The satellite CHAMP with its sensitive accelerometer on board provides the opportunity to investigate the thermospheric dynamics in great detail. In this study we concentrate on density structures in the cusp. During 25 Sep. 2000, the day we take as an example, air density enhancements of almost a factor of two are observed whenever the satellite passes the cusp region. For the interpretation of these events we consider also the concurrent ionospheric Hall and field-aligned currents (FACs). As expected, sizable currents are found in the regions of dense air. Small-scale FAC filaments (1-km size) seem to play an important role in the heating. Whenever these very intense FACs with amplitudes of several hundreds of μA/m2 show up, density enhancements occur.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
 The thermosphere is the top layer of the gravitationally bound part of the atmosphere. It is characterized by a large variability in density and temperature in response to enhanced solar extreme ultraviolet (EUV) radiation and to geomagnetic disturbances. In a number of studies the large-scale response of the thermosphere and ionosphere to magnetic disturbances have been addressed (see, e.g., Prölss  for a review). The morphology of these disturbance effects is rather complicated and highly variable thus difficult to describe. Simulations performed with the NCAR Thermosphere General Circulation Model (TGCM) indicate that there is a pattern consisting of two to four high and low-density regions, with diameters of 1000 to 2000 km, at high latitudes, which are rather fixed in the geomagnetic-local time frame [e.g., Schoendorf et al., 1996a, 1996b]. If confirmed, these patterns would provide a much needed framework to order and interpret high-latitude air density data. To test these model predictions Caspers and Prölss  have used data from the satellites ESRO 4 and DE 2 and found a reasonable agreement with the observations. The number of suitable passes (150) was, however, quite limited. The modeled density patterns and their mechanisms have yet to be confirmed, due in part to the lack of well distributed, high resolution neutral gas observations.
 The CHAMP satellite [Reigber et al., 2002] launched on 15 July 2000 may change this situation thanks to its well-suited, complementary set of instruments. On board the spacecraft there is a highly sensitive tri-axial electrostatic accelerometer which effectively senses the air drag experienced by the satellite. This is accompanied among others by state-of-the-art scalar and vector magnetometers. These instruments provide important information about ionospheric currents which may be responsible for heating the thermosphere.
 In this paper we present new observations of small-scale thermospheric density structures occurring within the polar cusp region at an altitude of about 400 km. We are open for any ionosphere-thermosphere interaction. For that reason we compare the air drag feature with the various kinds of ionospheric currents observed simultaneously. Such a direct comparison of ionospheric dynamics and thermospheric response has never been presented.
2. Measurement Technique and Data
 The CHAMP satellite orbits the Earth at an inclination of 87.3°. From its initial altitude of 456 km it has decayed to 400 km after three years. Further details of the CHAMP satellite and mission can be found at the website http://op.gfz-potsdam.de/champ/.
 The primary quantity of interest for this study is the air drag exerted on the spacecraft. This is measured by the highly-sensitive tri-axial accelerometer with a resolution of better than 3 · 10−9 m/s2. Since the accelerometer is precisely placed at the spacecraft's center of mass, all gravitational forces are balanced out. The obtained resolution in air density is better than 1 · 10−14 kg/m3 at a sample rate of 0.1 Hz. Another instrument considered here is the tri-axial Fluxgate Magnetometer. It delivers vector readings at a rate of 50 Hz and with a resolution of 0.1 nT. The data are calibrated with respect to the onboard scalar Overhauser Magnetometer. A dual-head star camera system mounted together with the magnetometer on an optical bench provides the orientation of the measured field vectors with arcseconds precision.
 The example of observation considered here in some details is from 25 Sep. 2000. On this day CHAMP is crossing the equator at local times (LT) of 08 and 20 LT on the day and night side, respectively. Figure 1 shows the air drag during the hours 01 to 13 UT of that day. The dominating oscillation at orbital period is caused by the differences in air density on the day and night side and is partly due to the slight eccentricity of the orbit. Superimposed on this harmonic signal are quite narrow spikes in air drag. These occur only in the northern polar region. Labels at each of the spikes list the corresponding corrected geomagnetic (cgm) latitude and the magnetic local time (MLT) of their occurrence. From the labeled numbers it can be suggested that all the enhanced air drag events are encountered in the ionospheric cusp region. At the south pole in the night-time sector there are also regions of enhanced air drag observed, but these are less prominent and are of larger scale size.
