historical and recent solar activity and geomagnetic storms affecting spacecraft operations l.
Skip this Video
Loading SlideShow in 5 Seconds..
Historical and Recent Solar Activity and Geomagnetic Storms Affecting Spacecraft Operations PowerPoint Presentation
Download Presentation
Historical and Recent Solar Activity and Geomagnetic Storms Affecting Spacecraft Operations

Loading in 2 Seconds...

play fullscreen
1 / 42

Historical and Recent Solar Activity and Geomagnetic Storms Affecting Spacecraft Operations - PowerPoint PPT Presentation

  • Uploaded on

I am the owner, or an agent authorized to act on behalf of the owner, of the copyrighted work described.
Download Presentation

PowerPoint Slideshow about 'Historical and Recent Solar Activity and Geomagnetic Storms Affecting Spacecraft Operations' - Sophia

An Image/Link below is provided (as is) to download presentation

Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author.While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.

- - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - -
Presentation Transcript
historical and recent solar activity and geomagnetic storms affecting spacecraft operations

INTRODUCTION: I have collected information on satellite anomalies related to disturbed solar-terrestrial conditions since about 1972. A more complete historical presentation on this topic is available on-line at the SCOSTEP website (110 slides and 70 comment pages). It is so large that anyone wanting a copy should ask for the CD version. The CD also contains related talks by others from past symposia.

Many people have contributed data and information summarized here or were sources of slides. Some older images were first used as viewgraphs, were written on at different meetings, and were scanned by a student worker, Marissa Rusinek. Some sources wish to remain anonymous, and to keep the name of their particular satellite(s) or company protected; others are noted on the slides or in comments. I gratefully acknowledge all sources of material used here, and the cooperation of many colleagues over the years. In particular, Dan Wilkinson, Dan Baker, Marsha Korose, Gordon Wrenn, Joe Fennel, Chris Kunstadter, James Heirtzler, and P. C. Klanowski.

Historical and Recent Solar Activity and Geomagnetic Storms Affecting Spacecraft Operations

Joe H. Allen, SCOSTEP

GOMAC 2002

Session 14: Modern Space Systems Issues

11-14 March 2002

Monterey, California


Dan Baker had Marissa prepare this illustration for an article in a book. It lists the three basic types/causes of satellite anomalies:

(a) Single Event Upset (SEU) caused by direct circuit element penetration of high energy protons or heavier ions.

(b) Deep Dielectric Charging (bulk charging) when relativistic electrons (> 1~2 MeV) penetrate and accumulate in dielectrics either outside the satellite (cables, thermal blankets, or power panel structure) or inside (circuit boards), and discharge with destructive effect. And

(c) Surface Charging when differential voltages originate on the satellite outer surface due either to its being engulfed by a cloud of thermal energy electrons (~10-15 Kev), or some change that interrupts the balance of charge maintained by photoelectron burn off (orbital eclipse or structural shadowing). These can result in either changes in reference voltages that trigger circuits (Phantom Commands), or generate destructive electrostatic discharges.

I have added less generally appreciated causes of other types of “anomalies”:

(d) Magnetopause Crossing Events (MPE) at GEO when the geomagnetic reference field is suddenly reversed and the satellite becomes disoriented. These field changes can have a range as large as 400 nT.

(e) Ambient Geomagnetic (field) Changes at LEO due to large currents encountered by satellites transiting field-aligned current regions that connect partial ring-currents with auroral electrojets. These can confuse instruments and interfere with electromagnetic coupling between the satellite and its momentum transfer wheel.

(f) Optical Disorientation due to limb sensors or star trackers that lose references when energetic protons and heavier ions create sparks in the viewing circuitry that obscures the normal target. And

(g) Power Panel Degradation due to the destructive penetration of the panel active elements by protons of energy >= 10 MeV.

(d) MPE disorientation

(e) dB/dT tumbling

(f) Optical disorientation

(g) Power panel degradation


I first prepared this slide in 1977 for an IAGA talk. It has been updated annually since then and now covers almost 7 sunspot cycles of solar and geomagnetic activity.

