SEARCH

SEARCH BY CITATION

Keywords:

  • geomagnetic storms;
  • IMAGE FUV;
  • topside ionosphere;
  • tomography

Abstract

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data and Method
  5. 3. Observations
  6. 4. Discussion
  7. 5. Summary
  8. Acknowledgments
  9. References

[1] We analyze the effects on the Southern Hemisphere ionosphere and plasmasphere due to the 29–31 October 2003 geomagnetic storms (the so-called series of Halloween storms). Solar wind data from ACE and ionospheric data from the GPS (Global Position System) ground and LEO (Low Earth Orbit) receivers, the TOPEX/Poseidon altimeter, the IMAGE FUV camera, and the DMSP drift meter are used to understand the ionospheric dynamics as a function of the storm phase. The detailed structure of the ionosphere has been obtained using tomographic reconstruction applied to data from both ground- and space-based GPS receivers. The tomographic approach using LEO observations of signals received from GPS satellites above the LEO's horizon allows us to investigate the topside ionospheric and plasmaspheric density distribution in more detail than can be obtained using ground-based GPS receivers. This is because with ground-based receivers, the higher topside ionosphere and plasmasphere contribute only a small fraction to the total electron content (TEC) and so the measurements are dominated by the ionospheric structure at the F2 peak. In contrast, the Australian LEO satellite, FedSat, which has been used for this study, orbits at 800 km altitude, well above the F2 peak and hence the TEC measured is primarily due to the upper topside ionosphere and plasmasphere. This paper presents the tomographically reconstructed topside ionosphere and plasmasphere electron density distributions using LEO observations. The temporal and regional maps of TEC and the IMAGE FUV data show that the storm that commenced on 29 October dramatically decreased the plasma density in the Southern Hemisphere middle and high latitudes. The region remained depleted of plasma for more than 24 hours until 31 October, when the second severe storm began. TOPEX/Poseidon data shows a daytime localized density enhancement occurred above the middle of the Pacific Ocean. These results show large interhemispheric and longitudinally narrow storm-time structure in the ionosphere and topside ionosphere/plasmasphere.

1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data and Method
  5. 3. Observations
  6. 4. Discussion
  7. 5. Summary
  8. Acknowledgments
  9. References

[2] Magnetic storms dissipate large amount of energy in the polar regions, which leads to profound changes in the upper atmosphere on a global scale [Fuller-Rowell et al., 1997; Prölss, 1997; Buonsanto, 1999; Yizengaw et al., 2004b, and the reference therein]. Among these changes is a substantial increase in the Joule and particle heating rates of the ionosphere. Such high-latitude heating causes strong upwelling of the atmosphere around the auroral oval. This moves oxygen-depleted or nitrogen-rich air up from much lower in the thermosphere to F region altitudes [Prölss, 1997; Buonsanto, 1999], leading to a significant decrease of the column integrated concentration ratio of atomic-to-molecular ([O]/[N2] or [O]/[O2]) throughout the entire high-latitude thermosphere. If the thermospheric dynamical region stayed unchanged during magnetic disturbances, the zone of depleted [O]/[N2] would be limited to the high-latitude ionosphere (approximately the auroral oval). However, the heating induces its own circulation, which at the F2 layer heights tends to bring the air equatorward to lower latitudes [Buonsanto, 1999].

[3] The storm-induced circulation, which is directed equatorward, opposes the “regular” circulation [see Fuller-Rowell et al., 1997; Danilov, 2001; Buonsanto, 1999; Yizengaw et al., 2004b; Crowley et al., 1989]. It is stronger at night, since it adds to the background circulation (i.e., equatorward at night and poleward during day) and because they are reinforced by antisunward ion drag due to magnetospheric convection E × B drift. During a severe magnetic storm period, like the Halloween storm, such molecularly rich air expands to middle and lower latitudes, more strongly at night, and then corotates with the Earth to the morning sector, creating a daytime depletion zone at middle and low latitudes.

[4] The equatorward expansion of the depleted zone also has seasonal differences. It reaches to lower latitudes in summer than in winter, as the total wind field includes a prevailing transequatorial summer-to-winter flow that restricts the equatorward motion of the composition changes in winter while allowing them to reach lower latitudes in summer.

[5] Since the first suggestion by Seaton [1956], it was believed that the molecularly enriched air causes a reduction in the ionospheric electron density. The electron concentration is, roughly speaking, directly proportional to the [O]/[N2] ratio at the F2 layer maximum height [Rishbeth et al., 1987]. The close relation between the [O]/[N2] ratio and the ionospheric electron density concentration in the F2 region, measured at several ionospheric stations, are well documented by Prölss et al. [1991], Fuller-Rowell et al. [1994], Buonsanto [1999], Danilov [2001], Strickland et al. [2001], and Prölss and Craven [1998]. Furthermore, the electron density depletions observed during the 29–31 superstorm at F layer height is partly related to the effect of vibrationally excited molecular nitrogen (N2) and oxygen (O2) in the loss rate of O+(4S) ions [Pavlov and Foster, 2001; Pavlov, 1994]. Pavlov and Foster [2001], using ionospheric measurement at Millstone Hill during 15–16 July 2000 severe storm, have identified the effects of vibrationally exited N2 and O2 on F region electron density and found a decrease of daytime peak density by up to a factor of ∼3.

[6] In this paper we address the interesting features of the Southern Hemisphere ionospheric response to the extremely high-speed solar wind shock event and related magnetic storm effects around 29–31 October 2003. The resulting intensified high-latitude heating appear to control the major part of the observed plasma density features, leading to the unusual density distribution difference between the two conjugate hemispheres. Magnetic storm effects in both the storm main and recovery phases are discussed using data from remote sensing and satellite in situ instruments. We will use ground-based GPS data, IMAGE FUV data, and satellite altimeter (TOPEX) data. The DMSP ion drift data and topside ionospheric and plasmaspheric tomography are also incorporated in this research to detect the unusual ionospheric depletion at higher altitudes.

2. Data and Method

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data and Method
  5. 3. Observations
  6. 4. Discussion
  7. 5. Summary
  8. Acknowledgments
  9. References

2.1. GPS TEC

[7] Owing to the dispersive nature of the ionosphere, dual frequency (f1 = 1.57542 GHz and f2 = 1.2276 GHz) GPS measurements can provide integral information about the ionosphere by computing the differential phases of code and carrier phase measurements recorded at ground-based GPS receivers [Klobuchar, 1996; Makela and Kelley, 2001; Yizengaw et al., 2004b]. Methods of TEC calculation from GPS observations have been described in detail in several papers [Sardón et al., 1994; Horvath and Essex, 2003a, 2003b, and the references therein]. More specifically, Yizengaw [2004] and Yizengaw et al. [2004b] have clearly outlined the algorithm that has been used in this paper. For convenience, TEC is usually measured in TECU (1 TECU = 1 × 1016 el m−2). In order to describe the overall ionization of the ionosphere above a certain local area of interest, the integrated electron density, along the raypath from the satellite-to-ground, which is called the slant TEC, is then mapped to the vertical by using a single layer approximation for the ionosphere at hsp = 400 km height [Yizengaw et al., 2004b]. Since the GPS TEC at higher zenith angle is highly exposed to multipath effect [Klobuchar, 1996], data for zenith angles greater than 75° are not used for this study.

2.2. TOPEX Altimeter TEC

[8] TOPEX uses a dual-frequency altimeter, which transmits pulses in the Ku-band (13.8 GHz) and the C-band (5.6 GHz), to measure the range delay difference between the two frequencies and to compute the ionospheric range error, ΔRion (see equation (1)). Since the free electrons in the subsatellite path retard the electromagnetic waves by an amount inversely proportional to the square of the radio frequency, the ionospheric range errors that are the order of millimeters can be utilized to obtain unambiguous TEC values [Imel, 1994] over the oceans (see equation (2)).

  • equation image
  • equation image

where f is the Ku-band frequency (referred above) and K, related to the ionosphere refraction, is a constant K = 80.6 m3 s−2 [Klobuchar, 1996; Ma and Maruyama, 2003; Yizengaw et al., 2004b]. Horvath and Essex [2003b] provided detail procedure of TEC extraction from TOPEX satellite altimetry. The TOPEX altimeter provides the vertical electron content beneath the satellite. Since the TOPEX altimeter is designed only to measure ocean heights, there are data gaps over landmasses.

