Radiation and the International Space Station: Recommendations to Reduce Risk (2000)

Solar particle events (SPEs), the subject of Chapter 2, are one of the two high-latitude radiation hazards to which astronauts on ISS construction crews can be exposed. The other is relativistic electrons—electrons with energies above 500 keV, which penetrate space suits—in the outer radiation belt. To assess the electron radiation hazard to ISS, CSSP/CSTR first clarified the extent to which the ISS orbit penetrates the volume of space where relativistic electrons occur and then considered how fluxes of relativistic electrons vary over time and on what they depend. It has thereby evolved empirical rules that can serve to trigger alerts and warnings of imminent increases in relativistic electrons and, after the increase occurs, to predict how the flux subsequently decreases. We begin with the relevant geometry.

The outer electron belt forms an Earth-circling torus, crescent-shaped in cross section, with the crescent concave earthward (see Figure 1.1). The apex of the heart of the outer belt extends roughly from 2.5 to 4 Earth radii (Re), although significant traces of the belt extend out to 10 Re. (1 Re is the unit used by radiation-belt physicists to represent distance; it equals 6,370 km.) Also, instead of referring to a crescent, they refer to L shells, where L is the distance (in Re) from the center of Earth to the apex of the crescent (or shell). Roughly speaking, an L shell is an Earth-circling surface generated by all magnetic field lines that cross the equator L Re from Earth's center. The crescent cross section of the radiation belts conforms to the dipolar geometry of the geomagnetic field.

L shells that define the outer radiation belt reach down to touch the atmosphere in each hemisphere in an annulus centered on the magnetic poles. In each hemisphere, the magnetic pole is offset from the geographic pole by 11.4 degrees. The heart of the annulus at the altitude of ISS extends from about 28 degrees to 31 degrees in angular width from its polar center. Since the geometrical considerations here are similar to those described in Section 1.3, we may use Figure 1.6, which gives the fraction of the ISS orbit exposed to the SPE zones for different zone sizes, to estimate the fraction of ISS orbit exposed to outer belt electrons. For this application, the appropriate zone size is about 33 degrees. (During highly relativistic events—discussed in Section 3.2—the outer radiation belt can extend down to 45 degrees, so by taking 33 degrees as a zone size, we are being conservative in the sense of not overestimating the seriousness of the situation.) Figure 1.6 shows that ISS spends between 0 and about 20 percent of its orbit in the relativistic electron annulus cycles every 24 hours. Although 20 percent seems like a small number compared with the 40 percent by which the SPE zones can engulf the ISS orbit during major geomagnetic storms, the accumulative effect can nonetheless be significant. Because ISS passes through the

relativistic electron annulus every day, even during geomagnetic calms, as far as the geometry of the relativistic electron belt goes, the threat seldom goes away. Having established that during a part of nearly every day ISS spends about 20 percent of each orbit in the relativistic electron annulus, we look next at the dose rate during this time.

The flux of relativistic electrons in the outer radiation belt varies over time by many orders of magnitude. The variation is marked by events, called highly relativistic electron (HRE) events, that have a characteristic life cycle, rising quickly (on the order of 1 day) and decaying slowly (on the order of 5 to 10 days), although some events decay anomalously even more slowly. The fact that the onset of HRE events is well correlated with changes in solar wind conditions offers an opportunity to develop a protocol for HRE-event alerts and warnings that could demonstrate sufficient predictive value to be of operational use. The decay rate, which is faster on L shells closer to Earth,¹ seems to be stable enough to use to predict fluxes days ahead. To illustrate this point, we look at several HRE events.

