Solar Particle Events and the International Space Station
2.1 BACKGROUND TO AN ASSESSMENT OF SPE IMPACTS ON ISS CONSTRUCTION
Solar storms often accelerate ions to energies that can penetrate space suits and even spacecraft. Occasions when this occurs are often called solar particle events (SPEs) but may also be called solar cosmic ray events, solar proton events, solar energetic particle events, energetic storm particle events, ground-level events, proton showers, or polar cap absorption events. Many of these terms, which are for most purposes synonymous, are still in use in the scientific literature. In this report, CSSP/CSTR consistently used "solar particle event" and its abbreviation, SPE.
NOAA's Space Environment Center (SEC) declares an SPE to be in progress when the dose rate of particles with energies above 10 MeV (i.e., space-suit-penetrating) exceeds 10 particles cm–2s–1sr–1 (directional flux) for more than 15 minutes. When this happens, SEC alerts the Space Radiation Analysis Group (SRAG) at the Johnson Space Center (JSC), recalling SRAG to Mission Control Center (MCC). Should the dose rate of particles with energies above 100 MeV exceed 1 particle cm–2s–1sr–1, an "energetic SPE" is declared, which mandates that SRAG remain on console. Such events can last several days.
Two criteria can be used to define when a flight is significantly impacted by an SPE. Criterion 1 says that a flight is impacted if an SPE occurs that reaches the "alert" (10 particles cm–2s–1sr–1) stage. Criterion 2—the "significant dose" criterion—says that it is impacted if an SPE occurs with an accumulated free-space dose of 108 particles cm–2 (omnidirectional fluence) and energies above 10 MeV. Converted into a dose to tissue, this corresponds to about 0.6 Gy (see Section 1.3 for definitions of radiation units),1 the actual number depending on the energy spectrum of the SPE particles. The second, more stringent criterion marks a condition that in the worst-case orbit geometry relative to the SPE zones and worst-case EVA timing, would noticeably increase an astronaut's radiation dose. Criterion 1 can be met without criterion 2 being met, but rarely if ever can the reverse happen. When criterion 1 is met, the flight's ground support personnel must be placed in a state of radiation alert, affecting decisions on when to launch, which astronauts to assign to which tasks, EVA schedules, and when to return. When criterion 2 is met, the repercussions for a flight and its crew are more serious, because a radiation condition has occurred that could bring an astronaut significantly closer to the specified limits for personal radiation exposure.
The science of SPE phenomenology, upon which radiation risk assessments like those described below are based, started over 50 years ago. It was an offshoot of cosmic ray research, which began even earlier, after Victor Hess demonstrated that there was a constant flux of ionizing radiation entering the atmosphere from space. In 1925, Robert A. Millikan dubbed this radiation "cosmic rays." The study of cosmic rays blossomed in the early
1930s, when the invention of the ionization chamber allowed more sensitive experiments. Now we know that there are two kinds of cosmic rays, galactic and solar. Galactic cosmic rays (GCRs) are always present, although their intensity varies with solar activity. Solar cosmic rays are present only during very intense SPEs.
Scientists have recorded SPEs indirectly from ground observations since 1942 and directly from spacecraft since 1965. Between 1942 and 1953, the only way to detect SPEs was with ground-based instruments (ionization chambers and muon counters) designed to monitor the intensity of galactic cosmic rays. At that time only particles with extremely high energies (>4 GeV), high enough to penetrate Earth's magnetic shield to the top of the atmosphere, were detectable. As these cosmic rays passed through the atmosphere, they generated a nuclear cascade intense enough that some of the secondaries (primarily muons) reached the cosmic-ray recording instruments on the ground. When an SPE also brought such high-energy particles to Earth, the instruments recorded a transient rise in the count rate above the background set by galactic cosmic rays. From such signals, cosmic ray physicists estimated the dose rate (flux) and total dose (fluence) of the highest energy particles in the SPE. The early 1950s saw the development and deployment of cosmic-ray neutron monitors, which could observe particles with energies nearly an order of magnitude lower than those detected by muon counters or ionization chambers. The SPEs of solar cycle 19 (1954-1965), one of which (February 11, 1956) is famous for its high intensities at high energies, were recorded with neutron monitors.
