IONOSPHERE AND ATMOSPHERE RESEARCH WITH RADARS

IONOSPHERE AND ATMOSPHERE RESEARCH WITH RADARS Jürgen Röttger, Max-Planck-Institut, Lindau, Germany published in UNESCO Encyclopedia of Life Support Systems (EOLSS), Geophysics and Geochemistry, 6.16.5.3, Paris, 2004. Keywords Ionosphere, middle atmosphere, thermosphere, ionosonde, MST radar, incoherent scatter radar, meteor radar, MF radar, coherent scatter, incoherent scatter, reflection, interferometry, Doppler shift, instabilities, irregularities, turbulence, gravity waves, traveling ionospheric disturbances, ionospheric heating, electro jet, plasma drift, spread-f Contents 1. Introduction 2. Radar Observation Principles 3. Mesosphere, Lower Thermosphere and Meteor Observations 4. Studies of the Mesosphere with MST Radars 4.1 D-region irregularities 5. Vertical Profiling of the Ionosphere with Ionosondes 6. Ionosphere Modifications 7. Oblique Incidence Ionospheric Sounding 8. Coherent Scatter Observations of E- and F-Region Irregularities 8.1 E-region 8.1.1 Auroral latitudes 8.1.2 Middle latitudes 8.1.3 Equatorial latitudes 8.2 F-Region 8.2.1 High latitudes 8.2.2 Middle latitudes 8.2.3 Equatorial latitudes 9. Ionospheric Profiling with Incoherent Scatter Radars 10. Sounding of the Topside Ionosphere and Magnetosphere 11. Conclusion Glossary EISCAT: European Incoherent Scatter (Scientific Association) facility in Scandinavia. FAI: Field Aligned Irregularities elongated along geomagnetic field lines. FDI: Frequency Domain Interferometry. IRI: International Reference Ionosphere. IS: Incoherent Scatter radar. MLT: Mesosphere Lower Thermosphere radar. MST: Mesosphere Stratosphere Troposphere radar. MU: Middle and Upper atmosphere radar located near Kyoto, Japan. MUF: Maximum Usable Frequency for a given radio path. NLC: Noctilucent Clouds, silver clouds observable at high latitudes in summer below the mesopause. PMSE: Polar Mesosphere Summer Echoes, radar echoes observable at high latitudes in 1

summer below the mesopause. SDI: Spatial Domain Interferometry. STARE: Scandinavian Twin Auroral Radar Experiment. SuperDARN: Super Dual Auroral Radar Network located at high latitudes. TID: Travelling Ionospheric Disturbances in the ionosphere excited by gravity waves. LF: Low Frequency, 0.03 0.30 MHz. MF: Medium Frequency, 0.30 3.00 MHz. HF: High Frequency, 3.00 30.0 MHz. VHF: Very High Frequency, 30.0-300.0 MHz. UHF: Ultra High Frequency, 300.0 3000.0 MHz. Summary We describe the use of radars for the studies of the Earth s ionosphere and atmosphere. In particular, these techniques are most useful for continuous observations in certain regions, which is not possible with other techniques such as rockets or satellites. It is obvious that the combination of these radar observations with in-situ and remote sensing methods with spacecraft, combined with models and simulations, yield a most complete view of our Earth s middle and upper atmosphere and ionosphere environment. 1. Introduction Electromagnetic waves of frequencies between a few khz and some GHz are used with radio and radar methods for scientific studies of the Earth's atmosphere-ionosphere environment. The reasons for these studies are at least twofold, namely basic research to understand complex phenomena in our environment and the communication oriented applications. A major contribution to our knowledge of the ionosphere stems from contributions of radar applications, operating in this frequency range. These will be briefly summarized in this paper by particularly taking into account most recent developments of systems, methods and their capabilities for atmosphere-ionosphere research. The Earth s atmosphere and its ionized part, the ionosphere, are dispersive media determined by their refractive indices n (Fig.1). Radar and radio methods are applied on certain frequencies f in the electromagnetic frequency spectrum. The wave undergoes total or partial reflection in the ionospheric plasma when its frequency is equal to the critical or plasma frequency f N which is used with ionosondes or HF-radars. When the wave frequency is much larger than the plasma frequency, incoherent scatter from thermal motions of free electrons in the ionosphere takes place. This is called incoherent or Thomson scatter. Instabilities in the ionosphere generate plasma turbulence, i.e. a direct perturbation of the ionization structure, which cause coherent scatter on HF, VHF and UHF and is used to study the E- and F-region irregularities. Clear air turbulence in the neutral atmosphere cause small deviations of the refractive index n due to density, temperature and humidity variations, which result in tropospheric and stratospheric irregularities leading to coherent scatter as well. At mesospheric altitudes neutral air turbulence causes an indirect perturbation of the ionization (mesosphere and D-region irregularities), which in turn results in coherent scatter. This is used with the mesosphere-stratosphere-troposphere radars, operating in the VHF and HF bands. 2

Fig. 1. Schematics of the Earth s atmosphere with respect to radar observations, given by the refractive index n of its neutral and ionized partitions. Since the ionosphere is directly coupled to the neutral atmosphere and the magnetosphere, some radar systems, such as the incoherent scatter (IS) radars, are designed or optimized for such studies. Those radars and their descents are also used for studies of the neutral atmosphere, such as the mesosphere, stratosphere and troposphere (MST). The main categories of ionospheric radars are summarized in Table 1. Both, the ionosphere and the atmosphere, are partially studied with both techniques in the region where the ionosphere is strongly coupled to the neutral atmosphere, namely the D region and the mesosphere. It is admitted that these terms do not follow a unified standard definition; they just had evolved in due course of the development of these methods. We have used here the most common or perhaps acceptable terms to classify the relevant radars or radar methods. Included in these governing terms is a diversity of different techniques as we outline here. We will specify the radar applications for studies of ionospheric and atmospheric structure and dynamics, plasma physics and ionospheric electrodynamics in more detail in the subsequent sections. The Mesosphere-Lower-Thermosphere (MLT) radars, also called MF radars, encompass partial reflection, and wind-wave-turbulence measuring methods. They cover the altitude region of about 60 km to 100 km. The LF drift method can be regarded as a passive radar; it measures the drifting pattern of lower ionospheric reflection of LF radio transmitters. Meteor radars, covering 70 km to 120 km altitude, include techniques for studies of atmospheric/ionospheric as well as astrophysical parameters. HF radars include the wide variety of the ionosondes and their progressive digital evolutions to determine high-resolution ionospheric profiles and drift velocities. They are operated from the ground for studies of the D-, E- and bottomside F-region, and from satellites to observe the topside of the ionosphere. 3