3. Interpretation of Observation
 The deceleration, d, of a satellite due to air drag can be calculated by the well-known formula
where ρ is the local air density. All the other quantities are known and have the following values in case of the CHAMP satellite, A = 0.74m2 is the cross-section area in ram direction, v = 7.6 km/s the orbital velocity, cf = 2.2 the drag coefficient and m = 520 kg is the satellite mass. As can be seen from equation (1) there exists a linear relation between the deceleration and the air density. For the interpretation we convert the readings of the satellite drag, presented in Figure 1, into air density, ρ.
 Further corrections are required, e.g., removal of the solar radiation pressure and normalization to a common altitude (450 km) by using the local scale-height of about 60 km consistent with MSIS-90 [Hedin, 1991].
 We also had a look at the concurrent CHAMP magnetic field measurements, in order to estimate the local ionospheric current densities. It is possible to determine Hall and field-aligned currents (FAC) separately from the satellite data. The Hall currents are estimated exclusively from the deflections of the field magnitude [Ritter et al., 2004]. For the calculation of the FACs only the transverse components are employed. We assume perpendicular FAC sheets [Lühr et al., 1996].
Figure 2 shows for a crossing of the northern polar region around 06:30 UT air density and current estimates. The air density in the top panel exhibits an enhancement by a factor of 1.8 with a half-value width of about 4° in latitude. In the panel below the Hall current density is presented. Collocated with the density peak enhanced current densities up to −0.2 A/m are observed at 75.5° cgm-lat. and 10.5 MLT. Another Hall current peak, about equally strong, occurs at 68° cgm-lat. on the evening side (18.3 MLT), which is, however, not accompanied by an air density enhancement. In the frame used here, positive Hall currents are directed anti-sunward, thus both mentioned peaks indicate eastward electrojets.
 In the third panel field-aligned currents are shown. Plotted are averages over 20 s, which is equivalent to a spatial averaging over about 150 km. CHAMP is first crossing the cusp exhibiting a sequence of downward/upward FACs, and in the evening sector the expected Region 1 and 2, upward/downward, FAC pattern emerges. This is consistent with the observed two eastward electrojets. The amplitudes are of moderate size. The bottom panel contains FAC estimates determined in the same way as before, but from the high sampling rate (50 Hz) vector magnetometer measurements, without any filtering. There appears a clearly outstanding packet of fine-scale FAC filaments with amplitudes up to 150 μA/m2 concurrently with the air density enhancement. Nothing comparable in amplitude occurs at any other time during the displayed interval. Hereafter we term the fine-scale field-aligned currents “AC-FAC”.
Figure 3 presents neutral density and FAC for three additional polar passes on the same day as presented in Figure 1. There is a clear one-to-one correspondence between the appearance of AC-FACs and air density enhancements. Comparable results have been found in many cases when CHAMP crossed the cusp, suggesting that both phenomena are in some way associated.
 There is one caveat in determining the air density from along-track acceleration. Head and tail winds may modify the obtained results. It is known that winds in the polar thermosphere can be quite strong, up to the speed of sound, which is of the order of 1 km/s at an altitude of 400 km. As a first-order estimate of the wind effect on the results derived from equation (1) we get
where vW is the wind velocity component in the along-track direction and vSC the orbital velocity of the spacecraft. For the extreme case of vW = 1 km/s we obtain an uncertainty of 30% for the estimated air density.
4. Discussion and Conclusion
 We have presented CHAMP observations of the thermospheric density derived from high resolution accelerometer measurements. Of specific interest are the local enhancements of air density in the dayside auroral region. Such confined enhancements are encountered almost every time when crossing the cusp region. Their average width at half the peak value is found to be 350 ± 150 km, which fits quite well the north/south extent of the cusp proper. Another remarkable feature is the one-to-one correspondence between density peaks and the occurrence of fine-scale FAC filaments in this region. In a recent study Neubert and Christensen  have scanned several years of Ørsted satellite data searching for small-scale FACs events. They find a clear occurrence maximum of these AC-FACs in the polar cusp region. The confinement to that region is particularly evident during non-storm times. Also in our case we are dealing with a moderately active day (average Kp = 3) and the AC-FAC packets are limited to magnetic latitudes and local times of the cusp. Both the up-welling air and the AC-FACs may to be related features. Since CHAMP has encountered the enhanced air density over many consecutive orbits, it may be suggested that the atmospheric up-welling in the cusp is a continuous process lasting at least for several hours.
