The Geophysics of Venus

Sean C. Solomon

Citation: 46, (1993); doi: 10.1063/1.881359

View online: http://dx.doi.org/10.1063/1.881359
View Table of Contents: http://physicstoday.scitation.org/toc/pto/46/7
Published by the American Institute of Physics
THE GEOPHYSICS OF VENUS
Impact craters on Venus's surface are surprisingly unscarred
by tectonic deformation and volcanic flows. Data from the
Magellan mission may help discriminate between catastrophic
and evolutionary explanations of the planet's recent placidity.

Sean C. Solomon

Venus is the planet most similar to the Earth in mass, of oceanic plates at convergence zones and their renewal
radius and solar distance. Current theories of the early at midocean ridges serve to resurface the Earth's ocean
evolution of the solar system suggest that Earth and floor continuously, replacing the entire seafloor in about
Venus each formed by the accretion of planetesimals— 108 years. The Earth's continents, underlain by thick
small rocky or rock-metal objects—that collectively con- buoyant crust, do not participate significantly in that
stituted a well-mixed sample of material condensed from recycling and thus preserve rocks as old as 4 x 109 years
the inner solar nebula. The bulk compositions of the two as well as a long and complex history of deformation,
planets should thus be similar. The rates of internal igneous activity, erosion and sedimentation. To what
heat generation and the energy available to drive interior extent do the large-scale deformational, or tectonic, pat-
convection should also be similar. An important differ- terns of Venus, with a similar internal heat budget but
ence between the two planets, however, is the character with very different surface conditions, resemble those of
of their atmospheres. The mass of the dominantly CO2 the Earth?
atmosphere of Venus is two orders of magnitude greater, That question was among several that motivated the
as a fraction of planet mass, than that of Earth's atmo- Magellan mission to Venus. A thick cloud cover precludes
sphere, and the surface temperature is 450 K higher, a optical studies of the Venusian surface from Earth or
consequence of continuous global cloud cover and a run- from orbit, but a series of Earth-based and orbital radar
away greenhouse effect. The mass of H2O in a vertical experiments dating back nearly three decades demon-
column of unit area is four to five orders of magnitude strated that radar imaging could yield important infor-
less for Venus's atmosphere than for the atmosphere and mation on the planet's geology. The Magellan mission
hydrosphere on Earth. As a result, the surface of Venus was designed to image the surface by illuminating it with
lacks a water cycle, and the processes of weathering, radar and measuring the reflected brightness, at a hori-
erosion and sediment transport that dominate terrestrial zontal resolution of 100-300 meters, and to map the
landforms are comparatively unimportant. surface elevation at a vertical resolution of about 80 m
On Earth, the surface manifestation of interior con- and a horizontal resolution of about 10 km. A single
vection is the steady relative motion of the tectonic plates, radar system accomplished both objectives.1 For imaging,
which separate at midocean ridges, converge at deep-sea a high-gain antenna looked sideways in a "synthetic
trenches and active mountain belts, and slip horizontally aperture" mode: As the spacecraft moved along its
orbit, multiple returns of the transmitted pulsed signal
past one another along great fault zones. The recycling from the same spot on the surface were saved and
summed to simulate a larger antenna aperture, thus
Sean Solomon is the director of the Department of Terrestrial yielding a better resolution. For mapping the surface
Magnetism of the Carnegie Institution of Washington and a elevation, a second low-gain antenna transmitted ver-
member of the radar investigation group for the Magellan tically downward and measured the time it took for
mission. the signal to be reflected back.

1993 American Insfirure of Physics

48 PHY5IC5 TODAY JULY 1993
 Global view of the surface of
Venus. To produce this
image, Magellan radar
images taken over a two-year
period were digitally
combined into a mosaic that
was then projected onto a
simulated globe centered at
longitude 1 80° E. The use of
false color enhances
small-scale features. The
bright features that cross from
lower left to upper right are
the highlands, coronas and
rift valleys of Aphrodite
Terra. Figure 1

