VLF plasmaspheric emissions—‘wave’ signatures of inner magnetosphere boundaries

VLF quarter-gyrofrequency emissions have been observed in the plasmasphere by low-altitude (up to 2500 km) Intercosmos satellites. Their experimental characteristics indicate that the origin and the distribution function of resonant electrons and the type of instability leading to their production can differ during various phases of magnetospheric disturbances. Their close connection to the earthward edge of the lowest energy plasma sheet electron fluxes during active periods is shown. During recovery periods following magnetospheric substorms, they persist (together with other symptoms of a certain internal plasmaspheric structure—subauroral T_e-peaks, SAR arcs, etc.) at roughly the same L-value within the new, more distant plasmapause. A new kind of instability is mentioned which could lead to their production during the recovery phase of magnetospheric substorms.     

Boundary populations in the polar caps

The morphology of precipitating particles, measured at low altitude in the polar regions, varies systematically with the strength and direction of IMF B_z and with solar wind speed V_sw. We use particle data taken onboard the DMSP satellites to determine these variations. Both individual satellite passes during the storm/quieting period of 26 and 27 August 1990, and statistical maps compiled from a data base over 4.5 yr are presented. We focus attention on those magnetospheric populations that have magnetosheath characteristics, the boundary populations. We show that the precipitating ion boundary population, whose down-coming spectra can be fitted to streaming Maxwellians, expands from a region confined near the dayside cusp for southward IMF, to a thick, annular region, including the dayside cusp, for northward IMF. The expansion in local time is inhibited by increasing solar wind speed. Boundary electrons behave somewhat differently. They have easier access to the polar regions and their variations have shorter spatial/temporal scale lengths than the boundary ions. For strongly northward IMF, intense, agitated boundary electrons can be found over all or part of the polar cap. Broad regions (up to ~ 100 km) of strongly accelerated electrons (several keV) that produce visible arcs are embedded in this population. Two features of the ion boundary population help identify its source. (1) The spectra of the boundary ions expanding into the polar cap exhibit field-aligned streaming, which, downtail, is toward the Earth. (2) The region into which the boundary ions expand best maps magnetically to a dawn-dusk cut across the neutral sheet, rather than to the low-latitude boundary layer. Therefore, we conclude that the immediate source for boundary ions in the polar regions during northward IMF is the plasma sheet boundary layer. These ions reach tail lobe field lines by convection whose direction when mapped to the ionosphere is sunward. Significant change in the topology of the magnetospheric magnetic field, and, in particular, the closing of high-latitude field lines, is not required to explain the data.     

Trimpi events on low latitude paths: An investigation of gyroresonance interactions at low L-values

We report on Trimpi events observed at Durban (L = 1.69, 29°53′S, 31°00′E) and investigate the efficacy of gyroresonance scattering in precipitating electrons into the atmosphere at low L (<2). The rate of occurrence of Trimpis at Durban is less than one per day. Our observations include a number of daytime events on OMEGA signals from La Reunion. Using the full relativistic equations of motion, a test particle simulation is employed to find the region in parameter space where large pitch angle scattering occurs. We find that at low L the conditions for pitch angle scattering are less favourable than at higher L (L ∼ 4). Resonant electrons have high (relativistic) energies, interaction times are of the order of milliseconds (T_i ∼ 5 ms) and large wave amplitudes (B_w ∼ 200 pT) are required at whistler frequencies to produce pitch angle changes of greater than 1°. Large pitch angle scattering is needed near Durban since particles near the loss cone will have been lost in the South Atlantic Geomagnetic Anomaly. We note that the radio frequencies transmitted into the magnetosphere from lightning are too low to give effective electron scattering at low L. We suggest an explanation for the low rate of occurrence of Trimpis at Durban.     

