Search

The ionosphere, embedded in and tightly coupled to the thermosphere, is strongly influenced by couplings to other geophysical regions. For example, above it, both the magnetosphere and plasmasphere greatly affect the ionosphere by the precipitation of soft and energetic particles, by heat conduction, and by fluxes of thermal particles. Below, the middle atmosphere affects it with upwardly propagating waves (gravity waves, tides, and planetary waves). All the while, polar and auroral regions greatly affect the mid-latitudes by the equatorward penetration of electric fields and winds, and by the equatorward propagation of waves (traveling ionospheric disturbances or TIDs). Exploring these couplings effectively furthers our understanding of at least the dominant processes and interactions that play such an important role in determining the character of this part of the Earths environment. Significant progress during the Solar-Terrestrial Energy Program (STEP) has demanded that the observational, analytical, and theoretical thrusts of the international scientific community be global in all senses. Observationally, this has led to coordinated measurements from many regions, from the poles to the equator, and from ground- and space-based instruments. It has also led to many different instruments, including new ones, measuring an extensive variety of (related) geophysical parameters. Depending on the instrument, measurements have been made continuously or at appropriate intervals to sample different geomagnetic conditions, and diurnal, seasonal, inter-annual, and solar-cycle variations. Extensive analyses have been carried out on these observations. New empirical models have been developed and old ones improved. Theoretical work has led to new and improved first-principles models that are being used to test our understanding of the observations. Our intent is to review this progress and suggest some future directions. Our approach is to illustrate the broad front of progress with representative examples, rather than to do an exhaustive review. We apologize from the start, to those whose good work we have not noted, in order to allow a broad balanced overview of areas advanced.

Very accurate measurements of electron density can be made at Arecibo Observatory, Puerto Rico, by applying the coded long-pulse (CLP) radar technique [Sulzer, 1986a] to plasma line echoes from daytime photoelectrons [Djuth et al., 1994]. In the lower thermosphere above Arecibo, background neutral waves couple to the ionospheric plasma, typically yielding ∼1–3% electron density “imprints” of the waves. These imprints are present in all observations made to date; they are decisively detected at 30–60 standard deviations above the “noise level” imposed by the measurement technique. Complementary analysis and modeling efforts provide strong evidence that these fluctuations are caused by internal gravity waves. Properties of the neutral waves such as their period and vertical wavelength are closely mirrored by the electron density fluctuations. Frequency spectra of the fluctuations exhibit a high-frequency cutoff consistent with calculated values of the Brunt-Vaisala frequency. Vertical half wavelengths are typically in the range 2–25 km between 115- and 160-km altitude, and the corresponding phase velocities are always directed downward. Some waves have vertical wavelengths short enough to be quenched by kinematic viscosity. In general, the observed electron density imprints are relatively “clean” in that their vertical wavelength spectrum is characteristically narrow-banded. It is estimated that perturbations in the horizontal wind field as small as 2–4 m/s can give rise to the observed electron density fluctuations. However, the required wind speed can be significantly greater depending on the orientation of the neutral wave's horizontal wave vector relative to the geomagnetic field. Limited observations with extended altitude coverage indicate that wave imprints can be detected at thermospheric heights as high as 500 km.

We report on a wide variety of thermospheric and ionospheric observations from three consecutive January World Day campaign periods. Despite remarkably similar geophysical conditions characterizing the in situ forcing of the upper atmosphere during these solar minimum campaigns, we find significant variability in the observations of the ionosphere and thermosphere particularly at low latitudes in the American sector. In addition, we present further observational evidence of the unexpected exospheric temperature suppression at low latitudes initially reported by Hagan and Salah (1988). We discuss the lower and upper atmospheric coupling mechanisms of plausible importance to the interpretation of the observed thermospheric weather patterns. We report evidence that lower thermospheric [NO] (nitric oxide number density) and upward propagating atmospheric tides affected the thermospheric energy and momentum budgets during the campaign periods.

