Using TEC and radio scintillation data from the CITRIS radio beacon receiver to study low and midlatitude ionospheric irregularities

Transcription

1 RADIO SCIENCE, VOL. 46,, doi: /2010rs004585, 2011 Using TEC and radio scintillation data from the CITRIS radio beacon receiver to study low and midlatitude ionospheric irregularities Carl L. Siefring, 1 Paul A. Bernhardt, 1 Douglas E. Koch, 2 and Ivan J. Galysh 2 Received 31 January 2011; revised 15 May 2011; accepted 16 June 2011; published 15 October [1] Unique data on ionospheric plasma irregularities from the Naval Research Laboratory Scintillation and TEC Receiver in Space (CITRIS) instrument is presented. CITRIS is a multiband receiver that recorded Total Electron Content (TEC) and radio scintillations from Low Earth Orbit (LEO) on STPSat1. The 555 ± 5 km altitude 35 inclination orbit covers low and midlatitudes. The measurements require propagation from a transmitter to a receiver through the F region plasma. CITRIS used both 1) satellite beacons in LEO and 2) the French sponsored global network of ground based Doppler Orbitography and Radiopositioning Integrated by Satellite (DORIS) beacons. This paper is both a brief review of the CITRIS experiment and the first combined TEC and scintillation study of ionospheric irregularities using a satellite borne beacon receiver. It primarily focuses on CITRIS/ DORIS observations and is a case study of the ionospheric irregularities and associated scintillation characteristics at MHz during the 2008 equinox solar minimum. In addition, CITRIS was operated in a complementary fashion with the Communication/ Navigations Outages Forecasting System (C/NOFS) satellite during C/NOFS first year of operations and comparison with measured C/NOFS irregularity characteristics are made. Several types of irregularities have been studied including Spread F and the newly discovered dawn side depletions. Citation: Siefring, C. L., P. A. Bernhardt, D. E. Koch, and I. J. Galysh (2011), Using TEC and radio scintillation data from the CITRIS radio beacon receiver to study low and midlatitude ionospheric irregularities, Radio Sci., 46,, doi: /2010rs Introduction [2] The Scintillation and Tomography Receiver in Space (CITRIS) system was developed by the Naval Research Laboratory (NRL) to provide ionospheric data in the form of Total Electron Content (TEC) and radio scintillations at VHF, UHF, L band and S Band [Bernhardt et al., 2006; Bernhardt and Siefring, 2006]. CITRIS was originally planned as a oneyear proof of concept mission to demonstrate that existing radio beacons on the ground or in Low Earth Orbit (LEO) can be used to provide global space weather data and measure ionospheric irregularities that affect both communication and navigation instruments. CITRIS was launched March 9, 2007 on STPSat1 into a circular orbit 555 ± 5 km altitude and 35 inclination. After about one year and one month of operations CITRIS completed the proof of concept mission. On April 16, 2008, the Communications/Navigation Outage System (C/NOFS) satellite was launched in a 400 km by 850 km, 13 inclination elliptical orbit. The primary purpose of C/NOFS 1 Plasma Physics Division, Naval Research Laboratory, Washington, D. C., USA. 2 Space Systems Development Department, Naval Research Laboratory, Washington, D. C., USA. Copyright 2011 by the American Geophysical Union /11/2010RS is to forecast the presence of ionospheric irregularities that adversely impact communication and navigation systems as described by de La Beaujardière et al. [2004]. From May 2008 to April 2009 the CITRIS receiver was used as a complement to the C/NOFS mission. The orbital periods of the two satellites are relatively close sometimes allowing coordinated measurements over relatively long periods. Some good examples of these coordinated measurements can be found in the work of Siefring et al. [2009]. [3] Irregularity structures in the ionosphere, e.g., Spread F at low latitudes [see, e.g., Basu and Basu, 1985; Basu et al., 1996], cause some of the most serious communications and navigation effects in the form of radio scintillations. Ionospheric plasma structures can start at large scale sizes (100s of km) and cascade through complex processes to short scale sizes (10s of meters). Irregularities with scale features from 100 m to 1 km harm communication and navigation systems through scintillations [cf. Yeh and Liu, 1977; Rino, 1979 and references there in]. The CITRIS frequencies are sensitive to scintillations with these scale sizes. To detect scale size larger than 7.5 km, CITRIS was typically operated to measure TEC averaged over one second CITRIS TEC and Scintillation Measurements [4] CITRIS was operated in two different modes: 1) a ground to satellite mode using the French global network of 1of12

2 CERTO on C/NOFS (or Other LEO Satellites) CITRIS on STPSAT1 150, 400, and 1067 MHz RF Link RF Link 150, 400, and 1067 MHz RF Link 401 and 2036 MHz CERTO Beacon Receiver & Antenna DORIS Geodesy Beacon Figure 1. CITRIS operates in two modes: 1) DORIS to CITRIS (ground to satellite) and 2) CERTO to CITRIS (satellite to satellite). ground based Doppler Orbitography and Radiopositioning Integrated by Satellite (DORIS) beacons transmitting at and MHz beacons or 2) a Satellite to satellite mode using the NRL Coherent Electromagnetic Radio Tomography (CERTO) three frequency beacons transmitting at / / MHz [Bernhardt and Siefring, 2006] or legacy beacons transmitting only at the 150/400 MHz frequencies. The ground and space hardware associated with these modes are illustrated in Figure 1. [5] CITRIS measures differential phase delay for each frequency pair. The phase delay P ab is directly proportional to the number of electrons on the propagation path as given by equation (1) [cf., Bernhardt et al., 2006]. Z TEC ¼ Nds¼ cn an 2 b