Pkmet.Space Sci., Vol. 37, No. 4, pp. 37%384, 1989 Printedin Great Britain.
Cn332-0633/89 S3.00+0.00 Q 1989 PergamonFTea plc
IMAGING THE EARTH’S MAGNETOSPHERE D. W. SWIFT, R. W. SMITH d S.-l. AKASOPU Geophysical Institute and Department of Physics, University of Alaska Fairbanks, Fairbanks, AK 99775-0800. U.S.A.
(Received 11 Nouember 1988) AI&r&-The ability to image the Earth’s magnetosphere would constitute a major advance in our ability to formulate a comprehensive, time-dependent model of the magnetosphere and of substorm processes. Imaging technology is rapidly developing to the point where useful optical images could be made with sufficient spatial and temporal resolution for viewing substorm processes. We identify resonant scattering solar emissions near 834 A by O+ as the most promising means of viewing magnetospheric plasma and tracing processes within the magnetosphere. Simulations of images at 834 A observed from the Moon’s orbit, which include detail on the radiation belts and the plasma sheet, are presented and analyzed. They show that changes in the global structure are expected to be visible through changes in the brightness and
structure of features which can he identified on the image. The analysis also shows that, given expected performance of detectors and reflectors now under development for use in e.u.v., useful images could be made in 15-30 min with 1000pixels. This will give time resolution adequate for observation of structures which could change in the period of a typical magnetospheric substonn. In addition to the further development of optical devices,high resolution spectra of solar radiation are needed to make more accurate calculations of the scattering due to 0+ at 834 A.
The study of the Earth’s magnetosphere, beginning with the International Geophysical year and the discoveries of the Van Allen belts, is now over 30 y old. The intervening years have seen the discoveries and identification of the magnetopause (1963), the bow shock and neutral sheet (1964-65), the plasma sheet (1967-70), the cusp (1971-72), the mantle (1973-75), the low latitude boundary layer (1973-75), and the aurora1 potential structure (1975-76). In addition, many dynamic features of the magnetosphere have been revealed by satellites. Among them are : the flux transfer events (Russell and Elphic, 1978), plasmoids (Hones, 1979; Hones et al., 1986; Baker et al., 1987), up-flowing ions from the polar and aurora1 ionosphere (Yau et al., 1984), and the transient response of the plasma sheet (Huang et al., 1987). Our current picture of the magnetosphere has evolved as a result of the laborious process of assembling a multitude of time series records taken on satellite passes largely uncorrelated with substorm occurrence, through various parts of the magnetosphere. Since the magnetosphere is subject to large changes from one pass to the next, it has been necessary to use our emerging theoretical understanding in an attempt to extrapolate the information gained from one satellite pass to the next. As a result, after 30 y of space observations, gaps and uncertainties remain in our conceptual picture of the magnetosphere. Think how much easier it would have been if we
could “see” the magnetosphere and follow changes in magnetospheric weather, as we can see the weather in the troposphere through the formation and movement of clouds. For example, it would be highly instructive to observe the formation and detachment of plasmoids during substorms that have been described by Hones et al. (1986). Another feature of extreme interest is the response of the near-Earth plasma sheet to the growth and expansive phases of substorms (Kokobun and McPherron, 1981), since there is much in these processes that is not understood (Kaufmann, 1987). Also, the polar ionosphere is a likely source of the plasma sheet (Cladis and Francis, 1985), so it would be of interest to observe the outward streaming of ionospheric plasma detected by Yau et al. (1984). The purpose of this paper is to suggest that there are important tracers of plasma motion in the magnetosphere that may he subject to imaging. There have been many 2- and 3-D simulations (Feder and Lyon, 1987 ; Ogino, 1986) to explain these phenomena, which are capable of predicting the dynamic evolution of plasma populations in the magnetosphere. Computer graphics from these simulations are capable of producing images on a global scale which would be directly comparable with those which could be. obtained by global imaging. The ability to image the magnetosphere would therefore contribute significantly toward convergence of theory and observation in magnetospheric physics. In this paper we investigate the possibility of imag379
