Geomagnetic Activity Conditions Research Articles

Composition signatures in ion injections and its dependence on geomagnetic conditions

Ion composition in substorm injections are investigated in detail by using data obtained from the magnetospheric ion composition spectrometer (MICS) on board the Combined Release and Radiation Effects Satellite (CRRES). A total of 398 injection events are identified from 4 February to 10 October, 1991. It is found that the flux enhancements of O+ ions in the “with‐oxygen” injections (O+ ions occupy more than 20% of the total energy density) are attributed to the flux intensifications of high‐energy (E/q ≥ 100 keV/e) ions. In contrast, the flux of H+ and He++ ions in the “without‐oxygen” injections (O+ ions are less than 20% total energy density) are enhanced by the intensification of lower energy H+ and He++ (around 50 keV/e) ions. Statistical results also demonstrate that the abundance of O+ ions in an injection strongly depends on the geomagnetic activity. Without‐oxygen events take place most probably in relatively weak geomagnetic activity conditions (Dst index is around −50 nT and Kp is around 4). The averaged fractional energy density of O+ ions increases roughly linearly with Kp index. All injections are found to be rich of oxygen ions during the strong magnetic storms (Dst < −100 nT). The present study indicates that the ionospheric oxygen ions are energized more readily during the storm conditions. When the near‐Earth reconnection occurs and expands from the central plasma sheet to the plasma sheet boundary layer or even to the lobe region, much more ionosphere origin O+ ions could participate and be energized in the magnetospheric dynamic process. This leads to a with‐oxygen injection in the inner magnetosphere. As to the without‐oxygen injection, it seems that only preexisting oxygen ions in the near‐Earth plasma sheet are accelerated and become a part of the injected ions, thus a rather low abundance of O+ ions can be observed.

Case studies of space weather events from their launching on the Sun to their impacts on power systems on the Earth

Abstract. Active geomagnetic conditions on 12–13, 15–16, and 22–23 September 1999 resulted in geomagnetically induced currents (GIC) measurable in power systems in Canada and the United States. Different solar origins for these three events gave rise to dissimilar interplanetary signatures. We used these events to present three case studies, each tracing an entire space weather episode from its inception on the Sun, propagation through the interplanetary medium, manifestation on the ground as intense magnetic and electric fluctuations, and its eventual impact on technological systems.Key words. Geomagnetism and paleomagnetism (rapid time variations) – Interplanetary physics (interplanetary magnetic fields) – Solar physics, astrophysics, and astronomy (flares and mass ejections)

Studies of X-ray observations from PIXIE

The Polar Ionospheric X-ray Imaging Experiment (PIXIE) on board the NASA/GGS POLAR spacecraft has been making observations of ionospheric X-ray emissions from the vantage of space for more than 3 years. A wide variety of observations have been made by PIXIE, which are detailed in this work. These include the local time distribution of the auroral X-ray intensity as well as the dependence of auroral X-rays on geomagnetic activity and solar wind magnetic field conditions. The auroral X-rays are produced as energetic electrons within the magnetosphere precipitate and are stopped in the ionosphere. Comparisons of the X-ray auroral intensity with other instrument observations have been made, which enables us to distinguish between temporal and spatial processes. In addition, several other X-ray features (not of an auroral nature) have been observed by PIXIE, and are described.

Energetic ion outflow from the dayside ionosphere and its relationship to the interplanetary magnetic field and substorm activity

The energetic ion outflow from the dayside ionosphere has been investigated for different AE and IMF conditions. Viking ion data, 1-min values of the AE index, and 5-min averaged IMF values have been used. An automated algorithm using principal component analysis has been used to categorize and classify the upflowing ion events (UFIs) into beams and conics. The Viking data cover altitudes between 6000 and 13,500 km, 0600 to 1800 magnetic local time, and 65 to 90° invariant latitude. Beams dominate the outflow prenoon and postnoon outside the cusp region, while conics are responsible for most of the cusp region outflow. Generally, the highest average number flux is found in the cusp region. The outflow intensity is higher for southward than for northward IMF, especially in the cusp and postnoon sector. Also, the outflow region increases slightly for southward IMF. However, the difference in outflow intensity and active area is stronger when the data are sorted according to quiet and active geomagnetic conditions using the AE index. The highest average UFI number flux is observed during the expansion phase of substorms, when the flux increase from quiet conditions ( AE<100) is by a factor of 2, but even during growth and recovery phase the average UFI number flux is ∼30% higher than for quiet conditions. The results imply that the dayside energetic ion outflow is controlled partly by nightside (driven by the release of stored energy in the magnetotail, as observed in the AE index) and partly by dayside (directly driven, as observed in the polarity in the IMF B z component) activity.

