Space Plasma Physics: The Study of Solar-System Plasmas (1978)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

THE IONOSPHERIC PLASMA Donald T. Farley School of Electrical Engineering Cornell University Ithaca NY 14853 ABSTRACT - We consider the ionosphere here as a plasma,' and only incidentally as part of the Earth's upper atmosphere. The long term stability and lack of walls in the ionosphere provide advantages over laboratory plasmas for some investigations. We shall concentrate here on three areas of research: incoherent scatter, instabilities in the equatorial and auroral E regions, and instabilities in the equatorial F region. Several other topics will be mentioned briefly. Incoherent scatter research is based upon and strongly confirms linear kinetic plasma theory. The E- and F-region instabilities lead to plasma turbulence which can be conveniently studied in the ionosphere. The linear theories of these instabilities are now at least partly understood, particularly in the long wavelength regime. The nonlinear processes by which the turbulence develops are being actively investigated via a variety of experimental, theoretical, and computer simulation programs. Once a good correspondence between the observations in space and the simulations has been achieved, the latter should guide the development of non- linear theories, which should have applications beyond those of space physics. 1117

1118 1. INTRODUCTION The ionosphere is by definition ionized and hence a plasma. It offers both advantages and disadvantage's for the study of plasma processes. It has no walls, is usually in a reasonably steady state (sometimes even complete thermal equili- brium), and covers a wide region of parameter space: the temperature varies over at least an order of magnitude, the ionization density varies-over 3 or 4 orders of ..magnitude, and the magnetic field and/or collisional processes are important for the ions and/or electrons in some regions but not in others. On the minus side, it is difficult (but not impossible) to do controlled experiments in the ionosphere. In most situations the parameters cannot be varied at will; one must settle for what nature provides. This will probably be less true in the future, however, with the advent of the Space Shuttle. The motivation of most ionospheric research is of course to gain an understanding of the important processes controlling the behavior of the upper atmosphere. The existence of the ionosphere and its importance for radio communications was recognized shortly after the turn of the century, and even now in the age of satellite communication systems, the ionosphere continues to have a considerable impact upon communications. We wish to understand the ionosphere, however, not only because of its effect upon radio propagation, but because it is a part of our environment. It is becoming increasingly clear that the whole atmosphere, including the uppermost parts, is more tightly coupled than was realized in the past. Processes in the lower atmosphere affect the upper and vice versa. The effect of tropospheric storms can be detected at altitudes of several hundred kilometers; the effect of magnetospheric and ionospheric processes on the lower atmosphere can apparently also be appreciable. The possible relationship between weather and the solar cycle is currently attracting considerable attention. Most of us who study the ionosphere are interested in plasma physics because the ionosphere is a plasma, and not vice versa. Here, however, we will adopt the opposite point of view and consider the ionosphere

1119 to be a plasma which just happens to be part of the upper atmosphere. We will not concern ourselves with 'such things as photochemistry and the production and loss of ionization, interactions with winds and gravity waves in the neutral atmosphere, general questions of morphology, etc. Instead we will examine three particular areas in which the fields of plasma physics and ionospheric physics have closely impinged upon each other. These are (1) incoherent (sometimes called Thomson) scatter, which was first developed as an electromagnetic probing technique for the ionosphere but which has also been successfully used to study high temperature plasmas in the laboratory, (2) E region (^100 km altitude) instabilities observed in the equatorial and auroral ionospheres, and (3) instabilities observed in the equatorial F region (£ 300 km altitude). A few other topics will also be briefly mentioned. Incoherent scatter, in addition to its usefullness as a probing technique, has provided perhaps the best detailed quantitative verification to date of the conventional linear kinetic theory of a "warm" plasma. Agreement between theory and observation is invariably within the experimental errors, which are sometimes less than one percent. Theoretical explanations of the E-and F-region instabilities are now emerging and are based on theories originally developed to explain laboratory plasma phenomena. In the ionosphere one can study these instabilities in their fully developed "turbulent" state at leisure, so to speak, since the main characteristics of the plasma change only slowly, usually on time scales of minutes or hours. 2. INCOHERENT SCATTER Incoherent scatter in its simplest form is simply the scattering from individual free electrons in a plasma. The fact that an electron will scatter electromagnetic radiation has of course been known for many decades, but the radar -28 2 cross section of an electron is so small (<" 10~ m ) that incoherent scatter (in which the signals from individual electrons are randomly distributed in phase so

1120 that one adds powers rather than voltages to get the total signal) from the iono- sphere was ignored for many years. After all, the total cross section of even a 10 km cube of ionosphere with a density of the order of 10 electrons/m , which 2 is a typical maximum density, is still only about 1 cm , which is a pretty small target at an altitude of 300 km. Furthermore, the electron velocities should apparently introduce large Doppler shifts in the returned signal, thereby broadening the bandwidth and making detection even more difficult. In 1958, how- ever, Gordon ~ pointed out that radars of that time had reached the state in which they could rather easily detect such weak scatter, provided that one used a really large antenna, an idea which led him to the construction of the Arecibo 9 radar in Puerto Rico. Stimulated by Gordon's idea, Bowles recorded the first measurement of incoherent scatter from the ionosphere using a BMW, 41MHz transmitter coupled to a hastily constructed 1024 dipole antenna array. From these beginnings incoherent scatter has been developed into an exceedingly powerful tool for probing the ionosphere from the ground and, perhaps to a somewhat lesser extent, has proved useful for studying laboratory plasma phenomena. Once the first ionospheric measurements were made, it was rapidly realized that one could not ignore the effect of the ions, as Gordon had done in his first preliminary calculations. For wavelengths longer than the Debye length (the usual case in ionospheric measurements), the ions play the main role in determining the shape of the spectrum of the scattered signal. A number of theorists ' ' ' soon had independently developed the basic theory, using a variety of approaches to the calculation (but all arriving at the same result, happily). The calculations involve finding <|AN(k,<1))| >, the expected mean square amplitude of density fluctuation waves with wave vector k and frequency a). The radar frequency and geometry determine k, and a) is the Doppler shift introduced by the scattering. At some point all the

1121 theories invoke the Vlasov or Boltzmann-Vlasov equation, which most 12 plasma physics textbooks rt point out is an approximation, since it involves a certain smoothing out of the microscopic electric fields. s The equation is then linearized in the perturbations and one is soon faced with integrals of the form v3F (v)/3v where F is the unperturbed distribution function of the ions or electrons, and v is the component of velocity in the direction of k. For any reasonable distribution function, including a Maxwellian for a plasma in thermal equilibrium, one must deal with the pole in this integral at v=w/k. The now universally accepted way to handle this pole was first described by Landau and leads to the familiar Landau damping of plasma waves. In the late 1950's and early I9601 s, however, there was still some discussion as to whether or not Landau damping was a real phenomenon, and at least one theory of incoherent scatter was developed on the assumption that it was not (by integrating through the pole and taking the principal part of the integral rather than using an indented contour). The results of the latter theory, while similar to the others, were sufficienty different that experimental measurements of the shape of the spectrum of the scattered signal were soon able to show that the "Landau damped" theories were correct. The relevance of incoherent scatter research to some other aspects of the development of plasma physics theory is mentioned in a review by Bauer . The theory leads to an expression for the radar scattering cross section a of the plasma which is given below, in the notation of Farley ' . Nre Here Tg and T. are the electron and ion temperatures, N is the electron density, r is the classical electron radius (2.82x10 m), k is 4TT/A for backscatter, e radar

