Operational impacts of space weather



[1] Space weather events on the Sun, such as coronal mass ejections and solar flares, can lead to a worldwide disturbance of the geomagnetic field and associated ionospheric and thermospheric disturbances. These events can, and do, impact the performance and reliability of space-based and ground-based operational systems. This paper presents specific examples of the operational impact of space weather events on space surveillance systems. The paper concentrates on the 14 July 2000 event during which the solar-terrestrial environment experienced its largest disturbance in the past 11 years. We report on effects that were detected during the July storm with the Space-Based Visible (SBV) sensor, a visible-band electro-optical system that operates as a contributing sensor for the U. S. Space Surveillance Network. We also discuss the impact of this space weather event on the Global Positioning System (GPS) and on satellite tracking observations by ground-based radars.

1. Introduction

[2] Space weather is defined as conditions on the Sun and in the solar wind, magnetosphere, ionosphere, and thermosphere that can influence the performance and reliability of space-based and ground-based technological systems [The National Space Weather Program, 2001]. The conditions referred to are solar flares, coronal mass ejections (CME), magnetospheric storms and substorms, and the low altitude effects of those disturbances (auroral activity, ionospheric scintillation and total electron content (TEC) variations, ionospheric joule heating, etc.). In this paper we assume the reader is familiar with these disturbances and their general effects on near-Earth space. Our goal is to document examples of space weather effects on operational space-based and ground-based space surveillance systems. We also wish to encourage users of these assets to utilize the current solar cycle to further understand and investigate the effects of space weather on current and future surveillance systems. This paper is concerned mainly with effects related to the July 14, 2000 solar flare and the associated solar proton event and geomagnetic storm. This was one of the largest solar-terrestrial disturbances in the past 11 years. We examine effects on space-based systems, specifically the Space-Based Visible (SBV) sensor on board the Midcourse Space Experiment (MSX) satellite, and then on ground-based systems, specifically satellite tracking radars and GPS receivers.

2. The SBV Sensor

[3] The SBV sensor is a visible band (0.3–0.9 μm) telescope on board the MSX spacecraft whose principal objective is to gather metric and photometric data on Earth-orbiting artificial satellites as an operational sensor for the US space surveillance network. The MSX was launched in April 1996 into a sun-synchronous polar orbit at an altitude of 898 km. The SBV sensor consists of a 15-cm aperture, high straylight rejection telescope and a focal plane populated with 4 abutted 420 × 420 pixel Lincoln Laboratory CCDs. The SBV is a staring sensor, which detects satellites as streaks against the star background. In addition to the optics, a signal processor (SP) and supporting electronics are contained onboard the spacecraft. The SP extracts the metric and photometric information on the stars and streaks. The SBV Processing Operations and Control Center at MIT Lincoln Laboratory correlates the observed streaks with the known catalog of Earth-orbiting satellites and determines the identity of each observed object [Stokes et al., 1998].

3. Space Weather Effects on the SBV

[4] We discuss three space weather effects on the SBV: 1) the South Atlantic Anomaly (SAA) [Spjeldvik and Rothwell, 1985], 2) transient effects from the July 2000 solar proton event, and 3) potential solar-cycle effects related to the variable radiation belt proton population. The MSX altitude is high enough that the spacecraft spends a portion of its time within the SAA. Energetic protons (E > 10 MeV) penetrate the telescope housing and impact the focal plane. These particles deposit energy in the CCD pixels producing streaks of varying lengths. These streaks are not satellite targets, but they mimic them, and are termed false streaks (the satellite streaks are termed valid streaks). While the MSX is within the SAA, the number of false streaks tends to overwhelm the SP and inhibit data collection. Therefore the SBV does not collect data while in the SAA. Figure 1 illustrates the degradation of the SBV data in the SAA.

Figure 1.

Example raw SBV frameset (left) and a frameset showing the effects of the SAA on the sensor (right). The arrow on the left denotes a satellite streak.

