeodesy,124 pp., Blaisdell,
and time interval, Nature, 176, 280-282, 1956. Waltham, Mass., 1966.
Feigl, K. L., R. W. King, T. A. Herring,andM. Rothacher,A Kellogg, J., and T. H. Dixon, Central and SouthAmericaGPS
schemefor reducingthe effect of selectiveavailabilityon geodesy-CASA Uno, Geophys. Res.Lett., 17, 195-198, 1990.
precise GPScarrierphasemeasurements EosTrans. Kellogg, J., T. H. Dixon, and R. Neilan, CASA: Central and
(abstract),
AGU, 71, 1268, 1990a. SouthAmerica GPS Geodesy-CASA,Eos Trans. AGU, 70,
Feigl, K. L., R. W. King, andT. H. Jordan,Geodeticmeasure- 649-656, 1989.
ment of tectonic deformation in the Santa Maria fold and Kellogg,J. N., et al., First GPS baselineresultsfrom the North
thrust belt, California, J. Geophys.Res., 95, 2679-2699, Andes,Geophys. Res.Lett.,17, 211-215, 1990.
1990b. King, N. E., P. Segall,andW. Prescott,Geodeticmeasurements
Fleigel,H. F., A. A. Feess,W. C. Layton,andN. W. Rhodus, near Parkfield,California,1959-1984, J. Geophys.Res.,92,
The GPS radiationforce model, paper presentedat First 2747-2766, 1987.
InternationalSymposium on PrecisePositioningWith Global King,R. W., andG. Blewitt,Presentcapabilities of GPSfor high
PositioningSystem,Int. Assoc.of Geod., Rockville,Md., precisionregionalsurveys,in Global PositioningSystem:An
1985. Overview,lnt. Assoc.Geod.Syrup.102, editedby Y. Bock and
Freymueller,J., andM. Golombek,Geometryandtreatmentof N. Leppard,pp. 24-39, Springer-Verlag, New York, 1990.
fiducialnetworks:Effect on GPS baselineprecisionin South King, R. W., E.G. Masters,C. Rizos, A. Stolz, and J. Collins,
America,Geophys. Res.Lett., 15, 1467-1469, 1988. SurveyingWith Global PositioningSystemGPS, 128 pp.,
Freymueller,J., and J. N. Kellogg, The extendedtracking Dummler, Bonn, 1985.
networkand indicationsof baselineprecisionandaccuracyin KornreichWolff S., T. H. Dixon, and J. Freymueller,The effect
theNorth Andes,Geophys. Res.Lett.,17, 207-210, 1990. of trackingnetwork configurationon GPS baselineestimates
Goad,C. C., Preciserelativepositiondetermination usingGlobal for the CASA Uno experiment,Geophys.Res. Lett., 17,
PositioningSystemcarrierphasemeasurements in a nondif- 647-650, 1990.
ferencemode, paper presentedat First InternationalSym- Lambeck, K., The Earth's Variable Rotation: Geophysical
posium on Precise PositioningWith Global Positioning Causesand Consequences, CambridgeUniversityPress,New
System,Int. Assoc.of Geod.,Rockville,Md., 1985. York, 1980.
Goad, C. C., and A. Mueller, An automatedprocedurefor Lambeck,K., GeophysicalGeodesy:The Slow Deformationsof
generating an optimumset of independent doubledifference theEarth, 718 pp., Clarendon,Oxford, 1988.
observables using Global PositioningSystemcarrierphase Lanyi,G. E., Tropospheric delayeffectsin radiointerferometry,
measurements,Man. Geod., 13, 365-369, 1988. JPL Prog.Rep.,42-78, 152-159, 1984.
Gordon,R. G., S. Stein,C. DeMets, and D. F. Argus,Statistical Larson,K., Precision,accuracyand tectonicsfrom the Global
testsfor closureof platemotioncircuits,Geophys. Res.Lett., PositioningSystem,Ph.D. thesis,269 pp., Univ. of Calif., San
14, 587-590, 1987. Diego, 1990a.
29, 2 / REVIEWSOF GEOPHYSICS Dixon:THE GLOBALPOSITIONINGSYSTEMß 275

Larson, K., Evaluationof GPS estimatesof relative positions Melosh, J., Nonlinear stresspropagationin the Earth's upper
from centralCalifornia,1986-1988, Geophys.Res.Lett., 17, mantle,J. Geophys. Res.,81, 5621-5632, 1976.
2433-2436, 1990b. Melosh,J., Shearstresson the baseof a lithosphericplate,Pure
Larsen, S., Global Positioning System measurementsof Appl.Geophys., 115, 429-439, 1977.
deformationin southernCalifornia,Ph.D. thesis,Calif. Inst. of Minster,J. B., T. H. Jordan,B. H. Hager, D.C. Agnew, and L.
