JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 114, A10309, doi: /2009ja014485, 2009

Transcription

1 JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 114,, doi: /2009ja014485, 2009 Topside ionospheric effective scale heights (H T ) derived with ROCSAT-1 and ground-based ionosonde observations at equatorial and midlatitude stations S. Tulasi Ram, 1 S.-Y. Su, 2 C. H. Liu, 3 B. W. Reinisch, 4 and Lee-Anne McKinnell 5 Received 25 May 2009; revised 22 June 2009; accepted 25 June 2009; published 16 October [1] In this study we propose the assimilation of topside in situ electron density data from the Republic of China Satellite (ROCSAT-1) along with the ionosonde measurements for accurate determination of topside ionospheric effective scale heights (H T ) using an a- Chapman function. The reconstructed topside electron density profiles using these scale heights exhibit an excellent similitude with Jicamarca incoherent scatter radar (ISR) profiles and are much better representations than the existing methods of Reinisch-Huang method and/or the empirical International Reference Ionosphere 2007 model. The main advantage with this method is that it allows the precise determination of the effective scale height (H T ) and the topside electron density profiles at a dense network of ionosonde/ Digisonde stations where no ISR facilities are available. The demonstration of the method is applied by investigating the diurnal, seasonal, and solar activity variations of H T over the dip-equatorial station Jicamarca and the midlatitude station Grahamstown. The diurnal variation of scale heights over Jicamarca consistently exhibits a morning time descent followed by a minimum around LT and a pronounced maximum at noon during all the seasons of both high and moderate solar activity periods. Further, the scale heights exhibit a secondary maximum during the postsunset hours of equinoctial and summer months, whereas the postsunset peak is absent during the winter months. These typical features are further investigated using the topside ion properties obtained by ROCSAT-1 as well as Sami2 is Another Model of the Ionosphere (SAMI2) model simulations. The results consistently indicate that the diurnal variation of the effective scale height (H T ) does not closely follow the plasma temperature variation and at equatorial latitudes is largely controlled by the vertical E B drift. Citation: Tulasi Ram, S., S.-Y. Su, C. H. Liu, B. W. Reinisch, and L.-A. McKinnell (2009), Topside ionospheric effective scale heights (H T ) derived with ROCSAT-1 and ground-based ionosonde observations at equatorial and midlatitude stations, J. Geophys. Res., 114,, doi: /2009ja Introduction 1 Institute of Astronomy and Astrophysics, Academia Sinica, Taipei, Taiwan. 2 Institute of Space Science and Center for Space and Remote Sensing Research, National Central University, Chung-Li, Taiwan. 3 Academia Sinica, Taipei, Taiwan. 4 Center for Atmospheric Research, University of Massachusetts, Lowell, Massachusetts, USA. 5 Hermanus Magnetic Observatory, Hermanus, South Africa. Copyright 2009 by the American Geophysical Union /09/2009JA [2] The knowledge of the vertical electron density distribution of the ionosphere is very important because of its potential applications in the estimation and correction of propagation delays in the Global Navigation Satellite System (GNSS), ionospheric storm studies, ion composition studies, space-weather effects on telecommunications, etc. The vertical electron density profile of the bottomside ionosphere (below the F 2 layer peak) can be obtained from ground-based ionosonde/digisonde or incoherent scatter radar (ISR) measurements. For the topside ionosphere (above the F 2 layer peak), the electron density profile can be obtained from the ISR, space-borne topside sounder measurements, or by empirical models. Since the observations by the ISR and topside sounders are limited both spatially and temporally, many workers mostly rely on the reconstruction of topside profiles using analytical functions. During the past two to three decades, great efforts have been made in the ionospheric empirical modeling and many analytical functions such as Chapman, exponential, parabolic, Epstein functions have been proposed to describe the vertical distribution of electron density [Booker, 1977; Rawer et al., 1985; Rawer, 1988; Di Giovanni and Radicella, 1990; Reinisch and Huang, 2001; Reinisch et al., 2004, 2007; Stankov et al., 2003; Bilitza et al., 2006, and references therein]. An important and inherent parameter in all these functions is the ionospheric scale height [Huang and Reinisch, 2001; Stankov et al., 2003; Belehaki et al., 2006]. The scale height is an important ionospheric characteristic that describes the shape (thickness) of the ionospheric electron density profile. A study of the scale height behavior provides valuable 1of13

2 information on the vertical distribution of ionization and can thus be used to answer many open questions in ionospheric physics, particularly those related to ionospheric composition and dynamics. [3] There are various definitions of the scale heights in the published literature. In order to avoid confusion, we adopt the definitions similar to those described by Liu et al. [2007a]. The plasma scale height (H p ) is defined as H p = k b T p /m i g, where k b is the Boltzmann constant, T p is the plasma temperature, m i is the ion mass, and g is the acceleration due to gravity. The vertical scale height (H v or VSH) is defined as the lowest value of dh/d(ln(ne)) [Kutiev et al., 2006; Stankov and Jakowski, 2006; Liu et al., 2007a], the vertical distance in which Ne changes by a factor of e(= ). The Chapman scale height (H T ), occasionally referred to as the effective scale height, is defined as the scale height in fitting the Ne(h) profiles with the a-chapman function [Lei et al., 2005; Liu et al., 2007a, 2007b]. The plasma scale height (H p ) is subject to theoretical considerations and linearly depends on the plasma temperature (T p = T e + T i where T e and T i are the electron and ion temperatures, respectively). Whereas, the VSH and H T are virtually the distribution heights of electron density profiles, measuring the altitudinal gradient or shape of the ionospheric Ne(h) profiles and are very important in various practical applications. Further, the effective scale height (H T ) is proportional to neutral temperature (T n ) and VSH is proportional to the plasma temperature (T p 2T n ), thus, H