Dayside ionosphere convection at high latitudes has been examined during a series of experiments using the Sondrestrom radar together with ancillary observations of the interplanetary magnetic field (IMF) by the IMP-8 spacecraft. The radar experiments obtained a latitude coverage of 67.6 to 81.3 deg Lambda and a temporal resolution of between 14 to 25 minutes. A total of 17 rotations through the dayside cleft region during April, June and July, 1983 have been examined. The observations show two convection cells with sunward flow at lower latitudes and antisunward flow at higher latitudes. The flow commonly rotates through a 180 deg angle resulting in the predominant appearance of east-west flows. Rapid temporal variations in the convection velocities are frequently observed. Many of the high latitude variations in convection velocity appear to be directly related to variations in the IMF By component, with eastward (westward) velocity associated with negative (positive) By. This is strong evidence for a direct electrical coupling between the solar wind and dayside high latitude ionosphere.

Sondrestrom observations show that a characteristic F‐region signature of the interaction between the magnetosphere and ionosphere is a narrow band of elevated electron temperatures. Its location is associated with the ion convection reversal in the morning and evening convection cells. Typically, near 500 km altitude, the temperature is 3500 to 4000 K. However, on 24 April 1983, a geomagnetically very active day with Kp values of 6‐ during the period of interest, the electron temperature reached 6000 K in the afternoon convection reversal. The ion velocities were between 1 and 2 km/s on both sides of the reversal. There was considerable soft particle precipitation and a large downward heat flux of 0.3 erg/cm²‐s at 450 km. These high temperature electrons then transferred 0.5 erg/cm²‐s to ions and neutrals between 175 and 550 km, which represents a very sizeable perturbation to the thermosphere.

First results are reported of a comprehensive campaign to measure ionospheric structure and dynamics of nighttime polar cap F‐layer or F‐region arcs. Airborne optical and digital ionosonde data were collected simultaneously with ground based incoherent scatter radar data, continuously over many hours in time and 500 to 1000 km in space. Polar cap F‐region sub‐visual auroral arcs, which are commonly observed during quiet magnetic conditions, are found to represent boundaries (or shears) in the polar cap plasma circulation pattern. F‐region electron concentration enhancements are found in these features and can be of significance to polar thermospheric circulation, composition, and thermal structure. The quiet polar ionosphere is so structured that high time resolution intensified optical images of thermospheric emmissions can make a major contribution to interpretation of incoherent scatter radar data.

By analyzing an observed storm time auroral electron temperature profile it is shown that anomalous transport effects strongly influence the thermal structure of the disturbed auroral ionosphere. Such anomalous transport effects are a consequence of plasma turbulence, the existence of which has been established by a large number of observations in the auroral ionosphere. The electron and composite ion energy equations are solved with anomalous electron thermal conductivity and parallel electrical resistivity coefficients. The solutions are parameterized with respect to a phenomenological altitude-dependent anomaly coefficient A and are compared with an observed storm time electron temperature profile above Chatanika. The calculated temperature profile for the classical case (A = 1) disagrees considerably with the measured profile over most of the altitude range up to 450 km. It is shown that an anomaly coefficient with a sharp peak of the order of 10,000 centered around the F2 peak is consistent with observations.

Sondrestrom radar observations reveal that the dawn-dusk (By) component of the interplanetary magnetic field (IMF) strongly influences the nightside polar convection. This effect is quite complex. The convection for one orientation of By is not the mirror image of the other orientation. A positive By (i.e., pointing toward dusk) seems to organize the velocities such that, at all local times, they are predominantly westward within the radar field-of-view (approximately 68 deg-to-82 deg invariant latitude). Between dusk and midnight, on one such occasion, sunward flow is observed within the polar cap. In the midnight and dawn sectors, when By is negative, the plasma velocities often appear shifted toward early hours such that large southward velocities are observed about 3 hours before midnight. These are the only times when the predominant velocity component is southward.