The yellow curve shows annual values of the smoothed sunspot number (see “Solar Geophysical Data Reports”, editor Ms. Helen Coffey, or NOAA’s NGDC/STP website).

The red curve shows the annual number of days when the global mid-latitude geomagnetic activity index Ap was 40 units or higher.

Large solar flares happen more often during years near or at sunspot maximum when more active regions populate the solar surface. Flare events are often accompanied by Coronal Mass Ejections (CME) of energetic particle clouds bursting out into interplanetary space and carrying along the heliomagnetic field. When these intersect Earth’s orbit, they distort the magnetosphere (Earth’s protective geomagnetic envelope), cause auroral zone “substorms” and global geomagnetic storms, and produce satellite anomalies. Their effectiveness is modulated by the orientation and intensity of the Interplanetary Magnetic Field (IMF) carried out from the Sun. Many CMEs are not directed at Earth, and those that arrive in near-Earth space may not find conditions optimum for coupling into the magnetosphere.

The time series of geomagnetic storms is more complex than the sunspot cycle due to a variety of different causes that produce different types and intensities of storms. The Ap index is based on activity during each UT-day, but since the Sun does not recognize the Greenwich Meridian, these indices truncate the maximum intensity measure of many storms. In 1975, I defined a simple modification, the Ap* index. It is based on the maximum 24-hour value attained by the running average of the 3-hourly ap indices (independent of the UT day). Ap* is used in several of the following slides to convey the relative intensity of the geomagnetic disturbance for different events. A fallout from consideration of different Ap* levels was the realization that lower amplitude, slow onset multi-day storms (40<Ap*<60) occur frequently during peak years on the declining side of the sunspot cycle and provide a background of enhanced radiation with which the more impulsive CME-related storms combine to produce some very large storms. However, these recurrent storms are clearly related to the yet-undefined processes that generate persistent relativistic electrons at GEO and lower altitudes.

The bottom line is that cycles of solar activity and geomagnetic storms vary due to a mix of causes and these same conditions can cause different types of satellite anomalies.

Major Magnetic Storms: 1932-2000

Annual Sunspot Number: 1930-2000

cumulative monthly magnetic storms 1932 2001 ap 40

Earth’s rotational axis is tilted relative to the ecliptic plane and changes its orientation to the Sun with a regular seasonal pattern. The geomagnetic dipole axis is further tilted (and offset) so that it describes an eccentric daily rotation that is amplified with the seasonal changes. At equinox the alignment between the geomagnetic field at its contact region with the IMF is optimal for merging between heliospheric field lines and geomagnetic field lines if they are oriented in opposite directions. Earth’s magnetopause then has an optimum “capture cross-section” to connect with a southward oriented IMF. Under these conditions, plasma from the solar wind can best enter the magnetopause and produce geomagnetic storms. In March-April and September-October more than twice the number of significant geomagnetic storms happen compared to the number during December-January and June-July. GEO satellite surface charging anomalies show the same pattern of seasonal highs and lows.



Of Storms

Cumulative Monthly Magnetic Storms: 1932-2001Ap* > 40










Jan Feb Mar Apr May June July Aug Sept Oct Nov Dec


Enhanced velocity co-rotating solar wind streams emerge from large, trans-equatorial coronal hole regions and spread outward from the relatively quiet Sun in the familiar “Parker Spiral” pattern. As the fronts of these zones sweep past Earth, they inject a greater volume of higher-speed, lower-energy electrons. These produce low-amplitude storms that typically last two or three days and recur with a 27-day periodicity (corresponding to the solar rotation period). These are most often seen during years of low or declining sunspot number when there are fewer CMEs. As the Sun rotates, these coronal hole regions may return two or more times with similar effects from each passage. When there are multiple coronal hole regions on the Sun at the same time, the 27-day pattern of activity from each region is intermixed. If a CME event happens during a time of enhanced solar wind conditions, then the result can be a very large, impulsive onset storm.