Technol., 1990. H. Royden,Implicationsof precisepositioning,in Geodesyin
Lawson, C. L., and R. J. Hanson, Solving Least Squares the Year 2000, editedby J. B. Rundle, pp. 23-45, National
Problems,340 pp., Prentice-Hall,EnglewoodCliffs, N.J., AcademyPress,Washington, D.C., 1990.
1974. Neilan, R., T. Dixon, T. Meehan, W. Melbourne, J. Scheid, J.
Leick, A., GPSSatelliteSurveying,352 pp., Wiley-Interscience, Kellogg,and J. Stowell,Operationalaspectsof CASA UNO
New York, 1990. '88raThe first largescaleinternationalGPS geodeticnetwork,
Li, V. C., and J. R. Rice, Crustaldeformationin greatCalifornia IEEE Trans. Inst. Measur., 38, 648-651, 1989.
earthquakecycles,J. Geophys.Res., 92, 11,533-11,551, Neilan,R. E., W. G. Melbourne,andG. L. Mader,The develop-
1987. mentof a globalGPS trackingsystemin supportof spaceand
Li, V. C., and H. S. Lim, Modelling surfacedeformationat ground-based GPS programs,in Global PositioningSystem:
complex strike-slipboundaries,J. Geophys.Res., 93, An Overview,lnt. Assoc.Geod.Syrup.,vol. 102, editedby Y.
7943-7954, 1988. Bock and N. Leppard,pp. 165-178, Springer-Verlag, New
Lichten,S. M., High accuracy
GlobalPositioning
Systemorbit York, 1990.
determination: in Global Positioning Nut, A., H. Ron, and O. Scotti, Kinematics and mechanics of
Progressandprospects,
System:An Overview,lnt. Assoc.Geod. Syrup.,vol. 102, tectonicblock rotations,in Slow Deformationand Transmis-
editedby Y. Bock andN. Leppard,pp. 146-164, Springer- sion of Stressin the Earth, Geophys.Monogr. Ser., vol. 49,
Verlag,New York, 1990a. edited by S.C. Cohen and P. Vanicek, pp. 31-46, AGU,
Lichten,S. M., Estimationand filtering for high precisionGPS Washington, D.C., 1989.
positioningapplications,
Man.Geod.,15, 159-176,1990b. Preston, R. A., R. Ergas, H. Hinteregger, C. Knight, D.
Lichten, S. M., and J. S. Border, Strategiesfor high precision Robertson,I. Shapiro,A. Whitney, A. Rogers,and T. Clark,
Global PositioningSystemorbit determination, J. Geophys. Interferometric observationsof an artificial satellite, Science,
Res., 92, 12,751-12,762, 1987. 178, 407-409, 1972.
Lichten,S. M., and W. I. Bettiger,Demonstration of sub-meter Rabbel,W., andH. Schuh,The influenceof atmospheric loading
GPS orbit determinationand 1.5 partsin 108 three dimen- on VLBI experiments, J. Geophys.,59,164-170, 1986.
sionalbaselineaccuracy, Bull. Geod.,63, 167-189, 1989. Reeber,E. E., and J. R. Swope,On the correlationbetweentotal
Lindlohr, W., and D. Wells, GPS designusing undifferenced precipitablewater in a colurn and absolutehumidityat the
carrierbeat phaseobservations, Man. Geod.,10, 255-295, surface,J. Appl. Meteorol.,11, 1322-1325, 1972.
1985. Remondi, B. W., Global PositioningSystem carrier phase:
Lorenzetti,E., and T. E. Tullis, Geodeticpredictionsof a strike Descriptionanduse,Bull. Geod.,59, 361-377, 1985.
slip faultmodel:Implications for intermediate andshortterm Rizos, C., and A. Stolz, Forcemodellingfor GPS satelliteorbits,
earthquake prediction,J. Geophys. Res.,94, 12,343-12,361, paperpresentedat First InternationalSymposiumon Precise
1989. PositioningWith Global PositioningSystem,Int. Assoc.of
Luyendyk,B. P., andJ. S. Homafius,Neogenecrustalrotation, Geod., Rockville, Md., 1985.
fault slip and basindevelopmentin southernCalifornia,in Robinson,S. E., The profile algorithmfor microwavedelay
CenozoicBasin Developmentof Coastal California, Rubey estimationfrom water vapor radiometerdata,Radio Sci., 23,
vol. VI, edited by R. V. Ingersoll and W. G. Ernst, pp. 401-408, 1988.
