D. Rees

A three-dimensional, time-dependent simulation of the global dynamical response of the thermosphere to a geomagnetic substorm

Abstract Some of the major features of the atmospheric circulation predicted to occur following a geomagnetic substorm in the earths thermosphere are described on the basis of a three-dimensional, time-dependent simulation. The circulation changes and the gravity wave oscillatory phenomena are traced at two levels of the thermosphere, 120 and 240 km, over a period of 4 1 2 h following the substorm onset. For both aspects of the response, and for each height, the characteristics, magnitudes and propagation are quite distinct, and the three-dimensional model shows that very large differences occur between one local time and another. The night-time period, and particularly that between 0000 and 0600 LT, shows, at middle latitudes, the largest wind and wave amplitudes, and the most complex response. In general, the response at any location to an individual substorm can be considered as a series of wave-like disturbances. This is due to the nature of the impulsive disturbance which is concentrated in an annular ring corresponding to the auroral oval, of which each part generates both poleward-travelling as well as equatorward-travelling components, with approximate circular symmetry. Observing the response at 240 km, it appears that focusing occurs of the initial poleward wavefront, and in subsequent poleward wavefronts, the latter arising from the apparent ‘reflection’ of waves initially propagating sunward as they interact with the mean anti-sunward flow on the dayside. Three major waves can therefore be identified in the night-time mid-latitude thermosphere—the initial wave from the nearby auroral oval, a second wave following about 90 min later originating in the dayside auroral oval. This wave is stronger than the first and, being focused and guided by the enhanced mean anti-sunward flow, is capable of generating larger waves. About another 90 min later a third wave arrives due to the dayward propagating wave being reflected, refocused and guided by the mean flow, which can also cause comparable disturbances throughout the thermosphere. There is a strong enhancement of the contra-rotating vortices of the high latitude region—providing sunward flow in the morning and evening auroral ovals with a strong return anti-sunward flow over the polar cap, which extends well into the mid-latitude regions in the midnight and early morning period. In general, amplitudes of the wind and wave disturbances increase continuously with altitude above about 100 km, even allowing for the omission of propagating tidal winds from our model (generated by the lower atmosphere). For the modest disturbance simulated here—a substorm we consider typical of a period when Kp ~ 6-winds of 200–250 m s−1 occur at 120 km, increasing to 500 m s−1 at 240 km, and near 400 km the winds exceed 600 m s−1. While at 240 km the response is dominated by oscillations superimposed on an enhancement of the general thermospheric circulation, at 120 km there is strong evidence that two long-lasting vortices may be created by the dynamical and energetic effects of the substorm, while there is yet no direct observational counterpart of such vortices—an anticyclone centred between 2200 and midnight, and a cyclone centred between 1000 and midday both in the auroral oval—they may be quite characteristic features of the slow recovery of the thermosphere from the impulsive effects of the additional energy input during geomagnetic substorms.

Derivation of a conservation equation for mean molecular weight for a two-constituent gas within a three-dimensional, time-dependent model of the thermosphere

Systematic circulation systems within the thermosphere create major departures of composition of both major and minor species from diffusive equilibrium. For example, latitudinal gradients in the mixing ratios of major and minor species in recent empirical models of the Earths thermosphere are inconsistent with changes of the thermal structure alone or with temporal or spatial changes of the turbopause altitude. A conservation equation describing the time rate of change of mean molecular weight is derived for a two-species gas, in the presence of molecular and turbulent diffusion and general global circulation. The equation is fully three-dimensional and time-dependent and is derived from a combination of the general diffusion equation and the time-dependent continuity equation. In the Earths thermosphere, the two species are [O] the light species and [N2,O2] the heavy species and the approach is valid since the time constants of dissociation of [O2] and recombination of [O] are long compared with typical dynamical time constants. One of the major effects of allowing a wind-driven departure from diffusive equilibrium is that, at the solstice, the pole to pole exospheric temperature difference is increased by more than 50%, while the prevailing summer to winter meridional wind actually decreases. A conservation equation of this kind has general application to any planetary atmosphere which may be considered to be predominantly comprised of two species. Results for a three-dimensional, time-dependent thermospheric model for solstice conditions are presented for the conditions of solar heating only. The model results are compared with previous model results with composition fixed at pressure levels and with empirical temperature and composition models of MSIS.

