Institute of Geophysics and Planetary Physics, University of California, Los Angeles, CA 90024, U.S.A.
Originally published in:
Adv. Space Res., Vol. 10, No. 5, pp. (5)125-(5)136, 1990.
Lightning in the clouds on Venus should most nearly resemble intra-cloud discharges on Earth. Intra-cloud discharges are weaker, shorter and more frequent than cloud-to-ground discharges and cause more slowly varying luminosity. Terrestrial lightning also has both geographic and local time orderings. At Venus there is much evidence for lightning similar to terrestrial lightning. The Venera landers saw electromagnetic radiation much like sferics from terrestrial lightning. Pioneer Venus also saw such signals leaking out into the night ionosphere. These signals have a strong local time dependence not unlike terrestrial lightning. The local time distribution helps explain the mixed results of optical surveys. The successful observation was on the evening side where there is much apparently lightning-generated electromagnetic radiation; the unsuccessful observations were on the morning side where such plasma waves and hence lightning discharges appear to be rare.
The possible existence of lightning on Venus has a long and controversial history. The speculation that lightning occurred on Venus first arose over 20 years ago, well before the in-situ study of Venus began /l/. The desire to confirm this speculation led to the installation on the Venera 11 and 12 landers, and was one of the objectives for the installation on the Pioneer Venus Orbiter, of instruments that could detect the electromagnetic radiation associated with lightning /2,3/.
The most common means of detecting lightning on earth, at least by the non-professional observer, is to detect the optical flash. Optical emissions, called the Ashen Light, have been detected on the nightside of Venus for over 3 centuries /4/. While some have speculated that the Ashen Light is caused by lightning /1,3,5/ others feel that any possible lightning on Venus would be too weak to power the Ashen Light /6/. Yet others have dismissed the Ashen Light observations as artifacts.
None of the early Venus missions were instrumented especially to detect the optical emissions of lightning. However, the Venera 9 visible spectrometer was capable of making such a measurement /5/ and the star sensor on Pioneer Venus could be adapted to such a search /7/. Finally, the VEGA balloon mission included a photometer that could detect optical pulses /8/.
The results of these optical investigations were initially apparently contradictory. Two of the three optical searches yielded no clear lightning observations /7,8/ while the remaining one yielded many flashes but from only one storm /5/. The electromagnetic investigations on the other hand provided abundant signals that had the characteristics expected for lightning /2,9/. To add to the confusion it was suggested that the Pioneer Venus signals were electrostatic and not electromagnetic /10/. However, this criticism did not apply to the Venera data and as we will see below was incorrect for the Pioneer Venus data as well.
The purpose of this paper is to review the evidence for lightning on Venus. To do this we will first examine the nature of terrestrial lightning so that we know what to expect at Venus. Then we will review the Venera electromagnetic data and the various optical searches. Finally, we examine the evidence from the Pioneer Venus plasma wave measurements that provide some clues for resolving the apparent paradoxes concerning the possible existence of lightning.
Terrestrial lightning is a powerful electromagnetic discharge occurring over a distance of a few kilometers most commonly taking place in cumulo-nimbus clouds but also occasionally present in snow, sand and dust storms, volcanic eruptions, earthquakes and nuclear explosions. Numerous authors have reviewed the lightning discharge /11, 12, 13, 14/. We will summarize but a few of its properties herein.
Terrestrial lightning occurs in two quite distinct forms, cloud-to-ground discharge and intra-cloud discharges. Such discharges occur when the electric field in a cloud exceeds the breakdown value which ranges from about 106 V/m in wet air to about 3 x 106 V/m in dry air. Usually cloud-to-ground discharges begin in the cloud with a stepped leader carrying about 200-300A and moving about 50-100 m per step. It travels with an average velocity of about 1.5 x 105 m/s. This is followed by a return stroke in the same channel at a velocity of about 0.6x108 m/s carrying 10-20 kA which decays in about 20 to 50 Ás. There are usually 3 or more of these strokes in any one flash separated by about 40 ms and lasting about 0.2 s. Figure 1 shows the electric field during 6 different ground-to-cloud return strokes /11/. The bottom two strokes show the first discharge of a series of strokes. The letter 'L' indicates the various steps of the stepped-leader. The upper four panels show subsequent strokes both with and without stepped leaders.