f 0 2" n 2 a hp ab i n2 b ¼ 1: n a n 2 b f 0 n 2 a hp ab i ð1þ n2 b where " ¼ q2 8 2 m e " 0 40:3 the TEC (integrated density) is in m 2, the differential phase P ab is in radians, q is the electron charge, m e is the electron mass, and " 0 is the vacuum permittivity. We note, the CITRIS instrument processes the signal on board the satellite and for the data presented in this paper measurements are recorded on a one second cadence. The h i in equation (1) (and elsewhere) indicates an average over one second that CITRIS compiles using 200 Sample/s data with the receiver filtering to 100 Hz ð2þ final bandwidth. Equation (1) assumes two beacon frequencies with the ratio n a :n b, where n a and n b are integers. Phase coherent frequencies are transmitted at f a =n a f 0, and f b =n b f 0, where f 0 is the base frequency. The DORIS frequencies and 2, MHz have n a = 107 and n b = 543 with a common base frequency of f 0 = 3.75 MHz, while the three CERTO frequencies MHz, MHz and MHz have integer ratios of 9:24:64 and a base frequency of f 0 = MHz. As with all phase measurements, there are 2p ambiguities and phase offsets. These make obtaining Absolute TEC (ATEC) difficult and analysis is commonly done using Relative TEC (RTEC) that is arbitrarily normalized to zero at its minimum value. This is the worst possible normalization choice for doing any sort of mathematical analysis of fluctuations in the TEC where a percentage variation would be a useful value to compare, but DTEC/TEC is undefined at the zero normalization point. Bernhardt et al. [2010] outlines an extensive history of attempts to generate absolute TEC from similar measurements and solves the problem for satellite to satellite measurements using close flybys of CITRIS with other LEO satellites. The close proximity allows for extrapolating the measured TEC to zero separation distance where the TEC must be zero. Siefring et al. [2009] explored CITRIS near passes with C/NOFS and compared radio TEC data with in situ measurements to find practical limits on the absolute TEC technique for satellite to satellite measurements. Siefring et al. [2009] also tested a different normalization for DORISto CITRIS measurements which will be used here and is outlined later in the text. 2of12

3 [6] CITRIS also measures scintillations in two forms; 1) amplitude scintillations and 2) phase scintillations. A measure of the amplitude scintillations is the scintillation index defined as sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi h Wa i 2 S 4 ¼ W 2 a hw a i 2 where W a is the received power at frequency f a, and again h i is a temporal average over one second for the CITRIS observations. The S 4 index is essentially the standard deviation of the received power divided by the average of the received power. S 4 can reach 1.4 and an S 4 of 1 means the variation in the power is as large as the average received power. An S 4 index of 0.5 represents a 12 db fluctuation in the received signal power and would be considered a strong scintillation event while an S 4 of 0.3 might be considered moderate scintillations. The phase scintillation index is the standard deviation of the phase given by qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ h a i 2 2 a where a is the received phase in radians. [7] Scintillation statistics were compiled using 200 Sample/s data (with 100 Hz bandwidth) with the receiver keeping track of the power and phase averages and standard deviations to generate the s and S 4 index at the same one second interval as the TEC data. The one second averaging time is shorter than typically used with ground receivers for scintillation measurements. For a satellite in LEO the averaging time needs to be relatively short compared a GPS or Medium Earth Orbit (MEO) satellite, because of the rapid satellite motion and short observing time. The one second averaging was chosen experimentally, considering data volume, signal to noise etc. once the CITRIS was in orbit. The lower noise environment in space and lack of local interferers may have allowed better performance in this respect than is typical for a ground based receiver. There are limits to the sample rates and averaging times that must be considered. Rino [2011] indicates the scattered spectrum peaks near the Fresnel scale p L F ¼ ffiffiffiffiffiffi D ð5þ where l is the radio wavelength and D is the distance from scintillating region to CITRIS. To measure a meaningful S 4, the sampling time must be short and the averaging time must be long compared to the time it takes for CITRIS to pass over a distance L F. These conditions are given by v sat t sample L F v sat t ave where v sat is the satellite velocity ( 7500 m/s), t sample is the sampling time (1/200 s), and t ave is the averaging time (1 s). For MHz the wavelength is approximately m and assuming D = 250 km the inequality yields 37.5 m 432 m 7500 m. We note that longer average times (combining statistics of each 1 s interval) can be obtained from the CITRIS processed data. The receiver accumulates the appropriate sums of the phases and amplitudes statistics ð3þ ð4þ ð6þ for each averaging period. The S 4 and s are calculated on the ground from these accumulated sums Ionospheric Measurements and the DORIS Network of Beacons [8] The DORIS system was developed by Centre National d Etudes Spatiales (CNES), Institut Géographique National (IGN) and Groupe de Recherche en Géodésie Spatiale (GRGS) to provide 1) very precise orbit determination [Nouël et al., 1994], 2) precise ground beacon position [Willis et al., 2006], 3) measurement of Earth center position [Feissel Vernier et al., 2006], 4) real time orbit determination [Jayles and Costes, 2004], and 5) gravity field determination [Nerem et al., 1994]. For these studies the Doppler phase delay induced by the ionosphere must be removed and is