During the first 24-hours of a typical recurrent geomagnetic storm, high energy (MeV) electrons measured at GEO altitude typically drop to minimum background levels. Then there is often a sudden two or three order of magnitude increase in counts of electrons at energies greater than 1~2 MeV at GEO monitors. This slide from Dan Baker shows the annual difference in counts of such energetic electrons as they changed during one sunspot cycle.

These are the energetic particles that cause bulk charging events and have recently come to be called “Killer Electrons”.


For a relatively fixed population of satellites, the number of reported anomalies follows the annual pattern of geomagnetic storms (both gradual onset and impulsive). However as the number of insured satellites in orbit continues to increase, the reported anomalies follow the same upward trend.

About half the satellites in orbit are not insured because they are research or national security assets. Reporting of their anomalies is less extensive or completely classified.

From C. Kunstadter

sat nd timeline in orbit failures in 2001 p c klanowski

P.C. Klanowski maintains the SAT-ND website with extensive information about satellite anomalies. His “Timeline” section gives dates of reported anomalies and active links to information about the satellite(s) involved. This slide shows reported anomalies from last year (2001). Note the clusters of events during March-April and September-October 2001.

The list is extensive, but not comprehensive. Many anomaly reports do not make it outside of the company operating the satellite, the aerospace company that made it, or the national security group which operates it.

I emphasize that not all anomalies are due to disturbed space environment conditions. Engineering design failures, bad components, and normal aging in orbit may be the cause of the problem. Often the best we can do is to infer a cause by comparison with reports from other satellites and consideration of the recorded conditions reported by the constellation of research and monitoring satellites then operating.

Even when conditions in space appear to be most threatening for correct satellite operations and anomalies occur, many aerospace engineers will argue that the cause was an engineering problem and not adverse space weather. To some extent, there is a semantic difference across the range of viewpoints on this subject. Some will point to study results that indicate a high proportion of tin in solder used for circuit boards from which “tin whiskers” are known to grow in the temperature and pressure conditions on orbit. When these whiskers establish a conductive contact with critical circuit elements, a short can result with destructive effects. They consider this an engineering design or product failure.

Other take the viewpoint that a growing tin whisker can establish a narrow gap between circuit elements that can channel a destructive electromagnetic arc when plasma and field conditions are at extreme levels. This makes an anomaly at such a time, at least in part, the result of the disturbed space environment (space weather).

Usually it is not possible to be definite about the exact cause, and even when it is known the facts may be “classified”. For better or worse, much of our reporting is voluntary and

SAT-ND Timelinein-orbit failures in 2001 – P.C. Klanowski
sat nd timeline in orbit failures in 2000

Satellite anomalies reported in 2000 are shown here from Klanowski’s list. Note the cluster of events in March-April and the single event given here for 15 July 2000. The latter was the “Bastille Day” flare, CME, and magnetic storm event. More slides follow about that time.

SAT-ND Timelinein-orbit failures in 2000

SOHO-EIT, 2000/07/14 @ 07:00 UT

The joint European and NASA solar monitoring satellite SOHO (Solar & Heliospheric Observatory) has been a great success. It provides a view of the Sun and in situ measurement of interplanetary conditions from a site along the Sun-Earth line upwind from Earth. In this whole disc view the corona is evident, several bright active regions clustered about sunspot groups are seen, several arch prominences rise about the solar limb, and dark coronal hole regions are evident. The magnetically complex bright region in the upper part of the Sun at Central Meridian is seen at 07:00 UT on 14 July 2000.

The video clip shown next gives a dynamic view of the eruption of a bright flare and the surge of relativistic plasma from a CME. The first energetic particles reached SOHO soon after the electromagnetic pulse of the flare and began to obscure the image seen by the infrared telescope. This condition persisted for several hours.