|Fig. 1. Electric field due to cloud-toground discharges a) first stroke with stepped leader; b) sub s equent stroke with leader; c) subsequent strokes. Time scale in microseconds with units of scale indicated beside each trace. cl est scale applies /11/.|
Discharges within a cloud are 2 to 4 times more frequent than cloud-to-ground discharges. There is no return stroke, per se, in an intra-cloud discharge but there is a recoil streamer which propagates about 2 x 106 m/s over channel lengths of 1 to 3 km with peak currents of 1 to 4 kA. Figure 2 shows the magnetic field pulses during five intra-cloud discharges. The pulses are much more narrow than those of the cloud-to-ground discharges and there are many more of them. The luminosity produced by intra-cloud discharges is rather continuous and more slowly varying than that produced by cloud-to-ground return strokes.
|Fig. 2. Magnetic field due to intracloud discharges /11/.|
A moderate cloud-to-round flash will generate about 4.5 x 108 J and a large one about 2 x 1010 J of which about 10-3 or 10-4 of the power goes into optical radiation. Most of the energy released in a lightning flash is expended in ionizing the air and heating it. Intra-cloud discharges generate less energy per stroke than cloud-to-ground discharges but there are more such discharges.
|Fig. 3. Global distribution of annual number of discharges per 100 km2 /12/.|
Terrestrial lightning has both geographic and local time correlations. Figure 3 shows the global distribution of lightning discharges /12/. There are three major centers of activity as a function of longitude; the Americas, Africa and Indonesia. At any one time there are about 2000 active storm cells each producing an average flash rate of about every 20 sec. Figure 4 shows the Universal Time distribution of lightning in each of these regions and the corresponding local time distribution. Each region has a similar local time profile maximizing about 1600 LT.
|Fig. 4. Universal time (left) and local time (right) distribution of lightning in the three main active zones /12/.|
In order to create lightning there must be an abundance of a substance that can be readily electrified, a process to electrify these particles and a large-scale charge-separation process. The electrification is proportional to the polarizability of a molecule which in turn is proportional to its dielectric constant. The dielectric constant for water is 80; and for H2SO4 which is present in the clouds of Venus, 110. The largest charges in terrestrial clouds at altitudes where the water has become supercooled to temperatures of from - 10 to - 40oC and where there are ice crystals present. H2SO4 freezes at temperatures similar to that of water. The terrestrial clouds in which lightning discharges arise vary in heightfrom 4 to 20 km. The distance between the effective centers of charge in a cloud is similar to the distance of the lower charge center to ground. On Venus the clouds in contrast occur at about 55 km which is much greater than the cloud layer thickness of about 10 km. Thus, it is much more difficult for a discharge to occur from the Venus cloud layers to ground. However, if a Venus discharge were due to vertical charge separation one might expect it to have a discharge length comparable to that on Earth.
The charge separation process is not well understood. It is generally believed that particles become charged differently according to size and then separate according to size when large particles fall more rapidly in updrafts under the action of gravity. One mechanism that has been proposed for the production of the charge gathered by these falling particles is ion production by cosmic rays in the atmosphere /15/. While this process may be too weak on the Earth, it may be much more productive on Venus which has no shielding magnetic field /16,17/ and is much closer to the sun. It is also difficult to compare the expected magnitude of updrafts on Earth and Venus. While strong updrafts were observed on the VEGA balloons /8/, no balloon or probe data are available at the local times where we would expect from our terrestrial experience that lightning would be present.