1122 A is the Uebye length, and y and y. are electron and ion "admittance" functions that can be expressed in terms of the plasma dispersion function for Maxwellian velocity distributions when the magnetic field is unimportant. For ion mixtures y. must be replaced by an appropriate summation. The effect of a magnetic field, collisions, relative drifts, and non-Maxwellian distributions can all be incorporated into the admittance terms. This theory shows, among other things, that when the Debye term is much smaller than unity (the usual case in ionospheric measurements) the ions play the .dominant role in determining the spectrum of the scattered signal, even though it is of course the electrons that do the scattering. The Doppler width of the spectrum is usually determined by the ion rather than electron velocities and hence is more than two orders of magnitude (for oxygen ions) narrower than Gordon's original prediction, making detection of the signal considerably simpler. Since 1960 the theory of incoherent scatter has been extended by the above mentioned authors and several others to include the effect of many other parameters of the ionosphere. In many situations the original thermal equilibrium, collision- less, single ion species, non-magnetic theory is directly applicable to the iono- spheric measurements, but in some it is not. The electron and ion temperatures may be different, collisions between ions and neutral particles may be important, the Earth's magnetic field may affect the scattering, more than one ion species may be present, the plasma as a whole may be moving, or the electrons and one or more species of ions may all be drifting at different mean velocities, the electrons may have a non-Maxwellian distribution, particularly a high energy "tail" caused by the production of photoelectrons, or the ions may even by non-Maxwellian, particularly in the auroral zone where the plasma drift velocities relative to the neutral particles can be very large. It seems inappropriate to list all the relevant references here; many of them can be found in reviews by Evans ' ' and Bauer .

1123 The theoretical work has been extremely successful. There is no unresolved / * disagreement between theory and experiment, and many of the theoretical results have been verified to a high degree of accuracy - better than one percent in some cases. For example, Figure 1, taken from Hagen and Hsu " , shows a series of auto-correlation functions, which are the Fourier transforms of the Doppler power spectra of the scattered signal, measured at the Arecibo Observatory in Puerto Rico at altitudes up to 1854 km. The fitted theoretical curves, for 0 -H ion mixtures, are plotted on the same axes, and in most cases the two cannot be distinguished. From such a fitting procedure the composition and temperature variations with time and altitude can be determined. Figure 2 T|t (from Hagen and Behnke ) shows an example of recent high altitude wide bandwidth measurements at Arecibo which succeeded in mc~suring the electronic component of the scatter from the ionosphere for the first time. Even •more recently, Behnke and Hagen have reported the first observation of the lower hybrid or whistler resonance predicted by theory. The theory of incoherent scatter is very rich in the sense that so many parameters of geophysical interest can affect the scattered signal, particularly the spectrum of the signal. This often makes the data somewhat difficult to analyze, but on the other hand when the analysis is carried out it yields a great deal of information that is obtainable in almost no other way. The ionized particles in the upper atmosphere also serve as tracers to provide information about ambient electric fields and the dynamics of the neutral atmosphere. The incoherent scatter measurements of temperature, drift velocity, and collision frequency are most important in this application. Incoherent scatter observations in the ionosphere complement rocket and satellite .in&-itu. measurements by providing very detailed monitoring of many parameters at a single geographical location over a range of altitudes extending from 80 km or below to several thousand km. • Incoherent scatter probing of laboratory plasmas is also carried out, but the technique does not seem to provide as powerful a tool in the laboratory as it

1124 does in the ionosphere. The measurements are usually made with lasers with wave- lengths much less than the- Debye length (in contrast to the ionospheric case), and so the ions have no effect on the scatter. The data give a measure of the electron velocity distribution (i.e., temperature and density). The lab spectral data are generally of poorer quality than ionospheric data due to statistical considerations. Since the scattering is a stochastic process, one must average many samples to get accurate estimates of the statistical parameters, even if the signal-to-noise ratio is very high. This averaging is generally easily done in the ionospheric case where changes in the plasma characteristics are slow, but this is often not possible in the laboratory. In spite of all this, however, the lab measurements do yield results which are difficult to obtain in any other way. To summarize, soIr.e of the important consequences of incoherent scatter research to data are: 1. The .extensive agreement between theory and ionospheric observations has provided a detailed quantitative verification of conventional kinetic plasma theory based on the linearized Vlasov equation. 2. Incoherent scatter radar observations have become by far the most powerful ground-based technique for studying ionospheric behavior. Most of the important parameters of the ionospheric plasma can be measured over a wide range of altitudes, usually with good time resolution as well. Unfortunately, because of the expense there are only five observatories in the world now making these observations (with a sixth under construction in northern Scandanavia). 3. Incoherent scatter probing of high temperature plasmas in the laboratory has been useful in measuring the electron temperatures. Looking towards the future, we can make the following predictions and recommendations: 1. Theoretical research on the theory of incoherent scatter will continue,

1125 but at a moderate level. The theory currently seems to be in very good shape, but new questions will undoubtedly arise, particularly in connection with obser- vations in the auroral zone and polar cap, regions in which we can expect at s times to find highly non-equilibrium, non-Maxwellian ion and/or electron distributions. 2. Incoherent scatter will continue to make important contributions to our understanding of ionospheric physics, particularly in polar latitudes. It is important to maintain and if possible upgrade the few existing observatories. It would also be highly desirable to build new facilities, such as the one 19 which has been proposed for a region near L=4 (see Evans ), probably in the Northeastern United States and Canada (where auxiliary receiving sites are planned). This project has been stalled in an advanced state of planning for several years now. Indian scientists are eager to establish an observatory in their country, but lack the nesessary funds. It should be emphasized that although incoherent scatter radars are certainly not cheap, they are not really expensive when compared to the cost of major satellite programs. We obviously learn a great deal from satellites, but some phenomena cannot be studied effectively by them. Tidal variations in the upper atmosphere, for example, are particularly suited to study by incoherent scatter, but several observatories (the more the better) distributed over the globe must be operating simultaneously. 3. Returning more to the area of plasma physics, incoherent scatter should continue to contribute to the theory of parametric instabilities in the 54 ionosphere driven by powerful HF transmitters. Measurements at Arecibo during the past few years have given interesting results, but the geometry there is not ideal for this sort of study. A heating facility is planned in conjunction with the new Scandanavian incoherent scatter observatory, and hopefully we should learn a great deal when these begin operation.