Processes which confine photoelectrons to the plasmasphere (e.g., collisional backscattering from the thermosphere and magnetic trapping due to pitch angle redistribution through Coulomb collisions in the plasmasphere) tend to increase the steady state photoelectron flux in the plasmasphere above the amplitude level that would otherwise have been attained. Theoretical calculations are presented of steady state photoelectron fluxes in the plasmasphere, for specified atmospheric and ionospheric conditions. (Observational plasma line intensity data for these conditions exist and will be compared elsewhere.) General features of the angular distribution are presented and compared with observations. The transparency of the plasmasphere and the backscattering properties of the thermosphere are investigated. The buildup effect due to collisional backscatter alone is calculated, and the combined buildup effect of pitch angle diffusion and backscatter is estimated. It is found that the inclusion of these effects increases the steady state photoelectron flux amplitude in the plasmasphere by about 50% over the value obtained when the buildup effects are neglected. The calculated steady state photoelectron fluxes in the plasmasphere are in good agreement with the available observations.

When sufficiently high power o-mode HF radio waves are transmitted at frequencies below the ionospheric maximum plasma frequency, not only are plasma instabilities excited, but ambient electrons are accelerated to energies of several electron volts (vs electron thermal energies of order 0.1 eV). Such fluxes have been experimentally inferred from enhancements of atomic oxygen 6300 A, and occasionally 5577 A airglow emission. Active theoretical work has led to conflicting theories for the basic electron acceleration mechanism responsible. We report here the most direct observations to date of such suprathermal electrons. During experiments at Arecibo, transmitting about 140 kW of o-mode power at 7.63 MHz, near but below the ƒ0F2 (exciting strong plasma wave instabilities near an altitude of 285 km) nighttime plasma line intensities were observed to be enhanced by a factor of 10–100. These enhancements extend over a much broader altitude range than that to which HF excited plasma instabilities are confined. When the transmitter was turned off, these plasma line intensities immediately relaxed to their thermal level. Data are presented which should help guide future theoretical development. One significant conclusion that the observations directly lead to is that electrons are accelerated up to at least 20 eV, an energy so high as to call for reassessment of existing published theories. The data also contradict a highly sensitive HF power dependence of the suprathermal electron flux. The production rate is relatively flat from 10 to 20 eV and relatively stable over several hours for constant HF power densities. Model calculations demonstrate that the observed altitude variation of the plasma line intensities is dominated by collisional energy degradation of the suprathermal electron flux to the neutral atmosphere and ionospheric plasma. The modeling also demonstrates that a valid treatment of the plasma physics involved must include some of the geophysical processes (e.g. scattering by the neutral atmosphere).

Simultaneous radar azimuth scans with the Millstone Hill, Sondrestrom, and EISCAT incoherent scatter radars produce maps of the ionospheric convection electric field at high latitudes which span 10 hours of magnetic local time. A series of convection “snapshots” made during an interval of increasing geomagnetic activity indicate that the large-scale convection pattern maintains a two-cell character during substorm onset and that average convection models derived from radar data provide a reasonable representation of the large-scale convection pattern during dynamically varying conditions.

Elevation scan observations made by the Millstone Hill and Sondrestrom incoherent scatter radars are combined to provide extended latitudinal coverage of thermospheric measurements. Maps of the latitudinal and temporal structure of the exospheric temperature are presented for two 24-hour periods over the latitude range 45 to 72 degrees north. On the magnetically quiet June day the Millstone and Sondrestrom data formed a consistent picture of thermospheric structure. On the disturbed July day the two radars observed very different behavior, with Millstone Hill observing strong, long-lived ion frictional heating events but Sondrestrom observing more quiescent behavior. Comparison with HILAT satellite data recorded on the July day suggest that the two radars were observing different ionospheric regions separated by a bright auroral arc with a very turbulent ionosphere to the south of the arc but smooth flow to the north.