T expected to be approximately one half of the VSH under diffusive equilibrium [Belehaki et al., 2006; Liu et al., 2007a]. [4] Using the simplified aeronomic equations, the topside electron density distribution can be represented by an a-chapman-type profile [Hargreaves, 1979, 1992] as Nh ðþ¼n m F2: exp 1 : ð1 z e zþ ; ð1þ 2 ð z ¼ h h mf2þ ; ð2þ H T where N m F 2 and h m F 2 are the F 2 layer peak density and height, respectively. The measured bottom side profile from the ionosonde directly gives the F 2 layer peak electron density parameters N m F 2 and h m F 2, and the only unknown parameter is the scale height, H T. [5] Recently, Reinisch and Huang [2001] and Huang and Reinisch [2001] introduced a technique to derive the scale height near the F 2 layer peak (H m ) from the shape of the bottomside profile using the a-chapman function. They have also shown that the variation of scale height directly above the F 2 layer peak is very small and assumed a constant scale height for the interpolation of the topside ionosphere (i.e., H m (h > h m F 2 ) H m (h m F 2 )). This method (hereinafter called the Reinisch-Huang (R-H) method) is very useful because of its deftness as well as simplicity, and with the availability of modern Digisondes along with automatic scaling algorithms, it became possible to derive the total vertical electron density profiles in real time. Subsequently using this method, the diurnal, seasonal and solar activity variations of H m was statistically investigated using ionosonde observations at Wuhan (30.6 N, E) and twelve other stations by Liu et al. [2006], at Hainan (19.4 N, 109 E) by Zhang et al. [2006], at Jicamarca (11.9 S, 76.8 W) by Lee and Reinisch [2008], and at Grahamstown (33.3 S, 26.5 E) by Nambala et al. [2008]. However, one drawback of this R-H method is that it does not include any additional measured parameter in the topside ionosphere; instead the scale height is determined at h m F 2 solely from the shape of the bottomside profile. This will be likely to cause errors at higher altitudes, particularly in the equatorial region where the zonal electric field plays a significant role in the vertical redistribution of plasma via the E B drift, as will be discussed more in this paper. Further, Reinisch et al. [2004] have also mentioned that the vertical electron density profiles reconstructed using the R-H method is found to exhibit significant errors when compared to Jicamarca ISR profiles. Lei et al.[2005] and Liu et al. [2007a, 2007b] have analyzed the Millstone Hill and Arecibo ISR observations to determine the scale height by least squares fitting of topside Ne(h) profiles with a-chapman functions and investigated the diurnal, seasonal, and solar activity variations of H T. However, the knowledge of the topside ionospheric effective scale height behavior remains insufficient particularly for equatorial latitudes as not much work has been done over this region except by Lee and Reinisch [2008] using the R-H method. It should be mentioned here that the Chapman scale height near F 2 layer peak (H m ) derived by the R-H method is different from the topside ionospheric effective scale height (H T ) derived by Lei et al. [2005] and Liu et al. [2007a, 2007b]. Even though both H m and H T are derived using the a-chapman function, the primary difference is that the H m is derived near the F 2 layer maximum while the H T is derived by fitting the topside electron density profile. [6] In the present work, we derive the topside ionospheric effective scale heights (H T ) by the assimilation of topside in situ electron density data from the Republic of China Satellite (ROCSAT-1) in conjunction with the Digisonde observations. Similarly to the work of Reinisch and Huang [2001], we assume that the topside profile from h m F 2 to 1000 km can be represented by an a-chapman function with a constant effective Chapman scale height H T. This may not be realistic for all conditions. As Reinisch et al. [2007] have shown, the best representation of the topside profile to plasmaspheric heights is obtained by using a Chapman function with continuously varying scale height. The ROCSAT-1 method, however, has only one measured density point at 600 km altitude sufficient to determine one effective scale height. In order to assess the efficiency of this method in the equatorial as well as midlatitudes, the diurnal, seasonal and solar activity variations of H T are statistically investigated over a dip-equatorial station Jicamarca (11.9 S, 76.8 W and 0.9 N dip.lat) and a midlatitude station Grahamstown (33.3 S, 26.5 E and 34.2 S dip.lat), and the results are discussed in detail. 2. Data, Methodology, and Validation [7] The first scientific satellite of the Republic of China (ROCSAT-1) orbits in a circular orbit at 600 km altitude (with inclination of 35 ) and measures the topside ionospheric ion properties with its major payload Ionospheric Plasma and Electrodynamics Instrument (IPEI) during the high to 2of13

3 Table 1. Monthly Mean F10.7 Solar Flux Values During the Periods of Observations High Solar Activity Period Moderate Solar Activity Period Month Monthly Mean F10.7 Solar Flux Month Monthly Mean F10.7 Solar Flux Equinoxes September March October April March September April October Seasonal average Seasonal average Summer November November December December January January February February Seasonal average Seasonal average Winter May May June June July July August August Seasonal average Seasonal average moderate solar activity period from March 1999 to June The in situ obtained ion density (equivalent to electron density due to charge neutrality) data from the ROCSAT-1 passes over Jicamarca (geographic latitude: 11.9 S ± 0.1, longitude: 76.8 W ± 10 ) and Grahamstown (geographic latitude: 33.3 S ± 0.1, longitude: 26.5 E ± 10 ) during the periods from September 2001 to August 2002 and from March 2003 to February 2004 are considered in the present analysis. [8] The ionogram data over Jicamarca and Grahamstown during the same periods were obtained from the UML Digital Ionogram DataBase (DIDBase) ( DIDBase/), and the bottomside electron density profiles are derived using the NHPC true height inversion algorithm embedded in the SAO Explorer software package ( umlcar.uml.edu/sao-x/sao-x.html). A limited number of vertical electron density profiles from the Jicamarca incoherent scatter radar (ISR) which are available in the MIT Haystack Madrigal database ( mit.edu/madrigal/) during the period of study are