1126 4. The advent of the Space Shuttle will provide opportunities for various active experiments in the ionosphere. For many of these, incoherent scatter from the ground will doubtless provide convenient diagnostics. The opinion of most incoherent scatter experts is that it does not make sense to try to do incoherent scatter observations from a satellite, even a very large one. 3. E-REGION ELECTROJET PLASMA INSTABILITIES There are substantial ambient electric fields in the ionosphere. At middle and low latitudes these are produced by the dynamo action of neutral winds driving the charged particles across the Earth's magnetic field through collisional interactions. The neutral winds, in turn, are controlled by solar induced pressure gradients, tidal forces, coriolis effects, and even ion drag (representing the work performed by the neutral wind in generating the electric fields). In polar and auroral latitudes, the electric fields are generated primarily through interactions between the magnetosphere and the solar wind and are transmitted down into the lower ionosphere along the highly conducting magnetic field lines. These electric fields generate a worldwide system of currents which flow primarily in the E region at altitudes of the order of 100-120 km. These currents are particularly intense in two regions: (1) the magnetic equator where the magnetic field geometry leads to an unusually high conductivity perpendicular to the magnetic field, and (2) the auroral zone where the electric fields perpendicular to B are often very large. In both regions the currents lead to plasma instabilities which generate electron density perturbations which have been observed by radar and-en-6-ctu probing. These observations have motivated a very substantial amount of theoretical work. The instabilities in the two regions have obvious similarities, but there seem to be some puzzling differences as well. Since the equatorial case seems to be the better understood at the moment, we will consider it first.

1127 Equator. At the equator typical east-west electric fields are of the order of 10~ V/m, which is fairly weak. However, at the equator the Hall current, which this field would normally produce, is vertical and inhibited, and a vertical polarization field builds up which is 10-20 times larger than the original driving field . This in turn drives a horizontal Hall current which is due almost entirely to electron drift. The important region is characterized by v «ti , v.»fi. , where v and ft . are the electron and ion e e i i e,i e,i collision and Larmpr frequencies. From another point of view, the effective east-west conductivity at the equator is approximately the Cowling conductivity 2 2 (a_+a.. )/a. where a_ is the Hall conductivity, which is considerably larger in the electrojet region than the Pedersen conductivity a, . The resulting electron drifts are generally westward during the day and eastward at night, at velocities that are typically of the order of a few hundred m/s. The currents at night are very weak because of the low electron densities in the E region, but the electric fields and electron velocities are as large as during the day. These electron velocities are much smaller than the electron thermal velocities but can at times exceed the ion acoustic velocity, which is essentially the ion thermal velocity and is roughly 350-400 m/s for typical temperature values. The close correlation between anomalous echoes from the E region seen at the equator with ionosondes (equatorial sporadic-E) and the strength of the electrojet current as measured with magnetometers was recognized many years ago . During the IGY (1957-58) a comprehensive study of equatorial scattering phenomena in both the E and F regions was carried out using several VHF forward scatter links in Peru (see Cohen and Bowles, and references therein). In the early 1960 's the work in Peru was resumed using frequency coherent VHF radars so that Doppler spectra of the electrojet echoes could be determined . The strong echoes showed a characteristic peak in the spectrum at a Doppler shift that corresponded closely to the ion-acoustic

1128 22 11 velocity in the medium. This result led Farley "* and Buneman to independently suggest that a type of two-stream plasma instability was the source of the irregularities responsible for the scattering. Because the current in this case is flowing perpendicular to the magnetic field, the instability threshold is greatly reduced from that of the ordinary collisionless plasma two- stream instability, for which the relative ion-electron drift velocity must exceed the electron thermal velocity when T #T., as in the ionosphere. In the C 1 electrojet case the drift must only exceed the ion-acoustic velocity by a moderate amount to trigger unstable waves traveling perpendicular (or nearly so) 49 to the magnetic field. There were also early suggestions by Maeda et al. and Knox that the gradient drift instability, first discussed by Simon in connection with laboratory plasmas, might also be important for the electrojet case. These suggestions did not receive much acceptance at first .because they did not seem to deal successfully with VHP data. Subsequent research, however, has shown that this instability is also important. This early work reached some important conclusions but left many unanswered questions which are only now beginning to be resolved. Research continued during the 1960's and has accelerated during the 1970's, which have already seen the publication of over 50 papers dealing with the electrojet instabilities. Radar diagnostic techniques have been greatly improved, some rocket observations have been made, the linear theory of the instability (in both fluid and kinetic theory forms) has been extended and refined, numerous theoretical studies of various nonlinear effects have been carried out (although with only mixed success), some of the effects have been simulated in laboratory experiments, and work on computer simulations of the instabilities has begun. This latter work, in particular, seems to hold great promise for the future.

1129 Even a brief discussion of all this work and/or a listing of all the references is not possible here. A short recent review of some of the material was given by Farley 24_ The most recent radar work in Peru and the current status of some of the theoretical problems are described in a series of papers by *\ f r\ Q Feier et al • ^ucn °f tne theoretical work on nonlinear effects done in the last few years is discussed and referenced in papers by Lee etal. °t Register and Jamin 56 and Sato • . The main observed characteristics of the equatorial electrojet instability are as follows: 1. There is a high correlation with the magnitude and direction of the electro- jet current. The medium is unstable when the electrons flow westward during the day, or in either direction at night. 2. The density irregularities are highly aligned with the magnetic field. 3. The irregularities are observed over a wider altitude range at night than during the day. 4. Echoes are obtained even with vertically pointing radars, implying unstable plasma waves with vertical k-vectors (not simply a vertical component of k), even though the electrojet flows horizontally. 5. Two distinct types of irregularities (plasma waves), with different thresholds of excitation, different k-spectra, and different radar Doppler spectra are observed. A few examples of the radar data aie shown in Figures 3-5. Figure 3 (from Cohen and Bowles, ) shows a series of power spectra measured near noon in Peru, pointing the radar east and west from the zenith as well as vertical. The electrojet was strong during this period and the oblique spectra are good examples of "type 1" (so-called because they correspond to the strongest radar echoes at 50 MHz and hence were the first detected). The east and west spectra are nearly mirror images of each other and the spectra peak at essentially the same Doppler

1130 shift for substantially different oblique zenith angles, in contrast to the behavior for "type 2" spectra for which the main Doppler shift is proportional to the sine of the zenith 'angle. Figure 4 (from Balsley and Farley ) shows data obtained at three radar frequencies, corresponding to plasma wavelengths of approximately 9, 3, and 1 meter, respectively. The 50 MHz data show most clearly the change from type 2 to type 1 spectra and back again as the electro- jet streaming velocity first increases and then decreases. Figure 5 (from Fejer et al. ) shows a series of spectra obtained looking vertically at Jicamarca using a narrow beam antenna, short pulses, and short integrations in order to get good spatial and temportal resolution. The "turbulent" character of the scatter is quite apparent. The shape of the spectra can change rapidly with both altitude and time. These and other data not shown indicate that the waves seen looking vertically can switch from predominantly up-going to down-going (or vice-versa) in times at least as short at 2 sees and altitude intervals at least as small as 800m. The linear theory of the instability has been developed by numerous authors. 27 In the notation of Fejer et al. , the results of the fluid development can be written as - 2ctNo (3-2) Here, u and y are the real frequency and growth rate of the plasma wave, V, is the electron drift velocity (the ions are essentially at rest), k is the wave vector, C is the ion-acoustic velocity, LN is the electron density gradient length (positive for density increasing with height), a is the recombination coefficient, and ijj is v v./R fl. for kiB but increases very rapidly when k departs from per- e ± e i ~ — — pendicularity by more than 1° or so. The first two terms on the RHS of (3-2) dominate at short wavelengths and set a drift velocity threshold for instability