Before 1957, cosmic ray physicists could infer the occurrence of large fluxes of lower energy particles (<100 MeV) only from the extra ionization they produced in the upper atmosphere. Such ionization absorbed radio signals, thus producing polar cap blackouts, which had (limited) use as a quantitative measure of particle flux or fluence. Since 1957, however, radio techniques have been developed that can measure the flux of lower energy particles through their ionospheric effects.2 More important, the advent of the space age enabled direct observations of solar proton events at lower energies than can be monitored from the ground. Spacecraft studies of SPEs started in the early 1960s, with Explorer 12 in 1961, Explorer 14 in 1962, and Interplanetary Monitoring Program (IMP) 1 in 1963. Routine satellite measurements of SPEs, with good data coverage, started in 1965 and have continued almost uninterrupted over the intervening decades. At present, the NOAA geostationary operational environmental satellites (GOES) are responsible for supplying real-time measurements of energetic particles to SEC for SPE monitoring, while NOAA's National Geophysics Data Center archives SPE data for research. The resulting database, combined with data from heliospheric probes and from ground-based and balloon-borne instruments, has enabled studies that have greatly increased our familiarity with and understanding of SPEs.
Most research on SPEs has focused on characterizing individual events and, as data increased, on statistical studies of those characteristics. The current state of models for SPEs is assessed in Appendix A, in the context of their potential for contributing to the risk assessment effort. Although these studies have not yet revealed how particles are accelerated to relativistic energies or how to predict the flux or fluence from an individual solar event, they are useful for making statistically realistic estimates of the likelihood that SPEs might impact ISS construction.
2.2 PROBABILITY OF SPE IMPACT ON ISS CONSTRUCTION
Turner and Baker3 first estimated the likelihood that SPEs will impact ISS construction using the significant-dose criterion (criterion 2, defined above). They statistically analyzed SPE data from the last four solar cycles to determine the average frequency of occurrence of class 2 SPEs (those capable of satisfying criterion 2) as a function of time from the solar minimum, then tabulated the intervals between the last solar minimum (near the beginning of 1997) and the times of the ISS construction flights, as specified in the then-current manifest, from June 1998 through June 2002. Assuming that the current solar cycle, cycle 23, will resemble the average of the previous four, they found that 11±6 class 2 SPEs would occur from June 1998 through June 2002. This means that between two and four class 2 SPEs probably would occur during an ISS construction flight. This result is easy to comprehend given that, according to Turner and Baker's statistics, the average probability that a class 2 SPE will occur during an arbitrary 2-week ISS construction mission is about 10 percent, and 34 shuttle missions are scheduled.
Turner and Baker's study raises an important question: Is it appropriate to apply statistics derived from averaging the previous four solar cycles (19 through 22) to the current solar cycle, considering the well-documented
extent of intercycle variability? For example, the number of class 2 SPEs ranged from 10 to 30 over cycles 19 to 22. To address this question, NASA funded a workshop in late 1996 (near the end of cycle 22) to "assess [solar-cycle] prediction techniques and arrive at a reasoned consensus, including uncertainty, on how [cycle 23] will develop."4 Because the record of sunspot numbers is longer than the record of SPEs, the workshop focused on predicting cycle 23's sunspot number history as an indirect measure of its potential for generating solar storms. It rated each technique on its relative ability to predict past cycles, then constructed a consensus forecast by taking an accuracy-weighted average over all techniques. The resulting consensus forecast predicted that cycle 23 should strongly resemble cycle 22 in amplitude and should peak in the spring of 2000. The panel revisited the question a year later, after the new cycle had begun, and confirmed the validity of this conclusion. Therefore, since cycle 22 produced stronger SPEs than the average of the last four cycles, 19 to 22, Turner and Baker's calculations probably underestimate the SPE impact on ISS construction flights.
To quantify for this report an exclusive cycle-22 perspective on the problem, Turner and Kemere5 advanced cycle 22's SPE history forward 10 years to represent cycle 23. They used a cycle-22 SPE database compiled mainly from data archives at NOAA's National Geophysical Data Center (Margaret A. Shea, Air Force Research Laboratory, and Don F. Smart, Air Force Research Laboratory (ret.), personal communication). Turner and Kemere then superimposed the ISS-construction flight manifest (released in May 1998) on their ersatz history of cycle 23 SPEs and recorded the number of coincidences between class 2 SPEs and shuttle flights. To build a statistically robust result, they shifted the flight schedule forward by 2 weeks 47 times. For each iteration, they tallied the number of flights that would have experienced a criterion 2 SPE impact and expressed the result as a percentage of iterations per number of flights impacted. As shown graphically in Figure 2.1, 23 percent had one or two impacted flights, 47 percent had three or four impacted flights, 23 percent had five or six impacted flights, and 7 percent had seven or eight impacted flights. Note that all iterations had one or more impacted flights. These results quantify the expectation that the number of impacted flights is somewhat greater based on cycle 22 SPE history than on average SPE history.
For further verification of these predictions, Shea and Smart conducted an independent study similar to Turner and Kemere's except it used the SPE alert criterion (criterion 1) instead of the significant dose criterion (criterion 2) to identify flight-impacting SPEs. The criterion-1 study used the same (Shea and Smart) database as the Turner and Kemere criterion-2 study, projected the cycle 22 SPE history to derive an ersatz history for cycle 23, but assumed 10-day flights instead of 2-week flights. To build a statistically robust result, the hit count was iterated 48 times, each time advancing the flight schedule by 15 days. Figure 2.1 shows the results of the criterion 1 (Shea and Smart) and criterion 2 (Turner and Kemere) studies.