The Magellan spacecraft was placed into a nearly smaller terrestrial planets, is in itself not remarkable for
polar orbit about Venus on 10 August 1990 and began a planet with a hot, dynamic interior, like the Earth,
mapping about one month later. In each successive albeit one on which erosion is unimportant.
"cycle" of 243 Earth days Venus turned once on its axis More remarkable is that the spatial distribution of
beneath the plane of the spacecraft orbit. By the end of craters of all sizes is indistinguishable from a random
the first three cycles in September of last year, Magellan population, and that most of them have not been signifi-
had imaged 98% of the Venus surface (see figure 1). At cantly modified by tectonic deformation or by volcanic
the start of the fourth cycle the elevation of the spacecraft flows external to the crater rim,23 despite evidence from
orbit at periapsis—the point of closest approach—was Magellan images that volcanic6 and tectonic6 features are
lowered to 180 km, and at regular intervals the high-gain widespread on Venus. (An example of such an impact
antenna was pointed toward the Earth through periapsis crater is Isabella, shown in Figure 2.) One interpretation
passage to permit the measurement of spacecraft accel- of these characteristics, championed by Gerald Schaber
erations produced by the gravitational field of Venus. and colleagues,3 is that most of the surface dates from
the end of a global resurfacing event that ceased about
Surface age 5 x 108 years ago, and that volcanic and tectonic activity
Information on the age of the surface provides a critical since then has been at much lower levels. A contrasting
context for interpreting the geological record of a planet. view, advanced by Roger Phillips and coworkers,2 is that
For Venus, in the absence of returned rock samples whose the Venusian surface exhibits a spectrum of ages. (Fig-
ages could be determined radiometrically by isotope geo- ure 3 compares the surface age distributions expected
chemistry, the only measure of surface age is the density under these two scenarios.) This view is supported by
of impact craters.2'3 As expected on the basis of earlier the observations that modified craters tend to be located
data and theoretical models, Magellan revealed that im- in areas of low crater density and that low crater density
pact craters smaller than about 30 km in diameter are appears to be correlated with increased radar backscat-
deficient relative to larger craters on Venus, because of ter—an indication of high topography and high rough-
the severe decrease in the kinetic energy of small mete- ness, both thought to be signatures of comparative geo-
oroids during transit through the dense Venusian atmo- logical youth. The paucity of small craters, however,
sphere. The areal density of craters larger than 30 km prevents one from using crater density to determine with
in diameter, together with estimates of cratering rate confidence the relative ages of geological units, as has
scaled from the Earth and Moon or taken from the known been done for the solid planets and satellites lacking a
population of Venus-crossing asteroids,4 indicates an av- significant atmosphere.
erage surface age of about 5 x 108 years, or 10% of the
age of the solar system. This age, greater than that of Tectonics
the Earth's ocean floor but less than the radiometric age Magellan has revealed that tectonic features of a wide
of the Earth's continental rocks or the surface age of the variety of styles and spatial scales are present over most

PHY5IC5 TODAY JULY 1993 49

 Impact crater Isabella in a Magellan radar image. Isabella
(centered near 30° S, 204° E) is the second largest crater on
Venus at approximately 175 km in diameter. Radar-bright,
rough-textured ejecta extending up to two crater radii from
the crater center and the remarkable bright flow features
extending hundreds of kilometers from the crater walls are
thought to date from the impact event3 and have not been
subsequently modified to any significant degree by
deformation or volcanism exterior to the crater. The
generally radar-dark, smooth floor of the crater may contain
younger volcanic deposits that were deformed after
emplacement to produce the concentric pattern of faults.
This and all subsequent radar images are in sinusoidal
equal-area projection; north is up, and the radar
illumination direction is from the left. The incidence angle
of the radar is about 33° for this image and in general is a
function of latitude.1 Figure 2