In short, our terrestrial experience suggests that because of the height of the cloud layer any lightning on Venus would be intra-cloud lightning which occurs more frequently and consists of more and shorter individual pulses than cloud-to-ground discharges. This height should also weaken any geographic correlations but should not affect local time correlations. On Earth these local time Correlations are strong and we might expect the same on Venus where the "4-day" winds transport the particulate matter in the clouds into cooler regions. Since the thickness of the clouds and breakdown voltage is similar in the terrestrial and the Venus atmospheres, the power and duration of an individual flash of terrestrial and Venus might be the same. However, the rate at which such flashes may occur should depend on the charging rate which might be controlled by the ion production rate and the velocity of updrafts which are unknown. Thus we cannot predict how often Venus lightning might occur.
|Fig. 5. The altitude distribution of VLF signals observed by Venera 11. The column on the left gives the wideband intensity. The next four columns give the intensity at 10, 18, 30 and 80 kHz. The last two columns give the spectral indices of the signals /31/.|
|Fig. 6. The altitude distribution of signals seen by Venera 12. The left-most column gives the amplitude at 10 kHz, the middle column gives the amplitude at 18 kHz and the right-most column gives the spectral index /31/.|
The Venera 11 and 12 landers carried both a high sensitivity loop antenna and an acoustic sensor. While the acoustic sensor was saturated during descent, the VLF measurements of the loop antenna provided the most unambiguous evidence of lightning of all the Venus lightning investigations /2/. The instrument returned amplitudes in 4 narrow band channels centered at 10, 18, 36 and 80 kHz and in a wide band channel. Figures 5 and 6 show the amplitude of the VLF signals as a function of altitude. The fact that the amplitude profiles look very different even though the descent trajectories were very similar suggests that the observed waves are not due to the interaction of the probe with the Venus atmosphere. Rather the signal variation is consistent with temporally varying lightning, Figure 7 shows high resolution VLF data during the weak activity at high altitudes on the Venera 12 landing. Here the pulses occur infrequently enough that they can be counted. In this stretch there is a burst about every 12- seconds on the average. Later during the descent the rate becomes too great to count. Venera 13 and 14 carried electromagnetic sensors similar to those on Venera 11 and 12 but no acoustic sensors. Instead they monitored the coronal discharge current from the spacecraft. No discharge currents were detected /18/. The activity on Venera 13 and 14 was similar to that on Venera 12 and significantly less than on Venera 11.
|Fig. 7. High resolution measurements of the wideband field intensity on Venera 12 /31/.|
The landing area for the probes was close to the subsolar point and far away (~ 8000 km) from the equivalent region in which terrestrial lightning is most frequent. This may be the reason why in Figures 5 and 6 the signal strength falls at low altitude. Simply, when the spacecraft reached the surface they were shadowed from the distant source of the VLF signals. On the surface the instruments continued to operate but only Venera 12 saw a burst of VLF noise and only for a brief period as shown in Figure 8. Lightning discharges close enough to be not shadowed appeared to be rare.
|Fig. 8. A burst of VLF noise while Venera 12 was sitting on the surface /31/.|
Ksanfomaliti /2/ assumed a cloud discharge that generated an energy of 109 J for 10-4 sec at a range of about 1800 km. However, if instead the duration of an individual stroke were 10-6 sec as suggested by Figure 2 at a range of 7000 km as appropriate for an late afternoon source, then the energy of a single stroke is 1.5xlO8 J. The total energy for a flash then depends on the number of these 'KI strokes in a flash. Figure 2 suggests that 20 is not an unreasonable number. If so, then the energy in a flash might be 3 x 109 J. However, this value depends linearly on the number of strokes per flash which actually occur on Venus.
The Pioneer Venus star sensor has also been used in an attempt to detect lightning /7/. In this search observations were made for a period of 20 minutes centered at periapsis for each of 36 orbits on the dawn side of midnight. The star sensor did not look directly at Venus because Venus was too bright and saturated the sensor. Rather it looked for scattered light in directions away from the planet. The total observing time amounted to the equivalent of 10 seconds over the whole night side. The upper limit derived from this search was consistent with terrestrial lightning, e.g., 6 flashes km-2 yr-l of 5 x 109 J over the area searched. Since, as we will discuss below, the occurrence rate in the area searched is not representative of other local times, i.e., lightning is perhaps a factor of 10 or more less frequent here than near dusk, this rate may easily be consistent with Krasnapolski's observations.
