Aeronomy of Ice in the Mesosphere receiver/communication lock analysis: When bad space weather is good
Abstract
[1] The Aeronomy of Ice in the Mesosphere (AIM) spacecraft (launched on 25 April 2007) is in low-Earth polar orbit. Some days after launch, AIM began to exhibit a problem in which it would not always achieve proper receiver uplink communications lock (termed “bitlock” throughout this paper). In this context, we examined solar conditions and geomagnetic activity to search for possible connections with periods of bitlock problems. We have found that higher solar wind speeds often lead to greater geomagnetic activity and this, in turn, seems to lead to improved AIM operations. Here we present analysis of AIM bitlock to show when relative improvements or diminutions in spacecraft operations have occurred. We conclude that the spacecraft bitlock problem is related, in part, to space environment conditions (along with a gradual secular trend toward lower performance). The best indicator of “good lock” state (of all we have evaluated) is associated with a shift from low (or quiet) geomagnetic and solar wind conditions toward more disturbed conditions.
1. Introduction
[2] It is increasingly understood and appreciated by the space operations community that complex spacecraft systems and subsystems can suffer degradation and even severe, abrupt anomalies while operating in the rigors of outer space. Many of the space environmental conditions or “space weather” factors that cause anomalies and/or catastrophic failures in space systems have been documented in the published literature [see Baker, 2002; Barbieri and Mahmot, 2004; Getley, 2004].
[3] The Aeronomy of Ice in the Mesosphere (AIM) spacecraft was launched on 25 April 2007. It was launched into a high-inclination (98°), low-altitude (600 km) sun-synchronous orbit to study and understand the origin and implications of Polar Mesospheric Clouds (PMCs) [Russell et al., 2009]. Shortly after its launch, however, AIM began to experience problems with its telemetry receiver [McCabe et al., 2008]. While studying the receiver bitlock problem, we found important relationships with the concurrent space weather conditions. This paper shows our analyses in this regard based on the first 1.5 years of AIM operations data. We conclude that there is an obvious secular degradation in the AIM receiver system, but there also is a clear modulation of the quality of bitlock performance that is apparently related to solar wind and geomagnetic conditions.
2. Background
[4] Past studies have shown that space weather impacts on spacecraft and on other human technology (such as aircraft and ground-based systems) often follow a pattern that is related to the 11-year cycle of solar activity. Figure 1 shows the annual smoothed sunspot number (SSN) for the period 1985 through 2007. Such data show the obvious maximum of solar activity that occurred in 1989–1991 for Solar Cycle 22. Similarly, sunspot numbers reached another maximum in the years 2000–2002 for Solar Cycle 23. At present (during the years 2007–2008) we are at the very minimum of Cycle 23 and are awaiting the onset of a new increase in sunspots and a return to increased solar activity of Cycle 24.
[5] As illustrated in Figure 1, there are specific features of solar and solar wind behavior in each solar cycle that generally are associated with particular space weather concerns. At sunspot maximum, the sun tends to produce strong, but episodic, outbursts of plasma, embedded magnetic field, and energetic particles called coronal mass ejections (CMEs). These CMEs, if emitted in the direction of the Earth and its magnetosphere, can cause the largest and most intense geomagnetic storms [see Gosling, 1993]. Large geomagnetic storms can, in turn, cause a variety of problems in technological systems including surface charging onboard spacecraft, impacts to the Global Positioning System (GPS), and disruptive induced currents in long electrical power lines [e.g., Baker, 2002; Coster and Komjathy, 2008; Huttunen et al., 2008].
[6] The several years following each sunspot maximum is a period called the “declining phase” of solar activity. This is a time on the sun when there often are long-lived transequatorial coronal holes that give rise to powerful streams of high-speed solar wind [Feldman et al., 1978]. When such high-speed streams strike the Earth's magnetosphere, they can initiate a sequence of processes that lead to very large enhancements of relativistic electrons in the Earth's radiation belts [Baker et al., 1986; Li et al., 2006]. Relativistic electrons in the magnetosphere have clearly been associated with deep dielectric charging and concomitant operational anomalies in a wide variety of spacecraft systems [e.g., Vampola, 1987].
