NASA Technical Reports Server (NTRS) 19980213327: Wind Streaks on Venus: Clues to Atmospheric Circulation

6 Pages·0.35 MB·English

by  NASA Technical Reports Server (NTRS)

NASA-CR-205116
SCIENCE
Reprint Series
21 January 1994, Volume 263, pp. 358-361
/ *
/ / _" /k-
L
Wind Streaks on Venus:
Clues to Atmospheric Circulation
Ronald Greeley,* Gerald Schubert, Daniel Limonadi, Kelly C. Bender, William I. Newman,
Peggy E. Thomas, Catherine M. Weitz, and Stephen D. Wall
Copyright © 1994 by the American Association for the Advancement of Science
Wind Streaks on Venus:
Clues to Atmospheric Circulation
Ronald Greeley,* Gerald Schubert, Daniel Limonadi,
Kelly C. Bender, William I. Newman, Peggy E. Thomas,
Catherine M. Weitz, Stephen D. Wall
Magellan images reveal surface features on Venus attributed to wind processes. Sand
dunes, wind-sculpted hills,and more than 5830 wind streaks have been identified. The
streaks serve as local "wind vanes," representing wind direction at the time of streak
formation and allowingthe first global mapping of near-surface wind patterns on Venus.
Wind streaks are oriented both toward the equator and toward the west. When streaks
associated with local transient events, such asimpact cratering, are deleted, thewestward
component is mostly lost but the equatorward component remains. This pattern is con-
sistent with a Hadley circulation of the lower atmosphere.
Earth-based observations, data from flybys, meters per second even at cloud heights
and measurements from landed spacecraft (14). Wind speeds at the surface are from
reveal that Venus has a rocky surface with 0.3 to 1.0 ms-I (15), well within the range
an average temperature of 753 K beneath necessary to move loose surface sand and
an acid-laden, predominantly CO2 atmo- dust (16).
sphere with a surface pressure of 90 bars. Accordingly, the principal mode of at-
Ideas about atmospheric circulation on Ve- mospheric circulation on Venus from just
nus are based on cloud motions (at _60 km above the lowest scale height (at =10 km
altitude) deduced from ultraviolet images altitude) to _100 km is a zonal retrograde
taken by flyby and orbiting spacecraft (1- (westward) superrotation of the atmosphere
5), wind speeds inferred from Doppler (14). However, global circulation models of
tracking of Venera (6, 7) and Pioneer the lower atmosphere (especially below = 10
Venus (8) atmospheric probes, radio track- km altitude) have hitherto been mostly un-
ing of balloons (at about 50 km altitude) constrained because of the paucity of rele-
during the VEGA mission (9, I0), and vant observations. Theoretical models ofthe
motions of features (at about 50 km alti- lower atmosphere circulation involve a Had-
tude) in infrared images of the planet (11- ley circulation driven by solar energy depo-
13). Zonal winds are westward with speeds sition in the deep atmosphere, preferentially
of =100 m s-z at cloud heights (=60 km at low latitudes (17-21). This circulation
altitude), decreasing approximately mono- involves equatorward surface winds, upflow
tonically with proximity to the surface to over the equator, poleward winds aloft, and
speeds of =10 m s-l at a height of 10 km downflow at high latitudes. A similar circu-
(14). No eastward winds have ever been lation is expected in both the northern and
seen in the atmosphere. Meridional winds southern hemispheres of Venus.
are both northward and southward hut with We report here the first observations that
speeds generally not exceeding several constrain the global circulation pattern of
the lower atmosphere of Venus. These are
R.Greely, K.C,Bender, P.E.Thomas, Department of
the wind streaks seen in Magellan radardata
Geology, Box 871404, Arizona State University,
Tempe, AZ 85287-1404, (22, 23) that result from the interaction of
G.Schubert, D.Limonadi, W. I.Newman, Department the atmosphere and the surface. These wind
ofEarth and Space Sciences, University ofCalifornia, streaks not only reveal the nature of the
Los Angeles, Los Angeles, CA 90024-1567.
C.M.Weitz and S.D.Wall, JetPropulsion Laboratory, lower atmospheric circulation but also reflect
4800 Oak Grove Drive, Pasadena, CA91109, the influence ofthe strong westward winds of
*Towhom correspondence should be addressed. the cloud-level atmosphere.