The middle atmospheric response to short and long term solar UV variations: analysis of observations and 2D model results

We have investigated the middle atmospheric response to the 27-day and 11-yr solar UV flux variations at low to middle latitudes using a two-dimensional photochemical model. The model reproduced most features of the observed 27-day sensitivity and phase lag of the profile ozone response in the upper stratosphere and lower mesosphere, with a maximum sensitivity of +0.51% per 1% change in 205 nm flux. The model also reproduced the observed transition to a negative phase lag above 2 mb, reflecting the increasing importance with height of the solar modulated HO_x chemistry on the ozone response above 45 km. The rnodel revealed the general anti-correlation of ozone and solar UV at 65–75 km, and simulated strong UV responses of water vapor and HO_x species in the mesosphere. Consistent with previous 1D model studies, the observed upper mesospheric positive ozone response averaged over ±40° was simulated only when the model water vapor concentrations above 75 km were significantly reduced relative to current observations. Including the observed temperature-UV response in the model to account for temperature-chemistry feedback improved the model agreement with observations in the middle mesosphere, but did not improve the overall agreement above 75 km or in the stratosphere for all time periods considered. Consistent with the short photochemical time scales in the upper stratosphere, the model computed ozone-UV sensitivity was similar for the 27-day and 11-yr variations in this region. However, unlike the 27-day variation, the model simulation of the 11-yr solar cycle revealed a positive ozone-UV response throughout the mesosphere due to the large depletion of water vapor and reduced HO_x-UV sensitivity. A small negative ozone response at 65–75 km was obtained in the 11-yr simulation when temperature-chemistry feedback was included,In agreement with observations, the model computed a low to middle latitude total ozone phase lag of +3 days and a sensitivity of +0.077% per 1% change in 205 nm flux for the 27-day solar variation, and a total ozone sensitivity of +0.27% for the 11-yr solar cycle. This factor of 3 sensitivity difference is indicative of the photochemical time constant for ozone in the lower stratosphere which is comparable to the 27-day solar rotation period but is much shorter than the 11-yr solar cycle.     

Studies of high latitude mesospheric turbulence by radar and rocket 1: energy deposition and wave structure

During early spring, 1985, the MAE-3 (Middle Atmospheric Electrodynamics) Program was conducted at Poker Flat Research Range, Alaska to study the origin of wintertime mesospheric echoes observed with the Poker Flat MST radar there, by probing the mesosphere with in situ rocket measurements when such echoes occurred. Pre-launch criteria required the appearance of echoes exhibiting some wave structure on the MST radar display; these could be met even under weak precipitation conditions with riometer absorption near or above 1.0 dB. Two morning rockets were launched under such conditions, the first (31.048) on 29 March 1985, at 1703 UT and the second (31.047) on 1 April 1985, at 1657 UT. Both payloads were deployed on a high altitude parachute near a 95 km apogee to provide a stable platform for data acquisition within the mesosphere (below 80 km). Each payload carried a solid state detector to measure energetic electrons between 0.1 and 1.0 MeV and an NaI crystal detector to measure x-rays from >5 to >80 keV. Payload 31.048 also carried a positive ion ‘turbulence’ probe which measured ion density changes (ΔN_i/N_i) during payload descent, whereas 31.047 carried a nose tip ‘turbulence’ probe designed to measure electron density changes (ΔN_e/N_e) during upleg ram conditions plus a Gerdien condenser for the measurement of bulk ion properties during downleg. The energy deposition curves for each event exhibited peak deposition rates between 75 and 80 km with a half width of 16–18 km, almost exclusively induced by precipitating relativistic electrons. They also showed a maximum bottomside gradient between 65 and 75 km. Radar echoes and atmospheric turbulence were observed in the same altitude domain, consistent with the anticipated need for adequate free thermal electron gradients to make such phenomena visible on the radar. The vertical wave structure from radar echoes was found to be consistent with that observed in horizontal wind and temperature profiles measured by Datasondes flown shortly after each large rocket. An analysis of the wave structure from radar data has shown that although large scale waves (λ_z ~ 7 km) were found to be present, a higher frequency shorter wavelength (∼ 1–3 km) component probably played a more significant role in modulating the signal-to-noise structure of the radar echoes.     

An alternative interpretation of auroral precipitation and luminosity observations from the DE,DMSP, AUREOL,and Viking satellites in terms of their mapping to the nightside magnetosphere

We present an interpretation, which differs from that commonly accepted, of several published case studies of the patterns of auroral electron precipitation into the high-latitude upper atmosphere in the near-midnight sector based on their mapping to the nightside magnetosphere. In our scheme bright discrete auroral structures of the oval and respective precipitation are considered to be on the field lines of the Central, or Main, Plasma Sheet at distances from 5–10 to 30–50 R_E, depending on activity. This auroral electron precipitation pattern was discussed in detail by Feldstein and Galperin [(1985) Rev. Geophys.23, 217] and Galperin and Feldstein [(1991) Auroral Physics, p. 207. Cambridge University Press. It is applied and shown to be consistent with the results of case studies based on selected transpolar passes of the DE, DMSP, AUREOL-3 and Viking satellites.A diagram summarising the polar precipitation regions and their mapping from the magnetospheric plasma domains is presented. It can be considered as a modification of the Lyons and Nishida (1988) scheme which characterizes the relationship between the gross magnetospheric structure and regions of nightside auroral precipitation. The modification takes into account non-adiabatic ion motions in the tail neutral sheet, so that the ion beams characteristic of the Boundary Plasma Sheet (BPS) originate on closed field lines of the distant Central Plasma Sheet (say, at distances more than ~30 R_E).     

