This section illustrates space weather events using the unusually large and geoeffective events observed during the periods 6 - 19 March 1989 and 6 - 11 January 1997. These events contained some of the largest substorms on record. The news article by Allen et al. [1989] describes the 1989 events in detail, some of which is paraphrased and other parts supplemented here. The ISTP web site contains a more detailed discussion of the January 1997 events.
The March 1989 events were associated with a unusually large and complex sunspot group, active region 5395. Figure 15.3 [Allen et al., 1989] shows one of the first X-ray flares from this region,
Figure 15.3: X-ray, energetic proton, and magnetic data from the GOES-7 spacecraft, the
horizontal component of the magnetic field measured at Boulder, CO, USA, and the cosmic
ray flux measured by the Deep River Neutron Monitor for the period 5 - 6 March,
1989 [Allen et al., 1989].
as well as the relatively high level of cosmic rays being detected at the
Deep River Neutron Monitor which indicated a fairly constant but unusually high number of
energetic solar particles reaching Earth. Figure 15.4 shows data from the available solar
wind monitoring spacecraft (IMP 8), as well as the measures 
and Kp for magnetospheric
activity.
Figure 15.4: Solar wind monitor data ( 
, 
, and 
) and geomagnetic
activity indices Kp and for the period 4 - 24 March 1989.
Note the arrival of a shock on 8 March, presumably associated
with the X-ray flare on 6 March, and the associated small increase in .
This increase in is probably a SSC/SI (sudden storm commencement / sudden impulse)
due to compression of the magnetopause and the
concomitant increase in the magnetic field of the magnetopuase current layer detectable on
Earth's surface. The rise in Kp and the large, localized decrease in near midnight
on 13 March are described as geomagnetic activity related to a magnetic substorm.
Figure 15.5 [Allen et al., 1989] show solar and magnetospheric data for the period 13 - 14 March.
Figure 15.5: Data for the period 13 -14 March 1989 in the same format as Figure 15.3 [Allen et
al., 1989].
Note the huge increase in the level of MeV protons at geosynchronous orbit (GOES-7 data) - the
increase is by 4 orders of magnitude relative to the level on 5-6 March. Numerous operational
difficulties were reported for geosnchronous and low-orbit spacecraft during this period.
The GOES-7 magnetometer data show negative values for the H-Parallel magnetic field component
several times on 13 March, corresponding to the magnetopause being pushed Earthwards of the
spacecraft and the resulting detection of Southwards in the magnetosheath. SSCs associated
with earlier compressions of the magnetopause are visible in the Boulder B-Horizontal data near
0128 and 0747 UT on March 13. Changes in the magnetospheric and ionospheric current systems
cause the negtive decreases in the Boulder B-Horizontal field from about 1000 UT until the
substorm itself starts near 2100 UT, as shown also in Figure 15.4's data. This is the
main phase of the substorm itself.
The changes in itself and the substorm are associated with injections of energetic
particles near geosynchronous orbit: Figure 15.6 shows the injection of energetic particles at
two of Los Alamos's DMSP satellites, which 
drift Earthwards and
undergo 
and curvature drifts around the Earth as they join the ring current.
Figure 15.6: Count rates of energetic electrons versus time and energy from
Los Alamos instruments on three geosynchronous spacecraft for 13 March, 1989. Note
the strong injections near 2100 UT for two of the spacecraft.
During this period the count rates of energetic particles precipitating into the auroral ionosphere and equatorial ionosphere increased by several orders of magnitude, typically in bursts with lifetimes of order half a day and not inconsistent with the theoretical loss times for particles from the ring current. The aurorae during the period of the substorm were unusually intense, unusually large in area, and moved to unusually low magnetic latitudes. For instance, aurorae were observed above the topic of Capricorn in far northern Australia and in Arizona, USA. Figure 15.7 shows these results [Allen et al., 1989].
Figure 15.7: Southern hemisphere auroral images taken by the Dynamics Explorer 1 (DE-1) spacecraft
Allen et al. [1989]. At left is a quiet time auroral ring (22 March, 1983) and at right
is the largest auroral zone recorded by DE-1 up to 1826 UT on 13 March 1989.
Perhaps the most important space weather effect of this period involved electric power failures due to intense auroral currents. The Hydro-Quebec Power Company experienced a massive power failure that darkened most of Quebec Province for nine hours and left over 6 million people without power and heat in their homes. Many people in more remote locations had no power for several weeks, even months. The failure occurred due to changing magnetic fields (due to substorm activity and associated auroral currents) inducing low frequency currents in power lines which saturated transformers and caused protective shutdowns that spread due to networking of the power grid. Similar, simultaneous (to within 1 second!) power losses occurred in central and southern Sweden. Power systems across much of the USA experienced difficulties but not widespread power outages.
