INTRODUCTION
The aurora is the most directly visible manifestation of space weather and has a beauty that is rarely equalled by any other sky display.
Aurorae are caused by particles from the Sun impinging on the Earth's atmosphere. These particles are guided down to the Earth from interplanetary space by the Earth's magnetic field. There is always a drizzle of particles from the solar wind that make it along field lines to the polar regions, both north and south. And thus there is always some form of aurora. However, this type of aurora can only be seen in the Arctic and Antartic. Australian scientists stationed at the Mawson Antartic base can see an aurora almost anytime the sky is dark, although this aurora is usually fairly faint.
Only when the Sun releases a large cloud of plasma (called a coronal mass ejection or CME) that reaches and enters the Earth's magnetic field does the aurora become more spectacular and increase in spatial extent so that it can be seen at lower latitudes. The larger the influx of particles the brighter the aurora becomes and the closer it moves to the equator.
These two space images from the Dynamics Explorer satellite show an 'ambient' (left image) and an active aurora. Aurorae always occur in the form of an oval centred on the magnetic poles. More active aurorae are caused when an increased flux of solar particles enters the Earth's magnetosphere. The auroral brightness increases and the oval expands, pushing the visible display closer to the equator. Image: Lou Frank, Univ of Iowa.
Small clouds of plasma are ejected from the Sun on a relatively frequent basis, but larger clouds are less common. People living in Invercargill in New Zealand or at the southern tip of Tasmania will see aurorae more frequently than people living in Sydney or Perth.
Because an aurora can present a spectacular visual display and a great photographic opportunity there is interest in a warning system to provide an alert of when an aurora is likely to occur or when one is in progress. Some people even travel hundreds of kilometres to view and photograph aurorae, and thus require accurate predictions if they are not to be disappointed.
This note discusses methods by which auroral alerts may be made.
[Note: More information on aurorae may be found in Guide to the Aurora]
OVERVIEW
Aurora can be detected from space or from the ground. Detection may be direct or indirect. Indirect methods measure parameters associated with aurorae, such as particle precipitation, atmospheric ionisation or geomagnetic activity.
Predictions of forthcoming aurorae can be made on the basis of CME prediction. However, this requires a knowledge of whether the CME will head toward the Earth and whether it will couple into the Earth's magnetic field when it arrives in our neighbourhood. This latter factor only occurs if the interplanetary magnetic field is pointing in the right direction (technically Bz must be negative).
Spacecraft stationed at the Lagrange point between the Earth and the Sun can give about one hour's warning of when a CME will hit the Earth.
Large aurorae also often show a 27 day recurrence (due the rotation of the solar region producing the CMEs) which may be used as a predictor.
However, for the purposes of viewing and photographing aurorae, predictions tend to have large error rates, and we usually are more interested in early detection of an aurora in progress. Large aurora generally last for many hours, and thus give us sufficient time to travel to a good observing location.
The diagram below gives an overview of various detection methods. These are discussed in more detail in the following sections.
SPACE-BASED AURORAL DETECTION
The detection of aurora from space relies on direct imagery or the sensing of particles which are precipitating down to auroral altitudes. Imagery from space has the big advantage that it is not affected by clouds and this can provide detection as long as the satellite is in range of the auroral regions.
Imaging
The Space Shuttle and the International Space Station (ISS) fly at low altitudes and often fly very close to or even through auroral displays.
Although images from the ISS are spectacular, its low orbit and inclination of 51 degrees means that it cannot provide a global picture of auroral activity.
 |
Meteorological satellites often record auroral activity. These satellites are in higher orbits than the ISS (generally from 700 to 1400 km altitudes), and have polar orbits which take them over either the southern or northern auroral region approximately every 90 minutes. The image at left is from a US Air Force Defense Meteorological Satellite Program satellite (DMSP). The problem with all visible or infrared satellite imagers is that they do not record aurora in the daytime. |
| Several specialised satellites have been launched into highly elliptical orbits with ultraviolet imagers to specifically record auroral activity. The most well known of these are the Dynamics Explorer satellites and the IMAGE satellite. When at their apogee these satellites give good continuous global coverage of either the north or south auroral regions. In this image the letter N represents the position of the north geographic pole whereas M is at the north geomagnetic pole. It can thus be seen that the auroral oval is centred around the magnetic pole. Image courtesy Lou Frank, University of Iowa. |  |
One of the problems with imagery is that of availability and interpretation. Even when available the extraction of auroral information is a fairly intensive operation. Automated extraction of a calibrated image that provides brightness and positional information for a specific ground would increase the utility of such imagery greatly.