An overview of the early November 1993 geomagnetic storm

This paper describes the development of a major space storm during November 2–11, 1993. We discuss the history of the contributing high‐speed stream, the powerful combination of solar wind transients and a corotating interaction region which initiated the storm, the high‐speed flow which prolonged the storm and the near‐Earth manifestations of the storm. The 8‐day storm period was unusually long; the result of a high‐speed stream (maximum speed 800 km/s) emanating from a distended coronal hole. Storm onset was accompanied by a compression of the entire dayside magnetopause to within geosynchronous Earth orbit (GEO). For nearly 12 hours the near‐Earth environment was in a state of tumult. A super‐dense plasma sheet was observed at GEO, and severe spacecraft charging was reported. The effects of electrons precipitating into the atmosphere penetrated into the stratosphere. Subauroral electron content varied by 100% and F layer heights oscillated by 200 km. Equatorial plasma irregularities extended in plumes to heights of 1400 km. Later, energetic particle fluxes at GEO recovered and rose by more than an order of magnitude. A satellite anomaly was reported during the interval of high energetic electron flux. Model results indicate an upper atmospheric temperature increase of 200°K within 24 hours of storm onset. Joule heating for the first 24 hours of the storm was more than 3 times that for typical active geomagnetic conditions. We estimate that total global ionospheric heating for the full storm interval was ∼190 PJ, with 30% ofthat generated within 24 hours of storm onset.

Relationship of topside ionospheric ion outflows to auroral forms and precipitation, plasma waves, and convection observed by Polar

The POLAR satellite often observes upflowing ionospheric ions (UFIs) in and near the auroral oval on southern perigee (∼5000 km altitude) passes. We present the UFI features observed by the thermal ion dynamics experiment (TIDE) and the toroidal imaging mass angle spectrograph (TIMAS) in the dusk‐dawn sector under two different geomagnetic activity conditions in order to elicit their relationships with auroral forms, wave emissions, and convection pattern from additional POLAR instruments. During the active interval, the ultraviolet imager (UVI) observed a bright discrete aurora on the duskside after the substorm onset and then observed a small isolated aurora form and diffuse auroras on the dawnside during the recovery phase. The UFIs showed clear conic distributions when the plasma wave instrument (PWI) detected strong broadband wave emissions below ∼ 10 kHz, while no significant auroral activities were observed by UVI. At higher latitudes, the low‐energy UFI conies gradually changed to the polar wind component with decreasing intensity of the broadband emissions. V‐shaped auroral kilometric radiation (AKR) signatures observed above ∼200 kHz by PWI coincided with the region where the discrete aurora and the UFI beams were detected. The latitude of these features was lower than that of the UFI conies. During the observations of the UFI beams and conies, the lower‐frequency fluctuations observed by the electric field instrument were also enhanced, and the convection directions exhibited large fluctuations. It is evident that large electrostatic potential drops produced the precipitating electrons and discrete auroras, the UFI beams, and the AKR, which is also supported by the energetic plasma data from HYDRA. Since the intense broadband emissions were also observed with the UFIs, the ionospheric ions could be energized transversely before or during the parallel acceleration due to the potential drops.

Correlation between core ion energization, suprathermal electron bursts, and broadband ELF plasma waves

Observations of the lowest energy or core ions provide a particularly sensitive measure of the early stages of auroral ion energization. Freja satellite observations of 0–20 eV core ions in the topside auroral ionosphere and cusp/cleft show signs of heating within both regions of VLF hiss and broadband ELF plasma waves. However, heating to several eV or more is associated predominantly with the ELF waves. A correlation analysis of wave and core ion data formed from orbital segments shows that, on average, correlations are highest for wave frequencies below several hundred Hz, and less at VLF hiss frequencies. A similar analysis shows a higher correlation between electron precipitation and ion heating for electron energies below several hundred eV (i.e., the energies associated with suprathermal electron bursts) and a lower correlation above the 1 keV energies associated with auroral inverted‐V's. Signs of core ion heating begin to appear when wave power at the O+ gyrofrequency exceeds about 10−3 (mV m−1)2/Hz, and when the integrated field‐aligned electron flux exceeds a few times 107 cm−2 s−1 sr−1. This electron energy flux threshold is at least an order of magnitude lower than previously inferred from earlier studies comparing suprathermal electron fluxes and energetic ions. Almost all observed heating events occur during enhanced or active geomagnetic conditions; i.e., Kp ≥ 4. While the most intense core ion heating is correlated with broadband ELF waves, we also present one example of weak ion heating of a few eV in a region of VLF auroral hiss.