2.3. IMAGE FUV

[9] The IMAGE satellite has a two-channel FUV spectrographic imaging instrument. One of the channels, called SI-13, detects the OI 135.6 nm emission [Mende et al., 2000]. The SI-13 has the following characteristics: it has a field of view of 16 × 16 degrees (6.8 Earth radii (RE) or ∼43,320 km altitude); its detector has 128 × 128 pixels; it has a single pixel prosection of 100 km × 100 km at apogee; its expected sensitivity is 1.3 × 10−2 counts per Rayleigh per pixel for each 5 s viewing exposure per satellite revolution, which is 120 s; its wavelength response has a peak response at 135.6 nm with a bandpass from 131.0 to 140.0 nm, and thus it mainly picks up 135.6 nm emission from atomic oxygen and some contribution from N2 Lyman-Birge-Hopfield (LBH) emissions. Further explanation of the SI-13 instrument can be found in the work of Mende et al. [2000]. Images from the SI-13 are used to study oxygen depletion due to magnetic storms. In this paper, IMAGE FUV data are used to verify the ground based GPS TEC regional map.

2.4. Topside Ionosphere and Plasmasphere Tomography

[10] FedSat is an Australian LEO microsatellite developed by the Cooperative Research Centre for Satellite System (CRCSS). FedSat, which weighs 50 kg, was launched on 14 December 2002, and successfully deployed at an initial altitude of ∼840 km. A dual-frequency GPS receiver is one many scientific instruments carried by FedSat. Details of FedSat and its payload instruments are well documented by Fraser et al. [2000] and Yizengaw [2004].

[11] It is convenient to divide the FedSat GPS data into two types: GPS below-the-horizon data that is recorded when FedSat detects GPS signals that have traversed the ionosphere below FedSat's orbiting height, GPS above-the-horizon data that is recorded when FedSat receives signals that have crossed the ionosphere above its orbiting height. FedSat's GPS above-the-horizon data can be used by itself to obtain the electron density distribution of the topside ionosphere and plasmasphere using tomographic reconstruction. For this paper, we have used only above-the-horizon data to obtain the topside ionospheric electron density distribution. Details of the tomographic algorithm can be found in the work of Yizengaw et al. [2004a], Yizengaw [2004], and references therein. The Global Plasmasphere Ionosphere Density (GPID) model [Webb and Essex, 2004] has been used as initial guess to our Algebraic Reconstruction Tomography (ART) tomographic algorithm.

[12] The TEC along the straight line between a GPS satellite and FedSat can be obtained using a simple linear combination of observations on the dual L-band frequencies used by GPS [Sardón et al., 1994; Yizengaw et al., 2004b]. With FedSat, a LEO satellite at ∼800 km altitude, the ray path descends (ascends) for a setting (rising) GPS satellite at speed of ∼3.87 km/s. This very fast speed of FedSat implies that it takes ∼10 min to sample the ionosphere tomographically for a single FedSat pass in the region of interest. To begin the tomographic analysis, the FedSat GPS data is preprocessed to remove outlier values and cycle slips, which are either removed or corrected. The TEC values are then obtained after correction for the receiver and satellite differential group delay biases. Details of the method of bias estimation are given by Breed et al. [1997]. Finally, the absolute slant TEC data from FedSat is used as input to the topside ionospheric and plasmasphereic tomographic reconstruction experiment.

3. Observations

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data and Method
  5. 3. Observations
  6. 4. Discussion
  7. 5. Summary
  8. Acknowledgments
  9. References

[13] A series of three severe geomagnetic storms were observed in the interval 29–31 October 2003 (the so-called series of Halloween storms). The nature of the Halloween storms are shown in Figure 1, which presents the most frequently used geomagnetic indices (Dst and Kp) and the Geocentric Solar Magnetospheric (GSM) z-component of the Interplanetary Magnetic Field (IMF) (obtained from Advanced Composition Explorer (ACE) satellite data set). The 34 min time lag of the IMF data has been added to correct for the propagation from L1 point to the magnetosphere. The storm sudden commencement (SSC), indicated by the vertical broken line, occurred on 29 October 2003 at 0611 UT, which is in the post midday sector for the far eastern part of the globe. The two coronal mass ejection (CME) events observed on 29 October and 30 October 2003 created two Dst decreases with minimum values −363 nT and −401 nT at 0100 UT and 2300 UT on 30 October, respectively. The IMF Bz component turned southward and remained south between 1832 UT on 29 October and 0310 UT on 30 October and between 1810 UT on 30 October and 0106 UT on 31 October as shown in Figure 1. The Kp index reached 9 between 0600–0900 UT and 1800–2400 UT on 29 and 30 October, respectively. Therefore as can be seen in the figure the main phases (minimum Dst value) of the three successive storms occurred at 0930 UT on 29 October and at 0100 UT and 2300 UT on 30 October, respectively. The AE index (not shown here) shows a dramatic increase and reached 3800 nT right after the SSC (0611 UT) and at 1800 UT on 29 October. The AE index was continue enhanced on 31 October as well. The solar activity index (F10.7 flux) reached to a maximum value (∼275 W m−2 Hz−2) of the month on 29 October 2003. The Solar Wind Electron Proton Alpha Monitor (SWEPAM) instrument on the Advanced Composition Explorer (ACE) spacecraft measured extremely high solar wind speeds of >1850 km/s, which is the highest speeds ever directly measured in the solar wind [Skoug et al., 2004]. These speeds were observed following two large coronal mass ejections (CME) driven shocks mentioned above.

image

Figure 1. Variation of geomagnetic indices: IMF Bz component, Dst and Kp indices are shown from top to bottom, respectively. The dashed vertical line depicts the Storm Sudden Commencement (SSC) time of the Halloween 2003 storm.

Download figure to PowerPoint

[14] Ground-based GPS TEC data from more than 65 GPS stations at the eastern region of the globe are studied in detail for the period of 28–31 October 2003. The GPS stations that we used, in which International GPS Service (IGS) operates, are spread from the Southern Hemisphere cusp region up to the Northern Hemisphere subauroral region as shown in Figure 2.

image

Figure 2. Geographic maps of the GPS receiver stations that provided data for this study. The circles around each station depict the shell area where the measurement subionospheric point will lie for elevation greater than 25°. The two dashed vertical lines indicate two designated common longitude meridians used to study the latitude versus local time variations of ionospheric electron density distribution. The big dots in a bolded circles depict the near by stations that have been used for the latitude-local time plots to the closest common longitude meridians.

Download figure to PowerPoint

[15] The GPS TEC responded to the storm in an unusual way. The ionospheric electron density was dramatically depleted in the Southern Hemisphere. However, at the time when the Southern Hemisphere is depleted, the Northern Hemisphere recorded a substantial density, compared with the density in the south, as shown in the left side of Figure 3. Figure 3 presents geographic latitude versus longitude regional maps of vertical GPS TEC. The horizontally curved dashed line depicts the location of the geomagnetic equator. Color bars on the right side of each plot show the scales for vertical GPS TEC in TECU (1 TECU = 1 × 1016 el m−2). The right side represents the corresponding TEC maps before the storm began, on 28 October 2003, for comparison.

image

Figure 3. The geographic latitude-longitude contour maps of ground-based GPS vertical TEC (solid circles) and TOPEX altimeter vertical TEC (solid squares) during disturbed period (left) and a day before storm began (right). The horizontal dashed line in each panel indicates the location of geomagnetic equator. The color scales at the right of each panel are given in TECU.

Download figure to PowerPoint

[16] Figure 3a shows a vertical GPS TEC map during the storm main phase on 29 October. Except for some density depletion at high latitudes and the eastern side of the image in the Southern Hemisphere, TEC during the storm main phase reveals normal density distributions with positive density gradients toward lower latitudes. However, the overall density distribution in Figure 3a shows significant enhancement compared with the corresponding density distribution before the storm began on 28 October shown in Figure 3b. The most significant enhancements, about 20% to 50% increases in TEC, are visible at mid-Australia and at the northwest region of the image. The equatorial and tropical regions also show similar enhancements.