358 SCIENCE " VOL. 263 ° 21 JANUARY 1994
Venusian wind streaks are radar back- in areas of low surface winds. Extremely speed (28). If similar flow patterns on Ve-
scatter patterns that contrast with the sur- rough surfaces, such as some lava flows, nus are assumed (29), according to the
rounding surface features (23, 24). Both may serve as traps for wind-transported simplest interpretation, radar-dark streaks
radar-bright (high radar backscatter, gener- particles. Each of these surfaces would probably consist of particle deposits that
ally caused by rough surfaces) and radar- have different radar backscatter properties either absorb radar energy or produce
dark (low radar backscatter) wind streaks (24), depending on several considerations smooth surfaces, whereas radar-bright
occur (Fig. 1). Although streaks range from including the areal extent and thickness of streaks represent areas where particles were
less than 5 km to several hundred kilome- surficial deposits, exposed bedrock and its swept from the surface to expose radar-
ters in length, typical streaks are about 20 surface roughness, and possible eolian bed- reflective bedrock.
km long. Streaks occur in several shapes, forms such as dunes. Venusian eolian features, including
including plume, fan, and long-narrow Many wind streaks on Earth are related wind streaks, require a supply of loose,
forms. The most abundant, informally to wind flow patterns over and around small particles and winds of sufficient
termed "zebra" streaks, consist of multiple, topographic features such as small hills. strength to move them. Although wind
alternating radar-dark and -bright streaks. Flow separation and reattachment, as well streaks occur at all latitudes (Fig. 2A) and
Nearly all zebra streaks are associated with as the generation of local vortices, lead to longitudes on Venus, most streaks are lo-
deposits inferred to be ejecta from young distinctive zones of surface erosion and cated in association with ejecta deposits
impact craters. deposition that are functions of the geom- from craters. In afew areas, material may be
Similar to wind streaks on Earth and etry of the topographic feature and the wind weathered from tectonically disrupted ter-
Mars (25-27), venusian streaks are
thought to be visible on radar images
because of differences in the distribution of
windblown particles in relation to surface - I °
wind patterns. Wind tunnel experiments
simulating conditions on Venus (16) sug-
gest that particles moved by the wind are
smaller than _-1 cm and that most would
be a few hundred micrometers in diame-
ter. Depending on the wind shear stress,
surfaces may be completely stripped of
loose grains (leaving exposed bedrock),
covered with large (> 1 cm) particles too
massive to be removed by the wind (form- 0
-90-64-53-44-36-30-23-17-11 -5 0 5 11 1723 30 3644 53 64 0 20 40 60 BO100120140160180200220240260280300320340
ing a lag deposit), or blanketed with grains Latitude(degrees) Azimuth(degrees)
transported from elsewhere and deposited
A B 4110
Fig. 1. (A) Radar-bright streak (top of image,
associated with cone-shaped hill) and "zebra" Azimuth(degrees) o 20 40 60 80 10Q120140160180200220 240260260300320340
Azimuth (degrees)
streaks at9°N, 67°E, north of Hestia Rupes. The
wind is inferred to have blown from north to Fig. 2. (A) Histogram of venusian wind streaks inbands of equal area latitude. (O)Azimuths of all
south at the time of streak formation. The larg- wind streaks in the global data base for the northern hemisphere, with north at 0°. (C) Azimuths of
est bright streak isabout 35 km long (Magellan allwind streaks inthe global database forthe southern hemisphere, with north at 0°. (D) Distribution
F-MIDR 10N065). (B) Radar-dark streaks at of wind streak orientations with respect to local slope, showing that streaks occur randomly with
46°N, 127.1°E, in Northern Niobe Planitia. Part regard to slope. Thedownslope isat 0°and the upslope isat 180°.Angle isthe difference indegrees
of Anake Terresa isvisible in the upper portion between downslope azimuth and streak azimuth. (E)Azimuths of subset of database (3666 streaks)
of the image. Thewind isinferred to have blown in which Type P wind streaks that may have formed in response to transient events have been
from north (top) to south (bottom) at the time of deleted (northern hemisphere). North isat 0°. (F) Azimuths of subset of data with deletion of Type
streak formation. The longest streak isabout 40 Pwind streaks that may have formed in response to transient events (southern hemisphere). North
km long (Magellan F-MIDR 45N126). isat 0°,
SCIENCE • VOL. 263 • 21JANUARY 1994 359
rains (23). Both of these geologic settings er, or in response to transient events such as though the equatorward component re-
could supply material appropriate for ero- impact cratering. An assessment of disturb- mains. The predominant trend in the
sion and deposition by wind. However, not ances to the atmosphere by cratering pro- southern hemisphere is for streaks to be
all impact craters and tectonically disrupted cesses on Venus suggests that turbulent aligned northeastward (average at the _50 °
areas have associated eolian features, sug- winds could be generated near the impact azimuth), with a smaller fraction of streaks