1131 slightly above the acoustic velocity (since ty in the region of most interest is 0.2-0.4). This is the so-called two-stream instability which is thought to correspond to the type 1 irregularities. The last two terms in (3-2) are most X important at long wavelengths and are thought to be mainly responsible for the type 2 irregularities; the third is the gradient-drift driving term and can be either stabilizing or destabilizing, depending upon the directions of the density gradient and drift velocity, and the last gives the chemical (recombination) damping rate, which is particularly small at night when the density is low. On the basis of this theory one can account for characteristics 1-3 listed above, as well as some of the properties of the type 2 irregularities. One cannot explain the detailed shape of the Doppler spectra of the radar echoes or predict the amplitude of the unstable waves; both of these obviously depend upon nonlinear saturation mechanisms. Nor can vertically propagating unstable waves, for which the two driving terms go to zero, be accounted for. In particular, the Doppler spectra of the type 1 irregularities show a sharp peak at kC , even when k-V, is apparently substantially larger than C . A perhaps even more fundamental problem is to account for the existence of two distinct types of irregulariites, since there is only one root of interest of the dispersion equation. Another question is that of anomalous resistivity: Is the conductivity of the electrojet region seriously affected when strong plasma waves are generated? In attempts to deal with some of these questions, various nonlinar processes such as quasi-linear effects, orbit diffusion, aid mode coupling, combined in some cases with the effects of refraction and convection, have been investigated. Some or all of these processes undoubtedly are important, but because of the inevitable simplifications which are introduced into the models to make them mathematically tractable, the results to date have not been very successful in explaining the observations. All the nonlinear models, for example, assume a

1132 horizontal laminar background electron drift, and hence cannot account for the 2Aa vertically propagating waves. To cope with this problem Farley and Balsley suggested a two (or more) 'stage process in which the medium , first develops large scale irregularities with associated "turbulent" gradients and electron drifts; these is turn can generate small scale irregularities (which are what the radar detects) with wave vectors in all directions. This idea was shown 64 to be quantitatively reasonable by Sudan et al. , and shortly thereafter Sato independently put forward a similar suggestion. This work stimulated interest in computer simulations of this turbulent coupling process. Numerical investigations of the electrojet processes were carried out as early as 1967 by Sato and Tsuda, but in my opinion at least, the first important computer simulations were those of McDonald et al. ' . These showed that the turbulent process described above really takes place; small scale irregularities moving in a wide variety of directions can be generated via the cascading process when the linear theory described by (3-1) and (3-2) predicts that only large scale waves moving primarily horizontally will be unstable. Figure 6 shows the simulated development of the irregularities, using a 50x50 point grid with a spacing of 1.5m and a drift velocity of 100 m/s. It can be seen that the initial perturbation develops into a highly chaotic state within 2.7 sees. This work which was begun at the Naval Resaerch Laboratory is now being extended at Cornell University. The early results of the Cornell simulations have already given good reproductions of radar observations of 44 the type 2 irregularities. Figure 7 (from Keskinen et al. ) shows a comparison between a simulated radar Doppler spectrum for 9.5 meter wave- length waves and one of the spectra shown in Figure 4. The importance of the simulations is that they will provide an opportunity to carry out "controlled experiments" on the ionospheric processes. Once we are convinced that the numerical results truly reproduce what is going on in the iono- sphere, we can vary the plasma parameters at will and study the effects, something

1133 that cannot be done in the real ionosphere. Furthermore, the simulations permit perfect "probing" of all the plasma variables. The catch to all this of course is that the simulations always involve approximations and often cannot completely reproduce the natural phenomena, particularly over a large range of wavelengths. Nevertheless, the equatorial electrojet observations and simulations seem to provide one of the "cleanest" opportunities available for studying in detail an example of fully developed two-dimensional plasma turbulence. What are the prospects for the future in this area; where should future efforts be directed? 1. As far as observations are concerned, joint comprehensive experiments will be of most value. We need rocket measurements of both the zero order and per- turbation electron densities and electric fields,, together with simultaneous radar measurements at as many angles and wavelengths as possible. Campaigns of this sort are planned for Peru. 2. The numerical simulation work should be continued and expanded. The most recent work referred to above neglects ion inertia, for example, and thus is not valid for drift velocities comparable to the acoustic velocity, the condition for generating type 1 irregularities. The simulations also need to be extended to shorter wavelengths. 3. If successful, the simulations should serve as a guide to more realistic semi-analytic theories of the nonlinear processes. Merely simulating a natural phenomenon is not of too much use if it does not lead to an improved understanding of the important physics involved. f 4. The simulations should be able to resolve the question of whether or not 53a anomalous resistivity ' js important in the equatorial E region. Auroral zone. Turning . now briefly to the auroral zone E region, we find that our understanding of the plasma processes there is still in quite a primitive state. Measurements of the Doppler spectra of VHP radar signals scattered from 8 auroral zone E region plasma waves were first obtained in the 1950's ,

1134 4 17 32-34 and recently a group from NOAA has been actively pursuing this sort of work ' ' Sometimes the radar spectral data look very similar to those obtained at the magnetic equator; it seems certain that the instability mechanisms are related. The- auroral effects are much more variable than those at the equator, • however, and the geometry is more difficult to untangle. Some spectra do not resemble anything seen at the equator. Furthermore, in contrast to the equatorial situation, in the auroral zone it is often difficult to know exactly what the electrojet current is doing. The theoretical situation is complicated by the fact that the magnetic field is nearly vertical, which means that differing altitudes with substantially differing plasma parameters (particularly the important collision to Larmor frequency ratios) are strongly coupled together. At the equator the altitudes are uncoupled. On the other hand, the induced polarization field that is an important factor in enhancing the equatorial current (and which may therefore perhaps be involved in the nonlinear saturation mechanisms) is not important in the auroral case. My feeling, which is not shared by everyone, is that a reasonable under- standing of the E-region auroral zone plasma instabilities will only be achieved after the equatorial zone processes are more fully understood. In the meantime, though, the radar, rocket, and satellite observational programs should be continued, and combined wherever possible. These coordinated programs are even more essential in the auroral than in the equatorial case, since in the auoral medium we often have only a poor idea of the ambient conditions in which the instability is developing. All-sky photographs, satellite photographs, magnetometer data, satellite and rocket electric field and density measurements, together with radar data all contribute useful information to the total picture. 53a Again, the question of what role anomalous resistivity plays in controlling the flow of current in the auroral zone is an important one that needs to be answered.