[7] In many ways, the most benign space weather period is around solar minimum when sunspots are at their lowest number. During this quiescent period, there tend to be few strong CME events and geomagnetic activity is also generally low. Past evidence has further suggested that solar coronal holes and associated high-speed solar wind streams are not very prevalent near sunspot minimum. Thus, during this phase of solar activity, we would expect space weather effects in operating spacecraft to be very minimal. AIM has operated entirely during this absolute solar minimum time period. However, the current solar minimum has been unusually deep and long lasting. As we will show below, there have also been unusually prominent, recurrent solar wind streams that were not fully expected in this phase of the solar cycle.
3. AIM Receiver Link History
[8] The AIM science mission is to study Polar Mesospheric Clouds that form in the upper part of the Earth's polar region mesosphere over an ∼100-day period centered around the time of the summer solstice in each hemisphere. AIM seeks to address the questions of how these clouds form and why they vary. AIM is the ninth small-class (SMEX) mission under NASA's Explorer Program.
[9] After launch, the commissioning of the spacecraft proceeded nominally through attaining normal pointing mode. Nine days after launch, however, the satellite failed to achieve “bitlock” which is the condition of locking on the command uplink subcarrier modulation. This was the beginning of the intermittent operation of the transceiver that has continued ever since. The downlinks for the AIM system occur anonymously and are always successful, so acquired science and state-of-health data can always be transmitted. The issue is the upload of commands to the spacecraft. Over the next several weeks many different uplink configurations were tested to characterize the performance of the receiver [McCabe et al., 2008].
[10] On passes when the spacecraft achieved bitlock nominally at the start of the pass, tests were performed to examine the sensitivity to center frequency of the subcarrier. Ground testing showed that the subcarrier link could be achieved at 16 kHz ± 90 Hz. These tests were run several times and showed the threshold remained consistent. On passes where bitlock could not be achieved, tests were run that extended the range of subcarrier frequencies. Subcarrier frequencies were stepped away from the nominal value without finding any shift in the frequency where the receiver would attain bitlock. Tests were also run to vary the main carrier frequency to search for a shifted window. During passes with no bitlock, a wide range of frequencies was searched, without ever finding one through which lock could be achieved [McCabe et al., 2008].
[11] Many other tests were preformed using both ground stations and the NASA TDRSS spacecraft. None of these configuration changes resulted in improved performance. Figure 2 illustrates how the uplink performance has degraded since launch. Not having found a reliable way to get commands into the satellite, the Flight Operations Team realized that there might be a day when the receiver would not allow for any commanding, and as a result a new operations concept was needed. The end goal of the new operations concept is to be able to continue to return science data even if the ability to uplink commands were to be totally lost [McCabe et al., 2008]. This new mode has been fully tested and implemented.
[12] As shown in Figure 2, the daily values of bitlock (the percentage of orbits per day when good bitlock was attained) has varied widely and rather dramatically. To be specific, AIM was launched on 25 April 2007 (day of year (DOY) 115). For the first nine days there was nearly 100% bitlock. From DOY 125 onward there has been a general downward trend in the daily bitlock, with many instances of two or more days in a row without any successful receiver uplink locks.
[13] Another feature in addition to the secular trend described above is that the AIM spacecraft had an extended outage of operation during February 2008. During this time (after day 400 in the extended day of year (2007) system used in Figure 2) the spacecraft was in “safe hold” mode for nearly two weeks and no communication was achieved. This interval is marked in the figure (and is not used in our analyses presented here).
[14] Because of the fluctuations in the day-to-day percentage of bitlock, we have explored different approaches to smoothing the data. In the end, we adopted the 7-day smoothing procedure used by the AIM flight operations team [McCabe et al., 2008]. On the basis of our analyses suggesting a 27-day (solar rotation) periodicity in the solar wind and magnetospheric activity, this smoothing is appropriate. Figure 3 shows the smoothed percentage bitlock as a function of time from early May 2007 to the end of November 2008.