Electric fields and currents in the equatorial electrojet deduced from VHF radar observations—III. Comparison of observed ΔH values with those estimated from measured electric fields

From VHF backscatter radar measurements at Thumba (dip: 56′S) of the phase velocities of type II irregularities in the equatorial electrojet (EEJ), electric field (E_y) values are estimated for different times of the day. Using the electric field values thus deduced and the Pedersen and Hall conductivities calculated using model values of electron densities and the collision frequencies of ions and electrons, the height integrated current intensity in the EEJ is estimated. The surface level geomagnetic field perturbation ΔH produced by this ionospheric current is then calculated. The calculated values of ΔH are compared with observed values of ΔH (after subtracting the magnetospheric contribution of D_st) for a number of days. The comparisons show good agreement between observed ΔH values and those calculated from measured electric fields. The agreement is found to be good even when type I irregularities are present at higher altitudes in the EEJ. This comparative study demonstrates the validity of estimating electric field values from VHF radar measurements and it indicates the possibility of deducing electric field values from ground level ΔH values, at least for statistical studies.     

The solar eclipse of 26 February 1979: analysis and modelling of primary pair production,electron density and ionic composition

The solar eclipse of 26 February 1979 was observed from Red Lake, Canada, (52 °N, 91 °W) where totality occurred at about 1053 local time. Several research groups and government agencies participated in an extensive ground- and rocket-based observational program directed at the middle atmosphere. At the time of the eclipse, an extensive geomagnetic storm was in progress and the middle atmosphere was undergoing temperature and circulation changes associated with a stratospheric warming. Concurrent observations of atmospheric constituents, solar radiation, electron flux and other middle atmosphere parameters were obtained as inputs for a D-region predictive chemical computer code, DAIRCHEM, tailored to eclipse conditions. Ion pair production rates were computed by an E-region infrared radiance model and were used as necessary source function input values for DAIRCHEM computations. The computations yielded predictions of electron and total positive ion densities about totality. The positive ion measurements of a supersonic Gerdien condenser and a subsonic blunt probe during the eclipse were in agreement with the model computations and provided normalizing summations of total positive ions for the interpretation of mass spectrometer measurements. The chemical computer code identified principal routes for increase and removal of key species such as O₂⁺, NO⁺, hydrated clusters and negative ions. The dominant precursor ion for pair production hydrates was O₂⁺ and the chemistry was characteristic of the disturbed D-region.     

Auroral rays: filamentation in the auroral accelerator

In situ measurements of particles, fields and optical emissions from a rocket that encountered auroral rays are reported. The measurements give insight into the production of rays, as well as the production of large fluctuations in electric fields perpendicular to the magnetic field. The fine structure and rapid variations of the electron energy flux associated with the rays are apparently produced by modulation in the degree of electron acceleration. Rays are produced when the energy flux increases in localized regions to values even higher than those normally encountered in bright auroral forms. Close and consistent similarities in the variations of the electron energy flux, the light and the electric fields suggest that the field variations were produced as a direct result of the changes in the stream of accelerated electrons. In examining possible causes of the velocity changes that produce the rays, two acceleration processes are considered; acceleration as a consequence of a potential difference between the magnetosphere and the atmosphere and acceleration by waves.     

Mechanisms and time-scales of the magnetospheric response to the interplanetary magnetic field changes during the 8 May 1986 substorm

On 8 May 1986, between 1113 and 1600 UT, an isolated magnetospheric substorm was observed, during which the AE-index exceeded 700 nT (CDAW 9E event). Three available sets of measurements (a) of the solar-wind parameters (IMP-8 satellite), (b) of the magnetotail energy flux (ISEE-1 spacecraft), and (c) of ground magnetic observatories, allowed us to make a detailed study of the overall magnetospheric response to changes of the interplanetary magnetic field (IMF) direction, during this event of weak solar-wind coupling.In order to study the mechanisms and time-delays of the magnetospheric response to the abrupt increase of the solar-wind energy input, we have evaluated the total magnetospheric energy output U_T following two different methods: (a) Akasofu's method, taking the ring current decay time τ_R constant, and (b) Vasyliunas' method where the values of u_t are independent of the solar-wind energy input as determined from the epsilon parameter. Both methods suggest that the driven system has been considerably developed during this substorm, while an unloading event has been superposed at the expansion onset.     