1135 4. EQUATORIAL F-REGION INSTABILITIES The fact that radio waves are often strongly scattered by irregularities in the nighttime equatorial ionosphere at altitudes of the order of 300 km or more was discussed as long ago as 1938 by Booker and Wells. These irregularities also cause the unusually strong radio star and satellite signal scintillations observed at equatorial latitudes. Somewhat unexpectedly, strong fading of satellite transmissions sometimes occurs at frequencies as high as 4 and 6 GHz , a result that must be considered in the design of communications systems. During the 30 some years following Booker and Wells' work the morphology of these irregularities was extensively studied and the propagation phenomena (how the irregularities might produce the observed radio effects, given that the irregularities exist) were reasonably well sorted out, but little progress was made in finding the source of the irregularities. Farley et 25 al. reviewed most of the ideas that had been proposed up to that date and pointed out that none of them could explain the various observations, particularly those of VHP scatter from short wavelength (3m) irregularities over a very wide range of altitudes, including regions where the mean vertical electron density gradient was positive, negative, or zero. This insensitivity to the density gradient seemed at the time to rule out gravitational instabilities as a cause. Haerendel ' ' pointed out that in any such theory one must really consider the total ionization within a magnetic flux tube, taking into account the curvature of the lines of force near the equator, and that this total content can have an upward gradient even above the region of maximum electron density in the F-region. Still, however, this idea could not explain the observations of irregularities as high as 1000 km or more. Until very recently, this F-region problem seemed particularly baffling because the equatorial F region is very well described and its relatively simple basic chemistry and dynamics are well understood. There are no significant currents flowing and apparently no important sources of free energy available

1136 except those involving the mean density gradients. How then to account for the instabilities in regions of positive, negative, and zero vertical gradient? The r • resolution of this apparent dilemma seems quite simple in retrospect. The basic idea is that a Rayleigh-Taylor instability first develops in the region of the sharp density gradient on the bottomside of the nighttime F region; this grows to large amplitudes and "bubbles" (i.e., regions of low density) eventually detach themselves and move upward into regions of increasing density. Unless the "bubbles" break up or are otherwise dissipated, they can continue to rise through the entire F region until they reach an altitude where their density is equal to the ambient background, an altitude which could be as high as 1000 km or more. Smaller scale irregularities are presumed to form on the sharp edges of the bubbles. These ideas, which a^e now generally accepted, emerged at the 1975 Gordon Conference on Space Plasma Physics, as a result of talks by M.C. Kelley and R.F. Woodman, and they are described in publications by Kelley et al. and Woodman and La Hoz . Figure 8, taken from the paper of Woodman and La Hoz, schematically illustrates the mechanism, using a three-fluid model in place of a continuous distribution of densities. The middle fluid is assumed to be the heaviest, and the bottom fluid the lightest. Thus the lower interface is gravitationally unstable, but the upper interface initially stable. When the bubble reaches the upper interface, however, it is lighter than the upper fluid; hence the instability continues and the bubble continues to rise. Recently the first numerical simulations of this Rayleigh-Taylor "bubble" f\*y mechanism have been carried out by Scannapieco and Ossakow , and their results, which are illustrated in Figure 9, seem to confirm the model, although the growth times shown are too long. Faster growth rates would have been obtained if the bottomside of the anbiant F region had been placed at a higher altitude Cas is observed^ where the ion-neutral collision frequency is lower.

1137 There were a number of new experimental observations in the 1970's that led to these new insights. The satellite observations reported ' ' 39 by Hanson and his colleagues "' showed that very substantial dis- continuities in electron density do exist in the nighttime equatorial F region. The "holes" or "bubbles" are quite apparent in the data. Spectral analysis of the observations showed that the irregularities spanned a range of wavelengths extending at least from tens of kilometers to hundreds of meters, _2 with roughly a k spectral power law, suggesting some sort of turbulent decay • /• 0 process. New radar observing techniques developed by Woodman and La Hoz and used for the 50 MHz studies in Peru provided dramatic pictures of the spatial and temporal development of the unstable region, at least for the short wave- length irregularities. An example of such a display is given in Figure 10. What is shown is echo strength obtained using a vertically directed radar and plotted as a function of time and altitude. Increasing darkness indicates increasing echo strength, on a logarithmic scale. Much of the detail unfortunately is difficult to reproduce; the original computer outputs and photographs cover about 50 dB of dynamic range. Nevertheless, the rising "plumes" of irregularities and the complicated structure associated with the development of the instability are well illustrated in the figure. Another stimulus to the resolution of this equatorial problem has been the series of barium cloud release experiments carried out in the ionosphere over the last decade. We will not discuss these here in any detail, but the observed phenomena have similarities to the equatorial processes we have been considering. The barium clouds often develop striations that are highly aligned with the magnetic field and frequently break up into quite small scale sizes. Analytical work and numerical simulations have shown that the primary cause of the striations is a gradient-drift instability operating at long wavelengths, with subsequent 43 decay to shorter scale sizes. The observations reported by Kelley et al. included a barium cloud release, the fortuitous observation of a natural "bubble"

1138 by the same rocket that released the barium, and radar observations that tracked the motion of the cloud and the natural irregularities. As a result of these various recent developments, it appears that at long last the general nature of the important instability processes operating in the equatorial F region is now largely understood, but many of the details are not. In particular, the mechanism for generating irregularities with wavelengths shorter than the ion Larmor radius still needs to be worked out. Hudson and Kennel have suggested that a collisional drift mode instability may be responsible, but more work needs to be done. As in the case of the E-region instabilities, more numerical simulation studies are planned which should help to guide future analytic theories. We have not even begun to try to sort out and interpret the complex velocity distributions within the unstable region 68 revealed by the 50 MHz ra-"ar spectral measurements of Woodman and La Hoz These Doppler spectra are sometimes simple and sometimes complex, and can vary rapidly with both altitude and time. In addition, an "explosive" instability with growth and decay times of tens of milliseconds is sometimes seen. What are the next steps to be taken in resolving the many remaining problems? First of all, a series of coordinated equatorial measurements is being planned for Peru, where the large Jicamarca radar is located near a rocket range and satellite tracking facilities. The plan is to measure satellite scintillations on as many frequencies as possible (using especially the "wideband" satellite) to study the long wavelength modes, use rocket probes (and passing satellite probes, as available) to make Jin 4<itu. measurements of the background and per- turbation densities and electric fields, and use the radar (perhaps with several beams) to map and study the regions of short wavelength instability. The radar will also be used as an on-line monitor to guide the timing of the rocket launches. Another idea currently in the planning state is to create an artificial "hole" or "bubble" in the nighttime F region over Peru by means of a rocket release