259-283, Prentice-Hall,EnglewoodCliffs, N.J., 1987. Rocken, C., J. Johnson,R. E. Neilan, M. Cerezo, J. Jordan, M.
MacDoran, P., Satellite emission radio interferometricEarth Falls, L. Nelson, R. Ware, and M. Hayes, The measurement
surveying:SERIES GPS geodeticsystems,Bull. Geod.,53, of atmosphericwater vapor: Radiometercomparisonand
117-138, 1979. spatial variation, IEEE Trans. Geosci. Remote Sens., in
Malys, S., and P. A. Jensen,Geodeticpoint positioning with press,1991.
carrier beat phase data from the CASA Uno experiment, Rocken, C., and C. M. Meettens, GPS antenna and receiver
Geophys. Res.Lett.,17, 651-654, 1990. tests: Multipath reduction and mixed receiver baselines,
McCaskill,T. B., andJ. A. Buisson,On-orbitfrequencystabilityof Proceedingsof Fifth International Symposiumof Satellite
NAVSTAR GPS clocks and the importanceof frequency Positioning,Las Cruces,New Mexico, pp. 375-385, 1989.
stability to precise positioning,paper presentedat First Rocken,C., and C. M. Meertens,MonitoringS/A and its effect
InternationalSymposiumon PrecisePositioning With Global on GPS geodesy(abstract),EosTrans.AGU, 71, 1268, 1990.
Positioning System, Int. Assoc. of Geod.,Rockville,Md., 1985. Rosenblatt,D., Y. Bock, J.-B. Minster, and D.C. Agnew,The
Melbourne,W. G., The casefor rangingin GPS-basedgeodetic impact of recent SelectiveAvailability on precise relative
systems, FirstInternationalSymposium on PrecisePositioning positioningwith GPS (abstract),Eos Trans.AGU, 71, 1268,
With Global PositioningSystem, Int. Assoc. of Geod., 1990.
Rockville, Md., 1985. Rundle,J. B., An approachto modellingpresent-daydeformation
Melbourne,W. G., GPS-basedgeodesytracking:Technology in southernCalifornia, J. Geophys.Res., 91, 1947-1959,
prospects for theyear2000,in Geodesy in theYear2000,edited 1986.
by J. Rundle, pp. 124-176, National Academy Press, Savage,J. C., A dislocationmodel of strain accumulationand
Washington,D.C., 1990. release at a subductionzone, J. Geophys. Res., 88,
Melbourne,W. G., R. Anderie,M. Feissel,R. King, D. McCar- 4984-4996, 1983.
thy, D. Smith, B. Tapley, and R. Vicente, Projectmerit Savage,J. C., Equivalentstrike-slipearthquakecycles in half
standards,Circ. 167, U.S. Nay. Observ.,Washington,D.C., space and lithosphere-asthenosphere Earth models, J.
1983. Geophys.Res.,95, 4873-4879, 1990.
Melbourne,W. G., R. Anderie,M. Feissel,R. King, D. McCarthy, Savage,J. C., and R. O. Burford, Geodeticdeterminationof
D. Smith,B. Tapley,R. Vicente,Projectmeritstandards, Circ. relativeplate motionin centralCalifornia,J. Geophys.Res.,
167, Update1, U.S. Nay. Observ.,Washington, D.C., 1985. 78, 832-845, 1973.
276 ß Dixon: THE GLOBAL POSITIONING SYSTEM 29, 2 / REVIEWSOF GEOPHYSICS

Savage,J. C., andW. H. Prescott,Precisionof Geodolitedistance Tralli,D. M., andT. H. Dixon,A fewpartsin l0s geodetic
measurements, J. Geophys.Res.,78, 6001-6008, 1973. baselinerepeatabilityin the Gulf of Califomiausingthe
Savage,J. C., and W. H. Prescott,Asthenosphere readjustment GlobalPositioning
System,Geophys.Res.Lett., 15, 353-356,
and the earthquakecycle,J. Geophys.Res.,83, 3369-3376, 1988.
1978. Tralli, D. M., and S. M. Lichten, Stochastic estimation of
Schaffrin, B., and E. W. Grafarend, Generatingclassesof tropospheric path delays in Global PositioningSystem
equivalentlinear modelsof nuisanceparameterelimination: geodeticmeasurements, Bull. Geod.,64, 127-159, 1990.
Applicationto GPS observations,Man. Geod.,11, 262-271, Tralli, D. M., T. H. Dixon, andS. A. Stephens, The effectof wet
1986. tropospheric pathdelayson estimationof geodeticbaselinesin
Schutz,B. E., C. S. Ho, P. A.M. Abusali,and B. D. Tapley, theGulf of CalifomiausingtheGlobalPositioning System,J.