also considered for the validation of reconstructed profiles. The two data periods considered in this study are corresponding to a high solar activity period (September 2001 to August 2002) and a moderate solar activity period (March 2003 to February 2004). The monthly mean F10.7 solar flux values of these periods and their seasonal classification are shown in Table 1. [9] Figure 1a illustrates the determination of the topside ionospheric effective scale height (H T )fromtheinsitu electron density at ROCSAT-1 altitude and F 2 layer peak using equations (1) and (2). The solid line indicates the bottomside profile obtained from the Digisonde measurements at Jicamarca around 1000 LT of 14 November 2001, and the solid circle represents the F 2 layer peak electron density (N m F 2 and h m F 2 ) point. The asterisk in Figure 1a indicates the in situ electron density measured by ROCSAT-1 over Jicamarca at 600 km altitude during the same time as that of bottomside profile. The a-chapman (equation (1)) function is fitted between these two points (F 2 peak and ROCSAT-1) to obtain the unknown scale height, H T. In the present case shown in Figure 1a, the unknown scale height is found to be km. Here, the scale height above the F 2 Figure 1. (a) Derivation of the effective scale height (H T ) by fitting the a-chapman function between the F 2 layer peak electron density point and the ROCSAT-1 in situ density point around 1000 LT on 14 November 2001 using an a- Chapman function. (b) Reconstruction of topside electron density profile using an a-chapman function with constant scale height (H T ). 3of13

4 layer peak is assumed to be height-independent [Chapman, 1931]. Now, using this scale height H T (78.77 km), the topside electron density profile is reconstructed using equations (1) and (2) between h m F 2 and 1000 km with a step size of 10 km as shown in Figure 1b. [10] The precise validation of the reconstructed electron density profiles can be made by comparison with the ISR measured vertical profiles. Figures 2a and 2b show the comparison of reconstructed profiles with Jicamarca ISR profiles around 1036 LTand 1000 LTon 13 and 14 November 2001, respectively. In Figures 2a and 2b, the solid red line is the vertical electron density profile from the Jicamarca ISR, and the dashed blue line is the reconstructed topside profile from H T. It can be seen from Figures 2a and 2b that the reconstructed profile follows very closely with the ISR profile up to 1000 km. Since the time resolution of Jicamarca ionogram data is 15 min, the maximum time difference among the coherent observations of ROCSAT-1 in situ electron density, ionosonde bottomside profile and ISR profiles is limited to a reasonable value of ±7.5 min. Several reconstructed electron density profiles are compared with Jicamarca ISR profiles during different local times (not shown here) and are found to have good similitude with negligible errors while the time difference among the observations from the three different instruments are taken into account. Further, the vertical electron density profiles derived from the International Reference Ionosphere 2007 (IRI- 2007) model (Figure 2, line with open circles) [Bilitza and Reinisch, 2008] as well as from the R-H method (Figure 2, line with solid circles) were also shown in Figures 2a and 2b for comparison. It should be noted here that the IRI-2007 profiles are derived by providing the measured N m F 2 and h m F 2 values as inputs and choosing the Ne-Quick model option [Radicella and Leitinger, 2001; Coisson et al., 2006] for the topside electron content. However, it can be observed from Figures 2a and 2b that the scale height derived by fitting the a-chapman curve between the ROCSAT-1 data and F 2 layer peak density point gives a better reconstruction of the topside profile when compared to the other two profiles. This indicates that the a-chapman or effective scale height (H T ) derived by the assimilation of topside in situ data in conjunction with the ground-based Digisonde measurements provides a much improved representation of the vertical density distribution in the topside ionosphere. Figure 2. Two examples showing the comparison of the reconstructed electron density profiles (dashed blue lines) with Jicamarca ISR profiles (solid red lines) along with the Reinisch-Huang (R-H) method (lines with solid circles) and the IRI-2007 model (lines with open circles) around 1000 LT on (a) 13 November 2001 and (b) 14 November Results and Discussion [11] In order to improve the local time coverage, we have combined four months of ROCSAT-1 data for each season corresponding to equinoxes, summer, and winter solstices as described in Table 1. The data in each season are further binned into 1 hour local time intervals, and the data points corresponding to geomagnetically disturbed periods (Kp 3+) are eliminated. The bottomside density profiles during the same periods (±7.5 min) as that of ROCSAT-1 revisits are obtained from the Digisonde observations at Jicamarca and Grahamstown. The effective scale height (H T ) is derived for each revisit of ROCSAT-1, and the vertical electron density profiles are reconstructed as described in section Observations at the Dip-Equatorial Station, Jicamarca [12] For example, Figure 3 shows the reconstructed topside electron density profiles (red lines) for each 1 hour local time intervals during the equinoctial period of September October 2001 and March April 2002 at Jicamarca. In Figure 3, the solid black circles represent the F 2 layer peak density points, and the solid green circles indicate the ROCSAT-1 in situ density points. The vertical electron density profiles derived from the R-H method (blue lines) are also shown in Figure 3 for comparison. It can be noticed from Figure 3 that the topside electron density profiles derived using the R-H method (blue lines) often depart significantly from the ROCSAT-1 in situ density points, particularly, during the noon and afternoon hours. This is mainly because of the vertical expansion of the topside ionosphere (increase in layer thickness and/or scale height) due to the E B drift at the equator. This feature can be better captured by the topside electron density profiles derived with the assimilation of topside (ROCSAT-1) in situ electron density points (Figure 3, red lines). Further, it can be clearly seen from Figure 3 that the thickness (scale heights) of the topside profiles derived from ROCSAT-1 data (red lines) is in general larger than that of the R-H method (blue lines) except during the morning hours ( LT). The rationale 4of13