1139 of H2 or water. The development of this artificial bubble would then be monitored by radar and perhaps additional rockets. The most ambitious and com- prehensive experimental program now in the planning stage is the proposed Equatorial Ionospheric Irregularity Study Satellite (EQUION). This satellite will be instrumented to measure plasma density, temperature, and composition, energetic particle fluxes, optical emissions, and electrostatic and electro- magnetic waves. The electron density and temperature measurements will have unusually good spatial resolution (^3m) in order to permit quantitative com- parison with 50 MHz radar observations in Peru. This satellite, if funded, should make very important contributions to our understanding of the F-region instabilities. On the theoretical front, more simulation work is needed, more effort is needed in the short wavelength regime, where now very little is really understood, and possible nonlinear saturation mechanisms need to be explored. Understanding the behavior of the k-spectrum of the irregularities is particularly important for understanding and predicting satellite scintillations and their dependence upon operating frequency (see Rufenach , for example). There is an additional aspect of equatorial ionospheric research that is perhaps worth mentioning. The magnetic equator passes through developing countries (with the possible exception of Brazil), and hence research in these countries has the benefit of bringing local scientists into the mainstream of research in geophysics and plasma physics. Peruvian and Indian scientists, in particular, have made substantial contributions to the E-and F-region equatorial programs. 5. OTHER TOPICS There are several additional topics that we should at least mention here. We have briefly alluded earlier to parametric instablities in the iono- 54 sphere induced by high power HF transmitters (see Perkins et al. for a comprehensive discussion) and to instabilities observed following barium cloud releases in the ionosphere (see Goldman et al. for recent work on simulating

1140 the development of striations and for references to earlier work). Ott and Farley pointed out that the anisotropic ion velocity distributions that can exist in the aurdr'al ionosphere, due to ion-neutral charge exhange collisions combined with a large ExB drift velocity, could be unstable to the Post-Rosenbluth instability. Recent satellite observations seem consistent with this prediction. Another important area of plasma research is concerned with high latitude processes which affect the ionosphere. Since these are really magnetospheric processes they will be discussed in detail elsewhere. Within a few thousand kilometers of the Earth's surface there is a complex and exciting region of plasma activity. Extremely strong radio emission occurs there with a power output 9 of the order of 10 watts, rivaling Jupiter as a radio source. The region is associated with auroral generation, and recent satellite observations and barium shaped charge releases indicate electric field components parallel to the magnetic field resulting in potential drops of many kilovolts. These releases, which have given graphic proof of the frozen~in field concept, have thus also revealed the. existence of decoupling mechanisms. Field aligned currents associated with the aurora are certainly related to these potential drops and associated regions of anomalous resistivity. In &*Jn ionospheric observations at high latitudes have shown that ion cyclotron waves are generated in regions of intense parallel 41 currents . Velocity shear instabilities have also been observed in the 42,45,67 F-region near auroral arcs 6. SUMMARY The ionosphere, besides being an interesting part of our environment which plays an important role in worldwide communications, is a convenient place to study certain aspects of plasma phsyics. In the last 15 years or so, there have

1141 been a number of fruitful interactions between ionospheric plasma research and plasma research originating in other areas. The technique of incoherent scatter t • has become an extremely powerful tool for ionospheric research because linear kinetic plasma theory has been able to successfully explain and interpret every aspect of the observations in terms of the ionospheric parameters. This success has of course also helped to put the kinetic theory itself on very firm quantitative grounds. We have seen that there are a number of interesting plasma instabilities in the ionosphere which lead to the development of turbulent conditions which persist for minutes or eveu hours in a reasonably steady state. This persistence and the absence of wall effects make the phenomena in some respects easier to study than comparable laboratory effects. If we can achieve a full understanding of the ionospheric instabilities and resulting turbulence, there will be in- evitable applications to other areas of plasma physics. Radar and -in 6-Ltu. studies of the ionospheric instabilities have until now, for the most part, been isolated from one another. Future work will hopefully involve more coordination, now that this is becoming feasible, due for example to the construction of rocket facilities in Peru. The future should see an increasing use of active experiments in the ionosphere. Some perturbing with high powered radio waves has been done, barium releases have been carried out, and some hydrogen releases are being tentatively planned, but with the advent of the Space Shuttle we should expect to see a great deal more activity on this front. The linear theories of the primary instabilities in the equatorial E and F regions seem to be in reasonably good shape, but these can only begin to account for the observations. Recent efforts to simulate the instabilities with computers have given some very encouraging results and will be actively pursued in the future. Hopefully these will lead to a better understanding of the important nonlinear limiting processes which determine the final saturation properties of

1142 the turbulence. Other processes not considered in detail here, such as parametric instabilities and high latitude phenomena, particularly those associated with the aurora, also hold out exciting prospects for progress in plasma physics in the next decade.

1143 10 FEB. 12.40 1972 l§84 7T4 SBt 342 km FIGURE 1 Incoherent scatter auto-correlation function measurements made at Arecibo. The pulse length used was 2 ms and the integration time was 20 min. The experimental and fItted theoretical curves are plotted together, but the agreement is so good that the two often cannot be distinguished (after Hagen and Hsu37).

1144 I11 o E I- U iii Q. ION COMPONENT 3- 30-1 20- 10' NOV 8, 1975 09:00 1o 14:00 AST 1460 km to 2486 km J L 0 ELECTRON COMPONENT 4301 ... Dalo — Theory 2- 429.9 430.0 / x / / ti / / a= .86 1.05 Te CK)= 30001150 ne (em-3)= 6100+400 ' \ I ERROR BAR \ ( 2cr ) ° ; ; .•.".". V. .. -. 4300 432.0 FREQUENCY (MHz) 434.0 FIGURE 2 Incoherent scatter data from the ionosphere at Arecibo showing, for the first time, the elec- tronic as well as ionic portions of the spectrum. Due to the wide bandwidth and resulting low signal-to- noise ratio, a great deal of averaging in both time and space is required (after Hagen and Behnke36). f =4992MHz POWER I •• 70° Wesl Mh 40m -?00 -100 0 100 20O DOPPLER SHIFT (Hz) -20O -I00 0 IOO 20O DOPPLER SHIFT( Hz) Verlical •20O POWER -IO0 0 IOO DOPPLER SHIFT( Hz) 200 POWER It 45* Wesl ll>M8m ,2hoem -2OO -IOO 0 IOO OOPPLER SHIFT(Hi) 200 -200 -IOO 0 IOO 200 DOPPLER SHIFT lHi) FIGURE 3 A series of power spectra of signals scattered from the equatorial electrojet near noon at Jica- marca, Peru at various angles east and west of vertical. The spectra are all normalized to unity maximum value and correspond to a relatively strong electrojet (from Cohen and Bowlesi4).

1145 I6 Moy , I969 i50 VMZ i544 I00 200 300 400 500 Phose Velocily l m/' ) 200 300 400 500 FIGURE 4 Doppler spectra from the equatorial electrojet obtained nearly simultaneously at Jicamarca at three radar frequencies. The mean local time of each 2.5 min integration is at the left of each curve. The phase velocity, rather than the Doppler shift, of the wave is plotted, and the spectra are all normalized as in Figure 3, although the maximum is not always shown. The spike that appears at the origin on several of the curves (and occasionally determines the normalization; e.g., 1416 at 146 MHz, although the spike is not shown) is due to detector bias and should be ignored (from Balsley and Farley2).