Table 1. Radar techniques for ionospheric research. The term HF radar encompasses also backscatter radars for distant observations of the ionospheric profile and irregularities, imaging of ionospheric structure and dynamics and its coupling to the magnetosphere as well as the over-the-horizon radars for detection and tracking of remote targets. HF radars, such as digital ionosondes and backscatter radars for profiling, are the only systems sweeping over a large portion of the frequency spectrum, encompassing typical ranges of ionospheric critical frequencies up to some 20 MHz. The HF-CW-Doppler sounding systems, sometimes misleadingly referred to as Doppler radars, are essentially applied for studies of ionosphere and atmosphere dynamics by measuring ionospheric reflection height variations at several distant locations. Ionospheric modifications, also called ionospheric heating, are done in ranges of ionospheric critical frequencies. It highly relies on radar methods for diagnostic purposes, operating in the HF, UHF and VHF bands. They include the investigations of the modification of ionospheric background parameters as well as the development of field-aligned and artificial periodic irregularities and plasma turbulence. Heating facilities are also applied as radars for magnetospheric, mesospheric and thermospheric observations. Mesosphere-Stratosphere-Troposphere (MST) radars detect echoes from turbulence induced ionization irregularities and dusty plasma layers in the D-region/mesosphere. The studies of the stratosphere and troposphere with MST radars are not subject of this article. The term coherent scatter radars, ambiguously also called coherent radars, is used for radars observing backscatter from ionospheric irregularities, which are aligned with Earth's magnetic field and for studies of the corresponding plasma instabilities. The term ionospheric 4

irregularity is frequently used, whereas most of them result from a spectrum of plasma waves, created by the instabilities. Incoherent scatter radars, detecting backscatter from in the ionospheric plasma, are the most powerful tools for studying the full profile of many ionospheric parameters, such as electron density, electron temperature, ion temperature and plasma velocity. Also incoherent scatter is basically coherent, since it results from the radar wavenumber component of the spatial spectrum of ion-acoustic waves existent in the ionospheric plasma. Transmitters used in radar applications are almost exclusively operated in pulsed-coded (binary phase or frequency) modulations, operate at peak power levels up to a few hundred Watts, such as the MLT radars, vertically beaming HF-radars and CW-Doppler systems. Meteor radars and coherent scatter radars have transmitter powers up to some ten kilowatts, whereas some trans-horizon and high-power MST radars apply up to some hundred kilowatts and incoherent scatter radars produce some megawatts. Antennas are ranging from simple dipoles to multi-yagi phased arrays and large dishes. The MST radars are descendents of the incoherent scatter radars, which measure ionospheric parameters, in particular of the E- and F- region. Their antenna gains and transmitter powers are more than an order of magnitude larger than those of the MST radars. The optimum frequencies of these radars are given by typical scales of the governing scattering and reflection processes. As Table 1 shows, the applied frequencies range from a few MHz (MF) to some GHz (UHF). 2. Radar Observation Principles Figure 2 shows a schematic diagram representing the main observation concepts applied in MLT-, meteor-, some HF-, MST- and coherent scatter radar systems. This figure explains the concepts for overhead scattering and reflection; in principal it applies also to all other radar applications, operating in the forward scatter, backscatter or total reflection oblique modes. To cover larger areas and to measure velocity vectors, the Doppler method (Fig. 2a) is utilized, where a narrow antenna beam is steered into different directions. It is also known as the Doppler beam swinging (DBS) method. This method is always applied by the incoherent scatter radar, where the full spectrum information is measured at different, horizontally separated positions in the ionosphere. Fig. 2. The three basic methods used by coherent scatter radars to observe the atmosphere and ionosphere: Doppler beam swinging (upper), spaced antenna drift (center) and interferometer technique (from Hocking, 1997). MST radars mostly use phased array antennas. Ionosondes apply dipole, rhombic or loop antennas, and incoherent scatter radars mostly apply large high-gain dish antennas. 5

The spaced antenna method (Fig. 2b), where the changing diffraction pattern on the ground caused by backscattering from aloft is measured, is applied to deduce the horizontal velocity, shape and coherence of the scattering layers by analyzing the cross correlation or cross spectra analysis of signals received at spaced antennas. This method is used with the MLT radars and becomes standard also in MST radars. The interferometer method (Fig. 2c) is a natural extension of the spaced antenna method. It utilizes the fact that the modern radar systems are phase-coherent and the amplitude and phase (or the quadrature signal components) are measured. Phase information at separated antennas allows the determination of the angle of arrival. It yields, with better accuracy, similar parameters as the diffraction pattern method and the Doppler beam swinging method. It also allows digital beam forming and high-resolution radar imaging by employing a larger number of receiving antennas. The advantages of this method are becoming more widely known and are accepted to constitute the features of contemporary radar systems for ionospheric and atmospheric research. The interferometer method is applied in many varieties in all radar systems, covered in this article. Such mutations are radar imaging to deduce the visibility and brightness functions of scattering regions, digital beam forming (also called post-beam steering) using spaced antennas (SDI). Interferometry in the frequency domain (FDI) is often used to overcome limitations in using very short pulses to detect thin layering. The exceptions are the incoherent scatter radars where the scattering process itself prevents the required coherency of the diffraction pattern at the receiving site. Here the Doppler method (Fig. 2a) with different antenna beam pointing angles is used. Another possibility measuring velocity and anisotropy is to apply multi-static observations using transmitting and receiving antennas at separated locations, such as the EISCAT tri-static UHF incoherent scatter radar system. 3. Mesosphere, Lower Thermosphere and Meteor Observations The mesosphere-lower-thermosphere MLT radars, also called MF radars, have their origin in the early applications of LF, MF and low HF radars to analyze partial reflections from the D region by means of the differential----------------------------------------------------------------------- ----------------------------------------------------------------------------------------------------------------- ------------------------------------------------------------------------ phase and amplitude methods to deduce the electron density profile. The partial reflection of the radar waves was assumed to be from irregularities in the ionization, and it was soon recognized that this technique could be extended by measuring the mean drift and variability of these irregularities, mostly using the spaced antenna method (Fig. 2b). Since the irregularities are strongly coupled to the neutral atmosphere due to the high collision frequency between neutrals and ions, this opened a possibility to study neutral dynamics of the mesosphere and lower thermosphere. A multitude of MF radars is used all over the globe, in particular to measure the mesospheric and lower thermospheric wind field, as shown in Figure 3. Echoes from meteors, ablating in the atmosphere between about 70 km and 120 km, had created interest, since they provide a means to deduce orbital information of the originating meteoroids as well as studying the atmosphere. Radars had been applied for these purposes including more detailed studies of the interactions of the meteor with the atmosphere and ionosphere, as well as for the modification of the ambient ionosphere by the meteors, such as plasma wave excitation and effects on the incoherent scatter process in the D-region. Meteor trails are called overdense when their plasma frequency is equal to or larger than the radar 6

