Large Scale Structures in the Magnetosheath: Exogenous or Endogenous in Origin?

X. X. Zhang1,2, P. Song1, S. S. Stahara3, J. R. Spreiter3, C. T. Russell4, and G. Le4


1. High Altitude Observatory, National Center for Atmospheric Research, Boulder, CO 80307, U.S.A.

2. Center for Space Science and Applied Research, Academia Sinica, Beijing 100080, P. R. C.

3. RMA Aerospace Inc., Mountain View, CA 94043, U. S. A.

4. Institute of Geophysics and Planetary Physics, University of California, Los Angeles, CA 90024, U. S. A.

Originally published in: Geophysical Research Letters, 23, 105-108, 1996
July 24, 1995; revised: October 16, 1995; November 7, 1995



          Observations of the solar wind and the interplanetary magnetic field from ISEE-3 are used as input to the gasdynamic convected field model, as implemented in a new space weather forecast model. Then the model output, for three case studies, is compared with the magnetosheath quantities observed at ISEE-2 in order to identify the sources of the observed variations of the magnetosheath. It is found that some variations in the magnetosheath plasma and magnetic field are well correlated with corresponding variations in the solar wind and hence have their sources in the solar wind. However, some variations in the magnetosheath magnetic field correlate well with those in the solar wind but not variations in plasma density. Finally, we find that other variations in both plasma and magnetic field in the magnetosheath do not have appreciable correlations with variations in the solar wind. Most of these latter variations occur in the inner magnetosheath, indicating that they are endogenous in origin. Our results show that the forecast model can provide an accurate estimate of the timeshift from the solar wind monitor to the magnetosheath, of the instantaneous locations of the bow shock and magnetopause, and of the properties of the plasma and magnetic field in the outer and middle magnetosheath.



          Changes in the solar wind plasma and the interplanetary magnetic field (IMF) influence the processes in the magnetosphere, and are important sources of many dynamic features observed at the magnetopause and in the magnetosphere ( Elphic and Southwood, 1987; Kivelson and Southwood, 1991; Song et al., 1992; Russell et al., 1992; Le et al., 1993).

          However, it is the nature of the plasma and magnetic field of the magnetosheath downstream of the bow shock that directly influences the magnetopause. Since the upstream variations in the solar wind plasma can be significantly modified upon traversing the bow shock and magnetosheath (Yan and Lee, 1994), the magnetosheath is an important region that any realistic space weather forecast model cannot ignore.

          Several theoretical models have been developed to understand the plasma and magnetic field properties in the transition region from the bow shock to the magnetopause. These models differ in the role that magnetic forces play in them. (Spreiter et al., 1966) proposed a gasdynamic model in which ordinary sound waves determine the properties of plasma flow and field draping around the magnetopause and magnetic forces play no role. Lees (1964) and Zwan and Wolf (1976) invoked slow mode processes inherent to magnetized fluids and found a depletion effect of the magnetosheath plasma. Wu (1992) investigated the formation of a depletion layer close to the magnetopause with a 3-D MHD calculation.

          Song et al. (1990, 1992) studied processes in the magnetosheath using data from ISEE-1 and 2 and discovered a region of plasma density enhancements and field depression near the magnetopause having a relatively large spatial scale. They also inferred that in some cases this slow-mode structure was locally generated in the magnetosheath as part of the interaction of the magnetosheath with the magnetosphere and was not one carried into the magnetosheath by the solar wind. Moreover the plasma depletion layer just outside the magnetopause often appears to be associated with a decline in density beginning at the slow-mode structure. Hammond et al. (1995) have recently reported a similar structure in the Jovian magnetosheath. Using a two-dimensional MHD simulation, Lee et al. (1991) showed that such a structure can be formed close to the stagnation region. Southwood and Kivelson (1992) suggested that slow-mode structures or slow wave fronts can be formed by sources of disturbances at the magnetopause. The structure may play an important role in establishing the flow and field pattern near the magnetopause (Song et al., 1992).