D. W. SWIFT et al.
ing the ma~etosphere in resonantly scattered sunlight. Hones (1987) suggested that dynamics of plasmoids could be observed from the Moon by imaging the release of Ca in the magnetotail. We present an analysis and simulation which shows that it may be possible to observe magnetospheric dynamics on a continuous basis by imaging emission from naturally occurring plasma species, namely O+. The next section will present an analysis of emission intensities from naturally occurring plasmas. The following section will present a discussion of possible instrument designs. Next we present synthetic images of what the magnetospheric plasma may look like to an observer at the distance of the Moon and show synthetic images of how the magnetosphere would look when viewed with a realizable instrument. 2. FEASIBILITY
The first requirement is that the trace emission be from plasma state. This is necessary because the medium is virtually collisiontess, so any neutrals present would be unaffected by plasma motion. Another requirement is that whatever plasma species is chosen be relatively abundant in the magnetosphere and present in only small quantities in the background solar wind. Otherwise light from the magnetospheric plasmas would be in poor contrast compared with the background light emanating from the solar wind of the much more extensive interplanetary medium. It is convenient that the emission be present at all times, and therefore scattered sunlight should be used. For efficiency, a resonant transition is most suitable. Noting these requirements, the obvious candidates are 0” and He+. The O+ ion is the second most abundant ionic species in the magnetosphere, sometimes approaching that of Hf (Lennartsson and Shelley, 1986). Moreover, it is below detectable levels in the solar wind, even though more highly ionized states of oxygen have been detected (Bame et al., 1968, 1970). The H+ ion is, of course, unsuitable for viewing, because it does not radiate photons. The only other possibility is He’, which has a strong Lyman-a resonance line at 305 A. However, this ion is far less abundant in the magnetosphere, with fractional abundances generally less than 0.01 (Lennartsson and Shelley, 1986). Bame et af. (1968) report fractional abundances of He+ of the order of 0.001 in the solar wind, but they (Bame et al., 1970) later suggested that instead of He+ they were viewing **Si+‘, so it seems unlikely that the solar wind contains a significant population of He+. Since we shall be relying on resonant scattering of sunlight, another consideration is the solar irradiance
in the resonance lines of He+ and O+. Hinteregger (1969) gives a 0.65 x lo9 photons cn? s-’ solar irradiance to the 0” and 02+ group at 843 A, while Timothy (1977) gives a 0.5 x 10’ irradiance for this same feature in the solar spectrum. These numbers are in reasonable agreement, considering the natural variability of solar emissions at this wavelength. The scattering cross-section of O+ summed over the three components of the triplet is about 5 x IO-l3 cm’. Taking Hinteregger’s value for the solar &radiance, we get a scattering rate of 3.3 x 10m4photons SK*per O+ ion. The solar irradiance at 304 8, is about 9 x 10’ photons crn2 s-’ (Timothy and Timothy, 1970). Taking the resonant scattering cross-section of He+ to be the same as Ii for the Lyman-a line, of 1.67 x lo-l3 cm2, we get a scattering rate of 1.5 x lo-’ photons s-’ per He+. When taking the relative magnetospheric abundances of O+ and He’ into account, the Of has an advantage. This analysis ignores the effect of the thermal Dop pler shift in the magnetospheric ions. This is a major un~rtainty. Assuming a 2 keV tem~rat~e, the fractional Doppler shift will be 5 x lop4 for O+ and 2 x 10s3 for He+. Since the number of scattered photons depends upon the match of the absorption lines of the plasma with the emission lines of the Sun, reduced scattering efficiency may occur if the plasma lines are broadened or shifted causing a mismatch. A definitive calculation of the scattering rate of O+ will require more detailed info~ation on the shape of the solar lines. This may require rocket observations targeted specifically at resolving the spectral features of the solar spectrum in the neighbourhood of 834 A. The present calculations assume that there is no loss of scattering efficiency due to Doppler broadening. We focus on the resonant scattering of the 834 8, solar radiation by the O+ triplet. We assume that each O+ will scatter 3.3 x 10m4photons s-‘. A typical response of a camera system (discussed in the following section) will allow it to register 1 photon SK’ for each 3.3 x lo6 O+ ions cme2 column along the line-of-sight. Assuming a density of O+ of 0.1 cm-3 (Lennartsson and Shelley, 1986) as typical of the plasma sheet and a path length of 3&, we find that total counts of the order 40 are achievable in the order of 1 min for pixels registering the brighter areas of the image. 3. INSTRUMENTAL