of the Venusian surface.6 Deformation is manifested in kilometers across. Rather, tectonic strain on Venus typi-
areally distributed strain of modest magnitude, accom- cally involves deformation distributed across broad zones
modated by families of faults and folds, spaced at a few tens to hundreds of kilometers wide, separated by com-
to a few tens of kilometers and often coherent over paratively undeformed blocks having dimensions of hun-
hundreds of kilometers, in many volcanic plains. Defor- dreds of kilometers. These characteristics are shared
mation is also commonly evident as zones of more intense with actively deforming continental regions on Earth.
horizontal shortening or extension of the crust. Ridge This similarity in tectonic styles does not imply that the
belts and mountain belts, marked by many closely spaced crust on Venus is similar in composition to that of the
folds and thrust faults, represent successive degrees of Earth's continents—only that both crusts display a
local shortening and crustal thickening. Ridge belts have broadly similar response to tectonic stress. In fact,
characteristic widths and spacings of hundreds of kilo- chemical analyses of surface samples to accuracies of a
meters and up to 1 km of relief. Mountain belts are few percent, made on site during several Soviet lander
comparable in relief and horizontal dimensions to those missions using x-ray fluorescence and gamma-ray spec-
on Earth, and likewise they often show evidence of having troscopy, indicate compositions generally nearer to that
undergone lateral extension both during and after active of Earth's oceanic crust than to terrestrial continental
crustal compression (because elevated terrain tends to material.8 On Earth, the continental plates are weaker
spread in response to gravitational stress). than oceanic plates because of the greater thickness of
Venus displays two principal geometric variations on crustal material, which at a given temperature deforms
large-scale extension: quasicircular corona structures 75- at significantly higher rates than does the underlying
2600 km in diameter and broad rises with linear rift mantle. On Venus, the high surface temperature is
zones hundreds to thousands of kilometers in length. The expected to lead to ductile behavior at significantly shal-
rift zones (see figure 4) have dimensions and relief similar lower levels in the middle to lower crust than on Earth
to intracontinental rift zones on Earth, such as the East and is probably responsible for the rich spectrum of
African or Rio Grande rifts, but the coronas (see figure deformational features.
5) have no evident terrestrial counterpart. Both are sites
of significant volcanic flux, but horizontal displacements Gravity anomalies
in the rift zones (inferred from rift valley geometry and The deviation of a planet's gravity field from that ex-
offsets of older features) may be limited to only a few pected for a rotating fluid body of radially varying density
tens of kilometers. Few large-offset strike-slip faults like is known as the gravity anomaly. Its correlation with
the San Andreas Fault are observed on Venus, but limited the topography provides clues to the internal density
local horizontal shear has been accommodated across structure of the planet. The mass of the topography
many zones of crustal stretching or shortening. Many exerts a pressure on the underlying rock. In the theory
elevated areas are characterized by extremely complex, of isostasy, there is a depth at which pressure is assumed
intersecting patterns of tectonic features at a range of to be uniform, because over geological time scales the
scales (see figure 6); these regions record multiple stages rock behaves as a fluid. Thus the excess pressure ex-
of strain of diverse geometries. Several large-scale tec- erted, for example, by a highland area must be "compen-
tonic features have topographic slopes in excess of 20°-30° sated" by a zone of lower-than-average density under-
over a 10-km horizontal scale. Numerical models simu- neath. Because crustal material is less dense than
lating the relaxation of such steep slopes by ductile flow underlying mantle material, a common form of such
in the middle to lower crust7 suggest that such regions compensation is a crust extending deeper into the mantle.
were tectonically active within the last 107 years. Regions with higher temperatures are also of lower den-
In general the preserved record of global tectonics of sity and can help to compensate. Isostasy is basically a
Venus does not resemble oceanic plate tectonics on Earth, static theory; in addition, there may be contributions to
wherein large rigid plates are separated by narrow zones gravity anomalies from convective stresses. An upwelling
of active deformation generally no more than a few region will have positive contributions to gravity from

50 PHY5ICS TODAY JULY 1993

 Surface age distributions on Venus under two scenarios. In
the catastrophic resurfacing scenario (blue histogram), most
of the surface is about 5 x 108 years old, but small areas are
younger. In an alternative, more gradual resurfacing
scenario2 (black histogram), there is a more widespread
distribution of surface ages, from 0 to approximately 109 tr
years. Both scenarios are broadly consistent with the U
distribution and states of preservation of impact craters on en
u.
Venus. (Courtesy of Roger Phillips, Washington University, O
O
St. Louis.) Figure 3