          Omidi and Winske (1995) pointed out that the slow-mode wave front may originate at the magnetopause and that upstream mirror mode waves may play some roles. On the other hand,Yan and Lee (1994) drew attention to the possibility that slow-mode structures can be formed through the interaction between interplanetary rotational discontinuities and the bow shock, and hence some slow-mode magnetosheath variations may have their sources in the solar wind. Therefore, in order to understand the properties of the slow-mode structure and other processes in the magnetosheath, it is important to identify the sources of the variations observed in the magnetosheath. This work will examine in addition the validity and limitations of the gasdynamic convected field model.



          In principle, a direct approach to study large scale structures in the magnetosheath would be to solve the three-dimensional magnetohydrodynamic (3-D MHD) equations in a realistic magnetosheath geometry. However, at present, there is no convenient time-dependent global 3-D MHD model available for routine use in comparative studies of magnetosheath observations. In this paper we use the gasdynamic convected field model (GDCFM) (Spreiter et al., 1966; 1968; Spreiter and Stahara, 1980) recently implemented in a space weather forecasting model (Stahara and Spreiter, manuscript in preparation) to study the correlation of variations in the magnetosheath with those in the solar wind. In this model the gasdynamic flow solution is calculated ignoring magnetic forces. Then the magnetic field lines are computed by convecting them through the fluid like dye lines or threads exerting no force on fluid. We have found that the gasdynamic forecast model can provide three important baselines for magnetosheath studies. First, the model can provide a relatively accurate timeshift between the solar wind monitor and the magnetosheath observer. Second, the model provides reference estimates of magnetosheath properties including the density, bulk velocity, temperature and magnetic field throughout the magnetosheath. Third, the model provides the expected locations of the bow shock and the magnetopause. Although the predictions are only an approximation to the fully time-dependent 3-D MHD solution, the physics included in the model is well understood, and exogenous variations are predicted systematically. Therefore, we use the model prediction as a baseline to study the frequently observed variations in the magnetosheath. We use the solar wind plasma and IMF measurements to drive the forecast model and select the point for the prediction according to the location of the satellite in the magnetosheath. Comparison of the results from the GDCFM (which approximates the fast but not Alf-ven and slow mode processes) with magnetosheath observations can separate the effects of solar wind variations from the magnetosheath Alf-ven and slow mode processes, and serve to identify the sources of large scale variations involving these structures in the magnetosheath.

          The magnetosheath data are from the ISEE-2 magnetometer (Russell 1978) and Fast Plasma Experiment (FPE) (Bame et al., 1978a). The solar wind data are obtained by the magnetometer (Frandsen et al., 1978) and solar wind analyzer (Bame et al., 1978b) on ISEE-3. In this letter, we present three crossings in detail. The first case involved strong variations in both the solar wind plasma density and IMF. The second involved significant changes only in the IMF, but not in the solar wind plasma properties. The third occurred during relatively steady solar wind and IMF conditions.



Case 1.
          Figure 1 shows an outbound crossing of the magnetosheath by ISEE-2 on September 12, 1978. Solid lines are plasma and magnetic field properties in GSE coordinates from ISEE-2 with 12 sec. solution and dashed lines are the forecast model predictions with a time solution of 1 min. ISEE-2 was located at (8.7, -1.9, 3.9)R GSE at 2000 UT and (13.3, -0.1, 5.4)R GSE at 2400 UT. ISEE-3 was upstream of the bow shock in the solar wind near (207, -68, 18)R GSE during this period. The timeshift between ISEE-3 and ISEE-2 is about 50 min.