In effect, Earth and near-Earth space was looking down the barrel of this event, and it would not be the quiet sort of recurrent enhanced plasma in the solar wind.

close up from soho eit sensor bastille day 2000 event minutes later

This detailed view of the Central Meridian active region observed by SOHO shows a close-up of emissions starting seconds after the previous slide. In the video, one sees the flare begin about an hour later. It is followed in less than half an hour by the arrival of high energy ions that start to cloud the optics with sparkling particle tracks. These conditions worsen and continue for about 9 more hours in this clip. Similar observations were made by other solar monitor satellites (YOHKOH and TRACE), by GEO satellites (GOES), and by research satellites (see following slides).

Close-up from SOHO EIT SensorBastille Day 2000 Event (minutes later)

SOHO full-disc infrared telescope image about 4 hours after the event onset.

Consider what effect this energetic particle cloud will have on satellites that rely on star tracker or limb tracker optics to maintain orbit orientation.

SOHO – EIT, 2000/07/14 @ 13:28 UT


NOAA’s GOES-8 satellite in GEO orbit monitored incoming electromagnetic (X-ray) emissions from the 14 July 2000 flare beginning around 10:00 UT. The next slide shows the arrival of the slower energetic protons from the CME.


Energetic protons measured by GOES-8 for the 14 July 2000 event. These particles began to arrive at GEO only minutes after arrival of the electromagnetic flare impulse. Such events have been monitored by SMS and GOES Space Environment Sensors since 1976 using a sequence of similar instruments. By definition (Space Environment Center), a solar proton event begins when the count of > 10 MeV protons exceeds 10 particles/sq cm/sec/sr (this graph shows 5-minute average values). This event was the 4th largest on record. This event also triggered a large geomagnetic storm with Ap*=192. The most disturbed 24-hour period of the magnetic storm began at 06:00 UT on the 15th (a day after onset of the particle event), and was the 28th largest magnetic storm recorded by the global observatory network since the start of determining Ap in 1932.

Maximum >10 MeV flux = 24,000 pfu,

4th highest intensity event since 1976

satellite anomalies 14 16 july 2000 proton event geomagnetic storm ap 192

Joe Allen:

On 14 July 2000, an impressive X5/3B X-ray flare was recorded by NOAA’s GOES-8 and -10. Within a few minutes after the flare detection, highly relativistic protons and heavier ions began to arrive at GOES in a very hard-spectrum proton event. Those who looked to the ACE real-time data plots were surprised to see that the solar wind velocity had dropped to background (as if the sensor was turned off). When the stream of very high energy protons arrived at GOES, the high energy electron sensors on both satellites failed, and stayed off for more than 24 hours. Magnetopause Crossing Events for GEO satellites near local noon happened on the 14th and again on the 15th for even longer periods. Data plots for these events are shown on the following slides.

Many interested persons began exchanging e-mail messages early on during this disturbed period, and continued over the next several days. Coincidentally, the COSPAR General Assembly was in progress in Warsaw, Poland, and many of those most likely to be concerned and most knowledgeable about such events were concentrated there. Executives of aerospace companies, satellite operators, and customers of GEO vendor services were busily exchanging e-mail and sharing near-panic experiences. Already a special session is being planned for the Fall AGU on the event. As Dan Baker has been heard to say, “It’s an ill solar wind that blows no good.”

The list of satellite anomalies here is known to be incomplete, largely because of vendor concerns.

Satellite Anomalies: 14-16 July 2000Proton Event & Geomagnetic Storm, Ap*=192
  • ASCA (Advanced Satellite for Cosmology and Astrophysics) – lost attitude fix resulting in solar array misalignment and power loss, satellite probably lost (in P.C. Klanowski list).
  • GOES-8 & -10 – SEM Electron sensor problems, power panels
  • ACE (Advanced Composition Explorer) – Temporary SW and other sensor problems
  • WIND – Permanent (25%) loss of primary transmitter power & Temporary loss of Sun and star sensors
  • SOHO (also YOHKOH & TRACE) – High energy protons obscure solar imagery
  • GEO and LEO Satellites – S/C orientation problems during MPE
  • GEO Satellites lost ~0.1 amp output from solar arrays
satellite anomalies at leo south atlantic anomaly saa auroral zone recent and historical

Low Earth Orbit (LEO) satellites, shuttles, and the International Space Station encounter different space environment conditions from those at GEO altitude but often meet extremes during the same disturbed times.