The Chatanika radar has been used to measure the ratio of atomic (O+) ions to molecular (O2 +, NO+) ions in the high-latitude ionosphere. The radar results agreed well with simultaneous in situ rocket data, giving confidence in the radar method of deducing ion composition. Measurements made over long periods of time show seasonal variations, diurnal variations, and variations due to auroral processes. The transition altitude, where the number densities of atomic and molecular ions are equal, is a convenient parameter for describing the composition variation with altitude or ‘composition altitude profile.’ The transition altitude occurs at ∼190 km at night and ∼170 km during the day, in agreement with midlatitude results. During the winter the daytime transition altitude is 15 km lower than in summer, a seasonal variation similar to that at midlatitudes. Energetic particle precipitation results in the lowering of the transition altitude, by 10 km in one case when energetic particles deposited ∼20 ergs/cm² s in the atmosphere. The largest variations in ion composition were found during periods of large joule heat input resulting from electric fields on the order of 50 mV/m. The transition altitude increased by 50 km in a case where the joule heat input rate was 30 ergs/cm² s. These observations were compared to calculations from a simple steady state model involving the principal consituents and reactions. The results indicate that the transition altitude during particle precipitation is most influenced by the increased ion production. There do not appear to be significant effects from possible increases of N2 vibrational temperature. A number of interrelated effects contribute to the increase in transition altitude during joule heating. The most important effect is the electric field contribution in raising the effective ion temperature. In addition, it appears that increased N2 density is also required to account for the observed change.

With the incoherent scatter radar at Chatanika, Alaska, a wide variety of measurements can be made related to the ionosphere, magnetosphere, and neutral atmosphere. A significant parameter is the amount of energy transferred from the magnetosphere into the ionosphere and neutral atmosphere during periods of auroral activity. In this report a procedure is examined whereby the incident energy flux of auroral electrons is ascertained from radar measurements. As part of the process radar-determined fluxes are compared with those ascertained from simultaneous photometric observations at 4278 A. The fluxes obtained by both techniques had similar magnitudes and time variations. If it is assumed that the largest uncertainty in the radar/photometer comparison is the effective recombination coefficient, then that coefficient can also be deduced. A value 3times10$sup -7$ cm$sup 3$/s at about 105 km is found, which is in good agreement with other recent determinations during active auroral conditions. This technique is combined with one to ascertain the Joule heating to determine the energy input from the magnetosphere to the ionosphere in a region localized above the radar on March 22, 1973, in the midnight sector. The energy input is continuous at a significant level, i.e., greater than the 3 ergs/cm$sup 2$more » that could be delivered by the sun, were it overhead. Moreover, at times, each of these inputs became as great as 30 ergs/cm$sup 2$ s. (AIP)« less

Observational evidence of the location of a dayside high‐latitude ionospheric current (DPY current) with respect to the different regimes of the high‐latitude magnetosphere is obtained by analyzing data from the magnetometer chain along the west coast of Greenland in conjunction with simultaneous measurements from the newly established incoherent‐scatter radar facility at Sondre Stromfjord. The latitudinal location of the DPY current is compared with the location of the maximum F‐region electron temperature and with the location of the plasma convection reversal from sunward to antisunward. The maximum in the F‐region electron temperature roughly coincides with the velocity reversal boundary, while the DPY current is always located more poleward, penetrating deep into the polar cap. When UT variations are examined, a correlation of 70 to 80 percent is found between the three locations.

Spectra of plasma wave intensities (kTp) in the ionosphere over Arecibo are calculated and compared with those from observations of the plasma line intensity. This approach involving directly observed quantities avoids the uncertainties that have plagued past comparisons with photoelectron theory. In addition, careful comparisons in physically relevant segments of the spectra show that any significant increase in the magnitude of the solar EUV flux would lead to a contradiction of the observed plasma wave intensities. Further, the comparisons indicate that resolution of the thermal electron‐gas heat balance problem must be sought through better heat transfer rates (e.g., heating and cooling rates, etc.), rather than in the solar EUV. This approach utilizes more fully the potential of the plasma line experiment as a diagnostic tool for aeronomical studies, (e.g., photoelectrons, auroral secondaries, ionosphere‐modification experiments, etc.).