Among the several satellites that currently record fluxes of relativistic electrons in the outer radiation belt are the Solar Anomalous Magnetospheric Particle Explorer (SAMPEX) and POLAR, two NASA research satellites. SAMPEX, which is in a low-altitude, 96-minute polar orbit, measures fluxes of electrons between 500 keV and 50 MeV in the low-altitude horns of the outer radiation belt. Plate 1 shows a color spectrogram from SAMPEX data of the average flux of electrons between 2 and 6 MeV as a function of L shell and day-of-year in 1997. During 1997, SAMPEX recorded 14 outer-belt, energetic-electron events during which the daily flux of energetic electrons jumped three orders of magnitude to values exceeding 10⁴ electrons cm^–2 s^–1 sr^–1. The bottom panel of Plate 1 shows similar data from the POLAR satellite, whose highly elliptical orbit, ranging from 2 Re at perigee to 9 Re at apogee, covers L values from 2 to 100. The same events are seen in both data sets. Note that the integral flux of four of the events exceeded 10⁵ electrons cm^–2 s^–1 sr^–1. Unfortunately for operational considerations, data from SAMPEX and POLAR satellites are not available in real time. Perhaps, however, a combination of LEO polar satellites that return data once per orbit and geostationary satellites that return data continuously in real time can assess electron flux in the outer belt completely enough to serve for critical operational purposes.

Two NOAA programs provide real-time satellite data on the space environment, including data on energetic electrons. One of these is the Polar-Orbiting Operational Environmental Satellites (POES) program, which collects data from satellites in low-altitude, polar orbits that cover the outer radiation belt, providing integral fluxes of electrons with energies >300 keV. POES data are taken at 840 km altitude, not far from the nominal 400 km altitude of ISS, and so are readily convertible to fluxes relevant to EVAs at ISS. The conversion must reduce (usually by a factor of less than 2) the >300 keV fluxes that POES provides to fluxes of >500 keV electrons relevant to EVAs, and it must compensate for the different altitudes of the two orbits (fluxes at ISS are about two-thirds of those at POES). The fluxes of >500 keV electrons at ISS are thus about 30 percent of >300 keV electron fluxes that POES measures on the same L shell.

Figure 3.1 illustrates POES energetic electron measurements. The first and third panels show 12-hour running means of fluxes in the outer belt (L from 4.5 to 4.8 Re) throughout 1997. The maximum average flux was often above 10⁴ electrons cm^–2 s^–1 sr^–1, and on one pass it exceeded 10⁵ electrons cm^–2 s^–1 sr^–1. It can be seen that the storm enhancements in Figure 3.1 decay at a rate that is about the same from event to event, forming a basis for the prediction capability described in Section 3.3.

The second NOAA program that provides real-time satellite data on outer belt energetic electrons is GOES. Satellites in this program also measure fluxes of SPE ions, solar X-ray flux, and the magnetic field at L = 6.6 (geosynchronous orbit). Daily averages of >2 MeV electrons recorded on GOES are shown in the second and fourth panels of Figure 3.1. Since L = 6.6 is beyond the core of the outer radiation belt, GOES fluxes are generally less than the core values that POES record. In 1997, the maximum daily value from GOES was about 10³ electrons cm^–2 s^–1 sr^–1, compared with 10⁵ electrons cm^–2 s^–1 sr^–1 from POES. Nonetheless, it can be seen that GOES fluxes track POES outer belt fluxes, which means that GOES can serve as a continuous proxy monitor of the intensity of outer belt electrons.

Figure 3.1 Average energetic electron fluxes recorded on POES (NOAA-12) and GOES: January to June 1997 (top half) and July to December 1997 (bottom half).

The HRE events of 1997 shown in Figure 3.2 fall into two categories: those with a sharp rise and exponential decay to background levels over several days and those that stay elevated for many days. These two categories also pertain to HRE events recorded by other satellites, such as OGO 5,² DMSP,³ POLAR,⁴ and SAMPEX.⁵ Decay rates of several days agree with theoretical calculations using diffusion and atmospheric loss processes. Longer decay rates have been attributed to an additional acceleration of ambient plasma around L = 6.⁶

The rapid rise and exponential decay events recorded by NOAA-12, a POES, permit developing and testing a prediction model that can be applied to such events. Figure 3.2 gives NOAA-12 data for the January 1997 event seen in Figure 3.1. The points in the top panel show the intensity of radiation in the core of the outer belt. The points are fluxes averaged between L = 4.5 and 4.8 (the core of the outer belt) for each orbit during the event. To smooth the points so that an exponential curve can be fitted to the decay phase, the second panel gives a running mean over 49 passes through the outer belt, which covers about 12 hours and 13 orbits. (NOAA-12 typically passes though the outer belt four times during every orbit.) The bottom panel, which shows >2 MeV electron fluxes recorded by GOES-8, illustrates the ability of a geostationary satellite to continuously monitor the electron intensity during such an event.