5 Figure 3. The reconstructed topside electron density profiles (red lines) using the ROCSAT-1 in situ data for each 1 hour local time interval during the equinoxes at Jicamarca. The solid black circles represent the F 2 layer peak density points, and the solid green circles represent the ROCSAT-1 in situ density points. The vertical electron density profiles derived from the R-H method (blue lines) are also shown for comparison. for this feature will be discussed in more detail later in this section. [13] Figure 4 illustrates the seasonal mean values of H T over Jicamarca along with their standard deviations as a function of local time for the equinoctial (Figure 4a), summer (Figure 4b), and winter (Figure 4c) solstice months during high (red lines) and moderate (blue lines) solar activity periods. The mean F10.7 solar flux values for each season are indicated in Figures 4a 4c, respectively. The interesting feature observed from Figure 4 is that the H T over the dipequatorial station Jicamarca has a distinctive diurnal variation with larger values during the day and smaller values during the nighttime for all seasons. The H T values during the high solar activity periods (Figure 4, red lines) exhibit a faint peak around LT and then reduces gradually to a minimum value around LT. After that, the H T value increases rapidly and shows a pronounced peak around local noon. During the afternoon hours, H T again starts descending and reduces gradually till the postmidnight hours. Further, the diurnal variation of H T during the equinoctial (Figure 4a) and summer (Figure 4b) months exhibits a clear secondary maximum during the postsunset hours. However, this postsunset secondary peak in H T is not evident in winter solstice (Figure 4c). The scale height values (H T ) during the moderate solar activity periods (Figure 4, blue lines) exhibit similar local time variations with a faint peak around the predawn hours, a minimum around LT, and a maximum during the local noon hours. The secondary maximum during the postsunset hours of equinoctial and summer months becomes weak or less significant during this moderate solar activity period. Further, it can be clearly noticed that the H T values are in general larger during high solar activity (Figure 4, red lines) than during moderate solar activity (Figure 4, blue lines) for all seasons. This can be understood from the classical definition of the plasma scale height (H p = k b T p /m i g), in which the scale height is positively correlated with the plasma temperature which in turn linearly depends on the solar activity. [14] The scale heights near the F 2 layer peak (H m ), derived for Jicamarca by Lee and Reinisch [2008] using the R-H method, also show similar diurnal variations over Jicamarca with a maximum during local noon, and a secondary peak during the postsunset hours of equinoctial and summer months. However, their results show that the smallest value 5of13

6 Figure 4. The diurnal variations of seasonal mean H T during (a) equinoctial, (b) summer, and (c) winter solstice months along with their standard deviations during high (red lines) and moderate (blue lines) solar activity at the dip-equatorial location, Jicamarca. in H m occurs at LT during solar minimum, and at LT during solar maximum. Further, the H m value increases immediately after sunrise during both solar minimum and maximum periods with a small peak at 0600 LT in solar minimum. They have attributed this increase in H m after sunrise to the increase in the neutral scale height (kt/mg), in which the scale height is positively correlated to the temperature. However, their results are in contrast with our results presented in Figure 4, where in fact a faint peak in H T occurs around h LT and then decreases gradually to a minimum around h LT. Also from Figure 3, it can be noticed that the thickness of the profiles (scale heights) derived from ROCSAT-1 in situ data (red lines) are smaller than those from the R-H method (blue lines) during the LT period. The vertical scale heights (VSH) derived by Liu et al. [2008] at equatorial latitudes using the vertical electron density profiles from the COSMIC radio occultation measurements during the low solar activity period ( ) also exhibit similar diurnal behavior with a weak peak around LT. Then the VSH descends and approaches a minimum around 0800 LT [Liu et al., 2008, Figure 4a]. Thus, the results presented in this study are consistent with the observations of Liu et al. [2008]. However, the causative mechanisms for these observed features in the diurnal variations of H T at equatorial latitudes remain unexplained. [15] According to the Chapman layer theory, the vertical electron density distribution at any given local time is controlled by the altitudinal variations in the production, loss and vertical transport of plasma [Rishbeth and Garriott, 6of13

7 Figure 5. The diurnal variations of various topside ionospheric parameters: (a) ion temperature, (b) electron density, (c) vertical E B drift measured by ROCSAT-1 satellite at 600 km altitude, and (d) the effective scale height (H T ) during the equinoctial months of high (solid lines) and moderate (dashed lines) solar activity periods. The blue lines in Figure 5b indicate the electron densities at the F 2 layer peak (N m F 2 ) obtained from the Jicamarca Digisonde. 1969]. Particularly, in the dip-equatorial stations, the vertical E B drift due to the zonal electric field largely influences the vertical distribution of electron density in addition to the diffusion due to temperature and ion composition variations. Liu et al. [2007a, 2008] have also pointed out that the vertical scale height is controlled by many factors such as plasma scale height (H p ), vertical drift, and the thermal structure in the topside ionosphere. In this context, the ROCSAT-1/IPEI data set is unique and provides many of the topside ionospheric parameters that influence the H T such as ion temperature, ion density, ion composition, and the vertical drift [Su et al., 1999; Yeh et al., 1999]. Hence, for a better understanding of the observed features in the diurnal variations of the H T, the ion temperature, ion density (equivalent to electron density), and the vertical drift parameters over Jicamarca at 600 km altitude from the ROCSAT-1 satellite are investigated. Figure 5 shows the diurnal variations of the topside ionospheric parameters such as ion