1146 Km I09- I0h56m26* ^k . JICAMARCA 15 JAN 1973 At = 5.2 sec ."k, ^ -* ;/. a: Q. O UJ > UJ cr -1. B '*vX X.. V. I0h 53*57* "V_ . _A , ^ .- , v. 103 V .. .• .-S\ /, . .:V ': J* X. I01.5 N. /. S. I00 .A. . -120 120 -I20 I20 DOPPLER SHIFT (HZ) FIGURE 5 Power spectra from vertically propagating 3 m electrojet irregularities. The scattering volume extended about 1.5 km in the horizontal direction and 0.8 km in the vertical. The integration time for each spectrum was 5.2 s, corresponding to an average of 32 spectral samples (from Fejer et al.28).

1147 t = .64s = l.28s t* 2.70s FIGURE 6 Computer simulation of the development of equatorial electrojet irregularities. The grid spac- ing (tic marks) is 1.5 m in both the vertical and horizontal directions and B is normal to the plane of the figure. Isodensity contours spaced by 2.5% of the ambient density are shown at four selected times (from McDonald etal.").

1148 i.o 3 JCI 5 05 -5 I0 I5 20 25 H2 FIGURE 7 Comparison between a radar Doppler spectrum derived from a computer simulation (a) and from measurements at Jicamarca (b) simi- lar to those shown in Figure 4 (from Keskinen et al.44). FIGURE 8 Schematic representation of a three density model of the ionosphere in which the mid- dle layer is the heaviest and the bottom layer the lightest. The bubble forms at the lower interface and propagates through the top layers (from Woodman and La Hoz68).

1149 00 t»o 4 SO 400 0* —I— 10 II tO T 1— T-j.ooom -10 0 tO OS 10 IS 10 titan) MO tso 4*0 T-S.OOOMC -to o OS 10 so to 350 100 '•lO.OOOJfC -20 0 (km) to FIGURE 9 Contour plots of 6n/n0 showing the simulation of the growth of the equatorial F region Rayleigh-Taylor instability and the rising of a "bubble." The growth rate here is unrealistically small, due to the choice of initial density profile, shown by the dashed line (after Scannapieco and Ossakow62).

1150 ! I I N £ J T3 O I O -3 « V •a c O 3 SO § 'Eb e u. 2 ea 3 CT o I f» o E I a. o o V a -a O o O

1151 REFERENCES 1. Baker, W.6., and D.F. Martyn, Electric currents in the ionosphere. I. Conductivities, Phil. Trans. R. Soc. London A, 246, 281-294, 1953. 2. Balsley, B.B., and D.T. Farley, Radar studies of the equatorial electrojet at three frequencies, J. Geophys. Res., 76, 8341-8351, 1971. 3. Balsley, B.B., G. Haerendel, and R.A. Greenwald, Equatorial spread F: Recent observations and a new interpretation, J. Geophys. Res., 77, 5625-5628, 1972. 4. Balsley, B.B., W.L. Ecklund, and R.A. Greenwald, VHF Doppler spectra of radar echoes associated with a visual auroral form: Observations and implications, J. Geophys. Res.. 78, 1681-1687, 1973. 5. Bauer, P., Theory of waves incoherently scattered, Phil. Trans. R. Soc. Lond. A^ 280, 167-191, 1975. 6. Behnke, R.A., and J.B. Hagen, Incoherent scattering of radio waves by whistler mode oscillations in the ionosphere, submitted to Radio Sci., 1977. 7. Booker, H.G., and H.W. Wells, Scattering of radio waves by the F region of the ionosphere, J. Geophys. Res.. 43, 249, 1938. 8. Bowles, K.L., Doppler shifted radio echoes from aurora, J. Geophys. Res., 59, 553-555, 1954. 9. Bowles, K.L., Observations of vertical incidence scatter from the ionosphere at 41 Me/sec, Phys. Rev. Lett.. 1, 454-455, 1958. 10. Bowles, K.L., B.B. Balsley, and R. Cohen, Field-aligned E-region irregularities identified with acoustic plasma waves, J. Geophys. Res., 68, 2485-2501, 1963. 11. Buneman, 0., Excitation of field aligned sound waves by electron streams, Phys. Rev. Lett.. 10, 285-287, 1963. 12. Clemmow, P.C., and J.P. Dougherty, Electrodynamics of Particles and Plasmas, Addison-Wesley, 1969. 13. Cohen, R., and K.L. Bowles, Ionospheric VHF scattering near the magnetic equator during the International Geophysical Year, J. Res. NBS, 67D, 459-480, 1963. 14. Cohen, R., and K.L. Bowles, Secondary irregularities in the equatorial electrojet, J. Geophys. Res.. 72, 885-894, 1967. 15. Denisse, J.F., C.R. Acad. Sc. Paris. 253. 1539, 1961. 16. Dougherty, J.P., and D.T. Farley, A theory of incoherent scattering of radio waves by a plasma, Proc. Roy. Soc. (London), A259, 79-99, 1960. 17. Ecklund, W.L., B.B. Balsley, and R.A. Greenwald, Crossed beam measurements of the diffuse radar aurora, J. Geophys. Res., 80, 1805-1809, 1975. 18. Evans, J.V., Theory and practice of ionosphere study by Thomson scatter radar, Proc. IEEE. 5_7, 496-530, 1969.

1152 19. Evans, J.V., The upper atmosphere observatory, Science, 176, 463-473, 1972. 20. Evans, J.V., Some post-war developments in ground based radiowave sounding of the ionosphere, J. Atmos. Terrest. Phys., 36_, 2185-2234, 1974. 21. Evans, J.V., High power radar studies of the ionosphere, Proc. IEEE, 63. 1636-1650, 1975. 22. Farley, D.T., A. plasma instability resulting in field-aligned irregularities in the ionosphere, J. Geophys. Res.. 68, 6083-6097, 1963. 23. Farley, D.T., Radio wave scattering from the ionosphere, Ch. 14 in Methods of Experimental Physics (R.H. Lovberg and H.R. Griem, eds.) 9B, Academic Press, 1971. 24. Farley, D.T., Irregularities in the equatorial ionosphere: The Berkner Symposium, Rev. Geophys. Space Phys., 12, 285-289, 1974. 24a. Farley, D.T., and B.B. Balsley, Instabilities in the equatorial electrojet, J. Geophys. Res., 78, 227-239, 1973. 25. Farley, D.T., B.B. Balsley, R.F. Woodman, and J.P. McClure, Equatorial spread F; Implications of VHF radar observations, J. Geophys. Res., 75, 7199-7216, 1970. 26. Fejer, B.G., D.T. Farley, B.B. Balsley, and R.F. Woodman, Oblique VHF radar spectral studies of the equatorial electrojet, J. Geophys. Res., 80. 1307-1312, 1975. 27. Fejer, B.G., D.T. Farley, B.B. Balsley, and R.F. Woodman, Vertical structure of the VHF backscattering region in the equatorial electrojet and the gradient-drift instability, J. Geophys. Res., 80_, 1313-1324, 1975. 28. Fejer, B.G., D.T. Farley, B.B. Balsley, and R.F. Woodman, Radar observations of two-dimensional turbulence in the equatorial electrojet, 2, J. Geophys. Res.. 81., 130-134, 1976. 29. Fejer, J., Scattering of radio waves by an ionized gas in thermal equilibrium, Can. J. Phys., 38, 1114-1133, 1960. 30. Goldman, S.R., L. Baker, S.L. Ossakow, and A.J. Scannapieco, Striation formation associated with barium clouds in an inhomogeneous ionosphere, J. Geophys. Res., 81, 5097-5113, 1976. 31. Gordon, W.E., Incoherent scattering of radio waves by free electrons with application to space exploration by radar, Proc. IRE. 46, 1824-1829, 1958. 32. Greenwald, R.A., W.L. Ecklund, and B.B. Balsley, Auroral currents, irregularities, and luminosity, J. Geophys. Res.. 78, 8193-8203, 1973. 33. Greenwald, R.A., W.L. Ecklund, and B.B. Balsley, Diffuse radar aurora: Spectral observations of non-two-stream irregularities, J. Geophys. Res., 80, 131-139, 1975. 34. Greenwald, R.A., W.L. Ecklund, and B.B. Balsley, Radar observations of auroral electrojet currents, J. Geophys. Res., 80, 3635-3641, 1975. 35. Haerendel, G., Theory of equatorial spread F, preprint, Max-Planck-Institute fur Physik und Astrophysik, 1974.