frequency. Here specular reflection-type echoes result. Specular-type reflection echoes occur when the radar wave vector is perpendicular to the trail. Trails are called underdense when their plasma frequency is smaller than the radar frequency. The echo strength is dominantly given by the electron density and the radius of the trail. The Doppler shift of the meteor radar echo is caused by the motion of the trail. Since the trail is carried by the neutral background wind, the Doppler shift is used to determine the wind velocity. Fig. 3. Mean zonal and meridional wind in the mesopause and lower thermosphere region observed with the LF drift method (from Jacobi et al., 1997). Similar results are obtained with MF/MLT and meteor radars. After the trail has fully developed, it begins to dissipate, essentially by ambipolar diffusion, eddy diffusion and recombination. The trail then becomes underdense resulting in nonspecular echoes. The backscattered amplitude from an underdense trail decreases exponentially with time mainly by ambipolar diffusion. The decay times are used to measure the ambipolar diffusion coefficient. It has been shown that the diffusion coefficient D is proportional to the ratio of the square of temperature T and pressure P. Thus, if either T or P is known, the other parameter can be deduced. When taking P from atmospheric models, the temperature T can be inferred from the meteor radar measurements of the diffusion coefficient. 4. Studies of the Mesosphere with MST Radars The technique of using sensitive clear air radars to investigate the upper atmosphere at altitudes from the troposphere to the mesosphere owes much of its success to the early days of ionospheric incoherent scatter observations. The lowest mesospheric height, from where MST radar echoes are returned, is given by the D-region electron density and is usually 7

around 60-70 km. At radar wavelengths of a few meters, corresponding to Bragg wavelengths of radars in the low VHF band, variations of the refractive index become very small due to viscous damping of neutral turbulence in the upper mesosphere. The largest height is around 90 km, where meter-scale irregularities, induced by neutral turbulence at the scales of a few meters, cease to exist. Above 80-90 km the incoherent scatter process would become dominant, provided that the radar is sensitive enough to detect such signals. The condition that sufficient electron density as well as neutral variations at the Bragg scales have to be present, cause the MST radar echoes from the mesosphere to be fairly variable and intermittent. Whereas these echoes are usually called "turbulence echoes" a new brand of echoes had attracted wide attention in the last decade, namely the Polar Mesosphere Summer Echoes, which are discussed in the next chapter. The basic parameters measured with MST radar are deduced from the Doppler spectrum, such as the signal power, the Doppler shift and the spectral width, and spaced antenna methods are also used for this purpose. Furthermore the anisotropy of turbulence scattering layers can be determined. These parameters are used to study turbulence itself, where two methods are applied. One relies on the dependence of the scatter cross section on turbulence intensity and the background gradient of electron density, the other on the spectral width. From these measurements the eddy diffusion coefficients can be deduced. Doppler shift or spaced antenna analysis yield estimates of the three-dimensional wind vector, which allows studies of mean winds, long-period waves, tides and gravity waves. Of particular importance is the deposition of momentum and energy by waves into the mesosphere, which strongly affect the mean mesospheric circulation. 4.1 D-region irregularities In the preceding chapter irregularities in the D region were introduced, which essentially result from neutral turbulence in the mesosphere, which we do not regard as ionospheric irregularities per se. We now briefly address peculiar D-region echoes of different origin, namely the Polar Mesosphere Summer Echoes (PMSE), which are occasionally also observed in mid-latitudes (Fig. 4). Fig. 4. Comparison of MST radar observations (upper panel) with temperature modeling (lower panel) showing the tidal modulation of radar echoes from the mid-latitude mesosphere (from Chilson et al., 1997). 8

After a long period of skeptics whether these echoes could be created by enhanced neutral turbulence, several more reasonable explanations were suggested, which base on the fact that the polar mesopause is very cold in summer (the relation to temperature can be seen in the tidal modulation of PMSE displayed in the lower panel of Fig. 4), and contains some water vapor. In the presence of meteoric dust and aerosols, this leads to the formation of heavy water cluster ions and ice particles, which affect the distribution of electrons in the ionospheric plasma. This results in particles dressed by electrons or which capture electrons and is often called a dusty plasma. The effects are depletions in electron density at a scale of some kilometer and small-scale electron density variations. The latter in particular give rise to a strong increase in the radar scatter cross section. PMSEs are regarded to result from coherent scatter from highly structured irregularities, embedded in the ionospheric plasma and partially distorted by neutral turbulence. The origin and characteristics of these irregularities is still being discussed. The relation to heavy ions, aerosols or ice particles in the mesopause region seems to be accepted. The extension of scatter from irregularities at scales, which are much smaller than scales of neutral turbulence can be explained by the extension of the inertial subrange in the electron gas in the presence of large particles (Fig. 5). These can reduce the ambipolar diffusion of electrons, which keep structures at shorter scales than neutral turbulence (the so-called Schmidt-number effect). It is also being discussed how charge accumulation on these large particles in the mesopause region can increase the scatter cross section. Fig. 5. Schematics of the spatial power density spectrum E(k) for turbulence decaying from the inertial-convective subrange through the viscous-diffusive subrange into thermal motions. The latter subrange can be extended to larger spatial wave numbers k in the ionospheric plasma during the presence of aerosols or heavy ions in the mesopause region. Coherent scatter occurs in the inertial subrange, whereas incoherent scatter results from the thermal motions beyond the viscous subrange or from spectral components of thermal motions with power density well below the turbulent fluctuations at smaller wave numbers. 9

The cross section of PMSE irregularities can be several orders of magnitude larger than that of incoherent scatter and also that of turbulence scatter in the low VHF band. Reliable estimates of the absolute scatter cross sections are difficult to obtain. It may be questioned, however, whether such estimates are useful when assuming that scattering may dominate on high radar frequencies and partial specular reflection at low radar frequencies. Rockets to measure electron, ion and neutral density as well as electric field fluctuations in PMSE layers have been used to compare in-situ with radar remote sensing observations. This has certainly increased the understanding of PMSE irregularities observed in the VHF band. However, for a finally conclusive understanding, the in-situ measurements by rockets need to be done in exactly the same volume, which the radar senses. For the purpose of an appropriate comparison, innovative radars also need to have capabilities for high spatial-resolution imaging. Coordinated observations with radar and lidar had been successful, where the lidar observes scatter from ice particles in Noctilucent Clouds (NLC) and the radar detects the coexisting plasma variations. Fig. 6 shows that PMSE are proper tracers of the layers, which are perturbed by waves and turbulence. Fig. 6. Polar Mesosphere Summer Echoes observed at 78 degrees N in the polar region. The main measured parameters, power (upper panel), vertical velocity (middle panel) and turbulent velocity (lower panel), show the typical features of thin layers, strong gravity wave oscillations, stable laminae and turbulence blobs (courtesy SOUSY-MPAe, 2000). Since the formation of these irregularities depend on temperature (Fig. 4), which varies seasonally and globally, the latitudinal variation of these mesosphere echoes as well as their potential hemispherical difference is now being studied with additional VHF radars in the Antarctic. It is assumed that long-term observations of PMSE would yield an indication of trends in mesospheric temperature and water vapor due to global climate change. 10