Balloon observations of stratospheric electricity above the South Pole: vertical electric field,conductivity, and conduction current

This paper summarizes the results of measurements of the electrical conductivity σ and vertical component of the vector electric field E_z acquired from eight stratospheric balloon flights launched from Amundsen-Scott Station, South Pole, in the austral summer of 1985–1986. The major findings of this research are as follows

• 1.(1) The data contribute to the set of global atmospheric electricity measurements and extend the work of COBB [(1977), Atmospheric electric measurements at the South Pole. In Electrical Processes in Atmospheres, Dolezalek H. and Reiter R. (eds), pp. 161–167. Steinkopf, Darmstadt, F.R.G.] to determine the electrical environment of the south polar region
• 2.(2) The average vertical profile of the conductivity at the South Pole, when compared with profiles obtained at other Antarctic locations, suggests that the conductivity scale height may increase with increasing geomagnetic latitude across the polar cap.
• 3.(3) The conductivity profiles measured at the South Pole and other Antarctic locations differ significantly from polar cap model profiles. On the basis of these measurements, the model profiles appear to require modification
• 4.(4) The magnitudes of the E_z profiles were observed to vary from day-to-day by a factor of > 2
• 5.(5) In all of the flights the air-Earth conduction current J_z, calculated as the product of E_z and σ, decreased with altitude in agreement with previous direct measurements of the air-Earth current by Cobb [( 1977), Atmospheric electric measurements at the South Pole. In Electrical Processes in Atmospheres, Dolezalek H. and Reiter R. (eds), pp. 161–167. Steinkopf, Darmstadt, F.R.G.]
• 6.(6) The magnitude of J_z was 2–3 times larger than the global average, which can be attributed to the lower columnar resistance of the atmosphere above the high-elevation Antarctic plateau. The magnitude of J_z agrees with that observed by Cobb, if the Cobb measurements are multiplied by the Few and Weinheimer [(1986), Factor of 2 error in balloon-borne atmospheric conduction current measurements. J. geophys. Res.91, 10937] correction factor of 2
• 7.(7) E_z from all of the flights during times of balloon float demonstrates characteristics of the classical ‘Carnegie’ diurnal variation, which is indicative of global influences on the ionospheric potential
• 8.(8) The influence of geomagnetic activity was observed as a decrease in the amplitude of the diurnal variation of E_z with increasing geomagnetic activity index K_p, which is the predicted effect at the South Pole of the magnetospheric polar-cap potential superimposed on the ‘Carnegie’ potential variation.

     

Ionospheric scintillation observations with radio interferometry

Radio astronomical interferometric observations are affected by atmospheric refraction, being particularly sensitive to inhomogeneities in the atmosphere. At frequencies below 2 GHz the influences of the ionosphere are significant in radio astronomy, especially for single dish observations and for connected element interferometry.Analytical expressions for the manifestations of weak ionospheric scintillation in radio interferometric observations, are derived. We indicate which ionospheric scintillation parameters can be derived from radio interferometric measurements. It is shown that the baseline dependence of the observed amplitude scintillation index implies a direct determination of the height of the region of random irregular electron distribution. Furthermore, the linear scale of the irregularities causing scintillation can be determined directly from the baseline dependence of the scintillation index S₄. From the mean square phase fluctuations as a function of interferometer baseline, the spatial scale of the irregularities responsible for this effect can also be determined. From a comparison with observational mid-latitude data we find indications that scintillation irregularities occur in the lower parts of the F2-layer. The spatial scale of irregularities causing amplitude scintillation is of the order of about 25 to about 500 metres. Phase scintillations are caused by irregularities with dimensions which are an order of magnitude larger.     