At face value, both curves in Figure 2.1 imply with near certainty that SPEs satisfying both criteria will impact at least two ISS construction flights. There is a 50 percent chance that significant-dose SPEs will impact five or more flights and that at least three additional flights will have an SPE alert. There is a 10 percent chance that significant-dose SPEs will impact eight or more flights and that at least four more flights will have an SPE alert. Although the exact number of impacted flights remains uncertain, it is inevitable that a flight director will have to respond to an SPE during a shuttle flight on more than one occasion, and probably on several occasions.
As is always the case with statistical predictions, there is a finite chance that the next solar cycle will produce considerably fewer SPEs than cycle 22. To make flight plan decisions based on this unrealistically sanguine hope, however, would be imprudent. In fact, if cycle 23 were not modeled on its predecessor, as was recommended by the solar physics community, the only reasonable alternative would be to evaluate a worst-case-cycle scenario, which would predict even higher risks of radiation. Radiation risk managers should plan, therefore, on significant-dose SPEs impacting more than one ISS construction flight.
2.3 CORRELATION BETWEEN SPES AND SIZE OF THE SPE ZONE
A major factor in determining whether a given SPE poses a threat to astronauts in ISS orbit is the size of the SPE zones to which SPE particles have access, a topic discussed in Sections 1.1 and 1.3. The angular width of the SPE zones is normally around 30 degrees, which poses little if any threat to astronauts in ISS orbit. When the zones widen to the order of 45 degrees, as can happen during a major geomagnetic storm, 40 percent of the ISS orbit falls within the SPE zones and the radiation threat becomes nonnegligible if an SPE is in progress. Clearly, there is a tendency for the SPE zones to widen during SPEs. The tendency is not perfect: SPEs can run their course wholly ignored by the SPE zones, and SPE zones can open when there are no SPE particles to ingest. Nonetheless, as CSSP/CSTR demonstrates, the SPE zones often widen during high-dose SPEs, thus statistically favoring the worst-case combination. SPEs do not physically widen the SPE zones; rather, one aspect of a solar storm generates SPEs while another separate but correlated aspect widens the SPE zones.
The particles in an SPE can come from either a solar flare or the shock wave driven by a coronal mass ejection (CME), with the latter being responsible for the largest total doses. The shock-accelerated SPE particles travel toward and away from the Sun along the interplanetary magnetic field (IMF) threading the solar wind, but a significant fraction are trapped near the propagating shock by wave-particle interactions.6 Thus, once the shock reaches Earth, the energetic proton flux can increase suddenly by as much as two orders of magnitude, making this shock spike the most dangerous portion of the solar particle event itself.
The SPE zones, on the other hand, are widened by the interaction between Earth's magnetosphere and the strong IMF associated with the CME-shock system. The spatial relations between the two sources of SPE particles and the regions of strong IMF accompanying a solar storm are illustrated in Figure 2.2. Consider, for simplicity, a situation in which Earth lies in the direct path of the oncoming. CME and shock. As noted above, some SPE particles arrive at Earth long before the shock, while the IMF still exhibits its normal strength and the SPE zones occupy their normal volume. Consequently, preshock SPE particles generally have minimal access to the orbit of ISS, unless an earlier geomagnetic storm widened the zones and is still in progress (a scenario characteristic of solar maxima and so not unlikely during the ISS construction period). Conditions change after the shock passes, for SPE particles are still present and so, too, is a strong and often turbulent IMF, which in many if not most cases induces a magnetic storm. The SPE zones then widen and fill with the solar energetic particles in the ambient solar wind.
From the perspective of radiation risk assessment, the double threat posed by high SPE fluence entering widened SPE zones is a flight director's nightmare. Aside from the amplifying effects of sequential solar storms, however, only a fraction of these events will combine the right elements to produce the worst-case scenario. These elements are primarily, though not exclusively, a fast CME launched near the central meridian of the Sun (the case discussed above), yielding the highest-dose SPEs and the biggest geomagnetic storms. Several situations can give rise to an SPE without a geomagnetic storm to widen the SPE zones: the IMF can point northward throughout the storm; the shock, CME, or both could miss Earth even though SPE particles reach it; or the SPE might come from a west-limb flare without an associated CME.