[5] Secondly, transient effects, such as solar proton events, can occasionally mimic the conditions in the SAA, creating high radiation environments over the polar caps. The July 14, 2000 solar flare initiated a solar proton event that raised the level of >100 MeV protons at geosynchronous orbit by a factor of >104 [NOAA Space Environment Center, 2000]. The effects of the solar proton event on the SBV are illustrated in Figure 2. The figure displays the average number of valid and false streaks detected per frameset for each data collection event (DCE; cf. Stokes et al., 1998) over a time frame of 10 days along with the GOES-8 geosynchronous energetic proton data for comparison. This time frame covers the period before and after the July 14 event. A frameset is made up of 8 consecutive frames of data, each of which is a 1.6-second exposure. The averages are segregated by MSX sub-latitude representing data collected over the poles (∣λ∣ > 60°) and over the equatorial regions (∣λ∣ < 60°). Vertical lines indicate the beginning of the flare and of the subsequent magnetic storm. There is a data gap that is due to tests of new processing algorithms. No data were collected when the MSX was over the equatorial regions during these tests; therefore the equatorial/polar comparison was not made. The SBV was unable to obtain data during the first DCE following the solar flare, and observed a significant decrease (increase) in the number of valid (false) streaks observed over the polar regions for the next 2 days. In the second DCE after the flare (Day 197, 01:32 UT in Figure 2), we returned to our previous operational procedures; however, the signal processor had difficulty detecting any streaks whatsoever. Over the polar regions, the signal processor had difficulty detecting stars. This is an unusual occurrence, and we investigated the health and status of the SBV sensor, it's supporting electronics, software, and the status of the spacecraft-to-ground links. We found no evidence that the lack of streaks and stars in the signal processor reports was due to a failure in the sensor or it's supporting systems and software. Since the sensor behaves in a similar fashion while in the SAA (cf. Figure 1), we interpret the data as a solar-proton induced effect. The near equatorial observations were apparently unaffected by the influx of solar protons. These data are consistent with an enhanced radiation environment over the poles, which is typical for strong solar proton events [Stassinopoulos and Raymond, 1988]. We conclude that the influx of solar protons degraded the sensor performance in the polar regions during this large solar proton event.

Figure 2.

Average number of valid (middle) and false (bottom) streaks per frameset: 8–18 July 2000 for the SBV sensor illustrating the effects of solar energetic protons. GOES-8 energetic proton flux data (top) is presented for comparison.

[6] We now discuss long-term effects on the performance of the SBV. In 1998, we noticed that a relatively large region south of the SAA appeared to have a reduced incidence of valid streaks and an increased incidence of false streaks. The number of objects detected by the sensor was reduced and this reduction was concentrated in MSX sub-latitude to a region south of the SAA, just above the Antarctic coast. The region of appears to follow lines of constant magnetic latitude centered about roughly −60° and extending north and south of −60° by a few degrees. The region extends from about −30° to 120° east magnetic longitude. We investigated the geographic distribution of valid and false streaks during the year 2000 and determined that this region of degraded performance had persisted for ∼2 years (cf. Figure 3); we are undertaking a more detailed examination of the data to determine whether the feature exhibits any substantial variability. The region appears to be correlated with an enhancement of trapped energetic protons that has been observed by the NOAA POES MEPED sensors (cf. Figure 3). The POES satellites reside in a ∼800–850 km sun-synchronous orbit similar to that of MSX and measure roughly the same environment that the MSX encounters. The trapped proton enhancement appears to map back to the equatorial plane at L > 4 [NOAA Space Environment Center, 2000] and we speculate that it is a solar-cycle related enhancement. We believe this to be the first observation of this trapped radiation region by a visible-band CCD surveillance sensor. The question of interest to operators/designers is whether or not this energetic proton enhancement is permanent or variable feature, and how it is characterized in terms of extent, and proton flux.

Figure 3.

The left plot shows the average number of false streaks per frameset for days 255–310 of 2000. Note the region of relatively high occurrence of false streaks south of the SAA. There are no data in the white regions. The “M” pattern is an artifact of the limited time span of the data, and the operational usage of the SBV sensor. For comparison, on the right is a one year baseline plot of E > 7 MeV proton counts from the POES MEPED 90° detectors (from http://sec.noaa.gov/tiger/index.html) showing trapped protons. Note the enhanced region south of the SAA.

4. Effects on Ground-Based Systems

[7] Ground-based surveillance radars and GPS receivers receive electromagnetic radiation that has propagated through the ionosphere. The ionosphere is a dispersive medium; its index of refraction is a function of both frequency and the total electron content (TEC). The ionosphere introduces range errors because the electromagnetic (EM) waves have propagated through it at speeds less than that of light in a vacuum. The ionospheric range errors are frequency and TEC dependent, and the TEC can vary with look angle, time of day, season, geomagnetic activity, and solar EUV flux level. At L-band (frequencies of 1–2 GHz), the range delay at zenith can be as large as 30 m at the equator. The range delay at low elevations is roughly 3 times this number. The range delay is inversely proportional to the square of the frequency [Coster et al., 1992], so that lower frequency radars experience greater range delay. This affects the accuracy of orbit determination using radar data. Therefore, the ionosphere is a major source of error for GPS users and satellite tracking radars (e.g., the FPS-85 UHF radar at Eglin AFB, FL, and the Millstone Hill L-band tracking radar in Westford, MA).