On NASA’s new Earth-looking satellite TERRA the MISR sensor signal was examined for days in February soon after launch. This was before the viewing port was opened to allow the sensor an unshielded look at Earth. Results are shown on the next slide and appear to paint an unintended picture of the region known as the “South Atlantic Anomaly” (SAA).

Other LEO satellites, the space shuttles, and (we suppose) the International Space Station also respond to enhanced radiation environments when passing through the SAA and the “horns” of the trapping regions.

Satellite Anomalies at LEOSouth Atlantic Anomaly (SAA) & Auroral ZoneRecent and Historical
  • TERRA-MISR 3-16 Feb 2000
  • TERRA-SFE off 26 Oct 2000
  • TERRA-MODIS 15 June 2001, 03:56 UT
  • TOPEX 1992-1998
  • NOAA-11 Sept 1988 - Aug 1990
  • STS-37, -39, -43, & -44 1991

The South Atlantic Anomaly (SAA) is a zone of lower geomagnetic field extending up to LEO altitudes over the region where Brazil bulges into the south Atlantic. It is there because the Earth’s internal dipole magnetic field is tilted relative to the rotational axis and offset from Earth’s center toward the Pacific region. The region is not a precisely defined locale, and it must shift slightly with the rotation of Earth and changing season of the year.

The SAA is rediscovered by instruments on many satellites (notably the Hubble Space Telescope), just as proton events and magnetic storm effects are found again and again by other satellites.

TERRA – MISR Data Before Shutter Opening

3-16 Feb 2000


Geomagnetic field contours near 1,300 Km altitude and superposed anomaly locations for TOPEX (red dots) and TERRA-MODIS temporary failure location in June 2001. Lower orbit inclination minimizes exposure to increased radiation in the cusps (“horns”) of the trapping regions. This figure is from an in-press paper by Heirtzler, Allen and Wilkinson.

TOPEX – 1992-1998 and TERRA-MODIS 2001


This figure was prepared in early 1991 by Dan Wilkinson from data selected by Herb Sauer from his sensor on the LEO polar-orbiting NOAA-10 and –11 satellites. The background radiation map of trapped electron flux is from a very active storm period in March 1989 and shows the effective merging of the SAA and the southern hemisphere cusp of the trapping regions at the nominal 850 Km altitude. The radiation gap in the northern region arises from the withdrawal of particles in the SAA as they interact with the higher density upper atmosphere and are no longer available to bounce between conjugate regions on opposite hemispheres.

The numbered blue triangles show locations of NOAA-11 when it experienced anomalies during a 2-year period roughly centered on the March 1989 Great Magnetic Storm. The triangles are pointed to show the direction in orbit of NOAA-11. The authors inferred that the radiation encounters produced anomalies related to a non-isotropic distribution of particles relative to the passage of the satellite.


Computer anomalies during four STS missions in 1991 also mapped the location of the SAA and the equatorward edges of the cusps. Because the STS orbits are usually low-inclination and low altitude, they are not normally subjected to the enhanced radiation of cusp passages. However, the relatively higher inclination of the International Space Station compels high inclination STS missions.

COMPUTER ANOMALIES: STS-37, -39, -43, & -44 1991 Missions


Klanowski’s list of reported satellite anomalies in 2000 included the TERRA and ECHOSTAR-IV events in late October 2000. Of possibly related interest was the announcement by SUN Computers that their surface-based systems had been shown to experience problems due to “cosmic ray” encounters. This was reported in “Fortune” magazine. I do not know whether any comprehensive study of these problems has been undertaken and the results reported in open media. During the period shown here of data captured by GOES-10 and made available on the NGDC/STP “SPIDR” on-line server, there were two energetic proton events (part of a 3rd was during the transition to November) separated by extended periods of high energy electrons at GEO orbit altitude.