temperature (Figure 5a), electron density (Figure 5b), vertical E B drift (Figure 5c), along with the scale height (H T ) (Figure 5d) for the equinoctial periods. The solid lines in Figures 5a 5d represent the high solar activity period, and the dashed lines represent the moderate solar activity periods. The seasonal mean values of the F10.7 solar flux values for both high and moderate solar activity periods are indicated at the top of Figure 5. In Figure 5b, the blue lines indicate the electron densities at the F 2 layer peak (N m F 2 ) obtained from the Jicamarca Digisonde, and the black lines represent the in situ electron densities at 600 km altitude measured by ROCSAT-1. [16] The weak predawn increase in the scale height can partly be related to a predawn increase in the plasma temperature [Oyama et al., 1996; Bhuyan et al., 2006] which being by definition related to plasma scale height, causes a corresponding increase in the effective scale height (H T ). Also from Figure 5a, it can be noticed that the iontemperature increases rapidly from 0400 h LT, which may 7of13

8 positively contribute to the observed increase in the scale height. However, it may be worth mentioning here that the diurnal variation of the effective scale height (H T ) does not tightly follow the classical plasma scale height (H p ) and/or plasma temperature (T p ) as reported earlier by Liu et al. [2007a, 2007b], Belehaki et al. [2006], and Stankov and Jakowski [2006] and discussed more later in this section. Another theory that was put forward by Stankov and Jakowski [2006] to explain this predawn increase is the lowering of the ionospheric F layer to regions of greater neutral density, leading to increased ion loss due to recombination. This loss effect is particularly strong in the lower side of the ionosphere including the F layer peak [Rastogi, 1988; Moffett et al., 1992]. As a result, the decrease in the N m F 2 and bottomside density is much faster than the topside ionosphere where the loss rate is lower as can be seen from Figure 5b. Thus, an increase in the effective scale height occurs. In addition, the downward plasmaspheric fluxes can also play a role in this increase [Stankov and Jakowski, 2006]. [17] Another remarkable feature in the diurnal variation of H T is the morning time descent and a minimum around LT. This can be explained as being due to the change in the shape of the electron density profile brought about by the difference in the relative production rates between the bottomside ionosphere and topside ionosphere. Soon after sunrise, the bottomside ionosphere is quickly ionized owing to the large quantities of ionizable gases whereas the production in the topside ionosphere is relatively small owing to low densities of ionizable gases at the higher altitudes [Rishbeth and Garriott, 1969]. Also, it can be clearly seen from Figure 5b that the electron density at the F 2 layer peak (N m F 2 ) increases rapidly after sunrise and reaches a maximum around 0900 LT, whereas the electron density at the ROCSAT-1 orbit (600 km) increases gradually and reaches a maximum around 1200 LT. Further, the difference between the F 2 layer peak electron density (N m F 2 ) and the topside electron density reaches a maximum around LT as may be seen from Figure 5b. This large difference in the relative production rates between the bottomside and topside ionospheres causes large negative gradients of electron density in the topside ionosphere. Hence, the thickness of the topside ionosphere becomes much smaller during these local times and the H T exhibits a minimum around LT as seen from Figures 3, 4, and 5d. However, as time progresses the vertical E B drift at the dip equator becomes more prominent and causes the vertical transport of plasma. It can be noticed from Figure 5c that the vertical E B drift increases after sunrise and reaches maximum around 1000 LT. This vertical drift measured by ROCSAT-1 at 600 km can be considered as the proxy to the vertical drift near the F 2 layer peak. Further, Scherliess and Fejer [1999] have also shown that the vertical drift over Jicamarca reaches a maximum around 1000 LT during the equinoxes. The vertical E B drift transports the plasma upward causing the reduction of plasma in the bottomside ionosphere including the F 2 layer peak and a simultaneous increase of plasma in the topside ionosphere. From Figure 5b it can be clearly observed that the F 2 layer peak density (N m F 2 ) decreases and the ROCSAT-1 density (at 600 km) increases and the difference between the two curves becomes smaller around local noon. As a result, the negative vertical electron gradient becomes small and the thickness of the topside ionosphere increases causing a strong maximum in H T around noon hours (Figure 5d). Similarly, the prereversal enhancement (PRE) in the E B drift causes a secondary maximum in the scale height during the postsunset hours as can be seen from Figures 5c and 5d. [18] With a view to further validate the hypothesis described above, the Sami2 is Another Model of the Ionosphere (SAMI2) model simulations [Huba et al., 2000; Huba and Joyce, 2002] have been used to obtain the vertical electron density distributions over Jicamarca. SAMI2 is a low-latitude to midlatitude ionospheric model that solves both ion continuity and momentum equations along the Earth s geomagnetic field in an offset, tilted dipole field. This model also solves the energy equations for H +, He+, O+, and the electrons. The SAMI2 model includes Scherliess-Fejer modeled E B drift of the plasma [Scherliess and Fejer, 1999], and essentially solves for the ion momentum equation for motion along the dipole field line [Huba and Joyce, 2002]. Figure 6 shows the vertical electron density distribution from the SAMI2 model simulations as a function of local time during the equinoxes (Figure 6a), summer (Figure 6b), and the winter solstice (Figure 6c) months for high solar activity. The white curves in Figures 6a 6c indicate the line joining the F 2 layer peak density (h m F 2 /N m F 2 ) points, whereas the red curves represents the line joining the points where the electron density reduced by a factor of e with respect to the F 2 layer peak density (i.e., the height where Ne(h) =N m F 2 /e). The altitude difference between these two curves is indicated by the vertical white lines. Thus, the length of the vertical lines between the white and red curves is a measure of the thickness of the topside ionosphere and is proportional to H T. It can be