5. Vertical Profiling of the Ionosphere with Ionosondes Vertical sounding of the ionosphere has been the predominant and quite indispensable technique yielding a wealth of information about the ionosphere and its coupling with the upper atmosphere and the magnetosphere. The pulsed transmitters, called ionosondes, sweeping from low frequencies around 1 MHz up to some 20 MHz in the HF band had essentially been used to measure profiles of virtual reflection heights from which the true height profiles of the ionospheric plasma frequency is deduced. The theory and the basic technique, yielding time of flight, polarization, absorption and Doppler shift of signals as function of frequency, had been developed extensively. Ionosondes operated from the ground can provide just profiles up to the ionospheric electron density maximum, since total reflection of the transmitted radio wave at the critical frequency prevents the transmission into the topside ionosphere (Fig. 7). For studies of this upper ionosphere topside sounders onboard spacecrafts, also operating according to the ionosonde principle, had been employed. Receiving in a spacecraft the signal transmitted from the ground, or v.v., at frequencies beyond the critical frequency is a natural step to combine the ground-based and space-borne techniques. This method is called trans-ionospheric radio sounding and provides for instance further information on horizontal structures in the ionosphere. In the recent decades advanced techniques had been applied to measure more characteristics of the ionosphere as function of frequency such as wave polarization, Doppler frequency spectra, amplitude and phase as well as the angle of arrival. Fig. 7. The ionospheric plasma frequency as function of altitude. Ground-based ionosondes measure the virtual range up to the maximum plasma frequency, which has to be converted into real heights. The topside can only be observed by ionosondes in satellites, so-called topside sounders (from Benson et al., 1998). 11

The new advantage of the modern systems is their capability to measure the velocity as well as the angle of arrival by means of interferometry. This is usually done after the Fourier analysis, which allows determining the angle of arrival for each individual spectral component and is called Doppler sorting. In the context of sounding the mesosphere and lower thermosphere with MF radars this method is called the Imaging Doppler Interferometer (IDI). The availability of complex signals from multi-receiver-channels also allows digital beam forming, which is antenna beam steering by transforming the digitized data. All these features, providing sky map images of the ionosphere are highly required, since the ionospherically reflected signals are mostly arriving from different parts of the ionosphere due its irregularity, patchiness and horizontal gradients. The Doppler sorting technique, providing angular and velocity information, is used to deduce the vertical and horizontal velocity of the ionospheric layers. Sophisticated routines, which are based on the standard methods of ionogram analysis, are automatically applied to calculate in real-time the true-height profiles of the ionospheric plasma frequency and control their quality (Fig. 7). Skymap imaging by digital sounders are most useful for studies of ionization troughs, patches and other kind of ionospheric irregularities such as traveling ionospheric disturbances resulting from atmospheric gravity waves. The applicability of digital ionosondes to contribute notably to observations of the high latitude convection pattern and its connection to the magnetosphere has been shown in many examples. Observations in mid- and low-latitudes are of equal importance for improving the International Reference Ionosphere (IRI) and the understanding of anomalies and irregularities in the equatorial E- and F-region. 6. Ionosphere Modifications Artificial modification of the ionosphere is not a main topic of this article but will be summarized briefly. This modification is done by high power radio waves in the HF range, which cause electron heating, parametric decay and many other kind of plasma instabilities resulting in plasma waves, striations, sheets, filaments and turbulence, i.e. irregularities. These in turn scatter radar waves, which are used to diagnose the plasma processes. This modification is also called heating. The ionospheric heating uses diagnostics with other instruments such as coherent scatter and incoherent scatter radars. Another research tool using ionospheric heating in radar applications is the creation of artificial periodic inhomogeneities, which cause resonance scattering. When the heater wave (pump wave) is reflected, a standing wave in density modulation results. A probing wave, which fulfills the Bragg condition with this standing wave, will be backscattered from this pattern. As long as the density irregularities decay at long enough time scales, a probing radar wave can be transmitted after the pump wave has been switched off and will be scattered from this Bragg pattern. Analyzing the scattered radar wave can yield several ionosphere parameters such as electron density, electron and ion temperature, neutral atmosphere density and temperature, ion chemistry and vertical velocity, depending on the probed altitude region. 7. Oblique Incidence Ionospheric Sounding In Part 5 vertical sounding of the ionosphere was described. Radio waves transmitted at lower elevation angles are usually applied for point-to-point communications. This technique can also provide remote sensing information about ionospheric parameters. In Figure 8 the basic propagation modes of this technique are sketched. 12

Fig. 8. Oblique sounding of the ionosphere: Point-to-point connections via reflection in the E- or F-region. Backscatter from the ground or from ionospheric irregularities, which are aligned in direction of the Earth s magnetic field B. This ionospheric backscatter occurs when the wave vector k r of the transmitted ray is perpendicular to the Earth s magnetic field (after Milan et al., 1997) Fig. 9. Oblique incidence ionograms recorded on a 2000 km propagation path showing the maximum (MUF) and minimum usable frequencies for multiple ionospheric reflections (lower panel). The center panel shows the mono-static ground-backscatter trace out to 2000 km, which matches the maximum usable frequency in the lower panel. The upper panel shows an extension of the standard MUF by forward scatter from ionospheric irregularities (from Dieminger, 1968). 13

An obliquely transmitted wave reaches the Earth's surface after reflection in the ionosphere. Besides forward reflection a small amount of the radar wave energy is scattered due to roughness of the Earth's surface, waves on the ocean surface or other targets on the ground or in the air. The back-scattered wave, which had propagated back to the transmitter, is received at that location and yields information about the scattering properties of the remote target as well as on the passed ionosphere. Figure 9 shows early examples of oblique incidence ionograms for point-to-point propagation, ground-backscatter and forward scatter from ionospheric irregularities. Figure 10 shows ground-backscatter with enhanced traces due to traveling ionospheric disturbances (TIDs), caused by atmospheric gravity waves. The lower panel shows the corresponding modeling of the traces, which allows deduction of gravity wave parameters. Fig. 10. Ground-backscatter ionograms and corresponding simulations showing focussing due to travelling ionospheric disturbances. 8. Coherent Scatter Observations of E- and F-Region Irregularities The backscatter from field-aligned irregularities in the F- and E-region need very careful consideration of refraction effects to minimize errors in aspect sensitivity and velocity measurements, when the radar frequency is close to the plasma frequency. The direct backscatter process from the ionosphere, i.e. from field-aligned E- and F-region irregularities, falls into the category of coherent scatter. The term "coherent scatter" is commonly used, since the scatter is from structures in the ionospheric plasma density, which exhibit some temporal and spatial coherency on the Bragg scale of the probing radar. This is often seen in contrast to incoherent scatter, which is in itself coherent, too, resulting from ion-acoustic waves in the ionospheric plasma. This is normally assumed to be in thermal equilibrium. In practice coherent scatter results from irregularities, created by structural modification of, or instabilities in the ionospheric plasma, which usually is not in thermal equilibrium. Plasma irregularities are a synonym for a spectrum of plasma waves or turbulence, which are created by the plasma instabilities in the E- and F-region. Structural modifications are variations related to breaking waves and turbulence in the neutral 14