 Such density peaks cannot be static features. They require a wind or heating system to maintain them for an appreciable amount of time. The upward motion could be driven by a local Joule heating fueled by ionospheric currents, as schematically shown in Figure 4. A similar concept has earlier been proposed by Prölss [1981, Figure 3]. Analyzing ESRO 4 data he found significant density enhancements in the auroral oval at altitudes around 250 km.
 The basic idea is that the atmosphere is heated by ionospheric currents at a lower level, say the E region and subsequently warm air is up-welling causing a density enhancement at 400 km and above. The local Joule heating can generally be described by the dot product of the current density, j and the electric field, E:
where σ∥ and σp are field-aligned and Pedersen conductivities, and E∥ and E⊥ field-aligned and transverse electric field components, respectively. ∂E⊥ is an additional small-scale E-field component superimposed on the background field. Generally, the first term on the right side of equation (3) is neglected due to the small field-aligned electric fields. The large-scale perpendicular electric field can be estimated from
where JH and ΣH are the height-integrated Hall current and conductivity, respectively. Since we do not know the Hall conductance, the E-field cannot be determined from our measurements, but the derived Hall currents (cf. Figure 2) may be taken as a relative measure for the electric field strength. In Figure 2 we see that the density peak at 06:26 UT is accompanied by an enhanced Hall current, but no air density enhancement is encountered when we pass another Hall current peak at 06:34 UT, in the evening sector. Obviously, the atmospheric heating cannot be explained exclusively by the large-scale current systems, also other current components seem to play a role.
 Our observations have shown that small-scale FAC can be very intense and the small-scale electric fields, ∂E⊥, associated with the closure currents may play an important role in atmospheric heating. Already Condescu et al.  had pointed out the importance of E-field variability. They concluded that the main Joule heating may take place in regions different from those deduced from classical observations. As can be seen from Figure 2, averaging the AC-FACs over 20 s reduces the amplitude by about two orders of magnitude. This gives an impression about the importance of high-resolution measurements for these kinds of studies.
 An event of small-scale heating, as observed by EISCAT, was modeled by Lanchester et al. . For the explanation of the measurements they had to assume Ohmic heating by very intense FACs. Similarly Otto et al.  used Ohmic heating by strong FAC filaments for the excitation of tall auroral rays. All this provides evidence for the relevance of the small-scale currents in atmospheric heating.
 Our observations show that the polar cusp is a preferred region for atmospheric up-welling. We may speculate that the electro-dynamic conditions in the cusp are favorable for a conversion of incident electro-magnetic energy into heat. Due to the lack of information about important quantities we cannot offer a conclusive explanation for the heating. Concurrent conductivity or high-resolution electric field measurements would be needed to determine the role of the various contributions in equation (3) to the heating. The combination of observations from ionospheric radars and/or sounding rockets with CHAMP measurements would help to answer the intriguing questions of thermospheric forcing and atmosphere-ionosphere coupling.
 Our observations may be related to model results of Schoendorf et al. [1996a, 1996b]. They find a four-cell pattern with high density structures at noon and midnight sectors. These cells are caused by a combination of Joule heating and neutral air circulation due to ion drag. The neutral winds have a dynamic effect on the density. Whether cyclonic flow support a high or low density cell depends on the neutral wind speed. The authors thus predict a variation of the cell structure with altitude and magnetic activity. Our data may help to verify the model results.
 In conclusion, CHAMP observations reveal the frequent occurrence of thermospheric high density structures in the ionospheric cusp region. The density enhancements are generally accompanied by very intense small-scale FAC filaments. These features occur independently of magnetic activity. Air density enhancements are more pronounced in the sun-lit than in the dark cusp. We suggest Joule heating as the prime cause for the air up-welling. The cusp seems to be a region of more or less continuous air up-flow and divergence into lower latitudes (cf. Figure 4).
 We would like to thank K. Schlegel for fruitful discussions in the context of this study. The operational support of the CHAMP mission by the German Aerospace Center (DLR) and the financial support for the data processing by the Federal Ministry of Education and Research (BMBF) are gratefully acknowledged.