clearly observed from Figures 6a 6c that during LT, large production of electron density near F 2 layer peak and large negative gradients above the F 2 layer peak limits the thickness of the topside ionosphere to smaller values. As time progresses to LT, the electron density near the F 2 layer peak reduces and the density in the topside ionosphere increases. This is mainly owing to the vertical transport of plasma via E B drift due to the zonal electric field at the dip equator. As a result, the thickness of the topside ionosphere increases rapidly and exhibits a strong maximum around noon hours as may be observed by the vertical lines in Figures 6a 6c. Further, the prereversal enhancement (PRE) in the E B drift at the equator lifts the equatorial ionosphere to higher altitudes and the corresponding changes in the vertical distribution lead to a secondary maximum in the layer thickness during the postsunset hours of equinoctial and summer solstice months. However, the PRE E B drift is small during the winter solstice months and the thickness of the topside ionosphere did not show any significant enhancement during the postsunset hours. Scherliess and Fejer [1999] and Fejer et al. [1999] have also shown that the PRE E B drift velocity over Jicamarca is weak or less prominent during the winter solstices. As a result, the effective scale height also did not exhibit secondary maximum during the postsunset hours of winter solstice months Observations at the Midlatitude Station, Grahamstown [19] Figure 7 is similar to Figure 3, which shows the reconstructed topside electron density profiles (red lines) 8of13

9 Figure 6. The vertical electron density distribution from the SAMI2 model simulations as a function of local time during the (a) equinoctial, (b) summer, and (c) winter solstice months of high solar activity periods over Jicamarca. The white curve indicates the line joining the F 2 layer peak density (h m F 2 /N m F 2 ) points, whereas the red curve represents the line joining the points where Ne(h) =N m F 2 /e. The altitude difference between these two curves is indicated by the vertical white lines. for each 1 hour local time intervals during the equinoctial period of September October 2001 and March April 2002 at a southern midlatitude station, Grahamstown. It can be observed from Figure 7 that the shape of the topside electron density profiles derived from the R-H method (blue lines) follows reasonably close to those derived from the ROCSAT-1 data (red lines) in the absence of E B drift effect at the midlatitudes. However, the thickness of the profile (scale heights) over Grahamstown derived from ROCSAT-1 data is somewhat larger than those derived from the R-H method at almost all local times. [20] Figure 8 illustrates the seasonal mean values of H T over Grahamstown along with their standard deviations as a function of local time for the equinoctial (Figure 8a), summer (Figure 8b), and winter (Figure 8c) solstice months during high solar activity. The mean F10.7 solar flux values for each season are indicated in Figures 8a 8c, respectively. It can be observed from Figure 8 that the diurnal variation of H T over the midlatitude station Grahamstown is relatively small when compared to that at Jicamarca. As expected, the rapid increase in the effective scale height values from 0900 to 1200 LT is not observed mainly because of the absence of the E B drift effect at midlatitudes. However, a small and gradual increase in the scale height is observed between 0900 and 1300 LT, which may be attributed to the increase in the plasma scale height and/or temperature (k b T p /m i g). The 9of13

10 Figure 7. The reconstructed topside electron density profiles (red lines) using the ROCSAT-1 in situ data for each 1 hour local time interval during the equinoxes at Grahamstown. The solid black circles represent the F 2 layer peak density points, and the solid green circles represent the ROCSAT-1 in situ density points. The vertical electron density profiles derived from the R-H method (blue lines) are also shown for comparison. predawn peak can be seen at 0400 LT during the equinoxes (Figure 8a) and around LT during the summer (Figure 8b) and winter (Figure 8c) solstice months. Also, the morning time descent in the scale height values from LT is consistent for all seasons even at midlatitudes, primarily owing to the difference between the production rates at the F 2 layer peak and 600 km above. [21] Hence, the results observed from Figures 7 and 8 indicates that the scale height values derived near the F 2 layer peak (H m ) using the R-H method agree reasonably well with the topside ionospheric effective scale heights (H T ) derived from the ROCSAT-1 in situ electron density data at midlatitudes. However, the diurnal variation of effective scale heights (H T ) derived in this study better captures the important features like predawn peak and morning time descent over the midlatitudes Role of Plasma Temperature at the Equatorial Latitudes [22] With a view to investigate the role of the plasma temperature in the observed diurnal variation of the effective scale heights (H T ), the vertical distribution of the plasma temperature over Jicamarca is derived from the SAMI2 model and presented as a function of local time in Figure 9a. Figure 9b shows the diurnal variation of the effective plasma temperature of the topside ionosphere as derived from the H T values (red line with error bars) and the ROCSAT-1 in situ ion temperature at 600 km (blue line with error bars) for the equinoctial months of Here, the effective plasma temperature (T eff ) is derived from the effective scale height (H T ) using the equation H T ¼ k bt eff mo ð Þ:g ; where the k b is the Boltzmann s constant ( m 2 kg s 1 K 1 ), m(o) is the oxygen mass ( amu), and g is the acceleration due to gravity (9.81 m s 2 ). That means the diurnal variation of T eff shown in Figure 9b is very similar to that of H T ; however, it is shown in temperature scales so as to make easy comparisons with the plasma temperature. In Figure 9, the ROCSAT-1 ion temperature is doubled so as to make it comparable to plasma temperature (T p = T i + T e ). It can be observed from Figures 9a and 9b that the SAMI2 model simulated plasma temperature and the ROCSAT-1 ion ð3þ 10 of 13