0.0-
0.4 0.6 0.8 1.0
AGE (10 9 years)

the elevated surface and a negative contribution from the permost mantle) is a layer of low viscosity that does not
lower density in the hot rising column of mantle. The transmit stresses from the bulk of the mantle underneath.
net gravity anomaly from dynamic processes in the man- (This low-strength zone probably accounts for the hori-
tle is a function of the radial (and lateral) variation in zontal mobility of the Earth's plates.) Researchers have
the viscosity of mantle material. interpreted the large compensation depths on Venus to
Measurements of the anomalies in the gravitational mean that Venus lacks such an upper-mantle low-viscos-
field made by the Pioneer Venus orbiter, and now con- ity zone,11 so that convective motions deep in the upper
firmed by the first eight months' worth of Magellan data, mantle are able to couple into the overlying lithosphere12
indicate that the interior dynamics of Venus differs from and cause long-wavelength vertical distortions of the
that of the Earth in important respects. In contrast to surface. The absence of a low-viscosity zone restricts
the situation on Earth, topography and gravity on Venus horizontal mobility and probably accounts for the large-
are strongly correlated on scales (or "wavelengths") of scale coherence of surface strains on Venus.6
several hundred to several thousand kilometers.9 Fur- The large ratios of long-wavelength gravity anomaly
ther, many major features have10a large ratio of gravity to long-wavelength topographic relief have been taken as
anomaly to topographic relief, indicating that topo- evidence that the crust and upper mantle of Venus may
graphic variations are compensated by interior density generally be stronger than one would infer from simple
variations at depths of up to several hundred kilometers. thermal models that extrapolate Earth's interior heat
Density variations associated with convective upwelling flow to Venus's 450-K-greater surface temperature.13
and downwelling deep in the upper mantle must be There is other evidence in support of this view. The
involved. On Earth, the compensation depths are much measured depths of impact craters are generally too great
smaller—tens of kilometers—and imply that beneath the to be consistent with significant relaxation of relief by
lithosphere (the mechanically strong outer layer that flow of crustal material,14 such as might be expected if
generally includes the crust and some thickness of up- temperatures in the lower crust were sufficiently great

Northern Beta Regio, a rifted highland region, in a Mag-

ellan radar image. The image is centered at about 33°
N, 283° E, and is 750 km wide. The radar-bright areas
consist of a fabric of closely spaced families of faults and
folds of various trends. This terrain has been extended in
the east-west direction, leading to the formation of north-
south-trending faults and of the steep-sided rift valley
(with east-facing walls nearly in radar shadow) visible in
the center of the image. The rift-related faults splay to
the northwest and northeast in the northern part of the re-
gion shown. Dark patches are smooth and are inferred
to be volcanic deposits overlying the older, bright ter-
rain. Figure 4