The generation of vertical thermospheric winds and gravity waves at auroral latitudes—I. Observations of vertical winds

Abstract Observations of vertical and horizontal thermospheric winds, using the OI (3P-1D) 630 nm emission line, by ground-based Fabry-Perot interferometers in Northern Scandinavia and in Svalbard (Spitzbergen) have identified sources of strong vertical winds in the high latitude thermosphere. Observations from Svalbard (78.2N 15.6E) indicate a systematic diurnal pattern of strong downward winds in the period 06.00 U.T. to about 18.00 U.T., with strong upward winds between 20.00 U.T. and 05.00 U.T. Typical velocities of 30 m s−1 downward and 50 m s−1 upward occur, and there is day to day variability in the magnitude (30–80 m s−1) and phase (+/- 3 h) in the basically diurnal variation. Strong and persistent downward winds may also occur for periods of several hours in the afternoon and evening parts of the auroral oval, associated with the eastward auroral electrojet (northward electric fields and westward ion drifts and winds), during periods of strong geomagnetic disturbances. Average downward values of 30–50 m s−1 have been observed for periods of 4–6 h at times of large and long-lasting positive bay disturbances in this region. It would appear that the strong vertical winds of the polar cap and disturbed dusk auroral oval are not in the main associated with propagating wave-like features of the wind field. A further identified source is strongly time-dependent and generates very rapid upward vertical motions for periods of 15–30 min as a result of intense local heating in the magnetic midnight region of the auroral oval during the expansion phase of geomagnetic disturbances, and accompanying intense magnetic and auroral disturbances. In the last events, the height-integrated vertical wind (associated with a mean altitude of about 240 km) may exceed 100–150 m s−1. These disturbances also invariably cause major time-dependent changes of the horizontal wind field with, for example, horizontal wind changes exceeding 500 m s−1 within 30 min. The changes of vertical winds and the horizontal wind field are highly correlated, and respond directly to the local geomagnetic energy input. In contrast to the behaviour observed in the polar cap or in the disturbed afternoon auroral oval, the ‘expansion phase’ source, which corresponds to the classical ‘auroral substorm’, generates strong time-dependent wind features which may propagate globally. This source thus directly generates one class of thermospheric gravity waves. In this first paper we will consider the experimental evidence for vertical winds. In a second paper we will use a three-dimensional time-dependent model to identify the respective roles of geomagnetic energy and momentum in the creation of both classes of vertical wind sources, and consider their propagation and effects on global thermospheric dynamics.

A theoretical and empirical study of the response of the high latitude thermosphere to the sense of the “Y” component of the interplanetary magnetic field

Abstract The strength and direction of the Interplanetary Magnetic Field (IMF) controls the transfer of solar wind momentum and energy to the high latitude thermosphere in a direct fashion. The sense of “ Y” component of the IMF (BY) creates a significant asymmetry of the magnetospheric convection pattern as mapped onto the high latitude thermosphere and ionosphere. The resulting response of the polar thermospheric winds during periods when BY is either positive or negative is quite distinct, with pronounced changes in the relative strength of thermospheric winds in the dusk-dawn parts of the polar cap and in the dawn part of the auroral oval. In a study of four periods when there was a clear signature of BY, observed by the ISEE-3 satellite, with observations of polar winds and electric fields from the Dynamics Explorer-2 satellite and with wind observations by a ground-based Fabry-Perot interferometer located in Kiruna, Northern Sweden, it is possible to explain features of the high latitude thermospheric circulation using three dimensional global models including BY dependent, asymmetric, polar convection fields. Ground-based Fabry-Perot interferometers often observe anomalously low zonal wind velocities in the (Northern) dawn auroral oval during periods of extremely high geomagnetic activity when BY is positive. Conversely, for BY negative, there is an early transition from westward to southward and eastward winds in the evening auroral oval (excluding the effects of auroral substorms), and extremely large eastward (sunward) winds may be driven in the auroral oval after magnetic midnight. These observations are matched by the observation of strong anti-sunward polar-cap wind jets from the DE-2 satellite, on the dusk side with BY negative, and on the dawn side with BY positive.