[16] We point out, however, that the smoothed bitlock data shows a rather orderly variation around the trend line. In fact, it tends to show a 27-day oscillation period that is reminiscent of a driver varying at the synodic solar rotation period. In Figure 4, we have subtracted the least squares trend from the bitlock data. This shows more clearly the large amplitude oscillations around the detrended zero reference level. This is especially evident after about DOY = 180.
[17] The results of our analysis suggest that the AIM receiver probably went through three “phases” in its lifetime (so far): (1) For the first 9 days (not shown here), the receiver was perfectly normal and nominal in its performance (100% bitlock); (2) For the period of the next ∼2 months (125 ≤ DOY ≤ 180), the receiver exhibited near-normal to nearly nonexistent bitlock; and (3) From DOY ∼ 180 to present, the AIM receiver has been showing a general secular degradation of bitlock, but with a systematic, rather orderly oscillation of high-to-low uplink success that is controlled by some other (probably external) factor. We have been quite interested in assessing what this factor is (or these factors are) in the possible context of space weather.
4. Dependences on Solar Wind/Geomagnetic Conditions
[18] The overall downward trend in the receiver performance seen in Figures 2 and 3 is undeniable. A reasonable question is whether this is due only to receiver degradation or whether other, external, factors are playing a role. We have examined several solar wind and geomagnetic parameters to see what underlies the changes and trends.
[19] Figure 5 shows as the red curve the detrended percentage bitlock (7-day smoothing) repeated from Figure 4. The blue curve is the concurrently measured solar wind speed (also smoothed by a 7-day running average). We see that the 7-day smoothing of the VSW data produces a very prominent and clear pattern of 27-day peaks in the solar wind speed. Note that these peaks in VSW often show an anticorrelation (or nearly so) with the smaller amplitude bitlock peaks. The overall correlation coefficient for this interval is a very weak value of r = 0.099.
[20] Solar wind variability would obviously be expected to drive geomagnetic variability. In Figure 6 we show a comparison of the detrended percentage bitlock (as in Figure 5), but in this case we have plotted the concurrent values of the geomagnetic storm index, Dst. (We have again done a 7-day smoothing of the Dst values). We see in Dst a pattern very similar to that in VSW (Figure 5): Geomagnetic activity increased (i.e., Dst became more negative) on a regular 27-day cadence. (By comparing Figure 5 with Figure 6, one can see that the Dst decreases occur when VSW is highest). The correlation coefficient is r = 0.129.
[21] Although not shown with figures here, we have examined several individual and composite solar wind and interplanetary magnetic field (IMF) parameters, in addition to the VSW values shown in Figure 5. We have studied the IMF BZ component, the IMF BS (rectified southward BZ) component, as well as the product VBS and the “epsilon” coupling parameter (ɛ = VB2sin4(θ/2)). The composite parameters (VBS and ɛ) show relationships to AIM bitlock quite similar to the VSW results shown here in Figure 5.
[22] We have also looked at several geomagnetic activity indicators in addition to Dst. These have included the NOAA planetary “A” index (Ap), the auroral electrojet index (AE), and the planetary “K” index (KP). On the scale of daily values and 7-day smoothing, all of these parameters show relationships to the percentage bitlock that are similar to what we show here for Dst.
[23] Given the suggestions of correlations between bitlock and solar wind and geomagnetic properties, we have expanded the time scales for the data overplots. Figure 7 shows DOYs 500–650, comparing bitlock and VSW. This more detailed plot hints at an anticorrelation between the parameters, but the linear correlation coefficient for the interval is r = 0.110.
[24] Figure 8 shows a more detailed plot for bitlock compared to smoothed Dst for DOY 200–400. Here we tend to see direct correlations between the curves (since Dst being more negative means increased geomagnetic activity). However, we again must conclude that the overall correlations are weak with r = 0.138 for the interval shown.
5. Role of Changing Activity
[25] Since there tend to be some indications that solar wind and corresponding geomagnetic changes play a role related to AIM bitlock, we have looked explicitly at time derivatives of various parameters. Figure 9 shows one of our key findings. The plot shows detrended bitlock (smoothed) for an extended period of time (DOY = 260–400). The blue curve shows the time derivative of the daily Dst value. In this case we have again taken the daily values of d(Dst)/dt and smoothed with a 7-day running average.