typically done so on a ten second cadence. Data from standard DORIS receivers have been used for ionospheric studies by Trigunait et al. [2004], Li and Parrot [2007], and other researchers. CITRIS, however, was specifically designed for making ionospheric measurements and has significantly more capability for ionospheric measurements than the standard DORIS receivers [Bernhardt and Siefring, 2010]. [9] Figure 2 shows the DORIS permanent beacon network that includes 54 beacons stations hosted by institutes from more than 30 countries. Also shown on the map are the geographic and magnetic equators and the latitude extent of the STPSat1 orbit. The DORIS stations highlighted with larger text and dots are those used in this study. These stations were selected because of good data coverage during the 2008 Aug Sept equinox and are representative of three different magnetic latitudes: 1) Near the magnetic equator, 2) Near the location where the Appelton anomalies typically develop ( 15 Magnetic Latitude) and 3) Midlatitudes. Table 1 gives the DORIS stations used in this study, their Geographic Latitude and Longitude and general latitude category. Three DORIS sites were near the magnetic equator, three sites were at 15 off the magnetic equator, and two sites were at midlatitudes. [10] This paper focuses on combined TEC and amplitude scintillation data (S 4 index) using the DORIS beacon measurements with CITRIS. Ionospheric irregularity observations are compared qualitatively with C/NOFS during the fall equinox of The measurements presented are taken in unusually quiet geomagnetic conditions and indicate equatorial irregularities are often found for nighttime measurements. The ionospheric conditions for these measurements are a little unusual. As described by Heelis et al. [2009], C/NOFS has been flying in an abnormally cold and compact atmosphere because of the long solar minimum with low F10.7 fluxes (average 67 for the months in this data). The altitudes of both the F peak and the transition height between O+ and H + are lower than usual. The F region electron densities also tend to be low. The phenomenologies of the equatorial irregularities observed by C/NOFS are varied and appear to be somewhat different from previous satellite missions flown in more active solar conditions. Several examples have been published including Spread F occurring post midnight instead of near dusk [Burke et al., 2009], broad plasma depletions [Huang et al., 2009] and dawn side depletions [de La Beaujardière et al., 2009]. Theoretical studies of these unique irregularity characteristics are just 3of12

4 Figure 2. Map of DORIS beacons adapted from International DORIS Site doris/maps.html. Ovals indicate 6 elevation observations of DORIS beacon by CITRIS. CITRIS/DORIS observations are presented for the highlighted sites, see also Table 1. Also shown are the STPSat1 orbit latitude limits and the geographic and magnetic equators. beginning with assimilative ionospheric models [Su et al., 2009] and physics based ionospheric models [Huba et al., 2010]. 2. Data Presentation 2.1. DORIS to CITRIS Data [11] Figure 3 shows sample TEC and scintillation data for the DORIS station at Betio, Kiribati in the Pacific Ocean. The left hand side is daytime data and the right hand side is nighttime data. The top row is the relative TEC plotted using standard Total Electron Content Units (TECU) where 1 TECU = m 2. The daytime data (top left) shows primarily the affect of the variation in path length (second row) from DORIS to CITRIS as the spacecraft overflies the transmitter site. Typically TEC traces have a U shape because of this variation in path length. For CITRIS, the antenna was oriented in the ram or wake direction depending on the required star tracker orientation. This orientation was a compromise to allow both ground to CITRIS and satelliteto CITRIS measurements. With the antenna beam pointing in the horizontal the system tends to lose the DORIS signal on the backside of the antenna, thus only about one half of the U curve is seen. The nighttime TEC trace in Figure 3 (top right) shows significant structure along with the path length variation. The third row shows the 401 MHz amplitude scintillation characteristics. There are relatively severe scintillations in the region of the nighttime pass where the TEC is structured. The maximum S 4 index approaches 0.7 with duration ( half width) of about 100 s. These results imply strong scintillations covering about 750 km along the orbital path (satellite velocity 7.5 km/s). From the STPSat1 orbit, the longitude extent of the strong scintillations can be inferred and is 4.9 degrees for this 100 s interval. The daytime S 4 index at 401 MHz shows a hint of some scintillation activity but values below 0.2 would not have a significant effect on a typical receiver. At 2036 MHz the S 4 characteristics are primarily due to receiver noise floor and consequent signal tonoise ratio (SNR) variation and not scintillations. At large distances the SNR is small because of the large path loss and the S 4 reflects this with a 0.2 to 0.3 level at the largest ranges. The 2036 MHz antenna has a narrower beam width than the 401 MHz antenna. The difference is apparent as the 2036 MHz S 4 index again increases as the DORIS station starts to move out of the main antenna beam. [12] The data seen in Figure 3 can be considered typical in many respects. However, the nighttime scintillations are Table 1. DORIS Beacon Sites Observed by CITRIS Site Station Code East Longitude (Degrees) North Latitude (Degrees) Magnetic Latitude Region Arequipa AREB Equatorial Ascension ASDB Magnetic Latitude Betio BETB Equatorial Cibinong CICB Magnetic Latitude Djibouti DJIB Equatorial Futuna FUTB Magnetic Latitude Hartebeesthoek HBKB Midlatitude Miami MIAB Midlatitude 4of12