Diffusive equilibrium and vertical motion in the thermosphere during a severe magnetic storm : A computational study

Vertical motions in the thermosphere, due to solar heating or high-latitude energy inputs, cause the distributions of atomic oxygen and molecular nitrogen to depart from diffusive equilibrium. To study this departure for a two-constituent model, a “diffusive equilibrium” parameter P is defined in terms of the O and N2 partial pressures. In general, regions of enhanced P are found where heating and strong upwelling are present, or have recently taken place, while regions of depressed P are produced by downwelling. It is shown that the relative vertical flow of O and N2, which acts to restore diffusive equilibrium, is proportional to the height gradient of P. The P-parameter is computed for the UCL three-dimensional time-dependent computational model, and its variations with height, latitude and local time are studied both for quiet-day conditions and for a simulated geomagnetic storm. It is suggested that the P-parameter might be applied to experimental data for the study of vertical flows.

A comparison of wind observations of the upper thermosphere from the dynamics explorer satellite with the predictions of a global time-dependent model

Abstract Seven polar passes of the NASA Dynamics Explorer 2 (DE-2) satellite during October and early December 1981 have been used to examine the high-latitude circulation in the upper thermosphere. Vector winds along the satellite track are derived by appropriate merging of the data from the remote-sensing Fabry-Perot interferometer (meridional wind) and the in situ wind and temperature spectrometer (zonal wind) and are compared with the predictions of a three-dimensional, time-dependent, global model of the thermosphere. Major features of the experimental winds, such as the mean day to night circulation caused by solar u.v. and e.u.v. heating, augmented by magnetospheric processes at high latitude and the sharp boundaries and flow reversals imposed on thermospheric winds by momentum transfer (ion drag) from the magnetosphere, are qualitatively explained by a version of the global model using a semi-empirical global model of polar electric fields (Volland Model 2 or Heppner Model A) and a model of global electron density which excludes the effects of high-latitude geomagnetic processes. A second version of the global dynamic model includes a theoretical model of the high-latitude ionosphere which is self-consistent and reflects the enhancement of ionization due to magnetospheric phenomena acting in addition to solar e.u.v. photo-ionization, including the interactive processes which occur between ionization and high latitude ion convection and thermospheric winds. This second dynamical model shows an improved comparison with the structure and magnitude of polar cap and auroral oval winds at times of other than extremely low geomagnetic activity, when the first model appears a better match. An improved empirical description of the complex magnetospheric processes exciting the thermosphere in the vicinity of the dayside polar cusp and an empirical description of storm-time electric fields will be required for a quantitative explanation of the polar thermospheric winds during geomagnetic substorm events.

Interpretation of an anticipated long-lived vortex in the lower thermosphere following simulation of an isolated substorm

Abstract Using a three-dimensional, time-dependent, global model, we have simulated the response of the thermosphere to an isolated substorm. The substorm is characterized by a time variance of the high latitude convective electric field with an associated enhancement of auroral E region electron density, from an initially quiet thermosphere. We have simulated such an impulsive energy input with both separated and co-incident geographic and geomagnetic poles and have found that, in both cases, in the lower thermosphere ( ∼ 120 km), a long-lived vortex phenomenon is generated. Initially, two contra-rotating vortices are generated by the effects of ion drag during the period of enhanced high latitude energy input centred on the polar cap/auroral oval boundary, one at dusk (18.00 L.T.) and the other at dawn (06.00 L.T.). After the end of the substorm, the cyclonic vortex (dawn) dissipates rapidly while the dusk anti-cyclonic vortex appears virtually self-sustaining and survives many hours after the substorm input has ceased. A theory is derived to explain and interpret the results and it appears that the effect is analogous to a meteorological weather system. In this case, however, the dusk anti-cyclonic vortex has, instead of pressure, the centrifugal acceleration balancing the Coriolis force. The equivalent anti-clockwise dawn vortex, unlike a low pressure system, has no balancing force, since Coriolis and the centrifugal term assist and this vortex rapidly disappears.