[17] Figures 3c and 3e present the TEC maps during the second storm's main phase, between 2300 UT on 29 October and 0300 UT on 30 October. At this time the Dst reached the lowest point of the second storm value (−363 nT), IMF Bz turned southward, and Kp reached ∼9. The continuous tracks denoted by solid squares represent the vertical TEC obtained from the TOPEX altimeter. A clear difference in the density distribution between the Northern and Southern Hemispheres is clearly visible in the figures. A deep density depletion is present over a large area in the Southern Hemisphere. At later times it expanded to lower latitudes and caused the equatorial region to be depleted by up to 70% in TEC compared with the corresponding density distribution during the prestorm interval shown in Figures 3d and 3f. Exceptionally, the TEC at McMurdo (far east cusp station in Southern Hemisphere), in Figures 3c and 3e, shows a slight increase, up to 30% in TEC, from the corresponding density distribution level shown in Figures 3d and 3f. The midlatitude and low-latitude region in the Northern Hemisphere (see Figure 3c) has also experienced moderate (up to 23% decrease in TEC from the quiet day value in Figure 3d) depletion but not as strong as the depletion in the south. In contrast to the density distribution in Figure 3f, the density distribution at the Northern Hemisphere (see Figure 3e) shows two features: significant depletion at high latitudes and almost no depletion at midlatitudes.

[18] The depletion remained in the region for more than 2 days. Figure 3g shows electron density distribution during the recovery phase of the third storm. Unlike Figures 3c and 3e, the electron density in the south, mainly at lower latitudes, shows significant recovery (see Figure 3g) from its previous long-lived density loss. It indicates that the unusual density distributions that have been imposed by the magnetic storm began to recover back to normal after almost 2 days of depletion. The TOPEX altimeter provided consistent TEC values with ground-based GPS vertical TEC values.

[19] To see the density depletion feature in latitude versus local time, two common median longitudes are identified: one at the east (150°E) and the other one at the west (110°E). The vertical dashed lines in Figure 2 indicate the locations of the two common longitudes. The longitudinal (local time) difference of GPS data from the nearby stations (indicated by big dots enclosed by bold circles in Figure 2) are corrected to these common median longitudes, using a simple relation described by Horvath and Essex [2003a]. Hence, the vertical GPS TEC (solid circles) and TOPEX altimeter TEC (solid squares) are presented as latitude-local time contour plots in Figure 4. The horizontal dashed lines in all panels and the vertical dashed lines only in Figures 4a and 4d, respectively, depict the geomagnetic equator and SSC time.

image

Figure 4. The latitude-local time slice ionosphere contour maps of vertical TEC at 110°E (left) and at 150°E (right) common longitudes. The solid circles and squares represent the ground-based GPS TEC and TOPEX altimeter TEC, respectively. The horizontal dashed line in each panel and the vertical dashed line only at the top indicate the geomagnetic equator location and SSC time of the Halloween storm, respectively. The color scales at the right of each panel are given in TECU.

Download figure to PowerPoint

[20] At the west side common longitude, deep density depletion began at high latitudes at about 1900 UT on 29 October (6 hours after the SSC time) as shown in Figure 4a which is a normal diurnal variation. It then expanded to lower latitudes, and on 30 October (see Figure 4b) the density depletion dominated the daytime solar plasma production, causing the Southern Hemisphere region to remain depleted throughout the day. As can be seen in Figure 4c, the daytime depletion still existed at high and middle latitudes in the Southern Hemisphere on 31 October. However, at this time its latitudinal coverage is reduced and the equatorial region started to recover. The local time density distribution in the Northern Hemisphere exhibits similar behavior as shown in Figures 4a–4c. However, in the Northern Hemisphere, the depletion did not penetrate to lower latitudes and remained a high-latitude phenomena during the disturbed periods. Compared with the electron density in Figure 4a, the daytime density depletion at the north expanded equatorward up to subauroral latitudes as shown in Figures 4b and 4c.

[21] The latitudinal gradient and local time density distribution at the east side common median longitude behaves as the west side density distribution does (see in Figures 4d–4f). However, the size of the depletion is slightly less than the west side, especially at the cusp and auroral stations. When Dst reached at its lowest values (−363 nT and −401 nT) at 0100 UT and 2300 UT on 30 October (see Figure 1), enhanced density value is obtained from GPS receiver at the cusp southern station (McMurdo) as shown in Figures 4e and 4f.

[22] Ionospheric electron density depletion can be due to a depletion of atomic oxygen and/or enhancement of molecular nitrogen. Likewise, a change in the ratio of thermospheric [O]/[N2] can reduce the atomic oxygen airglow brightness, which is observed by the FUV spectrographic imaging instrument aboard the NASA-IMAGE satellite (SI-13) [Mende et al., 2000]. The negative storm effect on F region ionosphere is partly related to effects of vibrationally excited N2 and O2 on the O+(4S) ion density [Pavlov, 1994; Pavlov and Foster, 2001]. Recent studies show that both positive [Immel et al., 2001] and negative [Strickland et al., 2001] ionospheric effects that are driven by composition variations can be observed in the FUV oxygen dayglow. Therefore depletion of TEC driven by composition disturbances should also exhibit corresponding reductions in OI dayglow brightness. Figures 5a–5d reveal the deviation of 135.6-nm dayglow brightness from quiet-time values during three successive orbits of IMAGE. Departure from the quiet-time brightness model, which is based on observational geometry and solar flux [Immel et al., 2000], is closely related to the change in thermospheric [O]/[N2] [Strickland et al., 1995]. The 135.6-nm oxygen dayglow signature, shown in Figure 5a (obtained a day before the storm began), shows almost no deviation from quiet-time levels, though a slight depletion (<35% decrease) around the southeast Australian continent is observed. During the storm main phase (Figure 5b), two special features are observed. These are moderate dayglow intensity depletions, up to 40% decrease, at the south of the African continent and very bright, active and wide auroral oval over auroral and subauroral latitudes. The latter indicates a strong magnetic storm was going on around 0811 UT (just 2 hours after SSC time) on 29 October. Later on the same day at 2341 UT, while the Australian region is the local morning sector, deep negative deviations, up to 60% decrease in OI dayglow, are evident over the Australian region and even in the cusp region (see Figure 5c). As can be seen in Figure 5c, the bright and active auroral oval lost its brightness at the Australian region side. Figure 5d also shows similar deep deviation over the same region as Figure 5c. In the cusp region, the deviation increased its magnitude compared to the deviation in Figure 5c. The bright and active auroral oval started to recover and fills up the breach.

image

Figure 5. The percentage deviations of oxygen dayglow intensities from the corresponding quite time values obtained from IMAGE FUV SI-13 imaging instrument. The images represents parentage deviation recorded (a) 1 day before the storm began, (b) during the first storm main phase, and (c and d) during the second storm main phase.

Download figure to PowerPoint

[23] Figure 6 presents the downward looking TOPEX altimeter TEC data during the disturbance periods (28, 30, and 31 October 2003). During these days, TOPEX was near the magnetic equator at ∼1100 LT. The curved dashed lines in each panel indicate the location of the magnetic equator. A day before the storm began, 28 October, the global TEC shows normal density distribution (i.e., intensity increasing with decreasing latitude) as shown in Figure 6a. The daytime equatorial anomaly (EA) is evident in the figure, which is the subject of another paper. After the storm began, 30 October, TOPEX detected a deep density depletion with a large meridional extent (>70°) over the Australian region (see Figure 6b). This daytime depletion was not seen on the previous day, 29 October. Another interesting feature in Figure 6b is the localized enhanced TEC magnitudes (>150 TECU) observed over the Pacific Ocean, which will be discussed later. The gap in these two successive TOPEX passes between 25°S and 30°S geographic is an artifact of the instrument, which resets the data when the ionospheric correction exceeds 150 TECU. During the recovery phase of the storm, at 1044 LT on 31 October shown in Figure 6c, the depletion started to recover and the localized enhanced TEC also decreased in magnitude.

image

Figure 6. The global TEC maps obtained from the down looking TOPEX altimeter (a) a day before the storm began and (b and c) during the disturbance periods. During these days, TOPEX was near the magnetic equator at ∼1100 LT. The curved dashed lines in each panel indicate the location of magnetic equator.