The VEGA balloons also searched for lightning using photometers and failed to detect a clearly positive signal /8/ but these balloons also flew over the dawn sector and not the dusk sector.
|Fig. 9. Spectral scan obtained by Veneta 9 over active lightning storm on October 26, 1975 at 1930 LT and 9o S /19/. Dark lines indicate portion of spectrum obtained with a factor of 8 less sensitivity. Dashed line gives radiant background. Dot-dash line shows airglow seen in absence of radiant background multiplied by a factor of 10.|
On October 26, 1975 the Venera 9 spectrometer detected one period of apparent optical flashes on the nightside of Venus /19/ at 1930 LT and 9o S latitude. Figure 9 shows a sample of the data obtained during this period. The flashes were observed over a period of 70 sec. Multiple strokes were not observed but since terrestrial intra-cloud lightning produces a nearly continuous luminosity this behavior is as expected. The optical energy was 3 x 107 J per flash which corresponds to a 1010 J total flash energy assuming an efficiency for optical production of 3 x 10-3. This is larger than the energy produced by a typical terrestrial intra-cloud flash but less than a strong return stroke from a strong cloud-to-ground discharge. It is also consistent within a factor of 3 of our estimate of the energy based on a recalculation of Ksanfomalitils measurement of the electromagnetic energy. Based on the statistics of this one event combined with the number of null observations, Krasnopolskii estimated that there were about 100 flashes per sec occurring on Venus with a global flash rate of about 45 flashes/km2/yr or about 7 times the terrestrial rate. We note that this observation was made in a region in which the rate of lightning occurrence is inferred to be greater than average as discussed in a later section.
Beginning with the initial observations of Pioneer Venus in the low altitude night ionosphere it was clear that the impulsive signals, or sferics, that were expected to be seen if there were lightning discharges in the Venus atmosphere were in fact observed /9/. There have been four separate studies of these data /20,21,22,23/, not including 2 reanalyses by H. A. Taylor and coworkers /10,24/. The first study /20/ examined all impulsive bursts at 100 Hz when the magnetic field was strong enough that the 100 Hz signals could propagate in the whistler mode, i.e., the waves could be electromagnetic. Electrostatic waves would be expected not to propagate. The magnetic field direction had to be directed toward (or away) from the planet and bursts had to be well above instrument threshold and occurring only at 100 Hz. The second study counted bursts closer to the instrument threshold and counted every peak in a set of closely spaced bursts in an attempt to more nearly reflect the occurrence of lightning flashes. As before if signals occurred at high frequencies, or if the magnetic field was weak or not pointed toward or away from the planet, no burst occurrence was noted. Figure 10 contrasts the new and old classification. The new classification results in a much higher count rate. The analysis of Taylor et al. /10/ was based on the new classification not the old one as he maintained. This accounts for many of the reported differences in conclusions.
|Fig. 10. Plasma wave amplitudes at 0.1 and .73 kHz illustrating difference between the two definitions of possible lightning generated waves.|
A third analysis was introduced by Singh and Russell /22/ who counted every burst at all frequencies independently without regard to field strength or direction. No list of events was maintained only the counts by season. This study showed that at low altitudes in the dark ionosphere bursts were present at all frequencies with similar occurrence rates. Some telemetry errors may have slipped through the classification procedure but the affects of these was at most minor.
The fourth analysis was by Russell and coworkers /23/ who attempted both to follow up on the study of Singh and Russell and to quantify better the rate of occurrence of the impulsive signals at all frequencies. In order to do this they noted when impulsive signals were present and when they were absent every 30-seconds above a threshold at each frequency. This study allowed a quantitative measure of the occurrence rate of lightning as to whether a flash occurred at least once somewhere in an area of perhaps 5 x 104 km in a 30-second interval. However, the flash rate could be much higher than this occurrence rate.