atmosphere. This is certainly the case in the mesosphere (D-region) due to the high collision frequency between ions and neutrals, but is also suggested that neutral turbulence affects the lower E region. It is also considered that currents and turbulence in mid-latitude sporadic-e layers can be induced by acoustic pulses giving rise to radar backscatter. From the radar systems point of view, we can simply say that coherent scatter has normally a much larger cross section and narrower spectral width than incoherent scatter and is mostly highly anisotropic. We also have to be aware that coherent scatter should not be confused with total reflection yielding strong signals as well. This was described in parts 5 and 7. 8.1 E-region Whereas neutral atmosphere and dusty plasma effects govern D-region irregularities, primarily electro-dynamic effects dictate irregularities in the E region. In contrast to horizontal layering of D-region irregularities essentially controlled by the neutral atmosphere, the E-region irregularities are predominantly elongated along the Earth's magnetic field. Radar observations of these irregularities are, thus, very sensitive to the pointing direction with respect to the magnetic field, i.e. the radar echo is strongest when the radar k-vector is perpendicular to the Earth's magnetic field (Fig. 8). Due to the connection of these irregularities to the direction of the Earth's magnetic field B, the instability mechanisms and the radar observation methods applied in low, middle and high latitudes have to be treated slightly different. The term irregularities most frequently refers correctly to plasma waves, created by the instabilities. The energy of these waves is distributed corresponding to spatial wave number spectra. The radar echoes result from those spectral wave components, which fulfill the Bragg scattering condition, i.e. their spatial scales have to be half the wavelength for mono-static radars. These E-region irregularities are basically created by two main plasma instabilities, the twostream instability occurring when electrons and ions drift with different velocity, and the gradient drift instability, which occurs when there is an electron density gradient component perpendicular to the drift direction. Differential drift of electrons and ions, i.e. electric currents, and gradients are caused by different mechanisms in equatorial and auroral regions. In the equatorial region polarization electric fields are created by tidal winds, giving rise to the equatorial electrojet in which irregularities grow. In auroral regions electric fields E map down from the magnetosphere and cause E x B plasma drifts, also known as the auroral electrojet. Energetic particles precipitating from the magnetosphere create the visible aurora and steep electron density gradients. Large electric fields and gradients give rise to instabilities creating irregularities, which scatter radar waves. This phenomenon was introduced as radar aurora. In mid-latitudes sporadic-e layers, which are created by wind shear due to gravity waves and tides, are characterized by strong electron density gradients, which in turn result in instabilities. Since all these instabilities depend on plasma drift, the term drift waves is frequently used to describe the scattering irregularities. Partially also the term plasma turbulence is used. Due to fairly large mean and spread velocities either high pulse repetition frequencies or, where this is not possible due to too long ranges such as in polar region, proper pulse codes, such as double or multi pulse, have to be used. 8.1.1 Auroral latitudes The high latitude E-region irregularity scatter is notably correlated with geomagnetic disturbances and the visual aurora, such that these radar echoes were even called "radio 15

Fig. 11. Spectra of the four different types of E-region irregularity coherent scatter echoes (from Fejer and Kelley, 1980). aurora" or auroral backscatter. Compared to the equator and midlatitudes there is a much larger complexity and variability in the scatter. This reflects the greater magnitude and variability of electric fields, currents and electron density due to the coupling with the magnetosphere. The radars, which had been used in high latitudes, are covering the frequency range from the HF to the high UHF band. Deduction of the full spectrum is always done as well as interferometer and bistatic measurements and comparisons of observations on different frequencies. In addition to the two types of spectra dominating the equatorial and mid-latitude E-region backscatter, two further types were later found in the auroral electrojet (Fig. 11). Type 3 characterizes a velocity lower than the ion-acoustic speed, whereas type 4 has a much larger velocity. We have to recognize that such a classification refers only to the observations of radar backscatter from the irregularities and that those spectral features are certainly affected by the radar system parameters. It was found that the Type 1, 2 and 3 echoes are mostly highly polarized, implying a small number of discrete scatterers located close to each other. Diffuse noon echoes were attributed to electric field enhancements in the vicinity of the magnetospheric cusp region, whereas discrete echoes were related to substorm features around midnight. At 50 MHz ionospheric refraction can still play a role. It was found that during the occurrence of irregularities the electron temperature was enhanced temporarily due to electron gas heating. The scattering cross section on VHF frequencies falls off with aspect angle by some 5-10 db per degree. It is also reported that the Doppler velocity decreases with increasing aspect angle, which could be due to ray spreading arising from refraction and diffraction while it could also be an inherent property of the irregularities. 16

Fig. 12. The STARE radars in Norway and Finland with the eight antenna beam directions overlapping in the auroral zone (from Nielsen et al., 1999). In northern Scandinavia a well-known system is operated near 140 MHz, namely the STARE (Scandinavian Twin Auroral Radar Experiment) radars. As Figure 12 shows, the antennareceiver configurations allow as many as eight adjacent azimuth sectors to be sampled simultaneously. The 2 x 8 receiver beams of the STARE system overlap in areas in northern Scandinavia. The combination of radial velocity measured in the common viewing areas of the twin radars yields the horizontal velocity component (assuming that the vertical velocity component is negligible). Provided that scattering irregularities exist (i.e. the electric field E exceeds the threshold for instability of some 20 mv m -1 ), this allows to measure the plasma drift in the auroral E-region and thus the electric field E, since the drift velocity is given by E x B. The measured drift velocity is limited to values close to the phase speed of ion acoustic waves, whereas the plasma drift can be larger. Figure 13 shows a typical plot of velocity vectors obtained with the new STARE radar. It indicates a swirl, which might be a sign of field-aligned currents. This shows that such STARE-type VHF radars can in a similar way be used to study E-region electrodynamics. The Millstone radar (440 MHz) and the EISCAT UHF (933 MHz) radar observations of irregularities at shorter spatial wavelength were performed in the monostatic and bistatic mode, which allow measurements at different aspect angles. In addition to coherent scatter echoes the incoherent scatter parameters of electron density, electron and ion temperature and plasma velocity (i.e. the electric field) were measured with these radars. It was shown that coherent scatter occurred at times when the electric field exceeded thresholds at the ion acoustic velocity. The scatter cross-section, extrapolated to zero aspect angle was in the order of 10-11 m -1 comparing well with other observations. A promising new approach for E-region irregularity studies was introduced in the recent years, which uses TV broadcast transmitters in the frequency range 88-108 MHz. Signals of these transmitters scattered back from irregularities can be analyzed by cross correlating with the directly received signals. First results obtained with this compact and low-cost system, also called parasitic radar system seem to be promising. 17