Auroral radio absorption as an indicator of magnetospheric electrons and of conditions in the disturbed auroral D-region

In a previous paper we demonstrated a method by which the auroral radio absorption measured with a riometer can be predicted from energetic electron measurements at geosynchronous orbit. The present paper enquires to what extent the process can be inverted: what levels of magnetospheric electron flux, and of D-region production rate, electron density and incremental absorption, are predicted by a given measurement of radio absorption and what reliance can be placed on such predictions?Using data from 45 precipitation features recorded with riometers in Scandinavia and at geosynchronous orbit with GEOS-2, it is shown that electron fluxes in the ranges 20–40,40–80 and 80–160 keV increase with increasing absorption and can be predicted to better than 50% for absorption events of 2 dB or greater. Electrons above 160 keV show little or no correlation with absorption. D-region production rates and electron densities can be predicted to within factors of 2 and √2, respectively.It is more difficult to specify the height of the absorbing region because of uncertainly in the profile of the effective recombination coefficient. Having regard to other data, an α_eff profile is proposed which satisfies rocket and incoherent scatter data as well as the present calculations. It is shown that any day-night variation in auroral absorption is associated with a change of spectrum rather than a change of recombination coefficient.     

Electrodynamic response of the middle atmosphere to auroral pulsations

On the nights of 21 and 28 October 1987, two Nike Orion payloads (NASA 31.066 and 31.067) were launched from Andøya, Norway, as part of the MAC/EPSILON campaign, to study the effect of auroral energetics on the middle atmosphere. Each payload carried detectors to measure relativistic electrons from 0.1 to 1.0MeV in 12 differential energy channels, and bremsstrahlung X-rays from >5 to >80keV in 5 integral channels. In addition, instrumentation to measure bulk ion properties and electric fields was also carried by these and/or near simultaneous flights. Flight 31.066 was launched during the recovery phase of a moderate magnetic substorm, during relatively stable auroral conditions. Flight 31.067 was launched during highly active post-break-up conditions during which Pc 5 pulsations (> 150s period) were in progress. The energetic radiation of the first event was composed almost entirely of relativistic electrons below 200 keV with negligible contributions from bremsstrahlung X-rays, while the radiation of the second event was dominated by much softer electrons ( < 100 kcV), which produced high X-ray fluxes that exceeded the cosmic ray background as an ionizing source down to altitudes below 30 km. Simultaneous conductivity measurements during both events show consistency with the ionizing radiations, with the pulsation event producing free electrons down to 55 km. far below their expected altitude range during night-time. These comparisons are discussed to evaluate the impact of such events on the middle atmosphere.     

Simultaneous ground-based observations of polar cap arcs and spacecraft measurements of particle precipitation

In order to investigate the particles which produce the polar cap aurora at the Vostok station in Antarctica, charged particle data obtained by the DMSP satellites for some days in a period from April to August 1985 were surveyed. Due to the satellite orbit the local time range in which the data were available was the morning sector. For all the events when sun-aligned arcs were observed on the ground the simultaneous DMSP measurements on almost the same field line showed an increased integral number flux J. > 10⁸ (cm⁸/s/sr)⁻¹ of the precipitating electrons with energy E_e > 200 eV. The electron spectra with double peaks are typical of intense electron precipitation in the polar cap arcs. The most noticeable feature of ion spectra in the polar cap arcs is the prominent minimum in ion flux in the energy range 0.1 < E_i < 1 keV in contrast with the oval precipitation ; this feature gives the possibility to separate the polar arcs from the aurora in the oval. In some events the satellite crossed the system of two widely separated arcs ; one of them was a sun-aligned arc whereas the other was circular at constant latitude according to the Vostok data. The analysis of the DMSP electron and ion precipitation data has shown that in these events the latitude-oriented arcs are located in the polar cap and not in the auroral oval.     

Electron temperature and electron density in the F-region of the ionosphere. I. Observed relationship

Ionospheric data from three incoherent scatter stations over the height range 225–450 km were studied for all daylight hours over a wide range of solar conditions. The relationship between electron temperature T_e, electron density Nand solar flux at 10.7 cm wavelength S_10.7 was expressed as T_e = A−B·(N−5 × 10¹¹) + C·(S_10.7−750), where N is in units of m⁻³ and S_10.7 in kJy.This provided a very satisfactory expression for all data taken at Malvern and St. Santin between 0800 and 1600 LT. For data taken at Arecibo, however, the linearity broke down at low electron densities. The data from all three stations were therefore divided into two sets according to electron density and reexamined.ForN < 5 × 10¹¹ m⁻³ B increased steadily with height and decreased steadily with latitude.For N > 5 × 10¹¹ m⁻³ B did not appear to vary with height, with season or with latitude. C was approximately constant for all sets of data.The different mechanisms involved in the heat balance of the electron population are discussed and a qualitative explanation for the relationship is proposed.     