Solar activity variations of meridional winds over King George Island, Antarctica

Abstract Diurnal variations of the magnetic meridional component of thermospheric neutral winds are derived for King George Island (62.2 °S; 58.8 °W). Calculations are made using a servo theory based algorithm. Input values are F-region peak heights determined from ionosonde data using a well known empirical equation. Ionosonde data correspond to conditions of low geomagnetic activity and both low and high solar activity level for all seasons. These diurnal variations are compared with corresponding variations calculated with the HWM90 model, and solar activity trends of the mean value and diurnal amplitude are examined.

Two types of ion energy dispersions observed in the nightside auroral regions during geomagnetically disturbed periods

The Akebono satellite has observed two types of energy dispersion signatures of discrete ion precipitation event in the nightside auroral regions during active geomagnetic conditions. The charged particle experiments and electric and magnetic field detectors on board Akebono provide us with essential clues to characterize the source regions and acceleration and/or injection processes associated with these two types of ion signatures. The magnetic field data obtained simultaneously by the geosynchronous GOES 6 and 7 satellites and the ground magnetograms are useful to examine their relationships with geomagnetic activity. Mass composition data and pitch angle distributions show that different sources and processes should be attributed to two types (Types I and II) of energy dispersion phenomena. Type I consists of multiple bouncing ion clusters constituted by H+. These H+ clusters tend to be detected at the expansion phase of substorms and have characteristic multiple energy‐dispersed signatures. Type II consists of O+ energy dispersion(s), which is often observed at the recovery phase. It is reasonable to consider that the H+ clusters of Type I are accelerated by dipolarization at the equator, are injected in the field‐aligned direction, and bounce on closed field lines after the substorm onset. We interpret these multiple energy dispersion events as mainly due to the time‐of‐flight (TOF) effect, although the convection may influence the energy‐dispersed traces. Based on the TOF model, we estimate the source distance to be 20–30 RE along the field lines. On the other hand, the O+ energy dispersion of Type II is a consequence of reprecipitation of terrestrial ions ejected as an upward flowing ion (UFI) beam from the upper ionosphere by a parallel electrostatic potential difference. The O+ energy dispersion is induced by the E × B drift during the field‐aligned transport from the source region to the observation point.

Spatial shape and development dynamics of auroral scintillation patches

Results of simultaneous measurements of altitude distribution of auroral small-scale irregularities and total electron content by the method of radio illumination of the ionosphere by signals from an orbital satellite at 150–400 MHz are presented. It is shown that under magnetically quiet conditions the small-scale irregularities (l⊥∼1km) tend to occupy local regions, which extend over ≲ 100– 200 km along the geomagnetic fieldB with ≲ 5– 20 km from the North to South. Characteristic times of formation and collapse of such local structures under the conditions of weak geomagnetic activity are of the order of more than one hour.

Global, collisional model of high‐energy photoelectrons

A previously‐developed colissional, interhemispheric flux tube model for photoelectrons (PE) [Khazanov et al., 1994] has been extended to three dimensions by including transport due to and magnetic gradient‐curvature drifts. Using this model, initial calculations of the high‐energy (≥50 eV) PE distribution as a function of time, energy, pitch angle, and spatial location in the equatorial plane, are reported for conditions of low geomagnetic activity. To explore both the dynamic and steady behaviors of the model, the simulation starts with the abrupt onset of photoelectron excitation, and is followed to steady state conditions. The results illustrate several features of the interaction of photoelectrons with typical magnetospheric plasmas and fields, including collisional diffusion of photoelectons in pitch angle with flux tube filling, diurnal intensity and pitch angle asymmetries introduced by directional sunlight, and energization of the photoelectron distribution in the evening sector. Cross‐field drift is shown to have a long time scale, taking 12 to 24 hours to reach a steady state distribution. Future applications of the model are briefly outlined.