Download figure to PowerPoint

[24] To examine if this large depletion in plasma density is also seen at higher altitudes, Figure 7 shows the DMSP's ion drift meter observations from ∼900 km. This DMSP satellite circles the Earth in a Sun-synchronous dawn-dusk orbit. Two passes of DMSP F13 are shown as it crossed over the region at 2007 UT on 28 October and at ∼1811 UT on 29 October. Data obtained on 28 October (Figure 7a) represent prestorm quiet time DMSP observations, whereas Figure 7b presents data obtained during the storm main phase on 29 October. In both figures, the top panel shows the plasma density observed at the satellite height, and the second and third panels depict the vertical plasma drift velocities and ion temperatures, respectively.

image

Figure 7. (a) Plots of the topside F region plasma density (top), up and down (sunward and antisunward) plasma drift velocity (second from top), ion temperature (third from top), and ions and electron flux (bottom) obtained from DMSP F13 satellite. (b) The same data set but during the disturbed period on 29 October.

Download figure to PowerPoint

[25] The quiet time density is found to be well above 0.35 × 105 cm−3, except for a localized depletion at ∼40°S geographic, as shown in Figure 7a (first panel). At about the same local time during the magnetic storm period (Figure 7b (first panel)), the density is down to ∼0.15 × 105 cm−3. This gives a depletion level of ∼40%, which is nearly the same depletion level as observed from GPS, FUV, and TOPEX data. The vertical velocity on 28 October (Figure 7a (second panel)) shows constant value around zero, except slight fluctuation between about 50°S and 55°S geographic latitudes. In contrast, the velocity during the disturbed period (Figure 7b (second panel)) shows strong upward velocities in the midlatitude region with a maximum peak at ∼45°S geographic. The ion temperature in the midlatitudes, during storm main phase, increased dramatically as shown in Figure 7b (third panel). While before the quiet time ion temperatures (Figure 7a (third panel)) varied little with latitude. Figure 7a (fourth panel) gives ion and electron density flux. The fluxes increased at higher latitudes compared to density fluxes at lower latitudes. Similarly, the density fluxes during the storm main phase, Figure 7b (fourth panel), enhanced between 25°S and 57°S geographic latitudes. The interesting thing in the figure is that high ion fluxes extended about 20–25° equatorward of electrons. Normally, the ion fluxes extends equatorward by a few degrees, about 2–5°.

[26] The integrated electron densities (TEC) along the LEO-to-GPS raypaths have been calculated during the north-to-south crossings of FedSat as shown in Figure 8a. Since the integrated density does not estimate the vertical density profile of the topside ionosphere and plasmasphere, the tomographic inversion of the slant TEC (integrated density along LEO-to-GPS raypaths) is introduced (see Yizengaw et al. [2005] and Heise et al. [2002] for details). The calculated slant TECs are then used as input for tomographic imaging of the topside ionospheric and plasmaspheric electron densities. FedSat takes only about 15 min to cover a latitude range of up to 60°. The pass, covering 9.52°S to 68.57°S, crossed the region at about 160°E longitude in between 2238 UT and 2255 UT on 29 October.

image

Figure 8. (a) Geographic maps of FedSat ground track, crossing the region at 160°E meridian. The horizontal dashed lines represent the geomagnetic equator, 30°S, and 60°S from top to bottom, respectively. (b) Tomographically reconstructed density distribution from data recorded by GPS receiver on FedSat during FedSat's pass shown in Figure 8a, (c) electron density distribution from GPID model. The curved white lines depict the magnetic field lines obtained from IGRF-2000 model.

Download figure to PowerPoint

[27] The tomographically reconstructed electron density distribution derived for this pass is shown in Figure 8b, while Figure 8c shows for comparison, the electron density from the Global Plasmasphere Ionosphere Density (GPID) model [Webb and Essex, 2004]. Figure 8b shows normal density distributions at higher and lower latitudes. However, at the midlatitudes a deep density depletion was observed during this pass, while at the same time the ground-based instruments detected an ionospheric density depletion. This indicates that the plasmaspheric density is completely gone. At the higher-latitude region, a magnetic field aligned density upward expansion is evident. The white lines indicate the magnetic field lines obtained from IGRF-2000 model [Langlais and Mandea, 2000]. The color codes at the right depict electron density (×104 el cm−3).

[28] To demonstrate the EA behavior from the TOPEX altimeter TEC data, Figures 9 and 10 examine the TEC versus geographic latitude. Figures 9 and 10 represent the daytime and nighttime TOPEX passes. The start and end times of each individual pass are indicated on the left and right sides of Figures 9 and 10. The vertical dashed line in each panel depicts the location of the geomagnetic equator. The bottom panels in Figures 9 and 10 present the corresponding (as identified by color codes) ground track of the satellite's passes.

image

Figure 9. The TOPEX altimeter vertical TEC versus latitude plot during TOPEX's daytime passes. The vertical dashed lines indicate the geomagnetic equator and the bottom panel depict the corresponding TOPEX ground track. The start and end time of each pass is shown at the left and right corner of each panels. The equatorial local time crossings of TOPEX are given at the top.

Download figure to PowerPoint

image

Figure 10. As for Figure 7 but for the nighttime TOPEX passes.

Download figure to PowerPoint

[29] During the storm recovery phase, 30 October at ∼1100 equatorial crossing local time, the five successive passes of TOPEX revealed interesting features as shown in Figure 9. The first pass, in Figure 9a, shows a broadened equatorial anomaly. The northern anomaly had a maximum value of ∼165 TECU at ∼21°N geomagnetic. However, the south peak had intensities higher than the northern anomaly peak. Owing to a data gap that occurred between 23°S and 32°S geographic, the exact value and location of the southern anomaly peak cannot be determined. Interestingly, the midlatitude (35°–55°S geographic) density in the Southern Hemisphere shows a dramatic increase compared with its conjugate midlatitude density value. For example, the intensity at ∼42°S geomagnetic is ∼170 TECU, whereas at ∼40°N geomagnetic the intensity is ∼40 TECU (∼70% difference). Figure 9b again shows a similar EA features but the magnitude of the northern anomaly decreases to ∼110 TECU and its location moved to lower latitudes (∼18°N geomagnetic). In addition to the overall midlatitude density enhancement in the region, an isolated strong density peak, at ∼51°S geographic, is observed. Figures 9c, 9d, and 9e that occurred to the west of the previous two passes give totally different features. The intensities at the south dramatically decreased down to ∼20 TECU at the EA peak region. However, the northern anomaly peak (see Figure 9c) shows relatively higher density, ∼90 TECU at ∼18°N geomagnetic. Owing to the data gaps over land masses, Figures 9d and 9e do not show the Northern Hemisphere density behavior.

[30] Another interesting feature that can be noted in the top five panels of Figure 9 is the high-latitude density peak, which is labeled with a “P.” The peak retreats back to the pole during the recovery phase of the second storm as shown in Figures 9c–9e. During the third storm main phase (1800–2400 UT on 30 October), the peak advances toward to the equator (see Figures 9a and 9b).

[31] Figure 10 shows TOPEX altimeter TEC obtained from three successive nighttime passes during the storm main phase on 29 October. During these passes, TOPEX was near to the magnetic equator at ∼2307 LT. A clear EA region with approximately symmetrical anomaly peaks was observed when TOPEX crossed the geographic equator at 2307 LT. This can be noted in the top three panels of Figure 10. As can be seen in Figure 10a, the maximum anomaly peak (∼75 TECU) occurred at ∼±10° geomagnetic. However, at the next pass, Figure 10b, the magnitude of the anomaly peaks dropped to ∼30 TECU (∼43% decrease from the previous pass peak value). The TEC anomaly peak value increased back to ∼60 TECU in the third pass (see Figure 10c). There is another feature present in the figure as well. Although the overall density magnitude shows depletion, compared to the other two passes, Figure 10b reveals short time density fluctuation feature over the Southern Hemisphere midlatitude region. The rate of change of TEC (not shown here), which is obtained by taking the difference of two TEC data sample per minute [Warnant, 1997], also shows similar fluctuations.

4. Discussion

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data and Method
  5. 3. Observations
  6. 4. Discussion
  7. 5. Summary
  8. Acknowledgments
  9. References

4.1. Asymmetric Density Distribution

[32] As indicated by the data collected from different instruments, the ionosphere has shown large plasma density distribution differences between the Southern and Northern Hemispheres in the aftermath of the initial storm during the Halloween 2003 storm interval. Deep plasma density depletion occurred at high latitudes of the Southern Hemisphere and then expanded to lower latitudes. The plasma density depletion (ionospheric storm negative phases) in TEC have been known and attracted much interest for many decades [Seaton, 1956; Prölss, 1997; Fuller-Rowell et al., 1997; Buonsanto, 1999]. Pavlov et al. [2004] has presented the first theoretical study of the role of variations in the neutral winds, temperature, and densities in producing the north-south asymmetry in the storm-time low latitude electron density.