The preceding analysis indicates the possibility of obtaining useful magnetospheric images from the Of emission at 834 A and possibly He at 304 A. The major challenge is to build instrumentation which will
Imaging the Earth’s magnetosphere
FIG.~.RESULTSOFMODELCALCULATIONSHOW~NGIMAGEOFTHEMAGNETOSPHEREINRESONANTLYSCATTERED SUNLlGHTFXOMO+. The scenes A-D are 40” x 40” views from a distance of 60Rr from the center of the Barth and
30” above the eliptic plane. (A) is viewed from 20:00 L.T. toward the center of the Earth, (B) is looking from 18:OO L.T. to a point IO& behind the Earth, while (C) and (D) show views looking toward the Barth from 1490 and 1690 L.T., respectively. The dark hole is caused by the Earth obscuring the plasma behind it. E shows the count rate, assuming a 10% efficiency. This is taken along the white trace shown in the image C, with the sweep from the magnetotail toward the Sun to the lower left.
Fro 2. (A) THE SAME
VIEW AS SHOWN
SWIFT et al.
IN FIG. lC, BUT REPRESENTED BY A 1.2"X 1.2"RESOLUTION.
(B) The same as A, except that the results of simulation of a Poisson distribution of the number of photon counts in each pixel is shown. (C) the same as B, except that a 2 pixel wide Gaussian filter has been applied. (D) Same as C, except that 0.67 pixel filter has been used ; and (E) is the same, but a 6 pixel wide filter has been used on B.
Imaging the Earth’s magnetosphere make an image of the magnetosphere sufficiently rapidly using these photons, i.e. on the scale of a few minutes so as to see a magnetospheric substorm lasting on the order of an hour. The e.u.v. region of the spectrum is difficult to use for reflectors, detectors and dispersing elements. It is impossible to use transmitting optics because of the high absorption at these wavelengths. Spectral imaging in the ultraviolet has been carried out by McCoy et al. (1986) and Kumar et al. (1983) using a concave grating spectrometer as a combined imaging and dispersing optical element. Such techniques require scanning of the object in 1-D at least since the 2-D images which could be obtained have one direction used for dispersion. The price for use of a single element high efficiency e.u.v. detector, such as a cone channeltron, is the requirement to scan in 2-D. Such a design was not found to be favorable. Traditional methods using grazing incidence are unacceptable here, and normal incidence reflectivity is less than 10% for metals. Recent developments in multilayer Bragg reflectors have now been proven useful in this region (Walker et al., 1988). Designs for these reflectors may be optimized for narrow band reflectivity (Meekins et al., 1986) and promise reflectivity near normal incidence reaching a theoretical 70%. The narrow band nature of their performance provides the predispersion required for the selection of the 834 A emission with a bandwidth of 10-20 A. Hence, it is possible to construct a “monochromatic” camera using a mirror made in this manner without the need for an additional predisperser. One design under consideration calls for a reflecting camera using a 50 cm’ reflector with a 20” half-angle field configured similarly to the Schmidt camera. No corrector plate is required for the 1000 pixel image which avoids using transmitting optics. Spectral predispersion is provided by the mirror as described above. The light reflected by the mirror will be imaged on a windowless microchannel plate detector having a position sensitive anode. Photons recorded by the detector are processed serially by computer to be stored to the appropriate pixel of the image matrix in memory. Assuming a combined reflectance for two surfaces of 30% and a detector efficiency of 8%, such a system, as mentioned in Section 2, will record 1 photon s-’ for each 3.3 x lo6 ions in a 1 cm2 column along the line-of-sight. Images of the magnetosphere containing 1000 pixels may be made in the order of minutes. 4. THE MAGNETOSPHERE
AS VIEWED IN 0+
To obtain some idea what the magnetosphere might look like when viewed by resonantly scattered