Fig. 13. Ionospheric flow velocities in the E-region measured with the STARE system, showing a vortex structure (from Nielsen et al., 1999). This summary of radar observations of E-region irregularities demonstrates progress in understanding but also quite obviously the complexity of these phenomena. The dependence of instability, which cause the plasma waves and irregularities, on the status of the background ionosphere and atmosphere is still to be resolved as is the unified nature and life history of the irregularities themselves. Radars covering wide frequency ranges and providing high time and space resolutions appear to remain essential tools, which need to be operated together with other diagnostic instruments, such as rockets and further ground-based optical remote sensing devices. The extensive radar observation has also created successful modeling and simulation and stimulated theoretical developments to understand the intricate plasma physics occurring in the ionospheric E region. 8.1.2 Middle latitudes Field-aligned E-region irregularities are also observed with VHF and HF radars in midlatitudes. During the past decade, several radars have been used for such mid-latitude E- region irregularity studies. Due to the Earth magnetic field geometry all these radar experiments are with obliquely pointing beam at some 30-50 degrees elevation angle. Since the scatter cross section of these echo types is moderately strong, a very high transmitter power and large antenna gain is usually not required. All the pulsed VHF radars apply spatial domain interferometry, and a few apply frequency domain interferometry as well. The observations of mid-latitude E-region irregularities proved, by means of combined ionosonde records, that most of these irregularities are closely related to sporadic-e layers. These sporadic-e layers often detected by ionosondes, are highly spread in height and azimuth caused by very irregular structures. Radar echoes obtained with VHF radars, which result from backscatter perpendicular to the magnetic field, were intensively studied by means of their Doppler spectra, multi-frequency applications, multi-beam and interferometer arrangements. Type 1 and type 2 echoes are detected (Fig. 11). They are regarded to be caused by the two-stream and the gradient-drift instability. 18

Fig. 14. Backscatter from mid-latitude E-region irregularities frequently show periodic downward moving striations, which are called quasi-periodic (QP) echoes. Using different horizontal antenna beam directions it is shown that the irregularity patches are also moving in East-West direction (from Rishbeth and Fukao, 1995). The VHF radar echoes in mid-latitudes appear particularly in the summer months and after sunset with maximum around midnight and also after sunrise. Dependence on geomagnetic activity was not proved. Continuous-type echoes occur between 90 km and 100 km height, but are not related to sporadic-e. The most striking feature is the frequently occurring quasiperiodic nature of the nighttime radar backscatter echoes (Fig. 14). These quasi-periodic (QP) echoes occur below 125 km and mostly move downward in narrow striations to at most 95 km altitude, but upward motions are observed as well. Periods are between a few minutes and about 20 minutes and their vertical motion is several meters per second. It was suggested that these periodic structures are resulting from instability depending on polarization electric fields due to altitude modulation by gravity waves with phase fronts parallel to the magnetic field. More insight into the generation mechanism is being achieved with combined experiments, using for instance radars at different frequencies and in-situ measurements with rockets. 8.1.3 Equatorial latitudes Radar observations contribute predominantly to the understanding of these E-region irregularities. In the equatorial region the Jicamarca VHF radar in Peru played a dominant role in these studies, since it is properly located close to the magnetic dip equator, allowing the antenna beam pointing almost vertical and perpendicular to the magnetic field. Its frequency 49.9 MHz is well adapted to prevalent irregularity scales and its large phased-array antenna, sub-dividable into modules, is most suitable for skillfully adapted, accurate beam pointing, incidence angle measurements and radar imaging. Figure 15 shows some examples of highly structured irregularities. 19

Fig. 15. Strong equatorial electrojet irregularities causing particular coherent scatter with amorphous structure and well-defined up- and downward moving striations. These observations were done by Swartz and Farley (1994) with the Jicamarca VHF radar with very high time and range resolution. These radar echoes are from plasma instabilities in the equatorial electrojet (EEJ). Their analyses first revealed the two main categories of echoes. Type 1 echoes comprise a large mean velocity of about the ion acoustic velocity (usually about 350-400 ms -1 ), a narrow spectral width and a large scatter cross section. They are due to the two-stream instability. Type 2 echoes show a mean velocity significantly smaller than the velocity of type 1, usually have a broad spectrum and a small cross section (see also Fig. 11). They are assumed to be due to secondary waves creating gradient-drift instabilities. There are also types 3 and 4, as found in the auroral electrojet. Interferometer applications were already introduced in Jicamarca in the very beginning to study the spatial fine structure of the irregularities and their motions. Plasma waves and turbulence in the equatorial electrojet show remarkable structures as measured with an altitude resolution better than 250 m (Fig. 15). In addition to determining the vertical structure and velocity it is possible to evaluate also horizontal motions by applying simultaneous interferometer measurements. Concentrated campaigns, combining radar and rocket observations of equatorial electrojet irregularities, had been performed in order to measure electric fields, current density, plasma waves and irregularity spectra in-situ with rockets and compare these with concurrent radar observations. 8.2 F-Region The F region is highly coupled to the E region, mainly due to the ionospheric electric field, which means there would be a certain relation of E-region to F-region irregularities. The background conditions, such as currents and electron density and their gradients are quite different, however. This causes differences in the instability mechanisms and the nature of field-aligned irregularities. The magnetic field geometry (see Fig. 8) dictates the use of HF radars for high latitude studies, since only HF waves are sufficiently bent by refraction to assure perpendicularity. At the equator vertically beaming HF and VHF radars can be used, 20

since the magnetic field is horizontal. High and low latitude field-aligned irregularities had been studied intensively by radars, whereas only little work is known for studies of midlatitude F-region irregularities. Irregularities in the upper F region, which are of different kind and related to field-aligned currents, will be discussed in Part 9 on incoherent scatter. 8.2.1 High latitudes It is assumed that short-scale irregularities in the F-region are produced by the gradient-drift interchange instability, where low density plasma is convectively mixed across mean plasma density gradients, by current-convective instabilities or by generation of electrostatic turbulence, where the limits between these processes appear not well separable. Required gradients are created by discrete electron precipitation and by larger scale plasma structures, which are known as propagating polar cap patches of electron density, or by blobs and troughs related to the auroral oval. These structures have scales of some 100 km and frequently steep gradients and polarization electric fields at their boundaries, which give rise to instability creating shorter scale irregularities. All these structures are in certain ways images of magnetospheric processes. Fig. 16. The SuperDARN system: Fans of antenna beams point into the polar cap ionosphere from where coherent backscatter from F-region irregularities is received. Each radar measures the radial velocity as function of range and azimuth. The combination of velocity fields from all radars yields the polar cap plasma convection pattern (from Lester et al., 1997). A network of radars around the polar cap (Fig. 16), called SuperDARN radars, is applied to study the coupling of the ionosphere to the magnetosphere. These radars measure the largescale ionospheric plasma motion. Typically, this flow consists of a two-cell pattern, which varies significantly due to solar wind - magnetosphere processes. SuperDARN radars are optimum instruments to monitor this plasma convection pattern. In Figure 17 an example of 21

the convection pattern is shown, which results from a spherical harmonics analysis of the composition of velocity measurements from all six northern hemisphere SuperDARN radars. A full coverage within the designated areas cannot usually be achieved because, alike other remote sensing methods, tracers are needed for these velocity measurements. The tracers are field-aligned F-region irregularities, which just do not exist at any time. It is also required that the background ionosphere allows the appropriate propagation and scattering conditions. However, the convincing results in Figure 17 prove that this is a uniquely attainable radar method to observe large areas of the polar cap convection pattern, which other methods cannot provide. A similar network with a smaller number of radars is operated around the Southern hemisphere polar cap. Fig. 17. Plasma convection pattern modeled from velocity fields (arrows) deduced from SuperDARN measurements (from Bristow et al., 1998). Detailed characteristics of Doppler spectra of high latitude F-region backscatter observed with radars are surprisingly rare as compared to E-region scatter. The SuperDARN radars usually measure the three first moments of the received backscattered signal, where particular emphasis is placed on the Doppler velocity to determine the convection pattern. To improve the understanding of the scattering process and the nature of the scattering irregularities it would be required to have the full spectrum information available as well as observations with high spatial and temporal resolution. Applying a collective scattering approach, taking into account that the radar echoes are scattered from a distributed, non-uniformly moving volume, it could be shown that the collectively scattered signal correlation time is a statistical characteristics of the fluctuating fluid motion of the scatterers. By this procedure it is possible to determine the plasma turbulent diffusion coefficient. 22