PHYSICS TODAY JULY 1993 51

 plate tectonics in which the entire lithosphere sinks into
the underlying mantle and is replaced by hot new mate-
rial—operates episodically on Venus and that for the last
5 x 108 years the lithosphere has been cooling and me-
chanically stable. He argues that a cool and thick litho-
sphere is in better agreement with the large values of
the flexural rigidity of the lithosphere and the large ratios
of long-wavelength gravity anomaly to long-wavelength
topographic relief than is a lithosphere in steady-state
conductive equilibrium with the long-term average heat
flow from the interior.
Marc Parmentier and Paul Hess16 have suggested
that melt extraction can decrease the density of the
lithospheric mantle and thus initially stabilize the float-
ing lithosphere; subsequent cooling may increase the
mantle density enough to make the lithosphere unstable.
Global overturn of this unstable layer would be followed
by widespread partial melting of the upper mantle, global
volcanic resurfacing and the gradual development of a
new buoyant layer of lithospheric mantle that gives rise
to another extended period of lithospheric stability. In
one-dimensional models of this process,16 the time be-
tween lithospheric instability events is 3-5 x 108 years.
One problem with these one-dimensional models is that
The Idem-Kuva corona structure in a they assume global synchroneity. Even if such instability
Magellan radar image. The corona, centered mechanisms are operative, it is likely that different parts
at 25° N, 358° E, is about 230 km in of the planet will be at different stages in the stabilization
diameter. Corona structures are distinguished and destabilization sequence and that such regional dif-
by an annulus of deformed terrain and ferences will smooth out global-scale temporal variations.
frequently by an elevated interior.6 In other catastrophic scenarios, time-variable mantle
Radar-bright volcanic flows emanate from convection rather than lithospheric instability serves as
topographic highs of more than 1 km relief in the mechanism for global resurfacing. An early effort by
the eastern and western portions of this Jafar Arkani-Hamed and Nafi Toksbz17 to simulate time-
structure. Figure 5 variable, three-dimensional mantle convection on Venus
led to models in which the characteristic flow speed and
mantle heat flux showed large oscillations (by factors of
to permit high rates of ductile strain. Estimates of the 2-10) at intervals of 1-2 x 108 years. In improved calcu-
flexural rigidity of the lithosphere—made from the shapes lations with better spatial resolution, mantle convection,
of topographic profiles across the curved troughs marking while still time varying, shows significantly lesser fluc-
the margins of several corona structures—are comparable tuations, by one order of magnitude.18
to the rigidity of oceanic lithosphere at deep-sea trenches Volker Steinbach and David Yuen19 have drawn at-
on Earth.15 This result is surprising, because researchers tention to the role that pressure-induced changes in
think the base of the mechanically strong lithosphere is upper-mantle mineralogy may play in governing the ra-
defined by the temperature marking the onset of signifi- dial character of mantle convection in the large terrestrial
cant ductile flow over geological time scales. This tem- planets. They argue that such phase changes cause
perature, about 1000 K on Earth, would normally be separate convecting layers to form in the upper and lower
expected to occur at shallower levels on Venus, because mantle, but that as the Rayleigh number of the mantle
of its hotter surface. One possible explanation is that (which measures the strength of thermal buoyancy in
the topographic profiles are being interpreted incorrectly: driving flow relative to the tendencies for viscosity to
Ductile flow accompanying gravitational relaxation of resist flow and for heat conduction to reduce buoyancy)
relief, for instance, can deform the surface in a manner decreases in response to the cooling of the central metallic
that mimics elastic plate flexure.7 If real, a stronger- core, whole-mantle convection tends to become favored
than-expected lithosphere on Venus could result from over layered convection. If Venus cooled more rapidly
a lesser rate of heat loss (per planet mass) than on than Earth early in its history, it may now have a lesser
Earth, or it could indicate that under extremely anhy- mantle Rayleigh number and have gone through a tran-
drous conditions crustal and mantle rocks are signifi- sition to whole-mantle convection, while the Earth may
cantly more resistant to flow than they are in the still be characterized by layered mantle flow. Such a
Earth's crust and mantle. transition would have been accompanied by overturn of
the upper mantle, the upward transport of significant
Catastrophic resurfacing hypotheses heat, the generation of substantial melt and probably
Hypotheses advanced to explain the tectonics of Venus, global volcanic resurfacing.19
and in particular the resurfacing history, fall into two
categories: catastrophic and evolutionary. Those in the Evolutionary hypotheses
former category explain the small fraction of impact The second category of resurfacing scenarios involves only
craters modified by exterior volcanism or significant de- a gradual lessening of volcanic or tectonic activity rather
formation as the result of a global-scale resurfacing event than one or more global catastrophes. There are several
that ended about 5 x 108 years ago.3 In one class of such motivations for such evolutionary scenarios. While the
scenarios, catastrophic resurfacing occurs because of an known volcanic histories of Earth20 and Mars21 include
instability in the lithosphere. Donald Turcotte13 has periods of greater-than-average activity and presumably
proposed that global lithospheric overturn—a version of therefore time-variable mantle convective flux, no other

52 PHYSICS TODAY JULY 1993

 Heavily deformed terrain in an upland area known as
Tellus Regio. This Magellan radar image, 80 km wide, is
centered on 39.5° N, 85.7° E. The terrain is dominated
by extensional fault structures trending north to north-
west. Several northeast-trending topographic ridges may
be remnants of earlier products of horizontal shortening.
This type of complexly deformed terrain is typical of
many highland regions on Venus. Figure 6