Because this latter analysis technique provides a quantitative measure of occurrence rate we will concentrate on the results derived from it. However, before doing so, we will briefly describe the instrument, the results of the first three studies and the controversy resulting from these studies.
The Instrument. The Pioneer Venus plasma wave experiment /25/ measures the electric field in 4 narrow frequency bands centered around 0.1, 0.73, 5.4 and 30 kHz. The instrument has a rise time of about 0.05 sec and a decay time of about 0.7 sec so that it may be able to follow flashes if they are slow enough and the telemetry rate fast enough but it should not be able to follow the individual strokes. The antenna consists of two wire 5 cm radius spheres on the ends of meter long booms separated by about 1 meter. The instrument is capacitively coupled to the plasma so that when the Debye length is smaller than the radius of the spheres the "gap" between the spheres and the plasma becomes small enough to affect the calibration of the instrument.
Whistler mode signals are electromagnetic waves which propagate below the electron gyro frequency which is about 300 Hz or more in the night ionosphere of Venus. The energy in these waves is carried predominantly in the magnetic component of the waves. Thus, given a constant Poynting flux of electromagnetic radiation from below, the electric field amplitude will be greatest when the density is least. Thus we expect such waves to be most detectable by the Pioneer Venus plasma wave instrument in the low density holes in the Venus ionosphere and least detectable in the densest part of the ionosphere, i.e., at lowest altitudes.
|Fig. 11. Amplitude of 0.1 kHz signals at low altitude in night ionosphere plotted versus the angle in the spin plane. The amplitudes are largest in the directions perpendicular to the magnetic field as expected for an electromagnetic wave /26/.|
The Initial Studies. The signals identified by Scarf and coworkers were extrapolated back along the magnetic field to the surface of the planet. These source locations seemed to be correlated with topography /20,25,26/. This correlation was questioned by Taylor and coworkers /10/ who proposed that the waves were electrostatic. Scarf and coworkers in turn showed that the waves were electromagnetically polarized as shown in Figure 11 /26/ and did not have the Doppler shifts expected for electrostatic waves /21/. Figure 12 illustrates the absence of the expected Doppler shift.
|Fig. 12. The electric field amplitude, the plasma density and magnetic field strength near periapsis on orbit 66. The Doppler shifts shown at the top of the figure expected if these waves were electrostatic are not observed /26/.|
Singh and Russell /22/ pointed out that there were impulsive signals in all four frequency channels of the plasma wave instrument and that it was possible for signals to propagate a short distance into the irregular night ionosphere of Venus. However, this study presented at best only a qualitative rate of occurrence, since no account was taken of the observing rate.
Latest Study. The most recent study attempts to derive quantitative occurrence rates by noting both where signals occur and where they do not /23/. All data were examined when Pioneer Venus was in the unilluminated ionosphere at a solar zenith angle greater than 900 and within 1.05 Venus radii of the extended Sun-Venus line. To be characterized as impulsive or bursty the magnitude of the signal had to vary in 30s by an amount comparable to the mean of the signal and to exceed thresholds of 1 x 10-5 V/(m-Hz1/2) at 0.10 kHz, 1.5 x 10-5 V/(m-Hz1/2) at 0.73 kHz, 3 x 10-6 V/(m-Hz1/2) at 5.4 kHz and 9 x 10-7 V/(m-Hz1/2) at 30 kHz. The plots used were those available from the National Space Science Data Center. Each 30-second interval was classified. In addition to the 4 electric field measurements, the magnetometer data were examined to eliminate periods in which telemetry errors occurred. Figures 13a and b show typical signals classified in this study. The signals labelled 'b' are interference associated with the motion of the antenna into the spacecraft wake. The signals labelled 'a' and 'c' are typical of those thought to be associated with lightning. The 'a' signals are probably from a source immediately below the satellite and are seen at lowest altitudes. The 'c' signals probably have propagated a large distance to the spacecraft and came from an extended region. The signals labelled 'e' are probably associated with the solar wind interaction.
Plasma wave amplitudes versus time on orbit
515 showing types of signals observed. Type 'b' is due
to the electric field antenna entering the wake of the
spacecraft. Seconds were truncated in originally
published plot. These times are correct /23/.
