Although our sun is usually regarded as a benign provider of heat and light for our planet, it is underappreciated that solar storms originating from the surface of the sun pose some serious risks, especially to our ever more connected technology that we are now so dependent on. Dr Foster Langbein of Risk Frontiers, Macquarie University, shares the research done on solar flares, near misses, the vulnerability of power transformers, worst-case scenarios and why insurers must consider severe space weather risk.
The sun normally pumps out something like a million tonnes of charged particles every second, a phenomenon known as the solar wind, and its time variation around earth’s magnetosphere – a sheath of charged particles maintained by the earth’s magnetic field – is known as space weather.
Manifestations of the solar wind are seen on earth as auroras at high north and south latitudes (Arctic or Antarctic) where the charged particles of the solar wind are funnelled and accelerated by the more intense magnetic fields near the poles and, colliding with molecules of the upper atmosphere emit visible light.
The sun’s dynamic atmosphere will sometimes produce sudden, spectacular bright flashes – solar flares - often accompanied by it hurling massive clouds of charged particles into space - coronal mass ejections (CMEs). If a solar flare is directed towards us and especially if the earth gets in the path of a CME, then the interaction with our magnetosphere can cause severe electromagnetic disturbances known as a solar or geomagnetic storm. During a strong geomagnetic storm, auroras can be seen at lower latitudes.
The occurrence of solar flares and CMEs rises and falls with the solar cycle – the roughly 11-year variation in the sun’s activity visible to us as changes in the number of sunspots.
Although the current solar cycle has been a particularly quiet one, the quietest in nearly a hundred years, our position of just past maximum means a higher chance of severe space weather, since flares and CMEs are more common during the declining phase of the cycle. Moreover, CME intensity is not correlated with the amplitude of the solar cycle.
Severe space weather
There are three main modes in which solar storms can affect us: geomagnetic, radio and radiation emissions – a separate severity scale is used to describe each aspect (G1-G5, R1-R5 and S1-S5 respectively). Space vehicles are of course particularly vulnerable due to their location; radiation damage and a build-up of static charge likely causing anomalous operations and even failure. Low earth orbit satellites are also subject to much more rapid orbital decay during a solar storm.
Severe space weather can also degrade or even black out high frequency radio communication as well as disrupting signals from GPS satellites. These effects, more pronounced in Polar Regions, can cause significant navigation problems in shipping and aviation. During severe space weather, aircraft have been required to divert from their optimal transpolar route at a cost of the order of US$100,000 per diverted flight due to these effects as well as fears of an increased radiation dose to passengers and crew.
Ground induced currents
The most damaging terrestrial effects are probably ground induced currents (GICs), where anomalous current is induced in any long electrical conductor. Electricity grids, pipelines and rail can all be affected.
The geomagnetic storm of March 1989 caused the collapse of the Hydro-Québec power grid in Canada within 90 seconds leaving about 6 million people without power for at least nine hours and causing an estimated $2 billion dollar loss. Engineers at North American power companies speculated at the time that it could easily have turned into $6 billion catastrophe affecting 50 million people in many US East coast cities if only a few more capacitors had failed in an interconnected power network.
Carrington event of 1859
A useful index for characterising geomagnetic storms is the Disturbance storm time Index (Dst) with a negative value measured in nanotesla (nT) as an average of readings from four low-latitude geomagnetic measurement stations. It has been measured continuously since 1957 and a geomagnetic storm is considered extreme with a Dst of less than -500 nT. The only measured storm of this magnitude since then has been the March 1989 event with a Dst of -589 nT and an estimated return period of 60 years. The index has however been inferred for two other notable historic events.
Probably largest historical event and also the first ever recorded is the so called Carrington event of September 1859 named after the British astronomer who recorded the flare and correctly associated it with the geomagnetic storm a day later.
The event was so intense it caused auroras visible at geographic latitudes of 25° and lower according to eyewitness accounts reported widely around the world in newspapers of the day. A reporter in the New York Times wrote: “At that time almost the whole southern heavens were in a livid red flame, brightest still in the southeast and southwest.” and at 1 am it was so bright that one “could read a newspaper by” it.
In New Orleans “Crowds of people gathered at the street corners, admiring and commenting upon the singular spectacle.” GICs from this event caused widespread disruption to the extensive telegraph network in both Europe and North America with some telegraph poles throwing off sparks and telegraph operators receiving electric shocks from their equipment. An estimated Dst of -850 to -1050 has been derived for this event.
New York Railroad Storm of 1921
A second large historical event less well known but with an estimated Dst of -825 to -900 occurred in May 1921 and has been sometimes named the New York Railroad Storm. This storm knocked out the entire signal and switching system of the New York Central Railroad below 125th Street and caused a fire in a control tower.
Once again telegraph systems were brought to a halt, this time in both northern and southern hemispheres, and undersea cables were also damaged. It is thought that even though GICs would have been up to a magnitude greater than those in the 1989 storm, the effect on the electricity supply was limited to local areas because widespread interconnection of power networks was yet to occur.