The tendency for major (Kp ~ 9) geomagnetic events to coincide with intense SPEs has been observed for individual storms, for example, those of July 1959, August 1972, October 1989, March 1991, and June 1991.8 It is also evident from a correlation analysis between SPE intensity as measured by peak dose rate and geomagnetic disturbance intensity as measured by a standard magnetic activity index, Kp. The Kp index is for geomagnetism what the Richter scale is for seismology, a logarithmic proxy measure of the energy of disturbance. This 3-hour-averaged planetary index, which ranges from 0 (extremely quiet) to 9 (extremely disturbed), is a single-parameter
measure of geomagnetic storm intensity that has been computed according to a uniformly applied algorithm continuously since 1932. It has become a standard against which space physicists like to correlate all manner of space weather variables. Geomagnetic conditions can be considered stormy when the Kp index exceeds 5, which happened 6.2 percent of the time during solar cycle 22. During class 1 SPEs (those capable of satisfying criterion 1 for an SPE event), however, Kp exceeded 5 nearly four times as often: that is, if an SPE is in progress, there is about a 24 percent chance that a geomagnetic storm is also occurring and that the SPE zones are dilated. For the most intense cases, the probability of a geoeffective storm coinciding with an SPE is still higher.
Since SPE zones widen during geomagnetic storms and Kp measures storm intensity, one can estimate SPE-zone width from observed Kp values. (There is, however, no database of SPE-zone widths that would allow us to eliminate the Kp proxy for them.) A Kp proxy for SPE-zone size can be obtained from an empirical relation that specifies in terms of Kp the angular size of a circle defined by the equatorward edge of the auroral oval, which is a rough indicator of the width of the SPE zones.9 The equatorward edge of the auroral oval delimits the volume of the magnetosphere in direct plasma contact with the magnetotail, a volume into which energetic SPE particles can considerably (though not always completely) penetrate.
Figure 2.3 shows 3-hour averages of the directional dose rate from particles with energies above 30 MeV (which penetrate all parts of a space suit) recorded in 1989, a year around the maximum of solar cycle 22, during which several intense SPEs occurred. The figure shows that the most intense fluxes tended to occur when the SPE zones were wide, although the tendency was not always observed. For example, there were high fluxes (above 103 particles cm–2s–1sr–1) when the width of the SPE zones was less than 30 degrees. These cases illustrate that the solar energetic particles get to Earth before the shock that brings the magnetic storm that opens the SPE zones. Nonetheless, a tendency for high fluxes to occur when SPE zones are wide is discernable. The box marks a danger
area where high dose rates combine with wide SPE zones to pose a hazard to astronauts in ISS orbit who are shielded only by a space suit. The nine points inside the danger area come from two storms during 1989. If ISS were inside the danger area while it was in the part of its diurnal cycle when it most deeply penetrates the SPE zones, it would spend a quarter to a third of each orbit exposed to SPE particles. During this time, an astronaut on a 6-hour EVA could receive a radiation dose between 10 and 100 percent of the short-term limits for eyes and skin.
2.4 SUMMARY AND RECOMMENDATION
CSSP/CSTR finds that a concentrated effort aimed at reducing SPE radiation risk to astronauts during ISS construction is needed. Based on the assumption—the best now available—that the radiation characteristics of the current solar cycle will resemble those of the last cycle, there is nearly a 100 percent chance that at least 2 out of 43 planned ISS construction flights will overlap a significant SPE and a 50 percent chance that at least 5 flights will overlap such an event. Moreover, whenever SPEs are in progress, the SPE zones show a marked tendency to widen over the polar latitudes reached by the ISS orbit, a tendency that becomes stronger with SPE severity. Two storms during 1989, near the maximum of the last solar cycle, illustrate the point. The SPE zones widened until they engulfed more than a quarter of the planned ISS orbit, while the radiation intensified enough to have pushed an astronaut over the short-term limit for skin and eyes during a single, ill-timed 6-hour EVA. CSSP/CSTR therefore recommends the rapid implementation of the following scientific elements of a program to reduce SPE radiation risk:
Recommendation 2: For real-time SPE risk management, carry out the steps needed to make usable by SEC and SRAG the models that use real-time data to specify the intensity of SPE particles and the geographical size and shape of the zones accessible to them.
This recommendation could be implemented early enough to have an impact on SPE radiation risk management during ISS construction; CSSP/CSTR views it as a high-priority item for action by NASA, NOAA, and the USAF. Chapter 4 and Appendix A of this report document other existing and potential resources that could contribute to implementing recommendation 2. Appendix A also addresses institutional issues related to developing the requisite modeling tools. Finally, chapter 5 includes a discussion of unresolved issues surrounding the ''transitioning" of research models to operational use.
Recommendation 1 in Chapter 1 is relevant to Recommendation 2, set forth above. For these recommendations to succeed, flight directors must become involved in assessing the effectiveness of the new or improved tools for SPE risk management that could allow liberalizing the current (unofficial) flight rule on radiation that says changes in flight plans can be based only on real-time, on-site data.