[26] From Figure 9, we see that in essentially all cases, the values of BL tend to go from lower to higher relative values when d(Dst)/dt has gone from positive values toward more negative values. What this means, in general, is that the bitlock condition tends to improve when Dst becomes more negative, i.e., when geomagnetic activity increases. Thus, in a sense, this says that “bad” space weather is “good”! Conversely, bitlock conditions tend to become less favorable when geomagnetic activity changes toward becoming less disturbed.
6. Discussion and Conclusions
[27] We surmise from analysis of over 1.5 year's worth of AIM operational data that the daily percentage of receiver bitlock has generally tended to decline from about nine days after spacecraft launch to the present time. The most recent run of data shows a flattening out of the daily bitlock value at about 6 ± 5% per day. Other than the safe hold period in February 2008, this situation has persisted since the end of 2007. Neither the science team nor the operations team (nor the spacecraft engineering team or manufacturer) knows for sure the underlying physical cause of the AIM receiver problem.
[28] From the analysis presented here, we can assert that some aspect, or aspects, of solar wind and/or geomagnetic activity clearly vary in concert with the ups and downs of AIM bitlock. This indicates that some aspect of the space environment or space weather is playing a role in the onboard receiver's performance. As we have shown here, the best predictor of a “good bitlock” state is a shift from low (or more quiet) geomagnetic conditions toward more disturbed conditions. This has led us to speculate that during more disturbed conditions the AIM spacecraft (and its subsystems) are being subjected to greater fluxes of energetic particles (e.g., energetic electrons in the auroral and subauroral zones) during each orbital pass. Such increased irradiation could cause a surface (or more probably, a subsurface) charging in or near the spacecraft receiver.
[29] A failure mode that could result in the observed symptoms is attributed to an inadvertent floating input on one of the 4000 series complementary metal-oxide-semiconductor (CMOS) logic devices used in the bitlock circuits in the spacecraft receiver. In this device a MOSFET transistor on the CMOS input looks like a small capacitance (on the order of a few pF). The circuit's output logic level is determined by the voltage at the input to the gate and the gate voltage is determined by the charge on the gate's equivalent capacitance. If the gate is floating (i.e., not connected to a driving input circuit as would happen if the connection is broken) the logic output level is indeterminate and would vary depending on the charge stored on the input capacitance of the gate.
[30] Figure 10 shows an equivalent circuit for such a logic circuit. If the leakage current in the input diodes is balanced, the gate will float near its threshold resulting in an output that will sometimes be a high- or low-state dependent on the state of charge on the MOSFET input capacitance. Our speculation is that this circuit might be well balanced and that small changes in accumulated charge, due to incident electron flux, can change the output state of the logic circuit from a level that prevents bitlock to a level that allows bitlock to occur.
[31] We cannot say with absolute confidence that the above scenario is the correct explanation. However, this does fit the facts as we know them. We can say with considerable certainty that there have been many specific instances in the past months where geomagnetic conditions became very, very quiet for several days on end. During such times there were cases where bitlock was not achieved for over a week or ten days. Then, when geomagnetic activity (measured by Ap or Dst) picked up, bitlock was rather remarkably reestablished [Ryan et al., 2007].
[32] In order to try to assist the Flight Operations Team, we have used the space weather forecasting models that we have developed within the Center for Integrated Space Weather Modeling (CISM) (see CISM website http://www.bu.edu/cism). In this CISM approach, we have used the modeling tools to examine solar conditions and to issue forecasts several days in advance about likely subsequent geomagnetic activity [see Baker et al., 2004]. The CISM Forecast Model (now running at the NOAA Space Weather Prediction Center in a quasi-operational mode) can predict when propitious conditions should be available for AIM operations. We expect to continue to use such operational tools to optimize AIM performance.
Acknowledgments
[33] We thank the reviewers for very helpful comments and suggestions. We also thank the AIM Flight Operations Team for their efforts. This work was supported in part by the National Science Foundation Center for Integrated Space Weather Modeling (CISM).