Satellite Particle Sensors
Most aurorae occur at altitudes of 100 to 300 km above the Earths' surface, with special aurora extending to 600km. The light emission is produced when sub-atomic particles, mostly protons and electrons, precipitate from higher altitudes and ionise atoms of the upper atmosphere at the lower heights previously mentioned.
Most meteorological satellites carry particle detectors and can measure the flux of precipitating particles along their orbit. They can thus define the equatorward boundary of the auroral oval when the particle flux shows a sudden increase or decrease (depending on their direction of travel). Particle flux graphs for selected US satellites are available from the Space Weather Prediction Center web site. The graphs are updated at time intervals limited by the satellites passage over a download station. There6:15 AM 1/2/17 are also plots of flux versus time and an accurate knowledge of the satellite's orbit is required to turn these into flux versus geographical position.
Global precipitating electron flux detected over several orbits of the NOAA-18
satellite. Both north and south auroral electrons can be seen as well as
activity due to the South Atlantic Anomaly.
Data from satellite particle sensors can be fed to a model to indicate the geographical extent of the aurora. The images below show two different model outputs from the US Space Weather Prediction Center (SWPC) - both showing predicted auroral activity around the south polar region. The image at right was constructed from a single pass of a NOAA polar satellite using its particle precipitation data similar to that shown in the above graph. The satellite pass and the particle data is shown overlayed on the colour model.
GROUND-BASED AURORAL DETECTION
Radio Techniques
Radio waves can be used in several different ways to provide an auroral alert system.
HF Auroral Propagation Disturbance
High frequency (~ 10 to 15 MHz) signals that propagate through the polar regions, for instance from Brazil or Argentina to Perth, Australia, are subject to modification when they pass through the auroral oval.
When auroral activity is high the signal will be subject to rapid fading which is described as auroral flutter. If the aurora is really active the signal may be chopped up so much that it is unrecognisable.
HF/VHF Auroral Radar Reflections
An intense aurora will reflect radio signals. Thus an FM transmitter in Melbourne may be heard in Perth, Australia, when the auroral oval is enhanced and comes closer to the Australian continent. Note that reflection of 100 MHz is rather rare. More common is reflection of lower VHF (~50 MHz) frequencies.
The Australian JORN Over-The-Horizon radar system has even detected echoes from the northern hemisphere Aurora Borealis.
A research radar network called TIGER has been set up by a consortium led by Latrobe University. This mimics the SuperDARN radar networks of the northern hemisphere. Three radars comprise the network: one at Bruny Island in Tasmania, a second at Unwin in the south of the New Zealand south island, and a third at Buckland Park in South Australia. These radars sweep a range of azimuths to the south and with overlapping beams they can detect auroral activity. They normally operate around 12 or 14 MHz.
Data produced by the TIGER radar system.
Images: Physics & Engineering / Latrobe University
Both HF and VHF systems may be set up using continuous radio beacons to provide an auroral alert system.
Auroral Cosmic Noise Absorption
Ionisation in the D-region of the ionosphere (~70 - 90 km altitude) associated with the precipitating particles that produce the aurora will absorb cosmic radio signals transiting this region. Most of this cosmic noise comes from our galaxy. The absorption is measured by an instrument called a riometer (Relative Ionospheric Opacity meter). An imaging riometer can also plot the distribution of excess ionisation and thus auroral activity. The Southern Hemisphere Imaging Riometer Experiment (SHIRE) located at the Australian Antarctic Division base at Casey, and the data it produces it shown below.