terrestrial planet displays evidence for rapid and com-

plete global resurfacing during the past 4 x 109 years.
Furthermore, global resurfacing by large-scale overturn
of the lithosphere or upper mantle should40 result in effi-
cient outgassing of the upper mantle; the Ar abundance
of the Venusian atmosphere—which is a factor of 4 less
than that of the Earth's atmosphere as a fraction of planet
mass—suggests that any such widespread outgassing
should have been restricted to times significantly earlier
than 5 x 108 years ago.22 Phillips and coworkers2 have
described a model in which volcanic resurfacing occurs
episodically in small patches a few hundred kilometers
in extent, with a characteristic time of order 105 years
between events. They argue that such a model, involving
a gradual decline in the volcanic flux of the planet, is as
consistent with the characteristics of impact craters on
Venus as are the catastrophic scenarios.
I have recently advanced the hypothesis that, at least
in the geologically recent past, the primary resurfacing 5 x 108 years ago.
mechanism on Venus has been tectonic deformation This tectonic resurfacing hypothesis implies a rapid
rather than volcanism.23 An important difference be- change on a nearly planetary scale from high rates of
tween Venus and all of the other terrestrial planets is resurfacing to low rates, as is called for by the cata-
its high surface temperature. Characteristic time scales strophic resurfacing model,3 but it involves no true ca-
for ductile deformation of crustal and mantle material tastrophe—and certainly no global outpouring of magma.
are known to vary exponentially with reciprocal tempera- If Venus were laterally uniform in both crustal thickness
ture; so, for a given thermal gradient and stress field, and heat flow, then the transition would be expected to
high rates of ductile flow are expected at much shallower have occurred with global synchroneity. While the sur-
levels on Venus than on other terrestrial planets. Direct face today is not strictly uniform, more than 80% of it
coupling of mantle convective stresses12 should give rise stands at elevations within 1 km of the mean value. To
to lithospheric strains that are broadly coherent over the extent that regions at similar elevations have similar
large regions. For a sufficiently weak lower crust, such crustal thicknesses and thermal structures and that the
stresses should also cause high rates of lower-crustal principle of isostasy applies, the surface of Venus may
deformation and thus high rates of surface strain. not depart greatly from uniformity, and an apparently
Therefore if the surface temperature prior to the era catastrophic, nearly global transition is not a bad first
of Venus history now preserved was comparable to that approximation.
at present, the higher heat flow—associated with early Departures from global synchroneity are also to be
planetary cooling, enhanced radiogenic heat production expected. In particular, highland regions, whether they
and a mantle convective vigor at least that of the pre- owe their elevations primarily to greater crustal thickness
sent—should have led to geologically rapid rates of crustal or to enhanced temperatures at depth, should persist as
deformation over most, and perhaps all, of the surface. regions of high strain rate long after the rate of defor-
Such an era would have been characterized by a nearly mation in lowland plains regions has dropped to modest
global extent of intensely deformed terrain (such as that levels. Lowlands should thus be preferred sites for the
in figure 6) and few impact craters undeformed enough preservation of relatively undeformed volcanic deposits
to be recognizable from surface images. At some point and impact craters, as is observed.2 In light of the
in the evolution of Venus under this scenario, however, tectonic resurfacing hypothesis the unusual cratering
heat flow declined to levels sufficiently low that the record on Venus is thus seen to be a consequence pri-
ductile strength of the lower crust increased significantly, marily of the atmospheric greenhouse and its effect on
so that rates of deformation that had been high on the surface temperature and the rheology of the crust.
geological time scales became much lower. Following No catastrophic internal event or recent episode of ex-
such a transition, which may have been abrupt in the tensive interior outgassing is called for. By this hypothe-
geological record because of the exponential dependence sis the resurfacing history should nonetheless contain
of strain rate on temperature, both volcanic deposits and elements of both the "catastrophic" and "gradual" sce-
impact craters would have persisted for long intervals narios for crater removal, with approximately synchro-
with at most modest deformation. The characteristics of nous stabilization of lithosphere beneath plains regions
the Venus surface as revealed by Magellan are consistent of average elevation and more recent tectonic activity
with this hypothesis if such a transition occurred about
in highlands.

PHYSIC5 TODAY JULY 1993 53

 32.5 Tracking data from the Pioneer Venus
26.0 orbiter and Magellan used to determine the
gravity field of Venus. Each plot displays
19.5 measurements of the line-of-sight velocity of
V. the spacecraft, obtained from Doppler shifts
Pioneer Venus orbit 435
13.0 ~ 13 February 1980 in the transmitter carrier frequency, after
removing the velocity signal of the
6.5 best-fitting orbital arc. The two plots
compare orbits at nearly the same longitude
0
' . " i • • . • • • * " " • - ' • * * ' • ' ' - ' . • • • • •
(87° E) and spacecraft periapsis altitude
(175 km). The superior signal-to-noise ratio
-6.5
of the Magellan data is due to the
13.0 _ Magellan orbit 6336 ',-•• significantly reduced influence of plasma
2 December 1992 effects in the Venus ionosphere at
6.5 Magellan's higher carrier frequency (8.4
GHz versus Pioneer's 2.2 GHz). (Courtesy
0 of William L. Sjogren, Jet Propulsion
Laboratory.) Figure 7
-6.5 V
-13.0
•_ I
-19.5 -