Electron trajectories in a model magnetosphere: Simulation and observation under active conditions

Trajectories of electrons with energies of 10–2000 MeV were studied using Tsyganenko's global quantitative model of the magnetosphere. Results were compared with balloon observations of atmospheric, return albedo, and cosmic electrons made at Prince Albert on August 26, 1987. Active geomagnetic conditions were associated with unusual variations in the flux of electrons on that date, which we interpret as a manifestation of an anomalously low and highly variable geomagnetic cutoff. Increasing the ring current intensity in the Tsyganenko model reduces the calculated effective cutoff rigidity somewhat, but to explain all of our observations, we must postulate that trajectories traditionally assumed to be forbidden are sometimes systematically converted into allowed trajectories. Within this model we can also identify field lines on which bursts of auroral X rays were generated.

Flux dropouts of plasma and energetic particles at geosynchronous orbit during large geomagnetic storms: Entry into the lobes

Geosynchronous orbit spacecraft occasionally observe the magnetospheric tail lobe during disturbed geomagnetic conditions. These lobe intervals are characterized by an absence of detectable ion flux above 1 eV and an absence of electron flux above several hundred eV. This study is an extension of Thomsen et al. (1994) and examines geosynchronous orbit plasma measurements for lobe encounters during a 4‐year period involving three spacecraft, for a total of 85 spacecraft‐months of observations. During the interval surveyed, 160 lobe encounters or close encounters were observed. These events are divided into the following two classes: events observed near midnight and those observed along the flanks. Both classes of events were strongly associated with very active geomagnetic conditions as indicated by Kp (mean value of 5), Dst (mean value of −104 nT), and the Auroral Boundary Index, particularly the flank events. Fifty percent of the lobe encounters occurred within ±24 hours of a geosynchronous magnetopause crossing, which indicates that many of the lobe encounters occurred when there was severe compression and/or erosion of the magnetosphere. Most of the magnetopause‐crossing‐associated events were flank events. The flank‐lobe events were also generally associated with a large interplanetary magnetic field magnitude, with the By or Bz component being dominant. Dropouts in the energetic electron and ion fluxes were observed concurrently with the thermal ion plasma dropouts, though they often extended over a longer time interval. For most of the events a geosynchronous injection of energetic electrons and/or protons was observed following the beginning of the thermal ion dropout, often at a different satellite than the one observing the lobe. For a limited subset of the lobe events (17 events) in which a GOES satellite was nearby, the magnetic field data indicate that the field is often highly stretched or radial near the time of the lobe events. We conclude that the near‐midnight and flank‐lobe events are due to two different mechanisms as follows: the near‐midnight events are consistent with the growth phase picture of large geomagnetic substorms, while the flank events are due to a major reconfiguration of the magnetosphere during large geomagnetic storms which are triggered by the passage of a large solar wind disturbance such as a fast coronal mass ejection. These observations show that the magnetosphere can be highly distorted at synchronous orbit, even far from midnight.

A kinetic model of the formation of the hot oxygen geocorona: 1. Quiet geomagnetic conditions

A model of the hot oxygen geocorona in the transition region near the exobase is described. It is based on a Monte Carlo solution of the nonlinear Boltzmann equation for hot oxygen atoms produced by chemical processes usually considered as a source of hot oxygen (photodissociation of O2 and dissociative recombination of O2+ and NO+ ions). The evolution of the system is described stochastically as a series of random Markovian processes. The energy distribution function of the thermal and non‐thermal O(³P) atoms and of the nonthermal O(¹D) atoms is calculated from the thermospheric collision‐dominated region to the exosphere where the gas flow is virtually collisionless. The model is applied to equatorial latitudes for conditions of low solar and geomagnetic activity. Numerical simulations show that the distribution function of thermal oxygen is increasingly perturbed by collisions with the hot oxygen population at high altitudes and departs significantly from a Maxwellian distribution at all altitudes. The number density and temperature of the nonthermal oxygen atoms are derived from their microscopic distribution function and found to be in qualitative agreement with previous theoretical and experimental estimates.