Between May 1981 and June 1982 an intensive campaign of 33 coordinated observations was carried out using the three incoherent-scatter radars capable of probing the auroral zone. During this period the groups operating the Dynamic Explorer satellites and the STARE radar made special efforts to acquire data coincident with the radar observations. The objective of these MITHRAS experiments and subsequent analysis is to further our understanding of the interactions of the magnetosphere, the ionosphere, and the thermosphere, with special emphasis on local time/universal time variations. Three experimental modes with different time resolution and spatial coverage were used to examine different aspects of these interactions. The analysis of the extensive data set involves collaboration among groups of experimenters as well as between experimenters and theoreticians.

The auroral ionosphere often differs from the mid- and low-latitude ionosphere due to the influx of energetic particles (primarily electrons) and the mapping of convection electric fields down the magnetic field lines from the magnetosphere. The auroral ionosphere can significantly affect the neutral atmosphere above about 100 km. This paper presents Chatanika Radar measurements related to the physics and chemistry of the atmosphere.

Preliminary observations of dayside high latitude ionospheric plasma convection with the Sondrestrom incoherent-scatter radar indicate that plasma can be observed to enter the polar cap region through rotational reversals at most local times between dawn and dusk and not just in a narrow region around noon. Assuming that rotational reversals are signatures of a solar wind-magnetosphere interaction which drives magnetospheric convection, the observations indicate that this interaction occurs over a longitudinally wide area of the dayside magnetosphere. The observations also show that the distribution of F-region plasma in the polar cap is dependent on ionization sources anywhere between dawn and dusk in the dayside high latitude ionosphere.

The Atmosphere Explorer-C satellite (AE-C) is uniquely suited for correlative studies with ground-based stations because its on-board propulsion system enables a desired ground station overflight condition to be maintained for a period of several weeks. It also provides the first low-altitude (below 260 km) comparison of satellite and incoherent scatter electron and ion temperatures. More than 40 comparisons of remote and in situ measurements were made by using data from AE-C and four incoherent scatter stations (Arecibo, Chatanika, Millstone Hill, and St. Santin). The results indicate very good agreement between satellite and ground measurements of the ion temperature, the average satellite retarding potential analyzer temperatures differing from the average incoherent scatter temperatures by −2% at St. Santin, +3% at Millstone Hill, and +2% at Arecibo. The electron temperatures also agree well, the average satellite temperatures exceeding the average incoherent scatter temperatures by 3% at St. Santin, 2% at Arecibo, and 11% at Millstone Hill. Several temperature comparisons were made between AE-C and Chatanika. In spite of the highly variable ionosphere often encountered at this high-latitude location, good agreement was obtained between the in situ and remote measurements of electron and ion temperatures. Longitudinal variations are found to be very important in the comparisons of electron temperature in some locations. The agreement between the electron temperatures is considerably better than that found in some earlier comparisons involving satellites at higher altitudes.

MITHRAS is a coordinated multiradar program to study the upper atmosphere. Its purpose is to examine the interactions among the magnetosphere, ionosphere, and thermosphere, as well as the phenomena that result from these interactions. It is based around a data set acquired by the Chatanika, Millstone Hill, and EISCAT incoherent-scatter radars between May 1981 and June 1982. A larger portion of this period was unique because it was the only time that three radars, well separated in local and magnetic time, operated together to probe the high-latitude region. The period was also unique because it coincided closely with solar maximum and the DE-2 spacecraft was available for correlative observations. To learn as much as possible from the observations, several tools were developed to improve the handling and analysis: a data-exchange tape format was developed, new display methods using color were implemented to present the data, F-region analysis procedures were extended, and binning procedures were developed to facilitate statistical analysis of the data. In addition, workshops were held so that the observations could be discussed and interpreted by the MITHRAS participants.