11 Figure 8. The seasonal mean values of H T over Grahamstown along with their standard deviations as a function of local time during the (a) equinoctial, (b) summer, and (c) winter solstice months of temperature exhibit similar diurnal variation patterns with a sharp increase in the predawn hours. This increase in the plasma temperature may partly be responsible for a slight increase in the effective temperature (T eff ) and/or scale height (H T ) during the predawn hours. However, during the morning hours the plasma temperature in the topside ionosphere from SAMI2 model as well as ROCSAT-1 ion temperature exhibits a descent till the local noon hours. This is followed by a slight increase in the afternoon hours and then a gradual decrease to a steady nighttime value. However, from Figure 9b, it can be observed that the effective temperature derived from equation (3) exhibits a different local time behavior with a minimum during the morning hours (around h LT), a pronounced maximum during the local noon, and a secondary maximum during postsunset hours. Thus, this distinct behavior of T eff confirms that the diurnal variation of the effective scale height (H T ) does not closely follow the diurnal variation pattern of the plasma scale height (H p = k b T p /m i g) at equatorial latitudes and is largely controlled by the vertical E B drift due to the zonal electric field at the equator as discussed in section Conclusions [23] This paper presents the results of a systematic study on the diurnal, seasonal, and solar activity variations in the topside ionospheric effective scale heights (H T ) over a dipequatorial station, Jicamarca, and a midlatitude station, Grahamstown. The diurnal variations of H T over the dipequatorial station, Jicamarca, generally exhibits a faint peak 11 of 13

12 Figure 9. The (a) vertical distribution of the plasma temperature over Jicamarca derived from SAMI2 model as a function of local time and (b) diurnal variation of the effective plasma temperature (T eff ) of the topside ionosphere derived from the H T values (red line with error bars) along with the ROCSAT-1 in situ ion temperature (doubled) at 600 km (blue line with error bars) during the equinoctial months of during the predawn hours ( LT) followed by a morning time descent to a minimum around h LT. This morning time minimum in the scale heights is attributed to the large negative gradients in the vertical electron density distribution in the topside ionosphere brought about by the difference in the relative production rates between the bottomside ionosphere and topside ionosphere. As the local time progresses from 1000 to 1200 LT, the vertical E B drift at the equator causes the vertical transport of plasma leading to the simultaneous density decrease near the F 2 layer peak and an increase of plasma in the topside ionosphere. This vertical redistribution of plasma due to a strong vertical E B drift at the equator leads to an increase in the thickness of the topside ionosphere and the scale height exhibits a pronounced maximum during the local noon hours. Similarly, the intense prereversal enhancement (PRE) in the E B drift causes a secondary maximum in the scale height during the postsunset hours of equinoctial and summer solstice months. However, this postsunset peak is absent during the winter solstice months in both high and moderate solar activity periods owing to weak or less PRE E B drift over Jicamarca during the winter. Further, the diurnal variation of the effective scale height (H T ) does not closely follow the diurnal variation pattern of the plasma scale height and/or plasma temperature, and is largely controlled by the vertical E B drift due to zonal electric fields at equatorial latitudes. The diurnal variation of the effective scale heights (H T ) over the midlatitude station, Grahamstown, is relatively small when compared to that at Jicamarca mainly because of the absence of strong vertical E B drift at midlatitudes. However, the important features such as predawn peak and morning time descent exist also at the midlatitude station, Grahamstown. [24] In this study, we propose the assimilation of topside in situ electron density data from the ROCSAT-1 satellite in conjunction with the ionosonde measurements for accurate determination of scale heights using an a-chapman function. The reconstructed topside electron density profiles using these scale heights exhibits an excellent similitude with Jicamarca ISR profiles, and are a much better representation than the current Reinisch-Huang topside profiles and/or the IRI-2007 model. The main advantage of this method over ISR measurements is that it allows the determination of the effective scale height (H T ) at a dense network of ionosonde/ Digisonde stations. With the assimilation of topside in situ electron density data from ROCSAT-1, this method becomes very effective in precise determination of the topside effec- 12 of 13

13 tive ionospheric scale heights in the equatorial latitudes and the features like the predawn peak, morning time descent, and role of E B drift due to zonal electric field are clearly evident. ROCSAT-1 satellite with its low-inclination (35 ) orbital plane has a splendid coverage in the equatorial and low latitudes; however, inclusion of the topside in situ electron density data from other missions like CHAMP, DMSP, and C/NOFS can substantially improve the spatial and temporal coverage. [25] Acknowledgments. S. Tulasi Ram is a Postdoctoral Fellow at the Institute of Astronomy and Astrophysics (IAA), Academia Sinica. This work is also partly supported by NSPO project 98-NSPO(B)-IC-FA07-01(S). The authors thank the Digisonde operators at Jicamarca (Oscar Veliz) and at Grahamstown who supplied ionogram data to the UML DIDBase ( and the MIT Haystack Madrigal database ( for providing Jicamarca ISR data. [26] Wolfgang Baumjohann thanks Eurico de Paula, Ivan Kutiev, and another reviewer for their assistance in evaluating this paper. References Belehaki, A., P. Marinov, I. Kutiev, N. Jakowski, and S. Stankov (2006), Comparison of the topside ionosphere scale height determined by topside sounders model and bottomside Digisonde profiles, Adv. Space Res., 37, , doi: /j.asr Bhuyan, P. K., M. Chamua, P. Subrahmanyam, and S. C. Garg (2006), Effect of solar activity on diurnal and seasonal variations of electron temperature measured by the SROSS C2 over Indian low latitudes, Adv. Space Res., 37, , doi: /j.asr Bilitza, B., and B. W. Reinisch (2008), International Reference Ionosphere 2007: Improvements and new parameters, Adv. Space Res., 42, , doi: /j.asr Bilitza, D., B. W. Reinisch, S. M. Radicella, S. Pulinets, T. Gulyaeva, and L. Triskova (2006), Improvements of