The seasonal behavior of high-latitude thermospheric winds and ion velocities observed over one solar cycle

The seasonal variation of nighttime thermospheric winds observed at Kiruna has been found to be significantly dependent on solar activity. Of particular interest is the observation that there is a large difference between the March and September equinox winds, despite similar levels of solar insolation. The September equinox winds are more December solstice-like. The average March equinox meridional winds are up to 70% larger than for September. The existence of an equinoctial asymmetry has not been predicted by either thermospheric or ionospheric model simulations, which assume that the equinoxes are fundamentally the same, and use forcing functions which are symmetric about the solstices. The average ion velocities measured at EISCAT are larger during the March equinox than the September at solar maximum, while the converse is true at solar minimum. In contrast, the March equinox nighttime thermospheric winds are larger for both solar maximum and solar minimum. Furthermore, the asymmetry is greater at solar maximum.

The generation of vertical thermospheric winds and gravity waves at auroral latitudes—II. Theory and numerical modelling of vertical winds

Recent observations of strong vertical thermospheric winds and the associated horizontal wind structures, using the 01(3P-1D)nm emission line, by ground-based Fabry-Perot interferometers in Northern Scandinavia have been described in an accompanying paper (Paper I). The high latitude thermosphere at a height of 200–300 km displays strong vertical winds (30–50m ms−1)of a persistent nature in the vicinity of the auroral oval even during relatively quiet geomagnetic conditions. During an auroral substorm, the vertical (upward) wind in the active region, including that invaded by a Westward Travelling Surge, may briefly(10–30 min)exceed 150 m s−1. Very large and rapid changes of horizontal wind structure (up to 500 m−1 in 30 min) usually accompany such large impulsive vertical winds. Magnetospheric energy and momentum sources generate large vertical winds of both a quasi-steady nature and of a strongly time-dependent nature. The thermospheric effects of these sources can be evaluated using the UCL three-dimensional, time-dependent thermospheric model. The auroral oval is, under average geomagnetic conditions, a stationary source of significant vertical winds (10–40 m s−1). In large convective events (directly driven by a strong momentum coupling from the solar wind) the magnitude may increase considerably. Auroral substorms and Westward Travelling Surges appear to be associated with total energy disposition rates of several tens to more than 100 erg cm−2s−1, over regions of a few hours local time, and typically 2–5° of geomagnetic latitude (approximately centred on magnetic midnight). Such deposition rates are needed to drive observed time-dependent vertical (upward) winds of the order of 100–200m s−1.The response of the vertical winds to significant energy inputs is very rapid, and initially the vertical lifting of the atmosphere absorbs a large fraction (30% or more) of the total substorm input. Regions of strong upward winds tend to be accompanied in space (and time) by regions of rather lower downward winds, and the equatorward propagation of thermospheric waves launched by auroral substorms is extremely complex.

Modelling of thermospheric composition changes caused by a severe magnetic storm

Abstract The UCL 3-dimensional time-dependent thermospheric model, with atomic and molecular components, is used to study composition changes in the neutral gas at F-layer heights produced by a severe magnetic storm. The computations give the mean molecular weight (MW), temperature and winds as functions of latitude, longitude, height and time for a period of 30 h. Starting from quiet-day conditions, the simulation starts with a 6-h “substorm” period in which strong electric fields are imposed in the auroral ovals, accompanied by particle input. Weaker electric fields are imposed for the remaining 24 h of the simulation. The energy input causes upwelling of air in the northern and southern auroral ovals, accompanied by localized composition changes (increases of MW), which spread no more than a few hundred kilometres from the energy sources. There is a corresponding downward settling of air at winter midlatitudes and low latitudes, producing widespread decreases of MW at a fixed pressure-level. These storm effects are superimposed on the quiet-day summer-to-winter circulation, in which upwelling occurs in the summer hemisphere and down welling in the winter hemisphere. The composition changes seen at a fixed height differ somewhat from those at a fixed pressure-level, because of the expansion resulting from the storm heating. The results can be related to the well-known prevalence of “negative” F-layer storms (with decreases of F2-layer electron density) in summer, and “positive” F-layer storms in winter and at low latitudes. However, the modelled composition changes are not propagated far enough to account for the observed occurrence of negative storms at some distance from the auroral ovals. This difficulty might be overcome if particle heating occurs well equatorward of the auroral ovals during magnetic storms, producing composition changes and negative storm effects at midlatitudes. Winds do not seem a likely cause of negative storm effects, but other factors (such as increases of vibrationally-excited N2) are possibly important.