[8] Ionospheric scintillation - ionization irregularities in the E and F region ionosphere - also causes difficulty for radars and GPS receivers because the EM waves suffer refraction and diffraction effects when passing through the disturbed regions [Basu et al., 1985]. For radars, this can result in degradation of the coherent integration capabilities of the radar and thereby can reduce the signal-to-noise ratio at which targets are detected. This effectively degrades the sensitivity of the radar. For GPS users, scintillation can lead to loss of lock by the receivers on the satellites.

[9] Some GPS receivers can be used to derive the TEC [Coster et al., 1992; Komjathy, 1997; Mannucci et al., 1998]. The July 15, 2000 magnetic storm associated with the July 14, 2000 solar flare caused significant disturbances in the TEC as measured by GPS receivers at four sites in New England, and at one site at Eglin AFB in Florida (cf. Figure 4). Note also the loss of lock (lack of data) of some of the New England GPS receivers near 22 hours after the start of the day. The loss of lock at the Millstone radar site in Westford, MA, has also been correlated with enhanced westward ion velocities and with disturbed atmospheric electric fields [Coster et al., 2001]. Typically these effects are correlated with scintillation activity. The loss of lock at Millstone also occurred at roughly the same time as the large TEC variations observed at Eglin [Coster et al., 2001].

Figure 4.

Time history of ionospheric TEC during the 15 July 2000 magnetic storm derived from GPS receiver measurements. Note loss of lock at Millstone (cyan), Westford (pink), and enhancements at Eglin (orange).

[10] Figure 5 shows range residuals (measured range – true range) for a track of a radar calibration sphere during the July 2000 magnetic storm. This plot illustrates the degraded metric accuracy of radars (especially the UHF radars) within the Space Surveillance Network during geomagnetic storms. Typically, the range residuals are about 0 ± 30 m, as shown in the top plot. During the geomagnetic storm, the residuals were significantly higher, up to 350 m. Models exist that facilitate correction of measured range for ionospheric effects [e.g., Coster et al., 1992] and it is worth noting that these residuals have been corrected for ionospheric effects using a GPS-based real-time ionospheric model [Coster et al., 1992]. The total ionospheric range delay on this day was estimated to be well over 1 km at low elevations. This illustrates the limitations of basing an ionospheric model on observations from a single GPS receiver during time periods of steep ionospheric gradients.

Figure 5.

Range residuals on calibration sphere 7646 from the FPS-85 UHF radar at Eglin AFB, Florida during the 15 July 2000 magnetic storm showing degradation of capability during ionospheric TEC enhancements.

[11] Finally, the difficulties in computing satellite orbits during periods of high geomagnetic activity are illustrated. Figure 6 shows a plot of what is termed the scale factor versus day of year 2000 for the MSX satellite discussed above. The scale factor is a least squares fit parameter that is used for each orbital arc and that scales the atmospheric drag model. It can be interpreted as an indication of the adequacy of the drag, and thus of the thermospheric model used to compute the drag [Gaposchkin and Coster, 1988]. If S is less than unity, then the atmospheric density predicted by the model is too large; if S is greater than unity, then the atmospheric density predicted by the model is too small. The data used in computing these orbits were collected from the S-band transponder system of the Air Force Satellite Control Network [Abbot, 2000]. At S-band, the ionospheric effect is always less than 250 m. As can be seen in Figure 5, the scale factor for 15–16 July 2000 was approximately 5, one of the largest scale factors ever computed for MSX.

Figure 6.

Scale factor as a function of day of year computed for AFSCN observations on the MSX satellite. This is a proxy illustration of the severity of atmospheric drag on the MSX satellite.

5. Summary

[12] We have provided several examples of the effects of space weather on operational space-based and ground-based space surveillance systems. The SBV sensor on board the MSX satellite is sensitive to the energetic (E > 10 MeV) proton environment. The SBV endures daily passage through the SAA. Transient events, such as strong solar proton events can degrade the performance of the SBV by creating enhanced proton environments in other regions of its' orbit. In addition, apparent long-term changes in the radiation environment also appear to affect the sensor, and we have reported the first observation of space environment effects on a visible-band space surveillance system. For ground-based radars, the ionosphere introduces range errors and angle errors due to variations in TEC and scintillation. These ionospheric variations can also affect GPS users. In general, geomagnetic storms make these errors larger and harder to model, which in turn makes it harder to correct the measurements for the errors. We believe that developing an understanding of the operational effects of space weather facilitates communication between scientists and operators and facilitates space weather research. In addition we believe an understanding the space environment by operators also enhances current operations as well as the development and operation of future space and ground-based space surveillance systems.


[13] This work sponsored by the Dept. of the Air Force under Air Force Contract No. F19628-00-C-0002. Opinions, interpretations, conclusions, and recommendations are those of the authors and are not necessarily endorsed by the United States Air Force.