the International Reference Ionosphere model for the topside electron density profile, Radio Sci., 41, RS5S15, doi: /2005rs Booker, H. G. (1977), Fitting of multi-region ionospheric profiles of electron density by a single analytic function of height, J. Atmos. Terr. Phys., 39, , doi: / (77) Chapman, S. (1931), The absorption and dissociative or ionizing effect of monochromatic radiation in an atmosphere on a rotating earth, Proc. Phys. Soc. London, 43, 26 45, doi: / /43/1/305. Coisson, P., S. M. Radicella, R. Leitinger, and B. Nava (2006), Topside electron density in IRI and NeQuick: Features and limitations, Adv. Space Res., 37, , doi: /j.asr Di Giovanni, G., and S. M. Radicella (1990), An analytical model of the electron density profile in the ionosphere, Adv. Space Res., 10(11), 27 30, doi: / (90)90301-f. Fejer, B. G., L. Scherliess, and E. R. de Paula (1999), Effects of the vertical plasma drift velocity on the generation and evolution of equatorial spread F, J. Geophys. Res., 104, 19,859 19,869, doi: /1999ja Hargreaves, J. K. (1979), The Upper Atmosphere and Solar-Terrestrial Relations: An Introduction to the Aerospace Environment, pp , Van Nostrand Reinhold, New York. Hargreaves, J. K. (1992), The Solar-Terrestrial Environment, pp , Cambridge Univ. Press, New York. Huang, X., and B. W. Reinisch (2001), Vertical electron content from ionograms in real time, Radio Sci., 36(2), , doi: / 1999RS Huba, J. D., and G. Joyce (2002), Open Source Project to aid ionosphere physics research, Eos Trans. AGU, 83(17), 188, doi: / 2002EO Huba, J. D., G. Joyce, and J. A. Fedder (2000), Sami2 is Another Model of the Ionosphere (SAMI2): A new low-latitude ionosphere model, J. Geophys. Res., 105, 23,035 23,053, doi: /2000ja Kutiev, I. S., P. G. Marinov, and S. Watanabe (2006), Model of topside ionosphere scale height based on topside sounder data, Adv. Space Res., 37, , doi: /j.asr Lee, C.-C., and B. W. Reinisch (2008), Quiet-condition variations in the scale height at F2 layer peak at Jicamarca during solar minimum and maximum, Ann. Geophys., 25(12), Lei, J., L. Liu, W. Wan, and S.-R. Zhang (2005), Variations of electron density based on long-term incoherent scatter radar and ionosonde measurements over Millstone Hill, Radio Sci., 40, RS2008, doi: / 2004RS Liu, L., W. Wan, and B. Ning (2006), A study of the ionogram derived effective scale height around the ionospheric hmf2, Ann. Geophys., 24, Liu, L., H. Le, W. Wan, M. P. Sulzer, J. Lei, and M.-L. Zhang (2007a), An analysis of the scale heights in the lower topside ionosphere based on the Arecibo incoherent scatter radar measurements, J. Geophys. Res., 112, A06307, doi: /2007ja Liu, L., W. Wan, M.-L. Zhang, B. Ning, S.-R. Zhang, and J. M. Holt (2007b), Variations of topside ionospheric scale heights over Millstone Hill during the 30-day incoherent scatter radar experiment, Ann. Geophys., 25, Liu, L., M. He, W. Wan, and M.-L. Zhang (2008), Topside ionospheric scale heights retrieved from Constellation Observing System for Meteorology, Ionosphere, and Climate radio occultation measurements, J. Geophys. Res., 113, A10304, doi: /2008ja Moffett, R. J., G. J. Bailey, and B. Jenkins (1992), Effects of greatly increased O+ loss in the ionospheric F region, Planet. Space Sci., 40, , doi: / (92) Nambala, F.-J., L.-A. McKinnell, and E. Oyeyemi (2008), Variations in the ionospheric scale height parameter at the F2 peak over Grahamstown, South Africa, Adv. Space Res., 42, , doi: /j.asr Oyama, K.-I., S. Watanabe, Y. Su, T. Takahashi, and K. Hiro (1996), Seasonal, local time, and longitudinal variations of electron temperature at the height of 600 km in the low latitude region, Adv. Space Res., 18(6), , doi: / (95) Radicella, S. M., and R. Leitinger (2001), The evolution of the DGR approach to model electron density profiles, Adv. Space Res., 27, 35 40, doi: /s (00) Rastogi, R. G. (1988), Collapse of the equatorial ionosphere during the sunrise period, Ann. Geophys., 6, Rawer, K. (1988), Synthesis of ionospheric electron density profiles with Epstein functions, Adv. Space Res., 8(4), , doi: / (88) Rawer, K., D. Bilitza, and T. L. Gulyaeva (1985), New formulas for IRI electron density profile in the topside and middle ionosphere, Adv. Space Res., 5(7), 3 12, doi: / (85) Reinisch, B. W., and X. Huang (2001), Deducing topside profiles and total electron content from bottomside ionograms, Adv. Space Res., 27, 23 30, doi: /s (00) Reinisch, B. W., X. Huang, A. Belehaki, J. Shi, M. Zhang, and R. Ilma (2004), Modeling the IRI topside profile using scale height from groundbased ionosonde measurements, Adv. Space Res., 34, , doi: /j.asr Reinisch, B. W., P. Nsumei, X. Huang, and D. K. Bilitza (2007), Modeling the F2 topside and plasmasphere for IRI using IMAGE/RPI and ISIS data, Adv. Space Res., 39, , doi: /j.asr Rishbeth, H., and O. K. Garriott (1969), Introduction to Ionospheric Physics, pp , Academic, New York. Scherliess, L., and B. G. Fejer (1999), Radar and satellite global equatorial F region vertical drift model, J. Geophys. Res., 104, , doi: /1999ja Stankov, S. M., and N. Jakowski (2006), Topside ionospheric scale height analysis and modeling based on radio occultation measurements, J. Atmos. Sol. Terr. Phys., 68, , doi: /j.jastp Stankov, S. M., N. Jakowski, S. Heise, P. Muhtarov, I. Kutiev, and R. Warnant (2003), A new method for reconstruction of the vertical electron density distribution in the upper ionosphere and plasmasphere, J. Geophys. Res., 108(A5), 1164, doi: /2002ja Su, S. Y., H. C. Yeh, R. A. Heelis, J. M. Wu, S. C. Yang, L. F. Lee, and H. L. Chen (1999), The ROCSAT-1 IPEI preliminary results: Low-latitude ionospheric plasma and flow variations, Terr. Atmos. Oceanic Sci., 10, Yeh, H. C., S. Y. Su, R. A. Heelis, and J. M. Wu (1999), The ROCSAT-1 IPEI preliminary results: Vertical ion drift statistics, Terr. Atmos. Oceanic Sci., 10, Zhang, M.-L., B. W. Reinisch, J. S. Shi, S. Wu, and X. Wang (2006), Diurnal and seasonal variation of the ionogram-derived scale height at the F2 peak, Adv. Space Res., 37, , doi: /j.asr C. H. Liu, Academia Sinica, Taipei 115, Taiwan. L.-A. McKinnell, Hermanus Magnetic Observatory, Hermanus 7200, South Africa. B. W. Reinisch, Center for Atmospheric Research, University of Massachusetts, Lowell, MA 01854, USA. S.-Y. Su, Institute of Space Science, National Central University, Chung-Li, Taiwan. (sysu@jupiter.ss.ncu.edu.tw) S. Tulasi Ram, Institute of Astronomy and Astrophysics, Academia Sinica, Taipei 115, Taiwan. (tulasiram@asiaa.sinica.edu.tw) 13 of 13