Figure 3.2 Energetic electron fluxes recorded by NOAA-12 and GOES-8  during an HRE event in January 1997.

In the second panel of Figure 3.2, the line through the decay phase of the event is an empirical fit of the form J = J_oe^t/t, where J is the integral electron flux, t is time in days, and t (= 5.184) is the decay time in days. Others have reported a similar decay rate from analyses of similar data recorded on several different satellites.⁷ With the decay time set at 5.18 days, this procedure was applied to nine other rapid rise and exponential decay events captured in the NOAA-12 data for 1997. The results are shown in Table 3.1, with event A being the event taken here as the standard, the January event of Figure 3.2. The columns of numbers under each event give the percentages by which the measurements differ in absolute value from the predicted flux based on the standard exponential fit. As a rule, one can predict the average flux of energetic electrons up to 6 days after the onset of a major enhancement (day 1) before the percent deviation exceeds 100 percent. For particle fluxes that vary by four orders of magnitude, this degree of predictability is rather remarkable. It could probably be improved by taking account of the observed decrease in t for smaller L shells.⁸

To construct a worst-case scenario for assessing the risk astronauts face of being exposed to HRE events, CSSP/CSTR considered event C in Table 3.1, the extreme event of 1997, which was not an active year. The maximum running-average flux recorded between L = 4.5 and 4.8 was 1.3 ×10⁵ electrons cm^–2s^–1sr^–1. If an EVA of maximum duration (6 hours) were to coincide with the maximum duration (6 hours) of this event, the fluence of >300 keV electrons encountered by an astronaut in the NOAA-12 orbit would have been 1.5 ×10⁹ electrons cm^–2. Since the ISS inclination is less than the NOAA-12 inclination, the ISS spends about twice as long per orbit in the outer radiation belt and would experience a correspondingly larger fluence. Taking all factors into account, including the 30 percent reduction of NOAA-12 fluxes discussed earlier, the fluence from relativistic electrons

during an ill-timed EVA at the same time as this HRE event would have been about 10⁹ electrons cm^–2, which corresponds to a dose of about half a sievert⁹ (see Section 1.3 for definitions of radiation units). Several events during 1998 reached four times the dose rate of this 1997 event (J.B. Blake, Space Sciences Department, Aerospace Corp., personal communication, 1999). Although the shielding provided by a space suit will reduce this number by a factor of about two, the dose would nonetheless be great enough to force an astronaut over the short-term limit for skin and eyes.

Another threat posed by the high-energy electron environment is the electrical charging of solid objects in space, such as pieces of the ISS and the astronauts' space suits. This is probably not a health hazard, but it might result in an electrostatic discharge when an astronaut touches something. If that something is connected to an electrical circuit, the circuit might be damaged. Electrostatic discharges in the outer electron belt are a familiar source of problems for spacecraft circuits. Although such a threat falls outside the purview of this report, it is raised to illustrate the need for communication between NASA centers that deal with different aspects of the radiation problem. This is the subject of Chapter 6.

A strategy for developing an operational capability to predict high-energy electron events of concern to astronauts on the ISS should include the following elements:

• Real-Time Event Identification from GOES Data. An operational system to predict the flux of energetic electrons in the outer radiation belt using the scheme described above must monitor particle fluxes in a timely manner. The bottom panel of Figure 3.2 shows that the enhancements in GOES-8 electrons occur simultaneously with the enhancements recorded on NOAA-12. GOES data arrive at the Space Weather Operations Center (SWOC) of SEC within a few seconds of measurement. By contrast, NOAA-12, which spends only about 10 percent of its time in the outer belt, transmits its data to ground only once an orbit, or every 101.5 minutes.