8.2.2 Middle latitudes Mid-latitude irregularities have not been a subject of intensive radar studies although irregular features are frequently observed as spread-f on vertical incidence ionograms. Such irregularities mostly occur in connection with travelling ionospheric disturbances (TIDs), which are resulting from neutral atmosphere gravity waves. Field-aligned irregularities (FAI) of this origin had also been observed with radar at high latitudes and have been studied at mid-latitudes with the MU radar in Japan, also in combination with ionosondes. Some simultaneous occurrence of sporadic-e irregularities was reported. The field-aligned F-region irregularities occur at nighttime and have scatter cross-sections of some 20 db above the incoherent scatter level. The irregularities show high Doppler velocities up to a few 100 ms -1 and occur in patches. The unique multiple beam pointing capability of the MU radar allowed to track the patches, which usually move in zonal direction. It is argued that turbulent upwelling can evolve from preexisting undulations of the bottomside F layer caused by gravity waves. The upwelling seed conditions for secondary instabilities, which create fieldaligned irregularities on the east or west wall of these upwelling patchy structures. 8.2.3 Equatorial latitudes Irregularities in the equatorial F region are termed equatorial spread-f (ESF). This name stems from the appearance of spread traces on vertical incidence ionograms, which manifest strong horizontal and vertical gradients in electron density at some ten kilometers scales resulting from ionization depletions. Shorter-scale irregularities form on these gradients, which cause a blurring of the spread ionogram traces. Most of the research, leading to improved understanding of the instability mechanisms, the irregularities and the scattering of radio waves resulted from observations with VHF radars operated close to the magnetic equator. Intensive work had also been done using trans-equatorial radiowave propagation in the HF and VHF band as well as analyzing scintillation of satellite signals and satellite and rocket measurements. Besides some campaign operations with smaller VHF radars, the main radar instruments used for studies of ESF are the 50-MHz radar at the Jicamarca Observatory in Peru and the VHF radar of the National MST Radar Facility in Gadanki, India. Besides the dominating nighttime phenomenon of ESF, HF radar measurements also showed some daytime irregularities in the F region. The most spectacular phenomenon is the occurrence of strong coherently scattering irregularities in the topside F region. This was first detected with the Jicamarca VHF radar (Fig. 18), which is sensitive to 3m-scale field-aligned irregularities, which are superimposed on the much weaker background electron density. These observations clearly show the connection of the topside spread-f with the bottom-side spread-f. Simultaneous in-situ measurements with satellites and incoherent scatter radar showed that these upward moving plumes of irregularities are occurring in depletions of electron density. Such plumes can be created by the gravitational Rayleigh-Taylor instability in the bottom-side ionosphere from where these low-density plasma bubbles move due to buoyancy to the topside. When the bubbles rise, shorter-scale secondary irregularities are formed on density gradients, which coherently scatter VHF radar waves. The motion of the bubbles was successfully simulated by numerical methods. In addition to upward motion the bubbles move zonally following the neutral wind and the E x B drift. Significant quasi-periodic perturbations are needed at the bottom-side, which are assumed to be seeded by atmospheric gravity waves as for instance seen by transequatorial HF radars experiments. The bubbles extend meridionally along magnetic fieldlines, such that a bubble rising to 1000 km altitude can have its footprints far north and south of the magnetic equator at distances close to the crests of the equatorial anomaly. 23

Fig. 18. Incoherent scatter (weak signal) from the equatorial ionosphere, which is a direct measure of the electron density, and superimposed coherent scatter (strong signal) from EEJ irregularities (lower heights) and from equatorial spread-f plasma irregularities observed with the Jicamarca VHF radar in Peru. The coherently scattering F-region irregularities move in large-scale bubbles to the topside ionosphere. The arrows determine the plasma velocity (courtesy Kudeki et al., 2002). Since the bottom-side of the post-sunset equatorial F-layer has very large vertical electron density gradients, smaller-scale irregularities can be formed here without being connected to the creation of plumes. The spectra of the VHF radar returns from the bottom-side can be very narrow (equivalent to a few 10 ms -1 ). At higher altitudes in the topside spread-f multi-peak spectra as wide as several hundred Hertz are found, which corresponds to turbulent velocities up to 1000 ms -1. It appears difficult to find unifying generation mechanisms for these 3-m irregularities. It is assumed that some of these irregularities, which occur over such wide velocity ranges, result from larger-scale structures, which are steepend due to non-linear interactions. The small-scale ESF irregularities are highly field-aligned, such that they can best be observed with vertically beaming radars at the magnetic equator. Such as for E-region irregularities, the fine structure of these F-region irregularities has been investigated by spatial interferometry. The high anisotropy, which is consistent with a large horizontal coherence in magnetic west-east direction, allows fairly long baselines, performed with the VHF radar at Jicamarca. 24

9. Ionospheric Profiling with Incoherent Scatter Radars The term incoherent scatter (IS) stands for the mechanism of scattering from electrons in the ionospheric plasma when the radar operating frequency is much larger than the plasma frequency. It stems from the initial assumption that electrons scatter incoherently due to their random thermal motion. It is also called Thomson scatter, relating this term to the scatter of all the individual electrons in the volume probed by the radar. When the Debye length of the plasma is much smaller than the radar wavelength, ions are dominantly controlling the scattering process such that it becomes quasi-coherent. This generally holds for probing the ionosphere at radar wavelength of a few meters to some ten centimeters. The scatter is actually from the spatial component in the spectrum of the ion-acoustic and electron-acoustic waves in the ionospheric plasma, which match the Bragg scatter condition. The phase velocities of these waves cause the scattered radar signals to be Doppler shifted. This leads to up- and down-shifted ion- and electron-lines in the incoherent scatter spectra. The latter are also called plasma lines, since they occur at offset frequencies approximately equal to the plasma frequency. The offset of the ion lines from the radar carrier frequency is given by the ion-acoustic velocity, which is dependent on the ion temperature. Due to Landau damping the width of the up- and down-shifted ion lines is wider than their frequency off-set. This causes an overlap of these two lines and the typical double-hump spectra of an incoherent scatter signal from the F- and E-region. In the lower E- and the D-region collisions between ions and molecules damp the ion-acoustic waves, such that the two humps merge into a single peak spectrum. In a first approximation the scattered power is proportional to the electron density in the probed volume. The electron density, which is proportional to the square of the plasma frequency, can also be deduced by measuring the offset frequencies of the plasma lines. The offset and the widths of the up- and down-shifted ion lines are given by the ion and the electron temperature and the mass of the ions. The mean shift of the total incoherent scatter spectrum with respect to the radar carrier frequency is given by the bulk velocity of the plasma in the scatter volume. Applying proper analysis routines to fit measured spectra to theoretical spectra, calculated from the universal plasma dispersion relation, thus allows to determining the main ionospheric parameters: Electron density, electron and ion temperature and ion velocity. For this purpose it is, however, required to include assumptions or models of ion mass and composition as well as the ion-neutral collision frequency. By introducing reasonable a-priori postulations or iteration procedures, multi-parameter analyses of the latter parameters can be done. Using a full-profile analysis for instance yields results as shown in Figure 19. The profiles cover the bottom-side and the top-side of the ionosphere as well, since the radar frequency is much larger than the plasma frequency (no total reflection). Another essential parameter is the electric field, which is deduced from the ion velocity measured in the F region, where collisions can be neglected. The ion drift velocity is directly proportional to the vector product of the electric field and the Earth's magnetic field, where the latter is a properly identified constant. Due to the collisions, the ion velocity in the E region is composed of the contribution by the drift due to the electric field and the neutral wind velocity. The combination of the electric field measurements at F-region heights and the ion drift in the E region is used to determine the neutral wind velocity at E-region altitudes. In the D region, where the ionospheric plasma is highly dominated by the collisions, the ion drift velocity is directly proportional to the neutral wind. 25