1153 36. Hagen, J.B., and R.A. Behnke, Detection of the electron component of the spectrum in incoherent scatter of radio waves by the ionosphere, J. Geophys. Res., 81., 3441-3443, 1976. 37. Hagen, J.B., and P.Y. Hsu, The structure of the protonosphere above Arecibo, J. Geophys.. Res., 79, 4269-4275, 1974. 38. Hagfors, T., Density fluctuation in a plasma in a magnetic field with applications to the ionosphere, J. Geophys. Res.. 66, 1699-1712, 1961. 39. Hanson, W.B., J.P. McClure, and D.L. Sterling, On the cause of equatorial spread F, J. Geophys. Res., 78, 2353-2356, 1973. 40. Hudson, M.K., and C.F. Kennel, Linear theory of equatorial spread F, J. Geophys. Res.. 80, 4581-4590, 1975. 41. Kelley, M.C., E.A. Bering, and F.S. Mozer, Evidence that the electrostatic ion cyclotron instability is saturated by ion heating, Phys. Fluids, 18. 1590, 1975. 42. Kelley, M.C. and C.W. Carlson, Observations of intense velocity shear and associated electrostatic waves near an auroral arc, J. Geophys. Res., 82, 2343-2348, 1977. 43. Kelley, M.C., G. Haerendel, H. Kappler, A. Valenzuela, B.B. Balsley, D.A. Carter, W.L. Ecklund, C.W. Carlson, B. Hausler, and R. Torbert, Evidence for a Rayleigh- Taylor type instability and upwelling of depleted density regions during equatorial spread F, Geophys. Res. Lett., 3, 448-450, 1976. 44. Keskinen, M., R.L. Ferch, and R.N. Sudan, Power spectrum studies of numerical simulation of ionospheric gradient drift turbulence, paper presented at meeting of American Geophysical Union, June 1977. 45. Kinter, P.M., Observations of velocity shear driven plasma turbulence, J. Geophys. Res.. 81, 5114-5122, 1976. 46. Knox, F.B., A contribution to the theory of the production of field-aligned ionization irregularities in the equatorial electrojet, J. Atmos. Terrest. Phys.. 26. 239-249, 1964. 47. Landau, L.D., On the vibrations of the electronic plasma, J. Phys. USSR, 10, 25-34, 1946. 48. Lee, K., C.F. Kennel, and F.V. Coroniti, On the marginally stable saturation spectrum of unstable type 1 equatorial electrojet irregularities, J. Geophys. Res., 7^, 249-266, 1974. 49. Maeda, K., T. Tsuda, and H. Maeda, Theoretical interpretation of the equatorial sporadic E layers, Rpt. lonos. Sp. Res. Japan. 17, 147-159, 1963. 50. Matsushita, S., Intense E ionization near the magnetic equator, J. Geomag. Geo., 1, 44-46, 1951. 51. McDonald, B.E., T.P. Coffey, S. Ossakow, and R.N. Sudan, Preliminary report of numerical simulation of type 2 irregularities in the equatorial electrojet, J. Geophys. Res.. 79, 2551-2554, 1974. 52. McDonald, B.E., T.P. Coffey, S.L. Ossakow, and R.N. Sudan, Numerical studies of type 2 equatorial electrojet irregularity development, Radio Sci., 10, 247-254, 1975.

1154 53. Ott, E., and D.T. Farley, Microinstabilities and the production of short-wavelength irregularities in the auroral F region, J. Geophys. Res., 80, 4599-4602, 1975. 53a. Papadopoulis, K., A review of anomalous resistivity for the ionosphere, Rev. Geophys. Space Phys.. 15., 113-127, 1977. 54. Perkins, F.W., C. Oberman, and E.J. Valeo, Parametric instabilities and ionospheric modification, J. Geophys. Res.. 79_, 1478-1496, 1974. 55. Post, R.F., and M.N. Rosenbluth, Electrostatic instabilities in finite mirror confined plasmas, Phys. Fluids, 9_, 730-749, 1966. 56. Register, A., and E. Jamin, Two-dimensional nonlinear processes associated with "type 1" irregularities in the equatorial electrojet, J. Geophys. Res., 80, 1820-1828, 1975. 57. Rufenach, C.L., Ionospheric scintillation by a random phase screen: Spectral approach, Radio Sci.. 10, 155-165, 1975. 58. Salpeter, E.E., Electron density fluctuation in a plasma, Phys. Rev., 120, 1528-1535, 1960. 59. Sato, T., Unified theory of type 1 and 2 irregularities in the equatorial electrojet, J. Geophys. Res.. 78, 2232-2243, 1973. 60. Sato, T., On mechanisms governing the electrojet plasma instabilities, J. Geophys. Res., 8JU 539-546, 1976. 61. Sato, T., and T. Tsuda, Computer study on nonlinear cross-field instability, Phys. Fluids, K), 1262-1268, 1967. 62. Scannapeico, A.J., and S.I. Ossakow, Nonlinear equatorial spread F, Geophys. Res. Letts., £, 451-454, 1976. 63. Simon, A., Instability of a partially ionized plasma in crossed electric and magnetic fields, Phys. Fluids. 6, 382-388, 1963. 64. Sudan, R.N., J. Akinrimisi, and D.T. Farley, Generation of small-scale irregularitiei in the equatorial electrojet, J. Geophys. Res.. 7JJ, 240-248, 1973. 65. Taur, R.R., Ionospheric scintillations at 4 and 6 GHz, Comsat Tech Rev., 3, 145-163, 1973. 66. Ungstrup, E., Observation of the Post-Rosenbluth instability in the polar cap F region, paper presented at meeting of the American Geophysical Union, December 1976. 67. Webster, H.F., and T.J. Hallinan, Instabilities in charge sheets and their possible occurrence in aurora, Radio Sci., 8_, 475, 1973. 68. Woodman, R.F. and C. La Hoz, Radar observations of F region equatorial irregularities, J. Geophys. Res., 8JU 5447-5466, 1976.