Major effects on space systems were reported during this period [Allen et al., 1989]. These included communication anomalies and outages for GOES-7 and other satellities. Perhaps more important were greatly increased difficulties in maintaining spacecraft attitude, pointing, and orbits for both commercial and military spacecraft: these difficulties were due to increased and variable ionospheric drag and also to the effects of changing magnetic fields. SMM's orbit, for instance, is believed to have dropped about 5 km over this period while one series of seven commerical geostationary satellites required more than one normal year's number of manual operator attitude adjustments in just the period 13-14 March. At least one spacecraft experienced a major component failure during this period, potentially due to space weather effects.
Severe communication difficulties were also experienced [Allen et al., 1989]. For instance, the US Coastguard's LORAN navigation system encountered numerous problems, especially on 6 and 13 March, while the US Navy's high-frequency radio network was out worldwide. In addition, geophysical exploration teams found their instruments essentially unusable for extended periods and microchip production factories in the northeastern USA were not operational for two periods, both due to changing geomagnetic conditions.
Figure 15.8 shows the solar wind data for the January 1997 space weather event, identifying the CME's shock and rotating magnetic field, as well as a CIR.
Figure 15.8: Solar wind data for the January 1997 storm period, as presented by L.F. Burlaga
on the ISTP web site. The CME shock and cloud, plus a subsequent CIR, are clearly
identified.
Note that the
magnetic field has a strong southwards component during the CME (magnetic cloud)
itself. Figure 15.9 shows the and 
indices for the period.
Figure 15.9: Geomagnetic activity indices 
and for the period
8 - 12 January 1997, obtained from the OmniWeb site.
A clear SSC is visible in near about 0200 UT on 10 January, while the main decrease
in occurs over about the next 12 hours on that same day, followed by a slower recovery
over another 24 hours. 
started to increase prior to these changes
in . These increases in geomagnetic activity started (and continued) while was
southwards. The peak in early on day 11 may be another SSC, this time associated with
compression of the magnetosphere by the very dense ``filament''at the end of the CME.
Figure 15.10 presents more detailed data for this period, showing the rather complex and strongly time varying nature of much geomagnetic activity.
Figure 15.10: Ground-based and spacecraft data showing different aspects of geomagnetic activity
for the January 1997 space weather events: (top) ground-based Canopus magnetometer data showing the
magnetic perturbations of the auroral electrojets as a proxy for AE, (second) energetic
electron fluxes measured at geosynchronous orbit, showing discrete injections on day 10
and before as well as the increase in flux level due to enhanced solar fluxes, (third)
ground-based SESAME magnetometer data, and (bottom) magnetometer data from the Wind spacecraft
showing how these magnetospheric variations are associated with southwards
The figure shows magnetic field perturbations associated with the auroral oval and auroral
electrojets (top panel), intense, time-localized injections of energetic electrons near geosynchronous
orbit (second panel), perturbations in the magnetic field observed by the SESAME magnetometer
chain (third panel), and the magnetic field observed in the solar wind by the
Wind magnetometer (bottom panel). Note that the most intense activity occurs during and is
part of the 10 January substorm. However, other, more localized disturbances occur during other
times when is southwards.
The aurorae during this storm covered an unusually large area, were unusually bright, and varied unusually quickly. Figure 15.11 shows these data.
Figure 15.11: Auroral displays on 10 January 1997; The top-left image shows a quiet auroral oval
before the storm. The top-right and bottom-left images are during the height of the storm,
showing the auriral oval to have expanded, broadened and brightened greatly. The final image
shows how quickly the auroral displays can disappear.
Before moving on to a general physical result, it is worth mentioning that the Telstar 401 satellite failed during this space weather event, possibly (but not certainly) due to radiation damage and dielectric charging. Space weather has also been responsible for several other satellite failures in the last few years, including the Galaxy 4 satellite whose loss led to most American pagers being unusable for about 1 day in either 1996 or 1997.
A necessary (but not sufficient) condition for geomagnetic activity to occur is that 
be southwards for at
least a 30 - 60 minute interval prior to an event.
This is shown conclusively, in a statistical sense at least, in Figure 15.12
[Muruyama et al., 1980].
Figure 15.12: Dependence of the AL index for substorm magnetic activity on and
[Muruyama et al., 1980; McPherron, 1995]. The abscissa is the value of AL normalized
by 
, while the ordinate is either the hourly average of when 
or else a duration-weighted value of called 
for 
. Statistically, substorm
activity is clearly associated with long duration and/or large values of southward .