Further amplitudes on orbit 723 showing
additional wave types. Type Id' is a telemetry dropout.
Seconds were truncated in originally published plot.
These times are correct /23/.|
The initial examination of these data showed that the occurrence rate with increasing altitude at all frequencies suggesting that at least some of the signals present were generated below the ionosphere and that the occurrence rates varied from season to season /23/. Figure 14 shows the altitude dependence of the signals at the 3 highest frequencies. The 100 Hz signals do not vary in occurrence with altitude below half the electron gyro frequency.
|Fig. 14. Altitude dependence of the percent occurrence rate signals seen at 0.73, 5.4 and 30 kHz /28/. The occurrence rate is obtained by dividing the number of 30-second intervals in which one or more impulsive signals occurred by the total number of 30-second observing intervals at that point in space.|
|Fig. 15. Dependence of signal occurrence rate versus frequency normalized by the local electron gyro frequency. Strong fields are to the left-hand side of each trace /27/.|
The magnetic field strength affects the occurrence rate of these signals but the direction of the magnetic field has only a small or negligible effect /27/. Figure 15 shows the occurrence rate versus the center frequency of each channel as normalized by the local gyro frequency. Strongest fields are to the left-hand side of each trace. At lowest frequencies below 1/4 of the electron gyro frequency the occurrence rate becomes less dependent on magnetic field strength but at higher frequencies the occurrence rate decreases rapidly as the frequency increases relative to the electron gyro frequency.
|Fig. 16. Local time latitude maps of the rate of occurrence of 0.73 kHz signals as a percent of the number of available 30-sec intervals for each of the 3 seasons /28/.|
The signals above the electron gyro frequency are useful for mapping the source locations precisely because they attenuate rapidly. Figure 16 shows the occurrence rate at 730 Hz as a function of latitude and local time for each of the first three seasons /28/. During the first half of the second season Venus passed behind the sun preventing telemetry transmission to Earth. Thus only 2 1/2 seasons of low altitude data are available. In the middle panel is also indicated the region covered by Borucki's star sensor survey /7/. The first and third seasons show a very strong local time dependence with a region of highest occurrence rate on the dusk side which is confirmed during the second season by the relative absence of signals on the dawn side. The star sensor search was clearly performed over a region of infrequent plasma wave activity.
|Fig. 17. Local time altitude map corrected for the altitude dependence and summed over the three frequencies /28/.|
We can correct for the altitude dependence of these signals as shown in Figure 14 at each frequency and combine them to get the corrected map of occurrence shown in Figure 17. This figure also shows the path of the VEGA balloons and the location of the reported sighting of lightning by Venera 9. At the peak of the disturbances at about 2100 LT impulsive signals occur within over 70% of the 30 sec observing periods, but on the dawn side impulsive signals about 10% of the time. Thus it is not surprising that the VEGA balloons mission and the Pioneer Venus star sensor search saw no significant activity while the Venera 9 spacecraft detected lightning. Both optical and plasma wave data are consistent with an evening maximum in lightning activity. Lightning activity is also more frequent in Figure 17 toward the equator /28/.
We believe that the local time variation seen on the dawn side of the peak activity in Figure 17 is due to variations in the source because the ionosphere does not vary much across midnight to dawn but the decrease in occurrence rate from 2100 LT to 1800 LT appears to be due to the increasing difficulty of the signals to propagate through an increasingly dense ionosphere. This is illustrated in Figure 18 which shows the occurrence rate from 2200 LT to 1800 LT plotted versus the density at 150 km of the VIRA ionosphere. There is a very strong anti-correlation of -0.99 /28/. It appears that the source of these waves extends well into the dayside but that the dense dayside ionosphere prevents them from reaching the spacecraft.
|Fig. 18. Occurrence rate versus ionospheric density from 2200 to 1800 IT /28/.|
We have also attempted to use these data to determine if there is a geographic correlation. Such a geographic map is shown for the 0.73 kHz signals in Figure 19 /28,29/. There appears to be some geographic control of the signals but this control is much weaker than the local time control and the association is not simply one with geographic highs /29/.