Solar Activity in 2000 affects SUN Computers and Satellites

geo anomaly types
  • PHANTOM COMMANDS – surface charging, seasonal, orbit time dependent, 10-50 KeV e-
  • DEEP DIELECTRIC CHARGING – surface or interior, cycle dependent, > 2 MeV e-
  • SINGLE EVENT UPSET (SEU) – proton/ion event, > 10 MeV p+
  • LOSS of ORIENTATION – magnetopause crossing events, star bursts

Joe Fennell (Aerospace Corp.) shared this old map of the local time of anomalies on a family of geostationary satellites from the early 1970’s. A variant of this graph was shown to Joe Allen by his visitor (Wayne Lejeune, TRW & UCLA) in August 1972. Joe Fennell told me that preparation of this diagram wan one of his earliest jobs at Aerospace. No distinction was attempted in this figure to associate different types of anomalies with the satellite local time of the anomalies. Also, no information was given here about seasonal distribution of the anomalies in the database.

GEO Communications Satellite Anomalies – J. Fennell (Aerospace)

Early 1970’s


Joe Allen:

This slide was prepared by Dan Wilkinson in 1984-85 to show the location in orbit of different types of anomalies experienced by NOAA’s Geostationary Operational Environmental Satellites (GOES) –4 and –5.

Phantom Commands are clearly clustered in the midnight to dawn sector. Identification of the type of satellite anomaly is often difficult and can only be inferred, so it is helpful to have a record of the time and location in orbit of anomalies of each type.

Single Event Upsets (SEU) are “penetration” events in which high energy charged particles (protons, alphas, or heavier ions) enter a chip and “burn” a destructive track or deposit charge that changes the contents of chip memory.

Random Part Failures (RPF) are those rarer events when a component on a satellite fails and, if the satellite is to keep operating, its controllers must switch operations to a redundant component.

The figure in this slide mimics one for USAF DSCS satellites shown to Joe Allen in August 1972 by Wayne LeJeune (TRW) during a visit to Boulder. A sudden catastrophic failure of a satellite when Wayne was in Boulder resulted in call from California to check out conditions. This was during the event that NASA christened the “Anomalous Large Proton Event”, although it later was fitted into a relatively continuous spectrum of event amplitudes.

Wayne came to Joe Allen’s office at NGDC and showed him a clock-face diagram of DSCS satellite anomalies. It looked much like the GOES plot seen here (see previous slide in this paper). It was interesting to overlay the anomaly map onto a similar sized diagram of the regions in which the eastward and westward auroral electrojets normally flow. The surface charging region (midnight to dawn) matched the westward electrojet region. We later learned that this corresponded to the results of Farthing, et al. in their internal NASA report.

GOES-4 and –5 Anomalies: Early 1980s


Several years ago, Harry Farthing (then NASA GSFC) shared this figure from the internal report they prepared as NASA analyzed the possible causes of anomalies on NOAA’s GOES-4 and –5 satellites. Engineering changes were made on GOES-5 after consideration of the earlier data. The result was a reduced sensitivity to lower energy electrons by GOES-5. However, the effect was not completely fixed.

The scale of increasing magnetic activity used for the upper panel is the Kp index. This is a semi-logarithmic 3-hour index that corresponds to the linerized ap index already mentioned by Allen. The UT day average, Ap, with a value of 40 or greater corresponds to the transition from Kp=4 to Kp=5.

In a separate study (1982), Joe Allen showed that the number of satellite anomalies happening when Kp > 3 is roughly double the comparable number for less disturbed times.

10 – 15 KeV electrons

drifting from midnight

Farthing, Brown & Bryant, 1982 NASA Report


Superposition of Wilkinson’s GOES-4 and –5 surface charging anomalies onto the map of the Rice University Magnetospheric Specification Model provided by John Freeman. The modeled zone of increased electron plasma injected from the magnetotail around midnight and moving toward dawn under the v X B force accounts for the surface charging. The distance of a GEO satellite from midnight at the start of an injection event determines the time required for the drifting electrons to catch up with the satellite and cause an anomaly (most often Phantom Commands) as per the results of Farthing, et al.