Fig. 19. Incoherent scatter radars measure, besides the electron density and plasma velocity, the electron and ion temperature profile in the E- and F-region (from Holt et al., 1992). The cross section of incoherent scatter is relatively small, such that transmitters with up to a few megawatt peak power and antennas with gain above 40 db need to be applied. Incoherent scatter radars (ISR) are located in Jicamarca near Lima, Peru, at Arecibo on Puerto Rico, at Millstone Hill near Boston, USA, in Sondreström on Greenland, near Tromsö, Norway, near Longyearbyen (500 MHz) on Svalbard, in Kharkov, Ukraine, and in Irkutsk, Russia. The Middle and Upper Atmosphere MU radar near Kyoto, Japan, is also used as incoherent scatter radar. The radars at Tromsö and near Longyearbyen are operated by EISCAT (EUropean Incoherent SCATter), an international association formed by six European countries and Japan. The EISCAT UHF radar is the only tri-static system; it has receivers in Tromsö, in Kiruna, Sweden, and in Sodankylä, Finland. This allows measure the three-dimensional vector velocity and the anisotropy of scattering processes. Plans exist to construct an incoherent scatter radar at the Polar Cap Observatory in Resolute Bay, Northern Canada. Its location close to the magnetic pole and the combination with the Sondreström and EISCAT radars, including the new EISCAT Svalbard Radar near Longyearbyen allows to observe a substantial portion of the northern polar cap ionosphere and its coupling with the exosphere and the magnetosphere. From the basic parameters, measured with the incoherent scatter radars, a multitude of deduced ionospheric parameters can be obtained, which usually requires further assumptions, models, independent measurements of further parameters with other instruments and 26

dedicated computational routines. Such deduced ionospheric, thermospheric or mesospheric parameters are for example: Neutral temperature, energy input and balance, E-region conductivities, precipitating particle energy spectra, recombination coefficients, field-aligned currents, ion composition and flux etc. Fig. 20 shows for instance the diurnal variation of the ion temperature and the fractional abundance of hydrogen and helium in the topside ionosphere over the Arecibo Observatory in Puerto Rica. The temporal and latitudinal variation of the electron density in the polar cap when an electron depletion, the so-called trough, moved over the EISCAT radars is shown in Figure 21. Fig. 20. Results of topside incoherent scatter measurements with the Arecibo incoherent scatter UHF radar on Puerto Rico: Diurnal variation of ion temperature, and the fractional abundance of hydrogen and helium (from Gonzales and Sulzer, 1996). Incoherent scatter radars have proved to be indispensable and powerful tools for studies of the ionosphere and its coupling and the interaction with the regions below (middle and lower atmosphere) and above it (exosphere and magnetosphere). 10. Sounding of the Topside Ionosphere and Magnetosphere It is well known that all ground-based ionosondes can only observe the ionosphere up to the height of the maximum electron density. The topside ionosphere has to be studied by other means. For this purpose ground-based incoherent scatter radars are applied as we have mentioned in the previous chapter. Also high-power HF transmitters are planned to be used for magnetospheric sounding. Such a facility is presently being constructed in the polar region of Svalbard. The other possibility is digital ionosondes in satellites orbiting above the height 27

Fig. 21. Temporal and spatial electron density variations measured with the EISCAT UHF incoherent scatter radar on Svalbard (78 degrees N), and the EISCAT VHF and UHF radar in Tromsö, Norway, providing a meridional cut over more than 10 degrees latitude (EISCAT Annual Report, 1995-1997). of the electron density maximum of the ionosphere. This radar technique of topside sounding has several advantages, such as the reduced man-made interference, since total reflection prevents reception of ground-based transmitters in the topside ionosphere. Topside sounding provides global coverage and allows studies of plasma physics in the local environment of the satellite. These studies, making use of the interaction of the radar waves with the space plasma should not be confused with in-situ space plasma probing by special satellite-borne sensors, which directly measure particles and fields in space. Basically topside sounders apply the same technical principle as ground-based ionosondes, but there are some differences. One has to consider that the sounder is in a plasma environment, which affects the antenna matching and causes plasma resonances and wave cutoffs seen in the ionogram. 28

High latitude ionosphere observations allow conclusions about magnetospheric processes and all the ionospheric radar techniques could be used for these purposes of "remote sensing" of the magnetosphere. The highly magnetized plasma and wave-particle interaction in the magnetosphere has almost exclusively been studied in-situ by spacecraft probes. A few attempts have been made to detect echoes from the magnetosphere with high power HF radars in Russia, and were assumed to be due ion-acoustic turbulence. Attempts for direct radar sounding of the plasma density, however, are scarce. Since the plasma density in the magnetosphere is as small as 10 6 m -3, low frequencies have to be used for this sounding. Radar signals at such low frequencies (VLF - MF range), transmitted from the ground, will usually be reflected by the ionosphere, unless they can enter a duct through an ionospheric trough. Consequently sounders above the ionosphere have to be applied for regular sounding. 11. Conclusion We have summarized the use of a variety radars for the studies of the Earth s ionosphere and atmosphere. In particular, these techniques are most useful for continuous observations in certain regions, which is not possible with other techniques such as rockets or satellites. It is obvious that the combination of these radar observations with in-situ and remote sensing methods with spacecraft, combined with models and simulations, yield a most complete view of our Earth s middle and upper atmosphere and ionosphere environment. This article can only provide a small overview on the most relevant methods using radar to study the Earthspace environment. More information can be obtained from a selection of other articles and books referenced in the following Bibliography. This article is the scientific extension of the article the author wrote for the issue Modern Radio Science of URSI, 1999. 29