sunlight, we have constructed a simple model of the magnetospheric O+ plasma number density. This consists of an asymmetric radiation belt on dipole magnetic field-lines. The lines intersect the Earth between 60” and 75” latitude on the night side and 70” and 75” latitude on the dayside. The density of the radiation belt is assumed to vary as n = 0.75/R, where n is in cm3, and R is the geocentric distance in Earth radii. There are assumed to be no particles earthward of R = 1.5. A simple slab model was chosen for the plasma sheet. The plasma sheet was chosen to be a slab 2Rr thick and extending from x = - 10 to -4ORE and y = - 15 to 15Rz with a uniform density of 0.1 cm-‘. The contribution of the ionospheric plasma was blanked out, because it would be so bright that it would overwhelm the contributions from the radiation belt and plasma sheet. The results of the model calculation are shown in Fig. 1. This figure shows four separate views of the magnetosphere at a distance of the orbit of the Moon, 60RE. The figure was made by computing the line-ofsight densities along the view directions. The field of view is 40” x 40”, and the resolution in this picture is 0.4”. All frames are viewed from 30” above the ecliptic. The trace shown in the bottom frame shows the counting rate for a 50 cm’ detector that could be expected along the scan line shown in the frame immediately above. The sweep starts from the magnetotail and proceeds sunward. In addition to blanking out the ionosphere, the Earth was allowed to hide from view the plasma behind it, but there was no provision for shadowing the plasma from sunlight. The images shown in Fig. 1 show the type of perspective that can be obtained from the distance of the Moon, and it can be seen that this offers a rather good vantage point to view most of the plasma in the Earth’s magnetosphere. However, these pictures show a level of resolution considerably better than is likely to be realized with instruments than can be built in the foreseeable future. Figure 2 provides an idea of image quality that can be obtained with an instrument employing a 50 cm2 collecting area and operating at a 10% efficiency. The images show the same view as shown in Panel C of Fig. 1. The frames in Fig. 2 show 33 x 33 pixel arrays, with an angle across each pixel of 1.2”. It is assumed that the detector spends 1 s collecting photons from each pixel, so that it takes about 18 min to assemble the entire frame. The image in Panel A was obtained by averaging over each nine pixels in each 3 x 3 pixel array in Fig. 1C. The total number of counts per pixel varies from about 300 in the brighter part of the radiation belt to about 40 in the plasma sheet to the upper right. At these counting rates, the image quality will be affected by counting
D. W. Swa
statistics. In order to evaluate this effect, a random number generator was used to generate for each pixel a Poisson distribution of counts. The result is shown in panel B of Fig. 2. Panels C-F show the results of smoothing in an attempt to enhance image quality. The smoothing was done by convoluting the 33 x 33 pixel images with the normalized Gaussian, P(x) = exp [ - x?/(Sx)*]. Panels C and D show the results of smoothing the images shown in Panels A and B, respectively, with 6x = 2, where the distance across each pixel is taken as unity. This level of smoothing completely removes the graininess of the image in A, while there is some indication of remaining graininess from the image B. Panels E and F show the results of smoothing the image in B with 6x = 0.67 and 6, respectively. From these examples we saw that attempts to completely remove the effect of the statistical fluctuation produces an unacceptable loss of resolution. The display shown in Fig. 2 suggests that an acceptable level of resolution for recognition of major magnetospheric features on a time scale of one third of an hour can be obtained with a collector placed on the Moon with a 50 cm* collecting area and an angular resolution of about a degree. Satisfactory images could be obtained with current designs only with the sacrifice of time resolution by an order of magnitude. 5. CONCLUDING REMARKS
Our preliminary analysis indicates that global imaging of the magnetosphere is not out of reach of even today’s technology. There is little doubt that both in situ satellite observations and the simultaneous images of the magnetosphere will considerably advance our understanding of magnetospheric substorm processes. REFERENCES
Baker, D. N., Bame, S. J., McComas, D. J., Zwickl, R. D., Slavin, J. A. and Smith, E. J. (1987) Plasma and magnetic field variations in the distant magnetotail associated with near-earth substorm effects, in Magnetotail Physics (Edited by Lui, A. T. Y.), p. 137. Johns Hopkins University Press, Baltimore. Bame, S. J., Asbridge, J. R., Hundhausen, A. J. and Montgomery, M. D. (1970) Solar wind ions: %Fe+* to %Fe+“, ‘%i+‘, **Si+s, 28Si+9,and ‘60+6. J. geophys. Res. 75,636O. Bame, S. J., Hundhausen, A. J., Asbridge, J. R. and Strong, I. B. (1968) Solar wind ion composition. Phys. Rev. L&t. 20,393.