• Routine Monitoring of Low-Altitude, Outer-Belt Electrons. Electrons exceeding 300 keV in the outer radiation belt recorded by the Space Environment Monitor on all POES should be routinely averaged and a running mean prepared, from which a regular 12-hour forecast can be made. This procedure would allow predictions for each L shell based on a decay rate appropriate to the shell.

• The Forecast Model. In general, until a realistic dynamic model can be developed, the following approach should yield useful results. If no event is in progress, persistence should dominate the forecast. If a rapid onset and exponential decay event resulting from reasonable pitch angle diffusion is in progress, then the equation described

in Section 3.3 (J = J_oe^t/τ) is a better choice than persistence. If a gradual decay event is in progress, then persistence again is the best option until a realistic model can be developed. Since the decay time is slow, the data from POES every 101.5 minutes are adequate to monitor progress of the decay of radiation fluxes to see if the prediction based on exponential decay is holding and to update the prediction.

The strategy described here applies to predicting the flux of relativistic electrons as it decays after the sudden rise (on the order of a day) that characterizes the onset of an HRE event. As mentioned earlier, the onset of HRE events is correlated with changes in solar wind conditions. A sudden, significant increase in solar wind speed (for example, more than 100 km/s in less than 1 day) and southward-pointing IMF are relatively good predictors of the onset of an HRE event.¹⁰ Such sudden, significant increases in solar wind speed are common at the leading fronts of fast solar wind streams, which might or might not be accompanied by a southward-pointing IMF. An upstream solar wind monitor can provide the input needed to implement a protocol for issuing alerts and warnings based on this correlation.

Nearly every day an ISS orbit transects the outer radiation belt, where relativistic electrons reside. When the length of transection is greatest, ISS spends about 20 percent of its time in the belt. During relativistic electron events, which happen on average about once per month, the intensity of relativistic electrons in the belt increases by as much as four orders of magnitude. During an ill-timed EVA, when the intensity of relativistic electrons is greatest, the dose received by an astronaut could be several tenths of a sievert, which could put him or her over a radiation dose limit. Procedures can be implemented to specify and forecast at least approximately the intensity of relativistic electrons in the outer belt. POES provide measurements of relativistic electron fluxes that are transferable to the ISS environment with reasonable accuracy. These measurements are available about every hour and a half. GOES provide relativistic electron measurements continuously, but they are not as directly transferable to the ISS orbit. They do, however, track POES measurements. Thus, in combination, POES and GOES measurements allow radiation risk managers to follow the variation of electron intensity in the outer belt. A crucial piece of hardware that the ISS project should provide is an electron dosimeter outside the station. This would allow SRAG to assess the correctness of the specifications and forecasts based on measurements by POES and GOES.

The spatial (as well as temporal) variability of energetic electrons must also be taken into account before the total dose on a particular EVA schedule can be predicted. To make accurate predictions, several other things must be done, including the following: (1) determining the decay time of HRE events as a function of the L shell; (2) converting >300 keV electron flux, which is measured, to >500 keV electron flux, which is desired (a rough conversion factor of 30 percent was discussed); (3) estimating the size of the "loss cone," which is needed to compute the flux at the position of ISS (latitude, longitude, and altitude) from the position of GOES or POES (to allow improving the 30 percent estimate and continuously updating it); (4) distinguishing between events with rapid rise and exponential decays and events with a more gradual decay; and (5) constructing a reasonable dynamic model of the radiation belts with physics included.

This chapter has cited evidence for a good correlation between changes in solar wind conditions and the onset of HRE events. A project to develop an HRE-event-onset protocol has a reasonable chance of producing a product reliable enough to be of operational use in issuing alerts and warnings.