CASA Uno orbit and baselineexperiments,Geophys.Res. Geophys. Res.,93, 6545-6557, 1988.
Lett., 17, 643-646, 1990. Turcotte,D. L., J. Y. Liu, and F. H. Kulhawy,The role of an
Seidelman,P. K., 1980 IAU theoryof nutation:The final report intracrustalasthenosphere on thebehaviorof majorstrike-slip
of the IAU workinggroupon nutation,CelestialMech., 27, faults,J. Geophys. Res.,89, 5801-5816, 1984.
79-106, 1982. Wahr,J. M., Theforcednutations
of anelliptical,rotating,elastic
Sibson, R. H., Continental fault structure and the shallow and oceanlessEarth, Geophys. J. R. Astron. Soc., 64,
earthquakesource,J. Geol. Soc. London, 140, 741-767, 705-727, 1981.
1983. Wahr, J. M., The Earth's rotation rate, Ann. Rev. Earth Planet.
Slade, M., and C. Yoder, 1960 Chile: New estimateof polar Sci., 16, 231-249, 1988.
motion excitation, Geophys. Res. Lett., 16, 1193-1196, Ware, R. H., C. Rocken, and J. B. Snider, Experimental
1989. verification
of improvedGPS-measuredbaselinerepeatability
Smith, D. E., D.C. Christodoulidis,R. Kolenkiewicz, P. Dunn, using water vapor radiometer corrections,IEEE Trans.
S. Klosko,M. Torence,S. Fricke,and S. Blackwell,A global Geosci.RemoteSens.,GE-23, 467-473, 1985.
geodeticreferenceframefrom LAGEOSranging,J. Geophys. Webb, F., Geodetic measurement of deformation in the offshore
Res., 90, 9221-9233, 1985. of southemCalifomia, Ph.D. thesis,Calif. Inst. of Technol.,
Sovers,O. J., and J. Border, Observationmodel and parameter Pasadena,1990.
partials for the JPL GPS geodetic modeling software Weldon, R., and E. Humphreys,A kinematicmodel of southem
"GPSOMC," JPL Publ., 87-21, Rev. 2, 1990. California, Tectonics,5, 33-48, 1986.
Spence,D. A., and D. L. Turcotte,Viscoelasticrelaxationof Wu, S.C., W. I. Bertiger,andJ. T. Wu, Minimizingselective
cyclic displacement
on the San Andreasfault, Proc. R. Soc. availability error on Topex GPS measurements, paper
London, Ser. A, 365, 121-144, 1979. presented at American Institute of Aeronautics and
Spilker,J. J., GPS signalstructureandperformancecharacteris- Astronautics
Conference,Portland,Oreg.,1990.
tics, Nay. J. Inst. Nay., 25, 121-146, 1978. Wubbena,G., Softwaredevelopments
for geodeticpositioning
Tapley,B. D., B. E. Schutz,andR. J. Eanes,Stationcoordinates, with GPS using TI-4100 code and cartier measurements,
baselinesand earth rotation from LAGEOS laser ranging: paperpresentedat First InternationalSymposiumon Precise
1976-1984,J. Geophys. Res.,90, 9235-9248, 1985. PositioningWith Global PositioningSystem,lnt. Assoc.of
Thatcher,W., Non-linearstrainbuildupand the earthquakecycle Geod., Rockville, Md., 1985.
on the San Andreasfault, J. Geophys.Res.,88, 5893-5902, Wyatt, F., Displacementsof surface monuments:Vertical
1983. motion,J. Geophys. Res.,94, 1655-1664, 1989.
Thatcher,W., and J. B. Rundle, A viscoelasticcouplingmodel Young, L. E., R. Neilan, N and F. Bletzacker, GPS satellite
for the cyclic deformation due to periodically repeated multipath:An experimental investigation,
paperpresented
at
earthquakesat subductionzones, J. Geophys.Res., 89, First IntemationalSymposium on PrecisePositioning
With
7631-7640, 1984. Global PositioningSystem,lnt. Assoc.of Geod., Rockville,
Thompson,A. R., J. M. Moran, and G. W. Swenson,Inter- Md., 1985.
ferometry and Synthesisin Radio Astronomy,534 pp.,
Wiley-Interscience,New York, 1986.
Thornton,
C. L., andG. J. Bierman,
UDUT covariance
factoriza-
tion for Kalman filtering, in Control and DynamicSystems, T. H. Dixon,JetPropulsion
Laboratory,
Divisionof Earth
Advancesin TheoryandApplications,vol. 16, editedby C. T. and PlanetarySciences,
4800 Oak Grove Drive,Pasadena,
Leondes,pp. 177-248, Academic,Troy, Mo., 1980. CA 91109.