Figure 1. An outbound magnetosheath crossing by ISEE-2 (solid lines) and corresponding GDCFM prediction (dashed lines). The positions of the magnetopause and the bow shock are indicated by arrows. The solar wind Mach number M=6.0, Alfven Mach number M=7.5 and =1.44

          Several strong plasma density enhancements with large time scales are observed in the magnetosheath. The enhancement between 2145 and 2217 UT is well predicted by the model. Furthermore, the model predicts that ISEE-2 is very close to the bow shock during this time as evidenced by a pair of predicted but not observed bow shock crossings. The density prediction in general is good near the bow shock but higher than observed near the magnetopause. The difference is as large as 60o . There is a prediction of multiple magnetopause crossings, which are not actually recorded by ISEE-2. The actual slowdown of the magnetosheath flow as evidenced in the V component is significantly less than the prediction. Since the FPE was not designed to measure the solar wind and has widely separated energy channels [Bame et al., 1978a], its velocity in the solar wind is different in magnitude from that of ISEE-3. The V component of the observed velocity is positive from 2145 UT to the bow shock while the y component of ISEE-2 position is negative. The difference in sign is due to the aberration effect that is properly included in the forecast model. Prediction of the three components and magnitude of magnetic field is in good agreement with the observation except for some small regions principally from the magnetopause to 2145 UT. Because the model is a single fluid model, the temperature is considered to be the sum of the proton and electron temperatures. The predicted temperature is about 50% lower than the observation, and the difference becomes somewhat greater corresponding to the slow mode structures.

Case 2.
          An outbound crossing of the magnetosheath on September 17, 1978 by ISEE-2 is shown in Figure 2 (in the same format as in Figure 1). This case has been previously studied by Song et al. [1992]. ISEE-2 was at (9.0, -2.4, 4.1)R GSE at 1500 UT and (12.5, -1.4, 5.2) R GSE at 1800UT. ISEE-3 was located at (211, -80, 17)R GSE. The timeshift between ISEE-3 and ISEE-2 observations is approximately 57 min. The solar wind was relatively steady, while the IMF rotated about 130o near 1510-1540UT and about 54o near 1550-1610UT.

Figure 2. An outbound magnetosheath crossing in the same format as Figure 1. The solar wind Mach number M=5.0, Alfven Mach number M=9.0 and 3.2

          The magnetic field prediction agrees very well with the observations throughout the magnetosheath except for a thin region near the magnetopause. ISEE-2 detected plasma density enhancements from 1533 UT to 1610 UT with the peaks more than double the background value. The model predictions of density during this time show no indication of this structure. Again, the magnitude of V is significantly larger than the perdiction. The V is reasonably well predicted, but the observation shows a significant additional deflection within the large-scale slow mode structure. There is little fluctuation in observed and predicted temperatures, but the prediction again is approximately 50% lower than the observation. Within the large-scale slow mode structure, the correlation with the solar wind is good for the direction of the magnetic field but the predicted plasma density shows little correlation with the observed density variations.

  Case 3.
          Figure 3 presents an inbound magnetosheath pass on September 5, 1978 by ISEE-2 (in the same format to Figure 1). ISEE-2 was at (8.6, 10.0, 2.0 )RGSE at 0200 UT and (7.6, 9.5,1.7) R GSE at 0500UT. ISEE-3 was located at (195.6, -48.2, 18.3)R GSE. The timeshift fom ISEE-3 to ISEE-2 is about 51 min. A major difference between this case and previous two is that this pass occurred at a much greater solar zenith angle, about 50o from the stagnation streamline.

Figure 3. An inbound magnetosheath crossing in a format similar to Figure 1, but with plasma the ratio of the thermal pressure to the magnetic pressure shown in the bottom panel. The solar wind Mach number M=5.5, Alfven Mach number
M=9.0 and =3.0

          Again, the model predictions for this pass are in good agreement with the observations in the magnetosheath except for a small region near the magnetopause. The changes of the plasma density and the magnetic field predicted by the model are very small for this pass because of the steady upstream condition. The slow mode structure with a plasma density enhancement and field decrease occurs from 0410 UT to 0440 UT and is not predicted by the model. The predictions of velocity components are much better than the other two cases and the aberration effect may become unimportant because the satellite was far from the stagnation streamline (nearly 10R). The plasma is a good indicator of slow mode processes because the thermal and magnetic pressures change out of phase. The predicted and observed s are about the same in the magnetosheath other than within the slow mode structure. There is no correlation between plasma density and magnetic field variations in the slow mode structure with those in the solar wind (after having examined the corresponding conditions in the solar wind).