|Fig. 19. Planetary longitude-latitude map of signal occurrence rate at 0.73 kHz for each of the first 3 seasons of Pioneer Venus data /28/.|
The discovery of the local time variation of the impulsive electromagnetic waves allows us to reconcile the somewhat confusing results on the possible occurrence of lightning on Venus. In the plasma waves the occurrence rate peaks at about 2100 LT but the anticorrelation of occurrence rate with ionospheric density suggests that the high occurrence rates continue onto the dayside. Thus, it is quite possible that, as on the Earth, lightning occurrence on Venus peaks in the late afternoon. If so, then it is quite understandable why the PVO star sensor and the VEGA balloon optical searches saw no or at most limited activity in the morning hours while Venera 9 saw clear activity in the evening hours. The Venera 10 and 11 landers also detected impulsive signals which are most readily explained as due to lightning. Given the local time of occurrence suggested by the terrestrial analog and the Pioneer Venus data, it is most likely that the signals observed were due to very distant activity about 7000 km away. Thus the signals at their source are probably much more intense than deduced by Ksanfomaliti /2/ but also they are probably shorter in duration than he assumed also which partially compensates for the increased distance.
The high electromagnetic impulse rates of up to 30/sec combined with the possible extreme distance of the source from Venera 11 and 12 suggest that the true impulse rate is very large perhaps exceeding the stroke rate on Earth. It is however, very difficult with the present data to determine the flash rate or stroke rates. We would need high temporal resolution data obtained under the ionosphere in the afternoon and evening hours to do this. Thus far we know very little about the properties of the clouds, the winds or the electromagnetic activity in this region of space.
The strongest evidence for lightning on Venus comes from the impulsive electromagnetic waves seen by the Venera 11 and 12 landers and the Pioneer Venus Orbiter. These observations suggest that Venus lightning may be similar to terrestrial lightning in many ways. The electromagnetic waves seen by PVO seem to be strongly correlated with local time occurring most frequently on the dusk side. The waves have at least weak geographic associations. The source appears to generate flashes at a rate that is comparable and possibly exceeds the terrestrial rate.
It is also possible that the Ashen Light is powered by lightning since it too has a similar local time pattern /30/. If so then, with some care Venus lightning may be monitored from Earth with ground-based telescopes.
This work was supported by the National Aeronautics and Space Administration by research grants NAG2-501 and NAGW-995.
1. A. B. Meinel and D. T. Hoxie, On the spectrum of lightning in the atmosphere of Venus, Comm. Lunar Planet. Lab., Univ. Arizona, 1, 35-38 (1962).
2. L. V. Ksanfomaliti, Lightning in the cloud layer at Venus, Kosm. Issled., 17, 747-762 (1979).
3. L. Colin and D. M. Hunten, Pioneer Venus experiment descriptions, Space Sci. Rev., 20, 451-525 (1977).
4. P. Moore, The Planet Venus, MacMillan, London (1956).
5. V. A. Krasnopol'ski, Lightning on Venus according to information obtained by the satellites Venera 9 and 10, Kosm. Issled., 18, 429-434 (1980).
6. M. A. Williams, L. W. Thomason and D. M. Hunten, The transmission to space of the light produced by lightning in the clouds of Venus, Icarus, 52, 166-170 (1982).
7. W. J. Borucki, J. W. Dyer, G. Z. Thomas, J. C. Jordan and D. A. Comstock, Optical search for lightning on Venus, Geophys. Res. Lett., 8, 233-236 (1981).
8. R. Z. Sagdeev et al., Overview of VEGA balloon in-situ meteorological measurements, Science, 231, 1411-1414 (1986).
9. W. W. L. Taylor, F. L. Scarf, C. T. Russell and L. H. Brace, Evidence for lightning on Venus, Nature, 279, 614-616 (1979).