-26.0

-32.5
o
TIME (hours)

High-resolution gravity measurements anomalies over those available from Pioneer Venus data
alone (see figure 8).
Measurements Magellan is now making of the gravity Because of the elliptical orbit of the Magellan space-
field of Venus may help us to choose among these hy- craft, the gravity data collected during the cycle that ended
potheses. The earlier Pioneer Venus mission determined in late May of this year have good spatial resolution only
the long-wavelength components of the gravity field, and near the periapsis latitude (10°N). Resolution degrades
Magellan has now confirmed them. Variations in the progressively with increasing latitude to the north and south
gravity field at shorter wavelengths are sensitive to the and is no better than 103 km at high latitudes, including
thickness and structure of the upper boundary layer of such areas as the Ishtar Terra highland, which contains
mantle convection, which in turn are related to both the the only examples of large mountain belts on Venus. A
average and local heat flux from the mantle. From the global set of gravity measurements of uniformly good reso-
relationship between gravity anomaly and topographic lution, and thus a global view of the thermal and mechanical
relief—both on a regional basis and across large features properties of the interior, requires tracking data from a
such as coronas, rift zones and mountain belts—it is spacecraft in low circular orbit.
possible in principle to distinguish between interior ther-
mal models that differ in the thickness and strength of While the Magellan spacecraft does not have suffi-
the mechanically strong lithospheric layer. Key wave- cient propellant to make its orbit circular by purely
lengths for this discrimination are several hundred kilo- propulsive means, Magellan project engineers have de-
meters and less; one can determine such variations in veloped a novel and ambitious scheme to circularize the
the gravity field by tracking an orbiting spacecraft at orbit by means of aerodynamic drag. This scheme, cur-
elevations comparable to or less than the wavelengths rently being carried out, involves lowering the periapsis
involved. elevation into the upper Venusian atmosphere; relying
on drag to reduce periapsis speed and lower apoapsis
The Magellan spacecraft is yielding gravity data that elevation (the farthest point in the orbit), while main-
are significantly better than those obtained by Pioneer taining periapsis elevation by means of small propulsive
Venus, for two reasons. First, while both the Pioneer maneuvers; and finally propulsively raising periapsis out
Venus and Magellan spacecraft were placed in eccentric of the atmosphere once the apoapsis elevation has been
orbits with near-equatorial latitudes at periapsis, the lowered sufficiently. This aerobraking scheme should
eccentricity of the Magellan orbit is far less, so the leave Magellan in an approximately circular orbit of about
spacecraft elevation at high latitudes is much lower. 200-300 km elevation by the end of this month. Tracking
Second, the transponder Magellan uses to determine the spacecraft for 360° of longitude in such a circular
gravitational accelerations operates with an X-band car- orbit will yield a gravity field of uniformly good coverage
rier frequency (8.4 GHz), which is much less susceptible and resolution by the end of this month. While NASA's
to plasma-induced noise in the Venus ionosphere than is continued support of data acquisition after orbit circu-
the S-band carrier frequency (2.2 GHz) used by Pioneer larization has from time to time been in doubt, there is
Venus. As a result the signal-to-noise ratio is much reason to hope that the measurement of a high-resolution
higher for the Magellan data (see figure 7). Early Ma- global gravity field following a successful aerobraking
gellan data are already yielding gravity field solutions maneuver will prove to be both scientifically and pro-
that show improved resolution and interpretability of grammatically compelling to the agency.