Freja's contribution to the ISTP event of October 27, 1992. A distorted magnetosphere

A unique arrangement of satellites and ground‐based magnetometers has provided the opportunity to study the distortion of the magnetosphere by conditions of elevated solar wind dynamic pressure and high geomagnetic activity on October 27, 1992. Data were acquired by the near‐Earth satellites Freja, UARS, DMSP‐F8 and DMSP‐F11, and by IMP‐8. This study will focus on patterns of large‐scale Birkeland currents deduced from the APL magnetic field experiments on Freja and UARS. Because of the relatively low inclinations of their orbits (63° for Freja and 57° for UARS) these satellites skim the current systems over a wide range of local time and provide a kind of “global image” of the Birkeland currents. Energetic particle data from the DMSP satellites are used in an attempt to complete the image of the auroral region. The principal findings of this study include the following. The dayside Birkeland current system is displaced 6° to 9° equatorward of the statistical pattern determined by Iijima and Potemra [1978] for undisturbed conditions (|AL|&lt; 100 nT) and 3° equatorward of their disturbed statistical pattern (|AL|≥ 100 nT). The densities of these daytime Birkeland currents are estimated to be 1 to 2 µA/m², comparable to, but not larger than the statistical values determined by Iijima and Potemra for disturbed conditions. The nightside Birkeland currents are located close to the Iijima and Potemra disturbed statistical pattern and they are less intense than the dayside currents.

Dynamics Explorer 2 observations of equatorial thermospheric winds and temperatures: Local time and longitudinal dependences

We have analyzed a large database of Dynamics Explorer 2 (DE 2) measurements of equatorial (&lt;±9° latitude) thermospheric neutral temperatures and horizontal wind components in the altitude range from 200 to 400 km for quiet and active geomagnetic conditions (Kp &lt; 3 and Kp &gt; 3, respectively). Local time variations and longitudinal dependences on the duskside, dayside, and nightside have been separately examined. The dayside hourly averaged neutral zonal winds are relatively insensitive to variations in geomagnetic activity and appear to show an inverse dependence (lower zonal wind speed for higher Kp) in the midnight‐dawn local time sector. The duskside neutral temperatures show a region of enhancement (by ∼60 K) over the South Atlantic magnetic anomaly. The quiet time zonal winds are weaker than the corresponding active time zonal winds on the dayside near the South Atlantic magnetic anomaly. Enhanced ionization in the region, which increases the ion drag on the neutral gas, is suggested as the causal mechanism for these effects. The DE 2 averaged measurements have also been compared with MSIS‐86/HWM90 and VSH models. The data and models are in good general agreement, though, the MSIS‐86/HWM90 model predicts stronger dayside zonal winds than are observed under active conditions and the VSH model appears to underestimate the magnitude of the dayside zonal winds for both high and low Kp cases.

A kinematic analysis of the high‐latitude thermospheric neutral circulation pattern

Averaged measurements from the Dynamics Explorer 2 (DE 2) satellite have been combined with theoretical model calculations to provide “synthesized” thermospheric neutral wind fields for solar maximum, December solstice conditions for both quiet (Kp ≤ 3) and active (3 + ≤ Kp ≤ 6) levels of geomagnetic activity. Specifically, high‐latitude DE 2 wind data obtained from multiorbit averages were merged with modeled winds from the National Center for Atmospheric Research (NCAR) thermosphere‐ionosphere general circulation model (NCAR‐TIGCM), using an objective analysis scheme based on the vector spherical harmonic (VSH) spectral technique of Killeen et al. (1987). A “kinematic analysis” was then performed to decompose the merged semiempirical wind fields into their respective divergent (irrotational) and nondivergent (rotational) components at high latitudes and to calculate the corresponding potential and stream functions. This decomposition provides quantitative insight into the sources of momentum that contribute to each of the wind components. The principal conclusions of the study are (1) the nondivergent component of the high‐latitude thermospheric neutral wind is representative of the convection‐driven component of the neutral wind at F region altitudes. This component of the neutral wind is primarily driven by the ion‐drag and Coriolis forces. (2) The nondivergent wind component makes up a large percentage of the total wind field for both quiet and active geomagnetic conditions. In fact, the nondivergent wind component has stronger sunward directed winds in the dawn and dusk sectors than observed in the total wind field but weaker antisunward directed winds over the polar cap. Vortex formation of the nondivergent wind component is consistently more dominant in the dusk sector than in the dawn sector. (3) The irrotational component of the high‐latitude thermospheric neutral wind is representative of the solar‐driven component of the neutral wind at F region altitudes, directed primarily along the 14 ‐ 2 MLT plane. This component of the neutral wind is mostly driven by the pressure‐gradient‐, Coriolis‐, and Pedersen‐drag (nonconvective‐ion drag) forces. (4) The irrotational component complements the nondivergent antisunward flow over the polar cap but inhibits the nondivergent sunward flow in the dawn and dusk sectors. The total wind field is therefore observed to have a strong antisunward surge over the polar cap with modified sunward flow in the flanks of the polar cap. The development of sunward winds in the dawn sector of the total wind field is capricious, owing to the near vector cancellation of the irrotational and nondivergent wind components in that sector. (5) The above results are also observed in the kinematic analysis of the NCAR‐TIGCM simulations. Poor formation of dawn sector sunward winds is observed in both simulations and indicates other nongeomagnetic forces are responsible, given knowledge of the input ion convection pattern having equal potential between convection cells. (6) Overall, the NCAR‐TIGCM simulations were in reasonable agreement with the DE 2 observations for the nondivergent wind component but typically underestimated the structure in the irrotational wind component. This analysis indicates the possible underestimate of matched power input to the convection patterns chosen to represent the two geomagnetic activity conditions in the NCAR‐TIGCM simulations, thus causing the underestimate of the irrotational wind component.