Cladis, J. B. and Francis, W. E. (1985) The polar ionosphere
as a source of the storm-time ring current. J. geophys. Res. 90,3465.
Feder, J. A. and Lyon, J. G. (1987) The solar wind-magnetosphereionosphere current relationship. Geophys. Res. L&t. 14, 880. Hinteregger, H. E. (1969) Effects of solar XUV radiation on the Earth’s atmosphere, in Annals ofthe ZQSY (Edited by Strickland, A. C.), Vol. 5, p. 305. MIT Press, Cambridge, Massachusetts. Hones, E. W., Jr. (1979) Transient phenomena in the magnetotail and their relation to substorms. Space Sci. Rev. 23, 393. Hones, E. W., Jr. (1987) Studying space plasmas from a lunar base. J. Spacecraft Rockets 24, 74. Hones, E. W., Jr., Fritz, T. A., Birn, J., Cooney, J. and Bame, S. J. (1986) Detailed observation of the plasma sheet during a substorm on April 24, 1969. J. geophys. Res. 91, 6845. Huang, C. Y., Frank, L. A., Peterson, W. K., Williams, D: J., Lennartsson, W., Mitchell, D. G., Elphic, R. C. and Russell, C. T. (1987) Filamentary structures in the magnetotail lobes, J. geophys. Res. 92, 2349. Kaufmann, R. L. (1987) Substorm currents : growth phase and onset. J. geophys. Res. 90,747 1. Kokobun, S. and McPherron, R. L. (1981) Substorm signatures at synchronous altitude. J. ge&hyi. Rex 86,11265. Kumar, S.. Chakarabarti. S. S.. Pearsce. F. and Bowver. S. (1983) The O+ 834 A dayglow : satellite observations and interpretation with a radiation transfer model. J. geophys. Res. 88, 927 1. Lennartsson, W. and Shelley, E. G. (1986) Survey of 0.1 to 15 krV plasma sheet ion composition. J. geophys. Res. 91, 3061. McCoy, P. P., Wolfram, K. D., Meier, R. R., Paxton, L. J., Cleary, D. D., Prinz, D. K., Anderson, D. E., Jr., Christensen, A. B., Pranke, J., Sivjee, G. G. and Kayser, D. (1986) The remote atmospheric ionospheric detection system, SPZE 687, 142. Meekins, J. F., Cruddace, R. G. and Gursky, H. (1986) Optimization of layered synthetic microstructures for narrowband reflectivity at soft X-ray and EUV wavelengths. Appl. Optics 25, 2757.
Ogino, T. (1986) A three-dimensional MHD simulation of the interaction of the solar wind with the Earth’s magnetosphere: the generation of field-aligned currents. J. geophys. Res. 91,679l. Russell, C. T. and Elphic, R. C. (1978) Initial ZSEE magnetometer results, magnetopause observations. Space Sci. Rev. 22,681. Timothy. J. G. (1977) The solar spectrum between 300 and 1200 A, in Solar Output and its Variation (Edited by White, 0. R.), p. 237. Colorado Associated University Press, Boulder. Timothy, A. F. and Timothy, J. G. (1970) Long-term intensity variations in the solar Helium II Lyman alpha line. J. geophys. Res. 75, 6950.
Walker, A. B. C., Jr., Barbee, T. W., Jr., Hoover, R. B. and Luindblom, J. F. (1988) Soft x-ray images of the solar corona with a normal incidence camgrain multilayer telescone. Science 241. 1781. Yau, A. W., Whalen, B. A., Peterson, W. K. and Shelley, E. G. (1984) Distribution of upflowing ions in the highaltitude polar cap and aurora1 ionosphere. J. geophys. Res. 89,5507.