gesting that (i) winds sufficiently strong to (31). Wind streaks formed in association aligned westward. In the northern hemi-
transport particles may not occur every- with these transient events may not repre- sphere, the streaks are mainly aligned south
where on the planet; (ii) some surfaces may sent general atmospheric circulation. We or equatorward (average at the 170° azi-
be too rough, preventing particle move- term these "Type P" streaks because they muth), with a smaller fraction of streaks
ment by wind; (iii) particles may not form, occur within the ejecta deposits of impact aligned westward. The predominantly
are cemented or bonded by some process, or craters defined by Campbell et al. (32) as equatorward orientation of the non-P Type
are too cohesive for movement by the wind; having "parabolic halo" ejecta blankets. We streaks in both hemispheres is consistent
or (iv) particles may be depleted at some separated Type P streaks in the database in with a classic lower atmosphere Hadley
sites of old impact craters because of their the following manner: A rectangular area circulation. The presence of equatorward-
removal and redistribution by surface covering the ejecta deposits was defined for oriented streaks in the highest latitude
winds. each of the 50 craters identified by Campbell bands of both hemispheres indicates the
Little is known about the time of forma- et al. The areas were outlined by the visual possibility of Hadley cell circulation reach-
tion of wind streaks and their lifetime. As inspection of potential ejecta deposits and ing to the poles. The Hadley circulation
part of this study, analysis of sequential air vary in size. Typical areas exceed 500 by 700 represents an average of the lower atmo-
photos of wind streaks in the Mojave Desert km, oriented parallel to the parabola axis. spheric wind patterns over the time period
and Bolivia showed little change over a Streaks in each area that have azimuths (unknown) recorded by the wind streaks.
14-year period. On the other hand, martian oriented +-20° radial to the crater were The differences in wind-streak azimuth
wind streaks were observed to appear, dis- "tagged" as Type P streaks. distributions between the northern and
appear, or change in shape in as little as 38 Most Type P streaks are > 100 km long southern hemispheres suggest differences in
days (25). A preliminary search of Magel- and have predominantly westward azimuths, the lower atmospheric circulation regimes
lan repetitive images has not revealed any parallel to the parabola axes. Type P streaks of the two hemispheres, hemispheric differ-
changes in venusian wind streaks in the = 1 typically are radar-dark or zebra forms, both ences in the supply or transportability of
year between observations, although fur- considered to result from the deposition of small particles, or as yet unrecognized geo-
ther analysis is in progress. sand and dust. We suggest that these streaks logic or topographic hemispheric influences
Regardless of the exact mode of forma- develop by the dynamic interaction of crater on wind streaks. There are no a priori
tion of the contrasts in radar backscatter, ejecta particles, the atmospheric response to theoretical reasons to expect hemispheric
wind streaks on Earth, Mars, and Venus the impact, and the westward zonal winds. differences in global atmospheric circula-
probably represent local wind directions at Streaks may form by the deposition of im- tion patterns on Venus. The northeastward
the time of streak formation and provide pact ejecta raised to great heights and trans- trend in the southern hemisphere and the
the opportunity to assess regional and glob- ported downwind (westward) by the high slightly east of south trend in the northern
al atmospheric circulation patterns (25, altitude, westward superrotation of the at- hemisphere are due to the Coriolis force,
30). To assess potential patterns on Venus, mosphere. The depositional pattern is influ- although the Rossby number (Ro) for Ve-
all wind streaks identified on Magellan im- enced by near-surface roll convection cells nus' atmosphere is estimated to be large
ages were mapped, measured, and classified generated by either the impact heating of compared with unity (14). The value of Ro
with the use of the scheme of Greeley et al. the surface and atmosphere or the impact- is the ratio of the inertial force to the
(23). Histograms of streak orientations induced upwelling of magma from beneath Coriolis force, and if Ro > > 1 then Cori-
were plotted for the northern and southern the surface and is oriented parallel to the olis effects should be relatively unimpor-
hemispheres (Fig. 2, B and C). Orienta- westward atmospheric circulation. This tant. The large estimate of Ro is based on
tions are given as azimuths in the down- model is supported by observations that the planetary rotation rate; Ro might be
wind direction. In the northern hemisphere there is a decrease in zebra streak length smaller and the Coriolis force might be
there is a bimodal distribution of azimuths; down-range from some craters (such as more important if the atmospheric superro-
one mode is toward the south-southeast and Stowe crater) and that zebra streaks do not tation rate is a more appropriate measure of
the other is toward the west. The azimuths occur in association with craters <30 km in dynamical influences.