[33] Since the first suggestion by Seaton [1956], it is believed that the negative phase is caused by the changes of the thermospheric composition due to the heating of the thermosphere during geomagnetic disturbances. In support of this suggestion, we presented the IMAGE FUV dayglow data to complement our results obtained from ground-based GPS and TOPEX altimeter TEC data. The increase (decrease) of oxygen dayglow implies the increase (decrease) of [O]/[N2] and hence the enhancement (depletion) of ionospheric plasma density. During severe magnetic storms, like the Halloween 2003 storms, strong energy deposition at high latitudes leads to a dramatic decrease of the atomic-to-molecular ratio and thus ionospheric plasma density at high latitudes as shown in Figures 38. The heating also induces its own circulation (storm-induced wind). The storm-induced winds then drive the depleted zone to lower latitudes or even to the conjugate hemisphere, depending on local time and season. The latitudinal distribution of the depleted zone is different in the sunlit and nighttime sector of the winter hemisphere. Buonsanto [1999] and Danilov and Lastovicka [2001] have clearly described the storm-induced wind circulation and have pointed out that in winter the equatorward directed storm-induced wind is opposite to the daytime background (quiet time) circulation, which is directed poleward, however, both have the same direction at night. In summer, the background circulation is equatorward throughout the day and thus coincides with the storm-induced winds.

[34] At the end of October or beginning of November, the Southern Hemisphere is in the summer season and the Northern Hemisphere is in winter. This implies that during the severe magnetic storms that occurred at the end of October 2003 the storm induced depleted thermospheric air could propagate to lower latitudes in the Southern Hemisphere but not on the winter hemisphere side. Hence, the large plasma density distribution differences observed between the Southern and Northern Hemispheres. The multi-instrument observations introduced in the previous sections consistently revealed such disproportional plasma density distributions between the Northern and Southern Hemispheres.

[35] In the northern winter hemisphere because of the background circulation and lower Joule heating due to lower conductance, the depleted thermospheric air could not penetrate to lower latitudes compared with its equatorward expansion in the conjugate hemisphere. Figures 3c and 3e show a clear hemispherical plasma density distribution differences. In the Northern Hemisphere the depletion extends only to midlatitudes, whereas in the Southern Hemisphere it expanded and covered the whole region, up to the magnetic equator.

[36] Furthermore, the northern wintertime equatorward expansion of the depleted thermospheric air (depleted zone) has a preference for the night and morning sectors due to the local time variation of the neutral winds. It penetrates to lower latitudes during nighttime than daytime. This can be noticed in Figures 4a–4c. Figure 4a shows, between 0230 and 0530 LT in northern midlatitude region (20°N–40°N, geographic), the TEC is ∼20 TECU. At the same local time and latitudes, in Figures 4b and 4c, the TEC decrease down to ∼8 TECU, about a 43% decrease. Although there is no data between 20°N and 30°N in Figures 4d–4f, all panels reflect similar nighttime latitudinal coverage of the depleted zone. Figures 4e and 4f even show the penetration of the depleted zone to the equatorial region (10°N–20°N, geographic) between 0500 and 0600 LT. This clearly shows that the depleted zone has penetrated to much lower latitudes in the nighttime sector, which is consistent with the current understanding about the storm-induced winds. However, during the daytime in winter, the depleted zone could not expand to lower latitudes and extended only from the high latitude to midlatitude as shown in Figures 4a–4f. The daytime poleward background circulation restricts the storm-induced wind from extending to lower latitudes. Figures 4b and 4e show that the depleted zone expanded down to ∼50°N in between 0800 and 1800 LT. In the same local time zone, in Figure 4c, the depletion extended to a lower-latitude region (∼40°N, geographic). However, in Figure 4f, it remained above ∼50°N geographic latitudes. Figure 4f also shows that the postsunset equatorial region plasma density is enhanced compared to the previous day TEC value (Figure 4e). This indicates that ionospheric processes are returning back to normal after more than 2 days of severe changes when the ionization budget was not controlled by the usual local time variation in production, loss and transport, but rather it was controlled and governed by the perturbations directly associated with magnetospheric storm effects.

[37] In Southern Hemisphere summer, when the background circulation and storm-induced wind coincide and are directed equatorward, the depleted zone penetrates to much lower latitudes, even to the conjugate hemisphere as shown in Figures 38. During the storm main phase the plasma density depletion is only visible at higher latitudes, more significantly at the eastern side high-latitude region (see Figure 3a). Eventually, the depleted zone expanded to lower latitudes as shown in Figures 3c and 3e. As can be seen in the figures the expansion has reached to much lower latitude in the Southern Hemisphere than in the Northern Hemisphere. The down looking TOPEX altimeter also revealed similar features as shown in Figures 6b and 6c. All of the observations are consistent with the interpretation of equatorward expansion of the depleted zone. Similarly, Figure 5 clearly shows that the plasma density depletions that are detected by ground-based GPS and TOPEX altimeter are due to the dramatic reduction of atomic oxygen and thus the depletion of atomic-to-molecular ratio. High-latitude Joule heating produces large-scale changes in thermospheric composition, directly affecting ionospheric F region densities. Enhancements (reductions) in the atomic-to-molecular ratio result in increases (decreases) in 135.6-nm brightness, which can be measured with IMAGE FUV SI-13 instrument as shown in Figure 5. A bright, expanded, and active auroral oval in Figure 5b indicates a strong magnetic storm was going on at about ∼0811 UT on 29 October. However, its deviation from quiet day values shows a weak [O]/[N2] depletion at the cusp region, and a moderate depletion [O]/[N2] at midlatitude region. Probably, this could be due to storm effects or substorms heating occurring much earlier [Danilov, 2001]. When the IMF turned south and more energy is input into the lower thermosphere [Buonsanto, 1999; Danilov and Lastovicka, 2001; Zhang et al., 2004], between 29 October at 1800 UT and 30 October at 0200 UT, the dayglow deviation in Figures 5c and 5d shows strong depletions. This progression is quite consistent with the plasma density depletion shown in Figures 3c, 3e, and 6b.

[38] As can be seen in Figure 4, the local time plasma density depletion features in the Southern Hemisphere (summer) shows quite different behavior compared with the Northern Hemisphere (winter), which is described above in detail. The plasma density in the Southern Hemisphere has been depleted throughout the day and expanded to low-latitudes as shown in Figures 4e and 4f. The deep plasma density depletion shown in between 2100 UT and 2400 UT on 29 October (see Figure 4d) eventually developed and expanded to lower latitudes in the postmidnight (0000–0600 UT) region as shown in Figure 4e. This nighttime phenomenon then corotates with the Earth to the morning sector, causing the daytime plasma density to deplete throughout the day [Zhang et al., 2003]. This could be the reason that the solar plasma density production was totally dominated and the entire hemisphere's plasma density depleted for more than 2 days as shown in Figures 4e and 4f.

[39] The plasma density enhancement (Figures 4e and 4f) that occurred in the cusp region (McMurdo) between 0430 LT and 1100 LT coincided with the southward turning of the IMF (see Figure 1). This could be due to the energetic particles that escaped during the nighttime reconnection and precipitated to cusp region when the IMF turned southward. Similarly, the slight plasma density enhancement (between 0500 LT and 1600 LT in Figures 4e and 4f) at Macquarie Island, which is located in the auroral oval region, may be due to auroral particle precipitation. The precipitating fluxes, shown in Figure 7b, increased its magnitude in the auroral and subauroral region, indicating the TEC enhancement shown in Figures 4e and 4f are clearly due to the precipitating particles.