Admissions of patients with epileptic seizures (E) and dizziness (D) related to geomagnetic and solar activity levels: Differences in female and male patients

Admissions of 762 patients for epileptic seizures and 1553 for dizziness were studied for the connection with the level of monthly and yearly solar activity in the 11-year solar cycle and with four levels of daily geomagnetic activity levels (400 epileptic patients and 802 patients suffering with dizziness). The most significant differences of the studied groups were in the male/female ratio in different conditions of solar and geomagnetic activity. Changes in the group of epileptic patients were accompanied by similar, but adverse changes in the group suffering from dizziness. Different conditions of responses of females and males to environmental changes can explain the demonstrated results.

Numerical modelling of the thermosphere-ionosphere-protonosphere system

The results of a numerical calculation of the thermospheric circulation, the ion and electron concentrations and temperatures as well as electric fields of magnetospheric and dynamo origin are presented for solar minimum, low geomagnetic activity and spring (equinox) conditions. The results are obtained from a numerical hydrodynamical model of the Earth's ionosphere and protonosphere for the height range from 80 km to a geocentric distance of 15 R E . The concentrations and temperature of the neutral atmosphere are computed from the MSIS-83 empirical model of the thermosphere. The effects of small electric fields on the global distributions of the thermospheric circulation, concentrations and temperatures of the charged particles in the ionosphere and protonosphere are discussed. It is shown that electric fields smaller than even 20 mV m −1 form all the main structures of the global plasma distribution at F-region altitudes and in the protonosphere, such as the main ionospheric trough, the light ion trough, the plasmapause and ‘hot spots’ in the ion and electron temperature distributions.

The combined effect of solar and geomagnetic activity on high latitude thermospheric neutral winds. Part I. Observations

Long-term averages of Fabry-Perot Interferometer (FPI) observations of the night-time OI ( 1 D) emission at 630 nm from the F-region thermosphere at a high latitude site are presented here. The data base contains measurements of thermospheric neutral winds for every winter period from November 1981 to April 1989 inclusive. This covers nearly one complete solar cycle in terms of the radio and UV/EUV flux variation, from the last solar maximum to the present solar maximum. The instrument is located in Kiruna, Sweden, which is situated at the equatorward edge of the auroral oval at quiet to moderate levels of geomagnetic activity and beneath the oval at higher levels of activity. From this location, the FPI has sampled the response of the upper thermosphere winds at night to a wide range of geomagnetic and solar activity conditions. The data presented here show a significant correlation between the solar radio or EUV fluxes and the response of thermospheric neutral winds to geomagnetic conditions in the auroral oval. For K p < 2 the auroral oval is polewards of Kiruna and neutral winds are similar at solar minimum and solar maximum. At higher levels of geomagnetic activity, for a given level of activity, the neutral winds are a factor of two greater at high levels of solar flux, than at low levels of solar flux. Alternately, the same winds are seen at K p = 3 at high solar flux levels, as at K p = 5 for low flux levels. Comparing the situation for high and low solar activity, for a given level of geomagnetic activity, in the dusk auroral oval, the sunward or westward winds are a factor of two larger at solar maximum. In the midnight period, the equatorward wind is a factor of two larger at high solar activity, compared with low solar activity. In this paper, the observations will be discussed, with the implication that K p is a rather poor indicator of momentum and energy coupling from the solar wind to the upper thermosphere. In a second paper, the solar-terrestrial processes which cause this phenomenon will be discussed and modelled.