Close miss in 2012
Recent analysis of data from the STEREO-A space probe – STEREO-A and STEREO-B are solar observatories which lead and trail the earth in its orbit - show that an extremely powerful CME crossed earth’s orbit in July 2012.
Fortunately for us, the earth was at a different point in its orbit and out of harm’s way but a week later and we would not have been so lucky. Indeed, were it not for the presence of the STEREO-A probe, we would not have known about the event at all. And soberingly, had the storm hit us, it has been estimated that this event would have generated a geomagnetic storm measuring a Dst of around -1200 nT, more than twice as large as the March 1989 storm and possibly larger than the Carrington event itself.
What are the chances of Carrington class storm hitting earth?
Research in 2012 by Peter Riley looked at this question by fitting a power-law (or Pareto) distribution to the historical record of Dst as well as other solar storm indices to extrapolate to the values estimated for the Carrington event. The likelihood was estimated as a 12% chance of a Carrington-class event occurring in the next 10 years.
A separate paper using a different method, a Poisson estimate based on the Dst value for the three recorded extreme geomagnetic storms (1859, 1921 and 1989), derived a value of 6.3% in line with the Riley figure given that it was based on a higher Dst estimate for the Carrington event and given the large uncertainties involved.
Corroborating evidence that these estimates are in the right ballpark have also recently been obtained by analysing the frequency of solar flares on other sun-like stars using data from the Kepler spacecraft. In fact, in the Kepler data, it would appear Carrington sized events are small fry with data showing some events orders of magnitude larger.
In today’s immeasurably more connected and electricity dependent world, the consequences of a Carrington sized event are likely to be much more severe than in 1859. A 2008 workshop run by the National Academy of Sciences (NAS) in the US on the Societal and Economic effects of Space Weather estimated a $1-2 trillion loss for such an event in the first year alone and recovery times of four to 10 years. A similar group dubbed SolarMAX gathered in Europe last year and studied the multiple technological infrastructure risks and knock-on societal effects.
Very large power transformers at risk
Of particular concern today are the vulnerability of the very large power transformers that step voltages in our power grids - unexpected GICs generated by a geomagnetic storm can damage them permanently.
Such transformers are generally custom designed for each location, cost on the order of $10 million each and have replacement times that are from many months to years. Power networks do not usually have backup transformers. The currents that caused damage in the Hydro-Québec grid have been estimated as 10 times weaker than those that would occur in a Carrington scale event. The loss of 14 transformers in South Africa during the “Halloween” solar storms in 2003 – with a Dst of around -400 nT, a large storm but still much smaller than the 1989 event - illustrates the potential threat to power grids if they are unprepared.
Some worst-case scenarios
The NAS report estimates up to 350 US transformers would be at risk of failure in a solar super storm which would leave something like 130 million people without power.
The knock-on effects of a long-term power outage – lighting, heating, food and medicine spoilage, water shortage and untreated sewage – would be a national scale disaster. During the event itself, disruption of telecommunication, GPS navigation and transport would hamper emergency and recovery efforts.
One of the SolarMAX collaborators, Ashley Dale, writing in a recent article in Physics World, explored a European scenario where the collapse of the power grid leaves the hundred-odd nuclear reactors struggling to find enough power or generator fuel to maintain the systems needed for the month long cool-down to prevent meltdown, such reactors rarely having more than a week’s worth of backup power.
These are of course worst-case scenarios. Some mitigation efforts are possible given enough warning: power companies can reduce or redirect system load and shutdown vulnerable components; satellite organisations can place satellites into safe mode; and aviation companies can assess and prepare alternate routing.
Still lots to learn about solar storms
Solar storm forecasting is however under resourced and still not well enough understood. Observations of major flare activity on the sun give only 15-30 minutes warning before the satellite-affecting radiation reaches earth, and although a geomagnetic storm inducing CME takes a longer time to reach earth, how long it takes – from 14 hours to several days - or whether one will even be generated at all from a flare region still needs better understanding.
Furthermore, the observational spacecraft are in great need of upgrade. The single space craft giving forecasts that helped mitigation efforts during the 2003 Halloween storms – the ACE space probe – is near the end of its operational life and a replacement has still not yet been launched. And a single craft gives us only a view from our vantage point, we cannot know about developments in the solar atmosphere on the far side of the sun until they rotate into view. The SolarMAX team propose a constellation of small solar observing craft able to provide a complete 3D view of the sun’s surface.
Insurers need to consider severe space weather risks
It is clear that solar super storms present considerable danger to our modern technological way of life. Due to our limited experience of very severe events – there does not seem enough to go on to construct a CAT model for pricing the risk for example – it is uncertain to what extent it would impact the insurance industry.
But despite lesser events in the past having relatively limited impact, the potential exists for major losses across all lines of business in a Carrington type event which is more probable than we might like. Moreover, such an event would be happening across the globe simultaneously, limiting any advantage in global diversification. The SolarMAX report speaks of a “risk gap” where insurance companies have traditionally omitted considerations of severe space weather. It seems clear that despite the difficulties in quantification, solar storm risk should not go unanticipated.
Dr Foster Langbein is CTO – Software Architect at Risk Frontiers, Macquarie University.