Modelling of transequatorial propagation of HF radio waves

In the geometrical optics approximation, a synthesis oblique ionogram of ionospheric and magnetospheric HF radio wave signals propagating between magnetic conjugate points has been carried out. The magnetospheric HF propagation is considered for a model of the waveguide formed by field-aligned irregularities with depleted electron density. The characteristic peculiarities of the magnetospheric mode have been determined: (i) strong disperion of the group delay with a frequency at 14–18 MHz, from − 1.4 to 0.6 ms/MHz for magnetically conjugate points at geomagnetic latitudes φ = 30°, 40° and 50°, respectively, (ii) spreading ∼ 1–2 ms, and (iii) a possibility of propagation between magnetic conjugates points at moderately low geomagnetic latitudes φ₀ ∼ 30–40° at frequencies exceeding 1.5 times the maximum usable frequency (MUF) of multi-hop ionospheric propagation.     

About the origin of high energy electrons in the inner radiation belt

A large flux of > 100 MeV electrons were registered in the inner radiation belt on low-altitude satellites. The origin of that flux is discussed. It appears that slow radial diffusion (D_o = 10⁻¹³ 1/s) gives a low probability for penetration of these electrons to small L from the boundary of magnetosphere because of synchrotron radiation energy losses. It is found that they can enter to the inner belt region without such losses after great magnetic storms when high speed radial diffusion sometimes takes place. Two great storms on 8–9 Feb.] 986 and 24 March 1991 are examples when one can directly observe a penetration of energetic electron fluxes into magnetosphere. The assumption about their Jovian origin is discussed.     

Simulations of the behaviour of the polar thermosphere and ionosphere for northward IMF

The dynamics and structure of the polar thermosphere and ionosphere within the polar regions are strongly influenced by the magnetospheric electric field. The convection of ionospheric plasma imposed by this electric field generates a large-scale thermospheric circulation which tends to follow the pattern of the ionospheric circulation itself. The magnetospheric electric field pattern is strongly influenced by the magnitude and direction of the interplanetary magnetic field (IMF), and by the dynamic pressure of the solar wind. Previous numerical simulations of the thermospheric response to magnetospheric activity have used available models of auroral precipitation and magnetospheric electric fields appropriate for a southward-directed IMF. In this study, the UCL/Sheffield coupled thermosphere/ionosphere model has been used, including convection electric field models for a northward IMF configuration. During periods of persistent strong northward IMF B_z, regions of sunward thermospheric winds (up to 200 m s⁻¹) may occur deep within the polar cap, reversing the generally anti-sunward polar cap winds driven by low-latitude solar EUV heating and enhanced by geomagnetic forcing under all conditions of southward IMF B_z. The development of sunward polar cap winds requires persistent northward IMF and enhanced solar wind dynamic pressure for at least 2–4 h, and the magnitude of the northward IMF component should exceed approximately 5 nT. Sunward winds will occur preferentially on the dawn (dusk) side of the polar cap for IMF B_y negative (positive) in the northern hemisphere (reverse in the southern hemisphere). The magnitude of sunward polar cap winds will be significantly modulated by UT and season, reflecting E-and F-region plasma densities. For example, in northern mid-winter, sunward polar cap winds will tend to be a factor of two stronger around 1800 UT, when the geomagnetic polar cusp is sunlit, then at 0600 UT, when the entire polar cap is in darkness.     

The E-region electron density diurnal asymmetry at Saint-Santin: observations and role of nitric oxide

Measurements of the E-region electron density were made with the Saint-Santin incoherent scatter radar during consecutive days in June 1978, March 1979 and December 1980. On the basis of a statistical study, the observations show the presence of a diurnal asymmetry of the electron density, with morning values usually exceeding the afternoon densities by 3–20%. Two possible causes of the dissymmetry are examined: the asymmetry in the diurnal variation of the neutral composition and the effect of nitric oxide. The presence of NO partly converts O₂⁺ into NO⁺ ions and increases the effective recombination rate of the electrons in the afternoon. Numerical simulations assessing the relative importance of the two factors are, in general, in good agreement with the measurements.