ANIK-E1 and –E2 failures in January 1994 happened toward the end of an extended number of days of high counts of relativistic electrons in GEO orbit. Four other anomaly events on an unnamed European GEO satellite were clustered around this time (see results from G. Wrenn in next slide). Also an INTELSAT GEO satellite experienced a temporary anomaly during the failure time of the ANIKs. This was interesting because the satellites were built by the same aerospace company at the same time, but to somewhat different consumer specifications.

Deep Dielectric Charging-Arcing: Jan 1994 ANIK-E1 & E2


The European GEO satellite series given the psuedonym “DRA-delta” by Gordon Wrenn was analyzed for 1992-1994 in relation to the changing monthly average flux of energetic electrons. His results showed a clear rise in number of anomalies with higher flux of electrons. Note, these years are on the declining side of the sunspot cycle shown on an earlier slide, and follow just after the maximum shown on the distribution slide from Baker. At the session on satellite anomalies in the IAGA part of IUGG-1995, Gordon said that “Allen’s conservative phrase ‘probably due to’ space environment conditions at GEO altitude should be changed to ‘certainly due to’.”

From Gordon Wrenn-1995


In the mid-1980s, NASA’s TDRS-1 satellite was launched into GEO to begin replacing the global tracking antenna network with an array of satellites to relay commands to LEO satellites and Shuttles. Eventually there were to be 4 such satellites operating in a constellation to provide uninterrupted communications. However, experiences with the first satellite of this series (see next two slides) were not good. Eventually (October 1985) convincing evidence was produced that the many anomalous events on TDRS-1 were a product of susceptibility to galactic cosmic rays and to solar energetic protons and heavier ions. One of the engineers involved in Albuquerque with operating TDRS-1 characterized the 93L422 RAM chip as “a flying solid state cosmic ray detector”. At some expense the chips were replaced in TDRSS-2 to –4 and greatly reduced the number of anomaly events (but did not eliminate all of them).

NASA continued using the RAM chip at least through the construction and launch of the Hubble Space Telescope (HST) which has problems similar to those of TDRS-1 when HST passes through the SAA.

Single Event Upsets due to protons, alphas, heavier ions


TDRS-1 anomalies during one of several hard-spectrum proton flare/CME events in October 1989. This figure was prepared by Dan Wilkinson (NOAA-NGDC).


The 5-year time-series of TDRS-1 SEU events shows both a susceptibility to galactic cosmic rays (upper panel - long period variation out of phase with the sunspot cycle) and to solar energetic particle events (many spikes in 1989).

During years of high sunspot number, particle clouds from solar activity create a shield that reduces access of galactic cosmic rays to Earth. This is seen in the record of neutron monitor values (middle panel).

Solar CMEs can happen anytime during the sunspot cycle (bottom panel), but are especially frequent during years around maximum.

TDRS-1 was an unintended probe of the energetic particle environment at GEO altitude.

satellite anomalies at geo recent and historical
Satellite Anomalies at GEORecent and Historical
  • GALAXY-VII 13 June 1998 SCP-1
  • BRAZILSAT 09 April 2000 TWTA
  • SOLIDARIDAD-I 27 Aug 2000 SCP
  • ECHOSTAR-IV 31 Oct 2000 Transponders lost
  • INSAT-2B 4 Nov 2000 Service Outage
  • GALAXY-VII* 22 Nov 2000 SCP-2
  • ECHOSTAR-VI ? April 2001 service
  • GALAXY-IIIR 21 April 2001 SCP
  • TELSTAR-6 22 April 2001 SCP
recent active time 2 11 april 2000
  • High levels >2MeV electrons, 04/02-12 except during storms
  • Proton event, 04/04-06, 55pfu
  • Large geomagnetic storm, Ap*=137, 04/06-07
  • Brazilsat-A2 lost TWTA on 04/09

According to the maximum 24-hour average of the 3-hourly ap magnetic activity indices for this period, the value of Ap* = 137 was the second largest magnetic storm of this solar cycle (see the Apstar document and list of events at the NGDC/STP website under geomagnetic indices). The Ap* index was introduced by Joe Allen in 1976 to provide an objective rank-order size classification of magnetic storms from 1932 to the present.