10. H. A. Taylor, Jr., I. M. Grebowsky and P. A. Cloutier, Venus nightside ionospheric troughs: Implications for evidence of lightning and volcanism, J. Geophys. Res., 90, 7415 (1985).
11. M. A. Uman, Lightning, Academic Press, New York (1969).
12. P. V. Bliokh, A. P. Nicholaenko and Yu. F. Fillippov, Schuma Resonances in the Earth-ionosphere Wave Guide, P. Peregrinns Ltd. (1980).
13.K. Rinnert, Lightning in planetary atmospheres, in Handbook of Atmospherics, Vol. 2, (ed. H. Volland), 100-133, CRC Press (1982).
14.M. A. Williams, E. P. Krider and D. M. Hunten, Planetary lightning: Earth, Jupiter and Venus, Rev. Geophys. Space Phys., 21, 892-902 (1983).
15. C. T. R. Wilson, Some thunderstorm problems, J. Franklin Inst., 208, 1-12 (1929).
16. C. T. Russell, R. C. Elphic and J. A. Slavin, Limits on the possible magnetic field of Venus, J. Geophys. Res., 85, 8319 (1980).
17. J. L. Phillips and C. T. Russell, Upper limit on the intrinsic magnetic field of Venus, J. Geophys. Res., 92, 2253 (1987).
18. L. V. Ksanfomaliti, Electrical activity in the atmosphere of Venus. I. Measurements on descending probes, Kosmich. Issled., 21, 279-296 (1983).
19. V. A. Krasnopolski, Venus spectroscopy in the 3000-8000 A region by Veneras 9 and 10, in Venus, (ed. by D. M. Hunten, L. Colin, T. M. Donahue and V. I. Moroz), 459-483, Univ. of Arizona Press, Tucson (1983).
20. F. L. Scarf, W. W. L. Taylor, C. T. Russell and L. H. Brace, Lightning on Venus: Orbiter detection of whistler signals, J. Geophys. Res., 85, 8158-8166 (1980).
21. F. L. Scarf, Comment on "Venus nightside ionospheric troughs: Implications for evidence of lightning and volcanism" by H. A. Taylor, Jr., J. M. Grebowski and P. A. Cloutier, J. Geophys. Res., 91, 4594-4598 (1986).
22. C. T. Russell and R. N. Singh, On the nature of impulsive VLF signals observed in the night ionosphere of Venus, Geophys. Res. Lett., submitted (1988).
23. C. T. Russell, M. von Dornum and F. L. Scarf, The altitude distribution of impulsive signals in the night ionosphere of Venus, J. Geophys. Res., 93, 5915-5921 (1988).
24. H. A. Taylor, Jr. and P. A. Cloutier, Telemetry interference incorrectly interpreted as evidence for lightning and present-day volcanism at Venus, Geophys. Res. Lett., 5 (1988).
25. F. L. Scarf and C. T. Russell, Lightning measurements from the Pioneer Venus orbiter, Geophys. Res. Lett., 10, 1192-1195 (1983).
26. F. L. Scarf and C. T. Russell, Evidence for lightning and volcanic activity on Venus, Science, 240, 222-224 (1988).
27. C. T. Russell, M. von Dornum and F. L. Scarf, VLF bursts in the night ionosphere of Venus: Effects of the magnetic field, Planet. Space Sci., in press (1988).
28. C. T. Russell, M. von Dornum and F. L. Scarf, Source locations for impulsive electric signals seen in the night ionosphere of Venus, Icarus, submitted (1988).
29. C. T. Russell, M. von Dornum, F. L. Scarf, Planetographic clustering of low-alitude impulsive electric signals in the night ionosphere of Venus, Nature, 331, 591-594 (1988).
30. C. T. Russell and J, L. Phillips, The Ashen Light, this issue (1988).
31. L. V. Ksanfomaliti, F. L. Scarf and W. W. L. Taylor, The electrical activity of the atmosphere of Venus, in Venus, (ed. by D. M. Hunten, L. Colin, T. M. Donahue and V. I. Moroz), 565-603, Univ. Arizona Press, Tucson (1983).