[40] The DMSP data shown in Figure 7b (first panel) reveals that the plasma density depletion extends to about 900 km altitude in the Southern Hemisphere during the magnetic storm period. For comparison we also presented the DMSP data obtained prior to the disturbed days (Figure 7a (first panel)). The vertical speed in Figure 7b (second panel) also shows the upward directed velocity increased and reached to its peak value at ∼45°S geographic latitudes. This clearly indicates that the depleted zone not only expanded meridionally to lower latitudes but also expanded vertically and created significant plasma density depletion (∼40% at ∼900 km height) at higher altitudes. The plasma density depleted is also detected at higher altitude as it is clearly seen from the topside ionospheric and plasmaspheric tomographic image in the Southern Hemisphere midlatitude region shown in Figure 8b. The enhanced DMSP upward velocity, shown in Figure 7b (second panel), indicates that the thermospheric depleted zone could be transported upward to higher altitudes by such strong the upward velocity. The strong DMSP upward velocity also confirms the penetration of magnetospheric eastward electric field when IMF Bz turns southward (see Figure 1).

[41] The increase of ion temperature between 30°S and 64°S geographic, shown in Figure 7b (third panel), provides another important feature of storm time plasma density distribution. Increased recombination rate due to the increased temperature is another cause for storm time ionospheric plasma depletion [Mikhailov and Förster, 1997]. It should be noted that the heated gas with depleted [O]/[N2] ratio has a higher temperature throughout the thermosphere. The increase of temperature leads at the F region heights to an increase of the linear recombination coefficient and thus to a further decrease of the electron concentration [Mikhailov et al., 1995]. This is shown in Figures 7b, where the plasma density depletion is consistent with the increase of ion temperature. The vibrationally excited nitrogen plays a significant role in increasing the temperature at the F2 layer. Pavlov and Buonsanto [1997] suggested that the chemical reactions of vibrationally excited N2 and O2 with O+(4S) ions play a significant role in changes the loss rate of these ions at F region altitudes. Evidence have been found that the O+(4S) loss rate enhancement due to these processes plays an important role in forming the plasma density depletion zone during geomagnetic storms [Pavlov, 1994; Pavlov and Buonsanto, 1997; Pavlov and Foster, 2001; Pavlov et al., 2004].

4.2. Localized Dayside Plasma Density Enhancement

[42] The other interesting feature of our observations results is the localized plasma density enhancement that occurred over the Pacific Ocean as detected by the TOPEX altimeter data shown in Figure 6b. In order to identify the feature clearly, the two successive passes with enhanced electron plasma density over the Pacific Ocean and another passes over the Australian and Indian Ocean region have been replotted in TEC versus geographic latitude mode as shown in Figure 9. Figure 10 also shows similar plots but for successive nighttime passes over the same region. The vertical dashed lines in all panels, except the bottom panels, of both plots depict the magnetic equator. Figures 9a and 9b show a broadened EA region with much greater plasma density peak at the Southern (summer) Hemisphere than Northern (winter) Hemisphere. Such peak difference is unusual and conflicts with the current understanding of summer and winter hemisphere plasma density peak differences. The EA peak is expected to be higher in the winter hemisphere because of the summer-to-winter transequatorial neutral winds that drive the plasma density to the winter hemisphere across the equator. In this case, the southern EA has a peak plasma density 70% greater than its northern counterpart density distribution. The southern midlatitudes' plasma density has even formed a second anomaly structure (finger-like structure) at midlatitude (Figure 9b), creating an enhanced plasma density peak at ∼51°S geographic which is far from the tropical EA peak region. There are two possible reasons that have been attributed to the cause of these unusual plasma density enhancements at the Southern Hemisphere midlatitude region. The first possibility is that the prompt penetration of magnetospheric origin electric field, which is eastward during the day and westward during the night. Tsurutani et al. [2004] have presented more extensive studies about the penetration of the magnetospheric origin electric field. During severe magnetic storm periods and when the IMF turns south, strong eastward electric fields can penetrate down to the middle and lower latitudes ionosphere before the shielding effect builds up [Jakowski et al., 1999; Yizengaw et al., 2004b; Tsurutani et al., 2004]. This prompt electric field not only affects the equatorial ionospheric region but the midlatitude region as well [Tsurutani et al., 2004]. Therefore the prompt electric field could be the possible force that drove ion plasma density to midlatitudes from equatorial region and has created an unusual plasma density enhancement in the southern hemisphere. Probably due to the satellite' inclined orbit geometry and the short lifetime of the feature, the Northern Hemisphere shows no enhancement at midlatitudes. As can be seen in Figure 9f, the satellite's ground track on the tropical region at the north and south have a difference of ∼30° longitude or of ∼2 hours local time. However, the tropical plasma density enhancement mostly has a longitudinal symmetric feature [Walker et al., 1994], like the one shown in Figure 5a. A localized oxygen dayglow enhancement (up to +40% deviations) is visible at the right edge of Figure 5a. Furthermore, the three TOPEX passes, which crossed the magnetic equator at 245°E, 215°E, and 185°E, in Figure 6b, shows similar observation at the northern tropical region. By comparing the 245°E and 215°E equatorial crossing passes, the electron plasma density at the northern tropical region shows significant increases at the latter pass. Again, by comparing the 215°E and 185°E equatorial crossing passes, TEC decreased at the latter pass. In both cases, the TEC at the south continuously increased toward the west up to ∼190°E. From this, it is possible that enhanced electron plasma density may be present in the gap between these passes at the northern tropical zone. Unfortunately, there is not any other instrument in the area to confirm this.

[43] The second possibility, especially for the case of Figure 9b, is the sunward plasmasphere drainage plume. Foster at al. [2002, 2004] have found finger-like midlatitude plasma density enhancements. They demonstrated that the enhancement has a direct association with the sunward plasmasphere drainage plume. Using multi-instrument data (ground-based GPS TEC, IMAGE EUV data, incoherent scattering (IS) radar observation data, and DMSP data), they found a total transfer of ∼1030 ions to the F region ionosphere from the sunward plasmasphere drainage within a duration of 5 hours. Although there is a lack of data to confirm this with other remote sensing instruments, such as ground-based GPS TEC and IS radar in the southern hemisphere, the result that we obtained from TOPEX (Figure 9b) has a similar feature like the one reported by Foster et al. [2004]. Unfortunately, the IMAGE satellite's orbit favored equatorial over polar imaging, denying us the opportunity to confirm at least the existence of the plume, though plumes routinely develop in the aftermath of a storm [Moldwin et al., 2004].

[44] The other possibility, which can be taken as a third possibility, is the positive enhancement in [O]/[N2]. The plasma density enhancement could be caused by a positive enhancement of the thermospheric [O]/[N2] driven by the large-scale redistribution of the thermosphere carried by traveling atmospheric disturbance (TIDs) [Immel et al., 2001; Förster et al., 1999].

[45] The plasma density obtained from TOPEX's pass over the Australian and Indian Ocean region (Figures 9c–9e) show a completely different feature from the previously seen TEC data. The Southern Hemisphere EA peak is completely washed out and probably driven by the transequatorial winds (mentioned above) to the conjugate hemisphere and created a significant EA peak at ∼18°N geomagnetic as shown in Figure 9c. This is in agreement with IMAGE FUV and GPS observations presented and with the current understanding the summer and winter EA features. Since TOPEX is mainly designed to study the sea surface anomaly, the TOPEX TEC data gap over the land is evident. Owing to such data gap, the northern hemisphere EA peaks are not included in Figures 9d and 9e.

[46] Another interesting feature that can be noted in Figure 9 is the plasma density peaks at the southern high latitudes during the first storm recovery phase and second storm main phase. This could be due to the auroral particle precipitation mentioned above. These plasma density peaks, which are labeled by “P” in the figure, retreated towards the pole during the storm (second storm) recovery phase (0100–1600 UT on 30 October) as shown in Figures 9c–9e. However, as can be seen in Figures 9a and 9b, during the third storm main phase (1800–2400 UT on 30 October), the peak electron densities advanced towards to the equator. The poleward retreating of the peak densities implies the contraction of the auroral oval during recovery phases, especially when Bz turned north. In the same way, the equatorward advance of the feature shows the auroral oval expansion to midlatitudes when Bz turned south and more energy was input into the high-latitude region. There is evidence in Figure 9c that during the second storm main phase the auroral oval has expanded down to ∼45°S geographic (∼57°S geomagnetic) and built a sharp electron plasma density peak centered at ∼46°S geographic.