5 Figure 3. CITRIS TEC and Scintillation data for the Betio DORIS site on August 18, The left hand column is a daytime pass (LT 1200) and the right hand column is a nighttime pass (LT 2300). more intense and longer lasting than most seen during solar minimum. Sometimes there are periods when the receiver will come in and out of lock (on the 2036 MHz for DORIS) at both large ranges and near the receiver sight. A couple of drop outs in the TEC can be seen in the daytime Figure 3 data. In these cases there was no cycle slip. Cycle slips can sometimes be corrected by using the time derivative of the TEC to adjust the next section of the TEC trace. This process for correcting cycle slips is used occasionally in the data presented here. In both of the passes there were short periods of data (not shown) 50 to 100 s after the receiver originally lost lock. Data after such long duration drops with unknown cycle slips are not included in any of the data presentation or analysis TEC Analysis [13] DORIS to CITRIS measurements suffer from being RTEC measurements. One issue working with RTEC data is doing concrete analysis even de trending can be less than satisfactory for quantifying irregularity variations. More quantitative analysis is possible if the measurements can be converted to even a rough estimate of absolute TEC. Siefring et al. [2009] explored using a more sensible normalization than setting the minimum to zero. There is information on the ATEC in the total RTEC variation and the steepness or 5of12

6 Figure 4. Geometry used for estimating path average electron density N pa and adjusting relative TEC to an estimate of absolute TEC. shallowness of the RTEC curve. To determine the normalization it is assumed the ionosphere only varies in altitude and a TEC offset is calculated under this assumption. This is consistent with the typical treatment for converting slant TEC to vertical TEC using only geometric considerations to compensate for the path length; a widely accepted practice for daytime measurements. [14] Figure 4 illustrates the geometry. D max and D min are the path length above 90 km (assuming no free electrons below 90 km) at maximum and minimum separations and TEC max and TEC min are the RTEC measurements at these points. A horizontally stratified ionosphere will yield a constant path average electron density, i.e., the idea is to assume the path average electron density is the same at the farthest and nearest distance. We note that the CITRIS receiver is in a nearly circular orbit (555 ± 5 km) so the altitude does not very significantly. The average electron density would be the ATEC divided by the path length and leads to the following equality TEC max þ Offset D max ¼ TEC min þ Offset D min and the TEC offset can be solved for with Offset ¼ TEC max D max TEC min D min 1 1 D min D max An estimate of ATEC is generated by adding the offset to the RTEC curves. This method avoids creating an artificial ð7þ ð8þ Figure 5. Sample TEC data for DORIS stations near the magnetic equator. The left hand side is a daytime case and the right hand side is a nighttime case. (top) The RTEC and estimated ATEC. (bottom) The pathaverage electron density N pa where the ATEC has been adjusted for the ionospheric path length. The total variation of the N pa from maximum to minimum is indicated and will be used as a measure of horizontal structure for each pass. 6of12

7 Figure 6. Two CITRIS passes by Arequipa on consecutive days at nearly the same local time and approximately the same path average electron densities N pa. (left) The 401 MHz scintillations associated with ±1% variations while (right) there are no scintillations associated with the D N pa /N pa ±0.2% variations. mathematical pole that is introduced by setting the minimum TEC to zero. [15] Figure 5 shows sample TEC analysis for the DORIS station at Arequipa, Peru. Again, the left hand side is a daytime pass and right hand side a nighttime pass. Figure 5 (top) shows both RTEC and estimated ATEC after solving for the offset. Using the ATEC, the path average electron density (N pa ) as CITRIS flies over the DORIS can be calculated and is presented in the bottom row of Figure 5. The variation of N pa over the pass can be quantified and is about ±1.5% for the daytime case and ±20% for the nighttime case. The daytime variation of only 1.5% indicates the validity of the calculated offset when the ionosphere is smooth in the horizontal direction and it is somewhat surprising that this technique has not been used in the past. The nighttime N pa shows signs that the ionosphere has significant structure present in the horizontal direction during each CITRIS/ DORIS pass. [16] As seen in Figure 5, the nighttime RTEC variation is below 0.5 TECU for the entire pass and the ATEC at the end of the pass is 0.22 TECU. The range at that point is 587 km from an altitude of 559 km, thus 0.22 TECU also represents the vertical TEC (VTEC). On CITRIS, it was common for nighttime RTEC total variation and ATEC to be near or below 1 TECU, which reveals the nighttime ionospheric electron density is extremely small for these low F10.7 solarminimum conditions. [17] The quantity DN pa /N pa can be generated using de trending to emphasize smaller scale irregularities. This procedure is useful for comparing features of the TEC and N pa with scintillation data is illustrated in Figure 6. Here the de trending is applied to the path averaged electron density using the Arequipa DORIS site on two consecutive days. In each case, the data are from about 15 to 20 min after 1 A.M. local time on different days. Figure 6, top, shows both the derived N pa and N pa a smoothed running average. The two sets of data lie almost on top of each other on the plotted scale. We can see that N pa has roughly the same order of magnitude on both days with the plots centered at cm 3 and cm 3. Figure 6, middle, shows the percent- 7of12