Details of the high-latitude ionosphere in the vicinity of the noon sector cleft have been examined using the Sondrestrom radar on April 18, 1983. An experiment combining azimuth and elevation scans was used to reveal the patterns of convection and F region temperature and density over a region of 1000 km diameter centered on the radar at 75°Λ. High-resolution line-of-sight observations have been used to produce detailed maps of the convection pattern and plasma transport each 20 minutes as the radar rotated through the noon sector during the 8 hour experiment. Patterns of electrostatic potential across the radar field of view reveal strong poleward and eastward convection in the morning sector several hours pre-noon, a region of low speed convection associated with the velocity convergence region, and high speed poleward and westward plasma transport from the afternoon sector into the region of cleft precipitation and polar cap entry near noon. The divergence of plasma flow at polar latitudes away from the cleft is seen in both the convection velocity and F region density data (poleward flow through the ionospheric cleft populates the polar cap with enhanced plasma density). A topside density trough is associated with the poleward edge of the strong afternoon convection toward noon where flow velocities approach 2000 m/s and a potential drop of 50 kV is seen across a 500 km region. The region of high-speed plasma entry into the polar cap was characterized by predominently northwestward directed flow on this day on which the IMF was directed away from the sun and Bz was near 0. Plasma entered polar latitudes across a rotational convection reversal which spanned at least 3 hours of local time in the noon sector. Low energy particle precipitation and an enhancement in the ionospheric electron temperature and density were observed at this convection reversal. High-resolution redar data revealed a region of multiple soft arcs aligned with the reversal and polar cap boundary in the post-noon sector.

The winds, temperatures, and densities predicted by the thermospheric GCM are compared with measurements from the Equinox Transition Study of September 17-24, 1984. Agreement between predictions and observation is good in many respects. The quiet day observations contain a strong semidiurnal wind variation which is mainly due to upward-propagating tides. The storm day wind behavior is significantly different and includes a surge of equatorward winds due to a global propagating disturbance associated with the storm onset. A quantitative statistical comparison of the predicted and measured winds indicates that the equatorward winds in the model are weaker than the observed winds, particularly during storm times. A quiet day phase anomaly in the measured F region winds which is not reproduced by the model suggests the occurrence of an important unmodeled interaction between upward propagating semidiurnal tides and high-latitude effects.

Coordinated EISCAT, Chatanika, and Millstone Hill incoherent scatter radar observations have revealed that in the auroral zone, the nighttime F region densities vary substantially with the longitude of the observing site: EISCAT’s densities are the largest and Millstone Hill’s are the lowest. The nighttime F region densities measured by the individual radars are not uniform: the regions where the densities are maximum are the so-called “blobs” or “patches” that have been reported previously. The observations are consistent with the hypothesis that the nighttime densities are produced in significant amounts not by particle precipitation, but by solar EUV radiation, and that they have been transported across the polar cap. The observed differences can be explained by the offset of the geographic and geomagnetic poles. A larger portion of the magnetospheric convection pattern is sunlit when EISCAT is in the midnight sector than when Chatanika is. In winter, when Millstone Hill is in the midnight sector, almost all the auroral oval is in darkness. This universal time effect, which was observed on all coordinated three-radar experiments (September 1981 to February 1982), is illustrated using two periods of coincident radar and satellite observations: November 18-19, and December 15-16, 1981. These two periods were selected because they corresponded to relatively steady conditions. Dynamics Explorer (DE) measurements are used to aid in interpreting the radar observations. DE 1 auroral images show what portion of the oval was sunlit. DE 2 data are used to measure the ion drift across the polar cap. Because the altitude of the ionization peak was high, the decay time of the F region density was substantially longer than the transit time across the polar cap. The southward meridional wind that was observed coincidentally with the ionization patches at Chatanika and EISCAT contributed to the maintenance of the F region by raising the altitude of the peak. DE 2 Langmuir probe measurements of electron density clearly showed a UT dependence, the same as that in the radar measurements.