in the southern hemisphere are also bimo- diameter, suggesting insufficient impact-gen- Type P streaks are associated with impact
dal, with one mode toward the north- erated heat or associated magma upflow to craters and are considered to result from the
northeast and the second mode toward the induce near-surface atmospheric convec- interaction of crater ejecta, convection cells
west. Thus, the global wind directions in- tion. In this model, Type P streaks are not in the atmosphere generated by impact, and
ferred from the streaks are generally equa- indicative of present general near-surface westward zonal winds in the upper atmo-
torward and toward the west. atmospheric circulation but reflect the west- sphere at the time of impact. The remaining
Before the Magellan mission, some pre- ward superrotation of the upper atmosphere 3666 streaks (excluding Type P streaks) have
dictions (29) suggested local atmospheric at the time of impact. Therefore, the west- orientations indicative of near-surface Had-
circulation involving slope winds. If these ward superrotation of the atmosphere ex- ley cell circulation averaged over the time
conditions existed, then wind streaks might tends at least several hundred million years recorded by the streaks. These Hadley cells
be more indicative of topography than of into the past, the inferred time of impact. could extend to polar latitudes.
general circulation. Histograms of the ori- The database contains 3666 streaks with
entation of wind streaks with regard to local the removal of Type P streaks. Histograms REFERENCES AND NOTES
slope (Fig. 2D) show random orientations, of streak azimuths in the northern and
suggesting that slope is not amajor factor m southern hemispheres (Fig. 2, E and F) 1 M. J Bellon, S Smith, G Schubert, A. D. Del
the determination of streak azimuths. show that the strong westward component Genio, J Atmos Sci 33, 1394 (1976)
2 S S Limaye, Icarus 73. 212 (1988)
Streaks may form in response to wind seen with the inclusion of Type P streaks is
3....... C Grassottil M J Kuelemeyer, ibid., p
circulation patterns over many years or hmg- mostly absent in both hemispheres, al- 193
360 SCIENC'E • VOL. 263 * 21 JANUARY 1994
4. A. D. DeJ GenJo and W. B Rossow, ,,ZAtmos. Sct 21.... Adv Geophys_ 28A, 347 (1986)
47, 293 (1990). 22. R S. Saunders et al., Science 252, 249 (1991).
5. W B Rossow, A. D Del Genio, T. P. Eichler, ibid., 23 R. Greeley et al, J Geophys Res 97, 13319
p. 2053. (1992).
6. M. Marov etal., /bid 30, 1210 (1973). 24 R E. Arvidson et aL, ibid., p. 13303.
7. V.V. Kerzhanovich and M Ya Marov, in Venus, D. 25. C. A. Sagan etaL, Icarus 17,346 (1972)
M Hunten, L. Colin, T M. Donahue, V. I. Moroz, 26. P. Thomas, J. Veverka, S Lee, A Bloom, ibid 45,
Eds. (Univ of Arizona Press, Tucson, 1983), pp, 124 (1981)
776.-778.
27. R Greeley, P. Christensen, R Carrasco, Geology
8. C. C, Counselman Iit, S. A. Gourevitch, R. W.
17, 665 (1989).
King, G. B Loriot, R. G Prinn, Science 205, 85 28. R. Greeley, J. D Iversen, J. B. Pollack, N. Udov-
(1979), ich, B. White, Science 183, 847 (1974)
9. J E Blamont etal, ib/d 231, 1422 (1986)
10. R A. Preston etaL, ibid., p, 1414. 29. R S, Saunders, A R Dobrovolskis, R Greeley, S
lt. M. J S. Be/ion eta/., tbid. 253, 1531 (1991_. D Wall. Geophys. Res Lett 17, 1365 (1990).
12. D. Crisp et al., ibid., p. 1538. 30. P. Thomas and J Veverka, J. Geophys Res. 84,
13 R W. Carlson etal., ibid., p. 1541. 8131 (1979),
14. G. Schubert, in (7), pp. 681-765 31. P. H. Schultz, ibid. 97, 16183 (1992).
15. V. S. Avduevsky et al., Cosmic Res. 14, 622 32. D B Campbell et al, ibid, p 16249.
(1976) 33. We thank R. D Baker and D L. Bindschadler for
16. R.Greeley etal., Icarus 57, 112 (1984). helpful discussions and E. Lo for developing
17. E. K&lnay de Rivas, J. Atmos ScL 30, 763 programs to manipulate the database. Supported
(1973). by the National Aeronautics and Space Adminis-
18. P. H. Stone, ibid. 31, 1681 (1974). tration through the Magellan Project and the Of-
19. G. Schubert et at, J. Geophys. Res 85, 8007 rice of Planetary Geoscience
(1980).
20. W. B. Rossow, J. Atmos. Sci. 40, 273 (1983). 17 August 1993; accepted 18 November 1993
SCIENCE • VOL. 263 • 21 JANUARY 1994 361