54 PHYSICS TODAY JULY 1993

 30 40 50 40 50 60
LONGITUDE (deg) LONGITUDE (deg)

Two gravity field solutions showing the improvement in resolution obtained with recent Magellan data.
a: Contours of the free-air gravity anomaly, in units of 10~5 m/sec2, for the Bell Regio (north of 25° N) and
eastern Eistla Regio (5°-25° N) areas from a solution to the global gravity field obtained from Pioneer Venus
orbiter tracking data expanded in spherical harmonics to degree and order 50. The contours are superposed on
a color-contoured map of elevations obtained from Magellan altimetry data, b: Contours obtained from a new
harmonic solution, to degree and order 60, obtained from a combination of Magellan and Pioneer Venus
tracking data. Note that the magnitudes of the peak anomalies are greater in b and that the gravity anomaly
contours show better agreement with the elevation contours. (Courtesy of Sjogren.) Figure 8

What is clear from the results of the Magellan mis- Grimm, W. M. Kaula, G. E. McGill, R. J. Phillips, R. S. Saun-
sion obtained to date is that the inner workings of Venus ders, G. Schubert, S. W. Squyres, E. R. Stofan, J. Geophys.
manifest themselves at the surface in a fashion very Res. 97, 13199(1992).
different from that on Earth. While Venus currently 7. S. E. Smrekar, S. C. Solomon, J. Geophys. Res. 97, 16121
lacks global plate tectonics, it has been subjected to a (1992).
unique resurfacing history that challenges our ability to 8. Yu. A. Surkov, V. L. Barsukov, L. P. Moskalyeva, V. P.
Kharyukova, A. L. Kemurdzhian, J. Geophys. Res. 89, B393
interpret and explain. Further study of Magellan images, (1986).
the acquisition of high-resolution gravity field data 9. W. L. Sjogren, B. G. Bills, P. W. Birkeland, P. B. Esposito,
through a full mission cycle following orbit circulariza- A. R. Konopliv, N. A. Mottinger, S. J. Ritke, R. J. Phillips, J.
tion, and continued development of interior dynamical Geophys. Res. 88, 1119 (1983).
models should sharpen the competing hypotheses and 10. S. E. Smrekar, R. J. Phillips, Earth Planet. Sci. Lett. 107, 582
may lead to an improved general understanding of mantle (1991).
convection and melting on all of the terrestrial planets, 11. W. M. Kaula, Science 247, 1191 (1990).
including Earth. 12. R. J. Phillips, J. Geophys. Res. 95, 1301 (1990).
13. D. L. Turcotte, Eos, Fall Suppl. 73, 329 (1992).
References 14. R. E.Grimm,S. C.Solomon,J.Geophys.Res.93,11911 (1988).
15. D. T. Sandwell, G. Schubert, J. Geophys. Res. 97,16069 (1992).
1. G. H. Pettengill, P. G. Ford, W. T. K. Johnson, R. K. Raney, 16. E. M. Parmentier, P. C. Hess, Geophys. Res. Lett. 19, 2015
L. A. Soderblom, Science 252, 260 (1991). (1992).
2. R. J. Phillips, R. F. Raubertas, R. E. Arvidson, I. C. Sarkar, 17. J. Arkani-Hamed, M. N. Toksbz, Phys. Earth Planet. Inter.
R. R. Herrick, N. Izenberg, R. E. Grimm, J. Geophys. Res. 97, 34,232(1984).
15923 (1992). 18. J. Arkani-Hamed, Eos, Fall Suppl. 73, 332 (1992).
3. G. G. Schaber, R. G. Strom, H. J. Moore, L. A. Soderblom, 19. V. Steinbach, D. A. Yuen, Geophys. Res. Lett. 19, 2243 (1992).
R. L. Kirk, D. J. Chadwick, D. D. Dawson, L. R. Gaddis, 20. R. L. Larson, Geology 19, 547 (1991).
J. M. Boyce, J. Russell, J. Geophys. Res. 97, 13257 (1992). 21. K. L.Tanaka.N. K. Isbell.D. H. Scott, R. Greeley, J. E. Guest,
4. E. M. Shoemaker, R. F. Wolfe, C. S. Shoemaker, Lunar Planet. in Proc. 18th Lunar and Planetary Science Conf., G. Ryder,
Sci. 22, 1253(1991). ed., Cambridge U. P., New York, and Lunar and Planetary
5. J. W. Head, L. S. Crumpler, J. C. Aubele, J. E. Guest, R. S. Inst., Houston, Tex. (1988), p. 665.
Saunders, J. Geophys. Res. 97, 13153 (1992). 22. T. Matsui, E. Tajika, Lunar Planet. Sci. 22, 863 (1991).
6. S. C. Solomon, S. E. Smrekar, D. L. Bindschadler, R. E. 23. S.C.Solomon, Lunar Planet. Sci. 24, 1331(1993). •

PHYSICS TODAY JULY 1993 55