8 Figure 7. N pa and scintillation data for three equatorial sites. Figure 8. N pa and scintillation data for three low latitude site near 15 Magnetic Latitude. These sites could be below the Appleton anomalies when it is present. 8of12

9 Figure 9. Comparison of the path average electron density for (top) a nighttime equatorial case associated with severe scintillations S4 = 0.6 and (bottom) a 15 o magnetic latitude case in the early evening with moderate scintillations S4 = 0.4. Smaller scale structures are apparent in the nighttime case. age change in the N pa as a function of time as defined by equation (9). DN pa N pa ¼ N pa N pa N pa This is similar to in situ measurements of irregularities where the intensity of the irregularities is quantified by DN pa /N pa. Of importance here, the amplitude on 9/6/2008 (left hand), with variations reaching 1% is significantly larger then on 9/5/2008, with variations below 0.2%. Comparing the S 4 index with the DN pa /N pa on 9/6/2008 (Figure 6, left) it is clear that there is a correlation between the peaks in the 401 MHz S 4 and the small scale variations in the path density. On the night of 9/5/2008 (Figure 6, right) there are no scintillations of any significance. This illustrates that both a high plasma density plus density fluctuations are required to generate the scintillations. It should be noted that the minimum TEC scale size measured in the CITRIS data presented is 7.5 km and it is smaller scale sizes irregularities near 450 m that cause the scintillations at 401 MHz. ð9þ 2.3. TEC and Scintillation Characteristics [18] Figure 7 shows processed path average electron density (top row) and scintillation data (bottom row) plotted versus Local Time (LT) for three DORIS beacons located near the magnetic equator. The DORIS sites are Arequipa, Betio and Djibouti; they are all at magnetic latitudes below about five degrees. The data displayed are taken from the months of August and September of Equinox was chosen for this initial analysis because of the equal lighting conditions. The data points in Figure 7, top, show the maximum, minimum and average value for the N pa for each pass, in a form similar to an error bar. For purposes of discussion we consider the variation about the average as a simple measure of structuring in the observed ionosphere. The expected LT trend is seen in the data with the daytime ionosphere having the highest electron densities and densities decreasing after sunset. We note that the CITRIS receiver is in a nearly circular orbit (555 ± 5 km) so the altitude variation is not significant between data points. There is a clear tendency for large variations to exist in the N pa (and TEC) when the density is low. The approximate sunrise at the satellite altitude is indicated by the vertical line on each plot. It is notable that the lowest densities are seen after sunrise at the satellite altitude. Other times of interest are post sunset ( 19 to 23 LT) where Spread F might normally be active and near or postmidnight (23 to 3 LT) where C/NOFS observations indicate the most structure in the ionosphere during this solar minimum time span [cf. Burke et al., 2009]. [19] Figure 7, bottom, summarizes CITRIS/DORIS 401 MHz amplitude scintillation data. Two aspects of the scintillations are presented, the peak S 4 and the duration of the scintillation disturbance categorized in five course bins, indicated by the dot size. The duration is an estimate. If the scintillation event was a simple peak, this duration is a half width. If the disturbance had a more complex structure a by eye judgment was made on the duration. All observations are plotted with a minimum S 4 index of 0.1 and 1 s duration. The two horizontal lines indicate S 4 index levels of 0.2 and 0.4. A level below 0.2 would not cause issues for a communications receiver. We might consider an S 4 of 0.2 to 0.4 to be moderate scintillations and >0.4 to represent significant or severe scintillations. There are a few points from Arequipa and the Betio ( 23 LT) example from Figure 3 where the level reaches >0.4. Comparing the N pa data with the S 4 data shows some trends. The peak scintillations tend to occur when there is both structure in the ionosphere and a relatively large density. Thus, even though the most structured passes occur from about 24 to 5 LT the scintillations are smaller. Long duration scintillations tend to occur between 20 to 23 LT. At Arequipa, there are some observations that are out of the normal trends. The scintillation event at 4:15 LT occurs when the N pa is low. Two significant scintillation events occur during the daytime near 7 LT where the ionosphere has started to build back up but the N pa observations still indicate significant structure. Finally, the Arequipa daytime scintillation event near 15 h is out of character. [20] Figure 8 shows similar path average electron density and scintillation data for three DORIS beacons located near 15 magnetic latitude. The Ascension Island, Cibinong, and Futuna sites were chosen because they would be located under the Appleton Anomaly (when it is present) and thus, 9of12