4.3. Nighttime Equatorial Anomaly

[47] Figure 10 also shows interesting nighttime ionospheric features. It presents unusual EA around the midnight sector over the Pacific Ocean, a region that experienced unusual daytime plasma density distributions, discussed earlier. The top three panels in Figure 10 show clear almost symmetrical nighttime EA peaks at 10°N and 15°S geomagnetic. The disturbance dynamo eastward electric field may be the cause for such a nighttime EA. Recently, Tsurutani et al. [2004], using global data coverage, have performed a comprehensive study about the electric field effect on ionospheric plasma density distribution. They found that the nighttime disturbed dynamo eastward electric fields have significant effects on equatorial plasma density enhancements around the post midnight sector.

[48] Besides the longitudinally localized plasma density depletion observed in the middle pass (Figure 10b) compared with the other two passes, it also observed a short-time plasma density fluctuations at the southern midlatitude region. This could be due to the impulsive Joule heating that generates equatorward traveling atmospheric disturbances, which, in turn, create plasma density fluctuations in the ionosphere region [Buonsanto, 1999].

5. Summary

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data and Method
  5. 3. Observations
  6. 4. Discussion
  7. 5. Summary
  8. Acknowledgments
  9. References

[49] Multi-instrument remote sensing and in situ satellite data have been used to study the response of the ionosphere to the extremely high-speed solar wind shock event and related magnetic storm activities that occurred during a series of three storm events between 29 and 31 October 2003. All data examined for this study reveal consistent observational results. We have investigated the remarkable ionospheric electron plasma density responses, of which most results we believe are unique to this event. The most significant results that we found are summarized as follows:

[50] 1. The ground-based GPS TEC reveals large differences in plasma density distribution between the Southern and Northern Hemispheres. A dramatic plasma density depletion (storm negative effect) at the south up to the magnetic equator has been observed. Such a dramatic plasma density depletion over the Australian region has not been reported before. The down looking TOPEX altimeter TEC data also consistently detected such hemispherical plasma density asymmetries.

[51] 2. The IMAGE FUV SI-13 instrument, which is sensitive to the changes of the atomic-to-molecular ratio, has provided a good tool for monitoring the thermospheric composition changes during Halloween storm. Not surprisingly, it detected deep oxygen dayglow depletion over the Australian region at the time when the electron plasma density depletion occurred.

[52] 3. The DMSP and topside ionospheric and plasmaspheric tomographic reconstruction showed the deep depletion detected by the ground-based instruments extended upward to 900 km with a maximum vertical velocity of 0.6 km/s as measured at DMSP orbit height (∼900 km).

[53] 4. The local time and seasonal effects of the storm negative phase observed in this storm are consistent with our current understanding of latitudinal motion during the negative phase of a storm. We found plasma density depletion at the southern latitudes throughout the day for more than 2 days, whereas in the north the 24-hour long depletion was restricted to only the higher latitudes.

[54] 5. TOPEX altimeter data showed an unusual localized daytime plasma density enhancement over the Pacific region, which could be associated with a sunward plasmaspheric plume and/or the prompt penetration of the magnetospheric origin electric field. The former has not been reported in the Southern Hemisphere before because of poor data coverage in the Southern Hemisphere. Although the result that we have shown in Figure 8 has not been confirmed with IMAGE EUV data, it certainly has a similar features with the results reported in the North American sector.

[55] 6. Ground-based and/or remote sensing instruments also detect the aurora retreating toward to the pole and advancing to the equator during storm recovery and main phases, respectively. The downlooking TOPEX altimeter clearly outlined this effect.

[56] In conclusion, we have shown the combined use of ground- and space-based instruments to understand the ionospheric and plasmasphereic features during the Halloween superstorm period. The two most prominent feature of the disturbance were the deep plasma density depletion in the Southern Hemisphere, leading to asymmetries plasma density distribution with conjugate hemisphere, and a localized strong plasma density enhancement over the Pacific Ocean. The former feature tells us that the severe Halloween storms dissipated large amounts of energy into the high-latitude ionosphere, leading to the depletion of atomic-to-molecular ratio and thus to depletion of electron plasma density at the southern middle latitudes. The later feature of the disturbance may be related to penetrating magnetospheric electric fields and the sunward plasmaspheric plume. At this time, we could not show the plasmaspheric plume feature from other independent measurement data due to a lack of data in the region. However, it could be taken as a hint that the plasmaspheric plume could trigger such strong localized electron plasma density enhancement. Finally, it certainly such kind of disproportional plasma density distribution that greatly impacts the Wide Area Augmentation System (WAAS) during such superstorms. Owing to this for the storm of 30 October 2003 the so called Precision Approach service of WAAS was down for more than 11 hours [Doherty et al., 2004].

Acknowledgments

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data and Method
  5. 3. Observations
  6. 4. Discussion
  7. 5. Summary
  8. Acknowledgments
  9. References

[57] This work has been supported partly by the NASA grant (grant NNGO4GG343G), and a JPL/UCLA partnership grant. Thanks are due to the International GPS Service (IGS) for the GPS data, JPL's Physical Oceanography Distributed Active Archive Centre (PO.DAAC) for TOPEX altimeter TEC data, and NOAA Satellite and information service for DMSP SSJ4 data. The ACE data are from CADWeb (http:///cdaweb.gsfc.nasa.gov/cdaweb/istp_public/) and the Dst,AE, and Kp data are from WDC-2 at Kyoto (http://swdcdb.kugi.kyoto-u.ac.jp). The research was carried out with financial support from the Commonwealth of Australia through the Cooperative Research Centers Program, and FedSat GPS data has been provided by the CRC for Satellite Systems.

[58] Arthur Richmond thanks Norbert Jakowski and A. V. Pavlov for their assistance in evaluating this paper.