Summary and Discussion

          In the three cases we have shown, the large time scale variations of the plasma density and magnetic field in the magnetosheath have different sources. Case 1 shows that some variations in the magnetosheath are driven by sources in the solar wind as evidenced by the good correlation of the variations of density and magnetic field between the solar wind and the magnetosheath. Such variations in the solar wind can penetrate the bow shock into the magnetosheath and may be modified in the nonuniform magnetosheath medium [Yan and Lee, 1994]. Case 2 shows that there is good correlation of the magnetic field in the solar wind and in the magnetosheath, but corresponding plasma density enhancements in the inner magnetosheath do not necessarily have a solar wind origin. Although large-scale slow-mode variations may be generated by variations in the IMF direction without plasma density variations through wave-wave interactions at the bow shock, the structure for this case does not seem to be generated this way because of the smallness of the field rotation (less than 54o) associated with the outer front of the slow-mode structure. For case 3, there seems to be no appreciable correlation in the plasma density or magnetic field between the solar wind and the inner magnetosheath variations from the observed structure although they are well-correlated in the outer and middle magnetosheath. This case shows the slow-mode structures to be locally generated in the magnetosheath rather than carried by the solar wind and their occurrence is independent of the variations of the IMF direction. This result confirms the interpretation by Song et al. [1992].

          For all these cases studied, there is an overestimate of the density prediction near the magnetopause. In some cases, the difference can be as large as 60%. Of the 60%, less than 10% may be accounted for by the compression included in the GDCFM. The remaining more than 50\% is most likely due to the plasma depletion effect. We note that the depletion effect appears to occur over the entire magnetosheath and not only in the small region near the magnetopause. We will further investigate this phenomenon statistically. Another interesting phenomenon is that the flow often slows down much less in the magnetosheath than predicted.

          Determination of the timeshift has been a crucial issue when correlating the variations in the solar wind with those in the magnetosheath, and then with subsequent variations on the magnetopause, in the magnetosphere and ionosphere and on the ground. In the past several methods have been used: (1) dividing the distance in the x direction between the two spacecraft by the measured solar wind velocity; (2) determining the propagation delay when the solar wind monitor is not on the Sun-Earth line, assuming that the surface of constant IMF and solar wind contain both ecliptic pole and the path spiral direction; (3) similar to (2) but determining the normal to the surface of constant solar wind conditions by some technique such as minimum variance at the upstream monitor; and (4) shifting the time in order to maximize the correlation between the clock angle of the IMF and that of the sheath field [Song et al., 1992]. There are large uncertainties in the time delay using each of methods (1),(2) and (3). While method (4) is significantly better, only one timeshift is provided and it cannot account for the changes in the time shift when the solar wind velocity changes. Our method is similar to method (1) but with the advantage that the variation of velocity along the streamline with magnetosheath is properly accounted for prediction.

          From the three presented cases, we have found that the thickness of magnetosheath is well predicted by the model, i.e., one may change either the magnetopause or the bow shock crossing time to improve its timing prediction, but it will worsen the other. Furthermore, the time shifts we used determine well the time of those field changes in the sheath whose origin was in the wind. Practically, one has no freedom in shifting the time. The bulk velocity, magnetic field and plasma density are in general well predicted by the model in the outer and middle magnetosheath while the predictions may become significantly different from the observations in the region near the magnetopause. As a whole, we find that the forecast model can provide a relatively accurate time shift with an uncertainty of less than 10 min., provide good reference locations of the bow shock and the magnetopause, and predict the magnitudes of the parameters reasonably well in the outer and middle magnetosheath. This justifies the use of the model predictions as a baseline to correlate the variations in the magnetosheath with those in the solar wind. The differences near the magnetopause and within the slow mode structure are due to the MHD effects not included in the model.