10 Figure 10. N pa and scintillation data for two midlatitude sites. at times, have higher electron densities and potentially more severe scintillations. The data shows some indications of the anomaly with higher peak densities during the day (note change in scale for the Cibinong and Futuna data), more variability in the absolute densities, and apparent structuring in some of the daylight passes especially in the early evening (15 18 LT). The Cibinong site shows numerous moderate scintillation events. There are no indications of severe scintillations at these sites, although there is poor sampling for the Ascension and Futuna sites for local times where higher scintillation might be expected and are seen in the equatorial data of Figure 7. Again the lowest electron densities appear to occur shortly after sunrise at the satellite orbit. [21] Figure 9 shows that there is quite a different character to the structure in the nighttime equatorial data and the 15 magnetic latitude data where structure is indicated during the day. The plots show DN pa /N pa (equation (9)) to emphasize the smaller scale irregularities for two sample cases. These illustrate that the structure in the presumed Appleton anomaly case have much larger scale sizes then the nighttime equatorial case. [22] Finally, Figure 10 shows path average electron density and scintillation data for two DORIS beacons located at midlatitudes. The densities are lower during the day, than the equatorial and low latitude examples of Figures 7 and 8 and they do not reach as low a level during the nighttime. Minimum densities still appear after 4 LT. The traces also do not show the large variations per pass as seen in the equatorial data of Figure 7. The scintillation data show mostly mild to moderate levels during the nighttime when equatorial irregularities might reach higher latitudes. There are some longer periods of scintillations, especially at the Hartebeesthoek site which appear to be scattered through all local times. 3. Discussion [23] On CITRIS, it is common for the nighttime RTEC total variation and estimated ATEC to be near or below 1 TECU. These results indicate that the TEC for the nighttime ionosphere is extremely low with these low F10.7 solarminimum measurements. The data plotted in Figure 7 shows that the nighttime equatorial ionosphere is very likely to be structured during solar minimum. This behavior is apparent from the significant number of data points with a wide spread between the minimum and maximum N pa between 2100 and We originally thought that the ionosphere showed structure every time the total TEC was low (<2 TECU) but the plot indicates this is true only about 80% of the time. We do not know of previous satellite data on the characteristics of the equatorial ionosphere in these low F10.7 conditions, so these data are very unique. [24] The data in Figures 7, 8 and 10 show the distinctive characteristic that the lowest average electron densities occur after sunrise at the satellite. These observations are in agreement with the exciting new C/NOFS data presented by de La Beaujardière et al. [2009]. The in situ data show deep depletions in the dawn side ionosphere when C/NOFS emerges at sunlit altitude. One possible issue with probe measurements is the change in the spacecraft local environment at the transition between darkness and sunlight. Although the evidence is good that the depletions reported by de La Beaujardière et al. [2009] are real and not instrumental effects, it is encouraging that the presented radio differentialphase measurement (which are not significantly effected by 10 of 12

11 the local plasma environment) indicates the same phenomenon. In addition, the TEC data indicate that the dawn side depletions affect the ionosphere at all altitudes below the STPSat1 orbit 555 km and possibly extend to a wider range of latitudes then expected, regularly reaching well into midlatitudes. de La Beaujardière et al. [2009] estimate, for one case, that the latitude extent of the depletions to be ±25 of the magnetic equator. Our data suggests that the phenomena reaches ±30 and likely further in both the American and African sectors. [25] Initial comparisons of the CITRIS N pa observations from the Arequipa and Betio equatorial sites with theoretical results for the SAMI3 Ionospheric model [Huba et al., 2010] show qualitative agreement. SAMI3 is a physicsbased model developed at NRL and has been described in several publications by Huba et al. [2000, 2008]. The model reproduces the overall LT dependence and the production of deep dawn side depletions. The explanation of the dawn side depletions proposed by Huba et al. [2010] lies in the neutral winds at the daytime terminator. Similar to the better known pre reversal enhancement which is thought to be an important factor in Spread f stability, these winds can lift the ionosphere causing low density plasma to be transported from low to high altitudes. [26] The winds at both terminators are not well characterized because of the difficulties of using optical instruments for making remote wind measurements under conditions of terminator lighting. The C/NOFS observations [de La Beaujardière et al., 2009] and the SAMI3 model results [Huba et al., 2010] show that there should be a strong longitudinal dependence. CITRIS TEC measurements should provide a climatologic database that will enhance our understanding of this critical coupling between the neutral winds and the ionosphere. In some respects, the TEC data that samples the ionosphere below a certain altitude (effectively averaging a large range of altitudes) is a better candidate for comparison to ionospheric models than in situ data from a specific altitude. [27] The observations indicate the nighttime ionosphere is regularly structured near the magnetic equator and this structure often extends to ±15 magnetic latitude. Burke et al. [2009] noted that C/NOFS in situ observations during the same time period show top side plasma depletions after midnight and considers these to be Equatorial Plasma Bubbles (EPB) from Spread F. Previously, post midnight EPB observations have been extremely rare from satellites [cf. Burke et al., 1979]. Other structures (as in the previously discussed Dawn side depletions) are also seen and these may not necessarily be due to a Spread F instability. The broad plasma decreases in the equatorial ionosphere described by Huang et al. [2009] are examples. The broad decreases also appear