References

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data and Method
  5. 3. Observations
  6. 4. Discussion
  7. 5. Summary
  8. Acknowledgments
  9. References
  • Breed, A. M., G. L. Goodwin, A. M. Vandenberg, E. A. Essex, K. J. W. Lynn, and J. H. Silby (1997), Ionospheric total electron content and slab thickness determination in Australia, Radio Sci., 32, 16351643.
  • Buonsanto, M. J. (1999), Ionospheric storms: A review, Space Sci. Rev., 88, 563601.
  • Crowley, G., B. A. Emery, R. G. Roble, H. C. Carlson, and D. J. Knipp (1989), Thermospheric dynamics during September 18–19, 1984: 1. Model simulations, J. Geophys. Res., 94, 16,92516,944.
  • Danilov, A. D. (2001), F2-region response to geomagnetic disturbances, J. Atmos. Terr. Phys., 63, 441449.
  • Danilov, A. D., and J. Lastovicka (2001), Effects of geomagnetic storms on the ionosphere and atmosphere, Int. J. Geomagn. Aeron., 2, 209224.
  • Doherty, P., A. J. Coster, and W. Murtagh (2004), Space weather effects of October–November 2003, GPS Sol., 8(3), doi:10.1007/s10291-004-0109-3.
  • Förster, M., A. A. Namgaladze, and R. Y. Yurik (1999), Thermospheric composition changes deduced from geomagnetic storm modeling, Geophys. Res. Lett., 16, 26252628.
  • Foster, J. C., P. J. Erickson, A. J. Coster, J. Goldstein, and F. J. Rich (2002), Ionospheric signatures of plasmaspheric tails, Geophys. Res. Lett., 29(13), 1623, doi:10.1029/2002GL015067.
  • Foster, J. C., A. J. Coster, P. J. Erickson, F. J. Rich, and B. R. Sandel (2004), Stormtime observations of the flux of plasmaspheric ions to the dayside cusp/magnetopause, Geophys. Res. Lett., 31, L08809, doi:10.1029/2004GL020082.
  • Fraser, B. J., C. T. Russell, J. D. Means, F. W. Menk, and C. L. Waters (2000), FedSat -an Australian research microsatellite, Adv. Space Res., 25, 13251336.
  • Fuller-Rowell, T. J., M. V. Codrescu, R. J. Moffett, and S. Quegan (1994), Response of the thermosphere and ionosphere to geomagnetic storms, J. Geophys. Res., 99, 38933914.
  • Fuller-Rowell, T. J., M. V. Codrescu, R. G. Roble, and A. D. Richmond (1997), How does the thermosphere and ionosphere react to a geomagnetic storm? in Magnetic Storms, Geophys. Monogr. Ser., vol. 98, edited by B. T. Tsurutani et al., pp. 203225, AGU, Washington, D.C.
  • Heise, S., N. Jakowski, A. Wehrenpfennig, C. H. Reigber, and H. Lühr (2002), Sounding of the Topside Ionosphere/Plasmasphere Based on GPS Measurements from CHAMP: Initial Results, Geophys. Res. Lett., 29(14), 1699, doi:10.1029/2002GL014738.
  • Horvath, I., and E. A. Essex (2003a), Vertical E × B drift velocity variations and associated low- latitude ionospheric irregularities investigated with the TOPEX and GPS satellite data, Ann. Geophys., 21, 10171030.
  • Horvath, I., and E. A. Essex (2003b), The southern-hemisphere mid-latitude day-time and nighttime trough at low-sunspot numbers, J. Atmos. Sol. Terr. Phys., 65, 917940.
  • Imel, D. A. (1994), Evaluation of TOPEX/Poseidon dual frequency ionospheric correction, J. Geophys. Res., 99, 24,89524,906.
  • Immel, T. J., J. D. Craven, and A. C. Nicholas (2000), The DE-1 auroral imager's response to the FUV dayglow for thermospheric studies, J. Atmos. Sol. Terr. Phys., 62, 4764.
  • Immel, T. J., G. Crowley, J. D. Craven, and R. G. Roble (2001), Dayside enhancements of thermospheric O/N2 following magnetic storm onset, J. Geophys. Res., 106, 15,47115,488.
  • Jakowski, N., S. Schluter, and E. Sardon (1999), Total electron content of the ionosphere during the geomagnetic storm on 10 January 1997, J. Atmos. Sol. Terr. Phys., 61, 299307.
  • Klobuchar, J. A. (1996), Ionospheric effects on GPS, in Global Positioning System: Theory and Application, vol. 1, edited by B. W. Parkinson, and J. J. Spilker, pp. 485515, Am. Inst. of Aeron. and Astron., Reston, Va.
  • Langlais, B., and M. Mandea (2000), An IGRF candidate main geomagnetic field model for epoch 2000 and a secular variation model for 2000–2005, Earth Planets Space, 52, 11371148.
  • Ma, G., and T. Maruyama (2003), Derivation of TEC and estimation of instrumental biases from GEONET in Japan, Ann. Geophys., 21, 20832093.
  • Makela, J. J., and M. C. Kelley (2001), GPS normalization and preliminary modelling results of total electron content during a midlatitude space weather event, Radio Sci., 36, 351361.
  • Mende, S. B., et al. (2000), Far ultraviolet imaging from the Image spacecraft: 3. Spectral imaging of Lyman-alpha and OI 135.6 nm, Space Sci. Rev., 91, 287318.
  • Mikhailov, A. V., and M. Förster (1997), Day-to-day thermosphere parameter variation as deduced from Millstone Hill incoherent scatter radar observations during March 16–22, 1990 magnetic storm period, Ann. Geophys., 15, 14291438.
  • Mikhailov, A. V., M. G. Skloblin, and M. Förster (1995), Daytime F2 layer positive storm effect at middle and lower latitudes, Ann. Geophys., 13, 532540.
  • Moldwin, M. B., J. Howard, J. Sanny, J. D. Bocchicchio, H. K. Rassoul, and R. R. Anderson (2004), Plasmaspheric plumes: CRRES observations of enhanced density beyond the plasmapause, J. Geophys. Res., 109, A05202, doi:10.1029/2003JA010320.
  • Pavlov, A. V. (1994), The role of vibrationally excited nitrogen in the formation of the mid-latitude negative ionospheric storms, Ann. Geophys., 12, 554564.
  • Pavlov, A. V., and M. J. Buonsanto (1997), Comparison of model electron densities and temperatures with Millstone Hill observations during undisturbed periods and the geomagnetic storms of 16–23 March and 6–12 April 1990, Ann. Geophys., 15, 327344.
  • Pavlov, A. V., and J. C. Foster (2001), Model/data comparison of F region ionospheric perturbation over Millstone Hill during the severe geomagnetic storm of 15–16 July 2000, J. Geophys. Res., 106, 29,05129,070.
  • Pavlov, A. V., S. Fukao, and S. Kawamura (2004), F region ionospheric perturbations in the low-latitude ionosphere during the geomagnetic storm of 25–27 August 1987, Ann. Geophys., 22, 34793501.
  • Prölss, G. W. (1997), Magnetic storm associated perturbations of the upper atmosphere, in Magnetic Storms, Geophys. Monogr. Ser., vol. 98, edited by B. T. Tsurutani et al., pp. 227241, AGU, Washington, D.C.
  • Prölss, G. W., and J. D. Craven (1998), Perturbations of the FUV dayglow and ionospheric storm effects, Adv. Space Res., 22, 129134.
  • Prölss, G. W., L. H. Brace, H. G. Mayer, G. R. Carignan, T. L. Killeen, and J. A. Klobuchar (1991), Ionospheric storm effects at subauroral latitude: A case study, J. Geophys. Res., 96, 12751288.
  • Rishbeth, H., T. J. Fuller-Rowell, and D. Rees (1987), Diffusive equilibrium and vertical motion in the thermosphere during a severe magnetic storm: A computational study, Planet. Space Sci., 35, 11571165.
  • Sardón, E., A. Rius, and N. Zarraoa (1994), Estimation of the receiver differential biases and the ionospheric total electron content from Global Positioning System observations, Radio Sci., 29, 577586.
  • Seaton, M. J. (1956), A possible explanation of the drop in F region critical densities accompanying major ionospheric storms, J. Atmos. Terr. Phys., 8, 122130.
  • Skoug, R. M., J. T. Gosling, J. T. Steinberg, D. J. McComas, C. W. Smith, N. F. Ness, Q. Hu, and L. F. Burlaga (2004), Extremely high speed solar wind: 29–30 October 2003, J. Geophys. Res., 109, A09102, doi:10.1029/2004JA010494.
  • Strickland, D. J., J. S. Evans, and L. J. Paxton (1995), Satellite remote sensing of thermospheric O/N2 and solar EUV: 1. Theory, J. Geophys. Res., 100, 12,21712,226.
  • Strickland, D. J., R. E. Daniell, and J. D. Craven (2001), Negative ionospheric storm concident with DE 1 observed thermospheric disturbance on October 14, 1981, J. Geophys. Res., 106, 21,04921,062.
  • Tsurutani, B., et al. (2004), Global dayside ionospheric uplift and enhancement associated with interplanetary electric fields, J. Geophys. Res., 109, A08302, doi:10.1029/2003JA010342.
  • Walker, G. O., J. H. K. Ma, and E. Golton (1994), The equatorial ionospheric in electron content from solar minimum to solar maximum for South East Asia, Ann. Geophys., 12, 195209.
  • Warnant, R. (1997), Detection of irregularities in the Total Electron Content using GPS measurements: Application to a mid-latitude station, paper presented at Beacon Satellite Symposium, Hungarian Acad. of Sci., Sopron, Hungary.
  • Webb, P. A., and E. A. Essex (2004), A dynamic global model of the plasmasphere, J. Atmos. Sol. Terr. Phys., 66, 10571073.
  • Yizengaw, E. (2004), Imaging the ionosphere, Ph.D. thesis, Dep. of Phys., La Trobe Univ., Bundoora, Victoria, Australia.
  • Yizengaw, E., P. L. Dyson, and E. A. Essex (2004a), Tomographic reconstruction of the ionosphere using ground-based GPS data in the Australian region, paper presented at Workshop on Applications of Radio Science (WARS), Natl. Comm. for Radio Sci., Hobart, Australia.
  • Yizengaw, E., E. A. Essex, and R. Birsa (2004b), The Southern Hemisphere and Equatorial region ionization response for September 22, 1999 severe magnetic storm, Ann. Geophys., 22, 27652773.
  • Yizengaw, E., P. L. Dyson, and E. A. Essex (2005), A study of the spatial density distribution in the topside ionosphere and plasmasphere using the FedSat GPS receiver, Adv. Space Res., in press.
  • Zhang, Y., L. J. Paxton, H. Kil, C.-I. Meng, S. B. Mende, H. U. Frey, and T. J. Immel (2003), Negative ionospheric storms seen by the IMAGE FUV instrument, J. Geophys. Res., 108(A9), 1343, doi:10.1029/2002JA009797.