Incoherent scatter observations performed in March and November 1978 at Chatanika have been used for studying the lower thermosphere in the auroral region. Neutral temperatures and densities have been found during periods without Joule heating. We present mean profiles of temperatures and collision frequencies (approximately proportional to neutral densities) for each month and profiles obtained during four specific nights. Between 93 km and 110 km the mean profiles of temperature are in good agreement with the Jacchia (1971) model, and the profiles of collision frequency are similar to those deduced from the model. Consistency checks between these collision frequencies and temperatures were performed, as were various simulations of the data. Neutral temperature profiles between 90 km and 140 km on the four specific nights show variations from one night to another that are not correlated to magnetic activity. However, there are systematic variations during each night that we suggest are due to atmospheric tides.

As part of the MITHRAS program, the Chatanika and Millstone Hill incoherent-scatter radars made coordinated observations of the polar ionosphere on June 27 and 28, 1981. We compare these data with predictions made by a high-latitude ionospheric model. Qualitatively, the same features are evident in both the model and the radar data: fairly constant densities on the dayside with a mid-latitude trough forming poleward of 65 degrees around 1900 MLT (magnetic local time). This trough is seen to extend equatorward with increasing MLT, such that the minimum densities occurring in the trough appear just after midnight around 60 degrees dipole latitude. These features are primarily understood in terms of different regions of convection, further influenced by photoionization and vertical transport. The only areas of major disagreement between the measurements and model are noted in the auroral oval and at a portion of the times during which substorms occurred. Quantitatively, equally good agreement is obtained between the model predictions and the radar data. The densities predicted by the model are usually within 25% of those measured by the radars, although appreciable differences occur in some regions of the ionosphere at certain times.

Radar observations of ion velocities in the magnetic zenith over Chatanika, Alaska, were used to determine the geomagnetic meridional component of the thermospheric neutral wind. Corrections for molecular diffusion and molecular ion contamination of the pure O+ composition assumed for the ionosphere were included in the analysis. Comparison of the averaged diurnal variation of the meridional wind showed good agreement between the two measurement techniques. Good agreement was also found for several cases of simultaneous observations. The evidence suggested that differences were caused by gravity waves. The 7 years of radar meridional wind results were examined with respect to magnetic activity, solar cycle phase, and season. During the day, the meridional component is poleward with a maximum of about 65 m/s between 1400 and 1600 local time. During the night, the wind is equatorward with a maximum of about 175 m/s between 0200 and 0500 local time. This maximum occurs after local magnetic midnight, which is about 0130 local time. When the neutral wind is averaged for 24 hours, there is a large net equatorward flow. During periods of increased magnetic activity, the nighttime wind between 2300 and 0600 local time becomes stronger toward the equator. The average increase between 0200 and 0600 local time is about 100 m/s; however, on individual days it can be as large as 400 m/s. These data pertain mostly to equinox, but the few summer and winter observations in the data set differ in the manner predicted by theory. Comparison of these results with theoretical models shows good agreement at most times, but suggests possible heating poleward of Chatanika during the morning hours. Observed exospheric temperature increases support this hypothesis.

As part of the MITHRAS program, the Chatanika and Millstone Hill incoherent scatter radars made coordinated observations of the polar ionosphere on June 27 and 28, 1991. The temperature data obtained during these days were compared with predictions made by a high-latitude ionospheric model. The comparison of the temperature measurements and the results of the ionospheric model depend on the assumptions made both in reducing the data and on the inputs that are needed by the model. The deduction of electron temperature from radar measurements depends upon a knowledge of the mean ion mass as a function of altitude. The model requires a knowledge of the heat flux at the upper boundary and the volume heating rate. The results of the model were compared with measurements for a variety of combinations of the required inputs. It was found that the best fits resulted with a heat flux of from 0 to {minus}0.7 {times} 10{sup 10} eV cm{sup {minus}2} s{sup {minus}1} at the upper boundary and a relatively high volume heating rate. These results also required that the model predictions for the average ion mass be used in the reduction of the radar data. However, other combinations of assumptions also producedmore » good fits. A systematic temperature difference of between 200 and 300 K was found between the Chatanika and Millstone Hill measurements of electron temperature at high altitudes.« less