to extend past midnight. For this study we do not attempt to classify the source or type of the ionospheric structure only to characterize them via the variations in electron density, the N pa /N pa and scintillations they cause. Although not completely understood, it is likely postmidnight density depletions are more common during solar minimum. The CITRIS/DORIS TEC data supports the existence and tendency to have the largest variations in density near or post midnight in the ionosphere below 555 km, especially near the magnetic equator. It is, however, interesting that the 401 MHz scintillation events are skewed to earlier in the evening (19 to 23 LT) and sometimes occur in the early morning ( 8 LT) when there are higher plasma densities present. [28] Overall, as expected, severe 401 MHz scintillations are not common (rarely reaching > 0.5 S 4 index) for the CITRIS/ DORIS observations during solar minimum. However, the data do regularly show detectable amplitude scintillations. In addition, when examining times when we expect scintillations, such as the dusk to midnight region at the equator, it is not difficult to find examples in the mild to moderate range (0.2 to 0.4 S 4 index). We note that in more active solar conditions it is not uncommon to observe S 4 index saturating around 1.0 with ground based measurements at UHF associated with Spread F. [29] The observations at 15 magnetic latitude do not show any tendency toward higher S 4 contrary to what might have been expected if under the higher plasma density Appleton anomaly. Under the Appleton anomalies the electron density can sometimes be structured even during daytime conditions (see Cibinong data). These structures seem to be mostly large scale gradients with 1000 km scale sizes, although they obviously contain smaller scales as they produce minor to moderate scintillations at 401 MHz. The three sites, Ascension Island, Cibinong, and Futuna did show the largest N pa and ATEC during the daytime as expected. The Appleton anomalies can be highly variable during solar minimum and this is apparent in Figure 8 with the large scatter in the N pa data between passes during the daytime. Characterizing this variability may require a larger climatological study then was performed here. [30] At midlatitudes the measured N pa show less LT variability than at lower latitudes. This is apparent in the N pa for each pass in Figure 10, which does not reach as high a daytime maximum density, or as low a minimum density as seen for the low latitude cases of Figures 7 and 8. It is also apparent from the N pa variation within each pass, i.e., the small spread between the maximum, average, and minimum values in Figure 10, top. The 401 MHz scintillation data have a different character than the low latitude sites. This difference may be because E Region phenomena, e.g., Sporadic E, may be contributing to the scintillations; Sporadic E is not common at the equator. It would be useful to examine ground based ionosonde observations near midlatitude DORIS sites to make further comparisons. This investigation is planned for future work. [31] Finally, scale sizes play an important role in determining if a structured ionosphere will cause scintillations. With the technique used here to estimate absolute TEC it may be possible to perform Fourier transform on the TEC data to test for the amplitudes of DN pa /N pa at smaller scale sizes. A spectrum that has high amplitudes near 7.5 km may indicate more structure near the 401 MHz Fresnel scale of 450m. As mentioned previously, CITRIS also measures S 4 and a and there are examples of amplitude scintillations in the 2036 MHz data and phase scintillations. Exploring these topics is planned for future work. 4. Summary [32] The first climatologic case study of ionospheric irregularities using a space based beacon receiver (CITRIS) has been presented. New analysis techniques were used in 11 of 12

12 this study. The results show good qualitative comparisons with both C/NOFS in situ observations during the same time period and with ionospheric models [Huba et al., 2010]. The CITRIS receiver archived TEC and radio scintillation data from both ground DORIS beacons and LEO CERTO beacons for 2 years from launch in March Future investigation of the archived DORIS CITRIS data will provide climatological trends for the occurrence of low latitude and midlatitude irregularities. This database is unique because the observation period was the weakest solar cycle for the past 200 years. The CITRIS operations on STPSat1 successfully demonstrated the utility of a space based receiver to use an existing network of ground based beacons for space weather measurements. [33] The TEC and scintillation data can provide useful information on a wide range of scale sizes of the observed irregularities. However, the physics behind the formation and development of ionospheric irregularities and consequent scattering of radio signals cannot be understood with a single type of measurement. More information is needed using multiple measurement techniques. An example of the utility of multiple measurements was presented by Bernhardt [2005] discussing combining a CITRIS receiver to measure predominantly horizontal irregularities with a GPS radio occultation receiver to measure vertical electron density profiles to fully describe the F region irregularities. The C/NOFS satellite carries a radio beacon transmitter, in situ ion and electron probes, electric field detectors, neutral density, mass, and wind instruments, GPS occultation receivers, and magnetometers. CITRIS on STPSat1 has already demonstrated the ability to provide cross calibration [Siefring et al., 2009]. In this study, it was demonstrated that CITRIS can contribute greatly to the observations of ionospheric irregularities on future multisensor