XXZ would like to thank the HAO/NCAR for support as a visiting scientist, and was sponsored by the Chinese Academy of Sciences and the NSF of China under grant 49404054. Work at HAO was sponsored by the NSF and supported by NASA under research grant W-18,582. Work at RMA was supported by the NSF under research grant ATM-9301022. Work at UCLA was supported by NASA under research grant NAGW-3948. We thank J. T. Gosling for providing us ISEE-2 data and NSSDC for ISEE-3 data.



Bame, S. J., et al., ISEE 1 and ISEE 2 fast plasma experiment and the ISEE1 solar wind experiment, IEEE Trans. Geosci. Electron., GE-16, 216, 1978a.

Bame, S. J., et al., ISEE-C solar wind plasma experiment, IEEE Trans. Geosci. Electron., GE-16, 160, 1978b.

Elphic, R. C., and D. J. Southwood, Simultaneous measurements of the magnetopause and flux transfer events at widely separated sites by AMPTE UKS and ISEE 1 and 2, J. Geophys. Res., 92, 13666, 1987.

Frandsen, A. M. A., et al., The ISEE-C vector helium magnetosheath, IEEE Trans. Geosci. Electron., GE-16, 195, 1978.

Hammond, C. M., J. L. Phillips, and A. Balogh, Ulysses observations of slow mode transitions in the Jovian magnetosheath, Eos Trans. AGU, 76(17), S251,1995 1995.

Kivelson, M. G., and D. J. Southwood, Ionospheric traveling vortex generation by solar wind buffeting of the magnetosphere. J. Geophys. Res., 96, 1661, 1991.

Le, G., C. T. Russell, S. M. Petrine, and M. Ginskey, Effect of sudden solar wind pressure changes at subaurora latitudes: Change in the magnetic field. J. Geophys. Res., 98, 3983, 1993.

Lee, L. C., M. Yan, and J. G. Hawkins, A study of slow mode structure in front of the dayside magnetopause, Geophys. Res. Lett, 18, 381, 1991.

Lees, L., Interaction between the solar wind and the geomagnetic cavity, AIAA J., 2, 1576, 1964.

Omidi, N., and D. Winske, Structure of the magnetopause inferred from the Kinetic Riemann problem, J. Geophys. Res. in press, 1995.

Russell, C. T., The ISEE 1 and 2 fluxgate magnetometers, IEEE Trans. Geosci. Electron., GE-16, 239, 1978.

Russell, C. T., M. Ginskey, S. M. Petrince and G. Le, Effect of solar wind pressure changes on low and mid-latitude magnetic records. Geophys. Res. Lett., 19, 1227, 1992.

Song, P., C. T. Russell, J. T. Gosling, M. F. Thomsen, R. C. Elphic, Observation of the density profile in the magnetosheath near the stagnation streamline, Geophys. Res. Lett., 17, 2035, 1990.

Song, P., C. T. Russell, and M. F. Thomsen, Slow mode transition in the front side magnetosheath, J. Geophys. Res., 97, 8295, 1992.

Southwood, D. J., and M. G. Kivelson, On the form of the flow in the magnetosheath, J. Geophys. Res., 97, 2873, 1992.

Spreiter, J. R., A. L. Summers, and A. Y. Alksne, Hydromagnetic flow around the magnetosphere, Planet. Space Sci., 14, 2230, 1966.

Spreiter, J. R., and S. S. Stahara, A new predictive model for determining solar wind-terrestrial planet interactions, J. Geophys. Res., 85, 6769, 1980.

Wu, C. C., MHD flow past an obstacle: Large scale flow in the magnetosheath, Geophys. Res. Lett., 19, 87, 1992.

Yan, M., and L. C. Lee, Generation of slow-mode waves in the front of the dayside magnetopause, Geophys. Res. Lett., 21, 629, 1994.

Zwan, B. J., and R. A. Wolf, Depletion of solar wind near a planetary boundary, J. Geophys. Res., 81, 1636, 1976.

Back to CT Russell's page More On-line Resources
Back to the SSC Home Page

Last Modified: July 22, 1996