  • Although the >10 MeV proton maximum count (55) from SEC was not particularly high compared to other events, David Evans (SEC) confirmed the event had some of the highest fluxes of auroral energy electrons seen in this solar cycle.
  • Information about Brazilsat-A2 from L.C.F. de Sousa (EMBRATEL), and later follow-up by Walter Gonzalez
  • See slide for modeling by Dan Moorer (LASP/Ball Bro’s.). 01/04/09 failure

The INSAT-2B and GALAXY-VII outages illustrate some of the difficulties arising from attempts to relate what happens on satellites in orbit and the environment in which they operate. In both cases, these anomalies (a few days after the ECHOSTAR outage – see list slide) came before proton events and after days of extended high energy electrons.

It isn’t always possible to make an association.




Max pfu = 31,700;

3rd largest since 1976



November2000 Space Environment at GOES-10



TUMBLING at LEO: 31 Mar-2 Apr 2001

Ap*=191 was 29th most disturbed 24-hour period (comparable to the Bastille Day storm in July 2000) since 1932. A number of commercial GEO satellites reportedly had problems with maintaining orientation during this event, and a LEO satellite began tumbling in orbit. Reportedly staff of NOAA’s Space Environment Center are preparing a list of MPEs to facilitate comparison between these conditions and satellite operations.

Major MPE


Magnetopause Crossing Event & Storm Disoriented Satellites


This very active period from about one year ago contained intense electromagnetic radiation from the Sun, multiple hard spectrum proton events, and high levels of energetic electrons.

Note – usually one notes that the energetic electron sensor can be polluted by high energy protons; however, we note here that as the proton events declined toward background levels, the electrons remained at uniform height and continued to show a smooth diurnal variation.

Most intense X-ray flare



2001 March 31-April 30: Space Environment at GOES-10


GALAXY-IIIR failed on 21 April and TELSTAR-6 on 22 April 2001. Companies operating these satellites attributed their problems to “space weather” disturbances. Geomagnetic conditions at Earth were relatively quiet on the 21st, but very disturbed on the 22nd with evidence of high-latitude particle injection (see next two slides).

Quiet pre-shock

No storms or substorms


Relatively quiet in auroral zone on 21st

AE(8) Quick-Look Indices (Kyoto)


Marsha Korose contributed this slide from one of her “briefings for Colonels and Generals”. All problems list here are attributed to Space Weather effects, and all required some system redesign. Usually fixes were found, but new satellite systems continue to encounter old problems from decade to decade.

Some events lead to new operational procedures to keep expensive satellites operating (e.g. the ANIKs).

Often the information about these problems and their fixes is not easily available, and there is legitimate concern about the impact that such information might have on potential customers for vendor services, on design costs of new satellites, on manufacturer reputations in a competitive environment, on insurance costs, and on the available of loans to build the next generation.

Sometimes the introduction of new technology to do increased tasks faster with more compact components that require less power makes the result more susceptible to conditions in orbit.

conclusions 1
  • Sunspot cycle decline and minimum years are ideal for “killer electrons” at GEO and lower orbit altitudes.
  • Sunspot cycle maximum years are ideal for energetic proton and heavier ion events that cause SEUs and sensor optics & power panel degradation.
  • Major magnetic storms may happen at any time and cause spectacular effects on satellites, technology and humans.
conclusions 2
  • Every satellite (or object) in orbit is a probe of the Space Environment from which to learn.
  • The history of satellites should be the basis for learning what causes operational problems.
  • Combining space environment data with satellite histories is necessary.
  • Solar Cycle # 23 still may be highly active. Are we any more ready today?
  • What about cycle # 24? Can engineering cope?