satellites. [34] Acknowledgments. This work was supported by the Office of Naval Research. The STPSat mission was supported by the Department of Defense Space Test Program. Collaborations with the C/NOFS mission were supported by the Air Force Research Laboratory, and the National Aeronautics and Space Administration. The authors acknowledge the contributions of M. Long, T. Rodilesso, T. MacDonald and M Wilkens from the Naval Research Laboratory and P. Roddy and D. Hunton of Air Force Research Laboratory. References Basu, S., and S. Basu (1985), Equatorial scintillations: Advances since ISEA 6, J. Atmos. Terr. Phys., 47, , doi: / (85) Basu, S., et al. (1996), Scintillations, plasma drifts, and neutral winds in the equatorial ionosphere after sunset, J. Geophys. Res., 101, 26,795 26,809, doi: /96ja Bernhardt, P. A. (2005), Eye on the ionosphere: Ionospheric profiling by GPS receiver occultations, taking advantage of existing Earth based infrastructure, GPS Solutions, 9(1), , doi: /s Bernhardt, P. A., and C. L. Siefring (2006), New satellite based systems for ionospheric tomography and scintillation region imaging, Radio Sci., 41, RS5S23, doi: /2005rs Bernhardt, P. A., and C. L. Siefring (2010), Low latitude ionospheric scintillations and total electron content obtained with the CITRIS instrument on STPSat1 using radio transmissions from DORIS ground beacons, Adv. Space Res., 45(12), , doi: /j.asr Bernhardt, P. A., C. L. Siefring, I. J. Galysh, T. F. Rodilosso, D. E. Koch, T. L. MacDonald, M. R. Wilkens, and G. P. Landis (2006), Ionospheric applications of the Scintillation and Tomography Receiver in Space (CITRIS) used with the DORIS radio beacon network, J. Geod., 80, , doi: /s Bernhardt, P. A., C. L. Siefring, I. J. Galysh, and D. E. Koch (2010), A new technique for absolute total electron content determination using the CITRIS instrument on STPSat1 and the CERTO beacons on COSMIC, Radio Sci., 45, RS3006, doi: /2009rs Burke, W. J., R. C. Sagalyn, R. G. Rastogi, M. Ahmed, F. J. Rich, D. E. Donatelli, and P. J. L. Wildman (1979), Postsunrise refilling of the low latitude topside ionosphere, J. Geophys. Res., 84, , doi: / JA084iA08p Burke, W. J., O. de La Beaujardière, L. C. Gentile, D. E. Hunton, R. F. Pfaff, P. A. Roddy, Y. J. Su, and G. R. Wilson (2009), C/NOFS observations of plasma density and electric field irregularities at post midnight local times, Geophys. Res. Lett., 36, L00C09, doi: /2009gl de La Beaujardière, O., et al. (2004), C/NOFS: A mission to forecast scintillations, J. Atmos. Sol. Terr. Phys., 66, , doi: /j.jastp de La Beaujardière, O., et al. (2009), C/NOFS observations of deep plasma depletions at dawn, Geophys. Res. Lett., 36, L00C06, doi: / 2009GL Feissel Vernier, M., et al. (2006), Geocentre motion measured with DORIS and SLR, and predicted by geophysical models, J. Geod., 80(8 11), , doi: /s z. Heelis, R. A., et al. (2009), Behavior of the O+/H+ transition height during the extreme solar minimum of 2008, Geophys. Res. Lett., 36, L00C03, doi: /2009gl Huang, C. Y., F. A. Marcos, P. A. Roddy, M. R. Hairston, W. R. Coley, C. Roth, S. Bruinsma, and D. E. Hunton (2009), Broad plasma decreases in the equatorial ionosphere, Geophys. Res. Lett., 36, L00C04, doi: /2009gl Huba, J., G. Joyce, and J. 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 Huba, J. D., G. Joyce, and J. Krall (2008), Three dimensional equatorial Spread F modeling, Geophys. Res. Lett., 35, L10102, doi: / 2008GL Huba, J. D., G. Joyce, J. Krall, C. L. Siefring, and P. A. Bernhardt (2010), Self consistent modeling of equatorial dawn density depletions with SAMI3, Geophys. Res. Lett., 37, L03104, doi: /2009gl [Correction, Geophys. Res. Lett., 37, L20104, doi: / 2010GL045004, 2010.] Jayles, C., and M. Costes (2004), Ten centimeters orbits in real time on board of a satellite: DORIS DIODE current status, Acta Astronaut., 54(5), , doi: /s (03) Li, F., and M. Parrot (2007), Study of the TEC data obtained from the DORIS stations in relation to seismic activity, Ann. Geophys., 50(1), Nerem, R. S., et al. (1994), Gravity field development for TOPEX/ Poseidon, joint gravity model 1 and model 2, J. Geophys. Res., 99, 24,421 24,447, doi: /94jc Nouël, F., et al. (1994), Precise Center National D etudes Spatiales orbits for TOPEX/Poseidon: Is reaching 2 cm still a challenge?, J. Geophys. Res., 99, 24,405 24,419, doi: /94jc Rino, C. L. (1979), A power law phase screen model for ionospheric scintillation: 1. Weak scatter, Radio Sci., 14(6), , doi: / RS014i006p Rino, C. L. (2011), The Theory of Scintillation With Applications in Remote Sensing, Wiley IEEE Press, Hoboken, N. J. Siefring, C. L., P. A. Bernhardt, P. A. Roddy, and D. E. Hunton (2009), Comparisons of equatorial irregularities measurements from C/NOFS: TEC using CERTO and CITRIS with in situ plasma density, Geophys. Res. Lett., 36, L00C08, doi: /2009gl Su, Y. J., J. M. Retterer, O. de La Beaujardière, W. J. Burke, P. A. Roddy, R.F.PfaffJr.,G.R.Wilson,andD.E.Hunton(2009),Assimilative modeling of equatorial plasma depletions observed by C/NOFS, Geophys. Res. Lett., 36, L00C02, doi: /2009gl Trigunait, A., M. Parrot, S. Pulinets, and F. Li (2004), Variations of the ionospheric electron density during the Bhuj seismic event, Ann. Geophys., 22(12), , doi: /angeo Willis, P., C. Jayles, and Y. Bar Sever (2006), DORIS, from altimetric missions orbit determination to geodesy, C. R. Geosci., 338(14 15), , doi: /j.crte Yeh, K. C., and C. H. Liu (1977), Diagnosis of the turbulent state of ionospheric plasma by propagation methods, Radio Sci., 12, , doi: /rs012i006p P. A. Bernhardt and C. L. Siefring, Plasma Physics Division, Naval Research Laboratory, Washington, DC 20375, USA. (siefring@ccs.nrl. navy.mil) I. J. Galysh and D. E. Koch, Space Systems Development Department, Naval Research Laboratory, Washington, DC 20375, USA. 12 of 12