SHIRE data. Image: SWS/WDC
Imaging
Ground based auroral cameras may be used to provide an indication of when auroral activity is occurring. Many times these cameras are based at remote locations and the images are relayed to a web site.
Two different types of cameras are used. If the auroral activity is likely to cover a significant portion of the sky an all sky camera is often employed.
 All sky camera (above) and all sky auroral image (right). Credits: OSU^ & SGO > |
|
When the site is located well to the north of expected auroral activity in the southern hemisphere, or to the south in the northern hemisphere, an horizon camera may be deployed. This condition is more often the case in the southern hemisphere where southerly land masses are not so common.
Keograms
A keogram is a time versus latitude plot of auroral activity and is a way of showing how visible auroral activity varies throughout the day. It is produced by taking a vertical slice through successive all sky auroral camera images, and putting then together to form a single image, as shown below.
Forming a keogram(above) Keogram from Svalbard, University of Oslo(right) |
 |
Specrographic Sensors
Auroral light is confined to quite narrow spectral lines in the visible, infrared and UV wavelengths. The figure below shows some of the auroral lines in the visible spectrum.
Of these lines two are generally the brightest. These are atomic oxygen lines at 577.7 nm and 630.0 nm as can be seen in the spectral intensity plot below.
These two lines can be used as the basis of a spectral alert system. Using two lines in coincidence can help reduce the possibility of false alarms.
Geomagnetic Sensors
Particles from the Sun not only precipitate into the atmosphere creating aurora, but they also increase the particle population in the Earth's magnetosphere. These increases cause a net decrease in the geomagnetic field, and they also cause large oscillations in the field. The larger the influx of particles the greater is the field decrease and the amplitude of the oscillations. These oscillations are a result of the magnetospheric particle population building up and then the dumping of these particles down toward the Earth (a little like flushing a cistern when it becomes full).
It is just this activity that produces both aurorae and geomagnetic activity (or storms). The correlation is not one hundred percent, particularly when the magnetic field sensors are located some distance from the auroral zone, but usually gives a reasonable indication of when auroral activity is occurring.
Geomagnetic activity is measured by magnetometers - instruments that measure variations in the Earth's magnetic field. The direct output of a magnetometer is termed a magnetogram. The figures below shows magnetograms on quiet and active days.
To eliminate diurnal and other long period variations in the geomagnetic field (which are not related to auroral activity), a filter is applied to the raw magnetic variation. This filter only passes variations that have periods less than 3 hours. Then, every 3 hours, a magnetic index called the K index is computed. This is a logarithm type index and ranges in value from K=0 (very quiet geomagnetic conditions) to K=9+ (a very large geomagnetic storm in progress).
Note that a logarithmic index has a constant multiplier between adjacent values. Thus if activity doubles from an index value of 1 to 2, it will also double between the index values of 8 and 9. The seismic Richter scale is another example of a logarithmic index.
Kp indices computed from Learmonth, Western Australia magnetograms
during a day when a minor magnetic storn was in progres. Images: SWS/BoM
In some countries, such as the US, where northern states tend to lie close to the quiet auroral oval, a K index of 5 is sometimes used as a threshold to indicate when aurorae may be visible. However, in Australia, the K index usually needs to be above 7 for there to be any chance of sighting an aurora (see the section below on Australian auroral alerts for more specific information).
Auroral models are often used in conjunction with magnetic indices to give a statistical estimate of the position of the auroral oval, and thus an indication of whether auroral activity may be visible from any particular location.
The output of an auroral model available on the Australian Space Weather Services web site for the southern hemisphere is shown below.
This model has only three inputs: the month, the time (UT) and the value of the K-index. Although the internationally defined K index only goes up to 9, during a very large storm the magnetic activity in fact can go well above 9 and this is indicated as 9+.
AURORAL ALERTS
Australian Auroral Alerts
Auroral alerts are issued by email and SMS message from the Space Weather Services of the Australian Bureau of Meteorology. For details see SWS Auroral Alerts. These alerts are based on intensity of geomagnetic activity reported by an Australian network of magnetometers.
SWS auroral alerts are first initiated when the K index reaches 7, and are re-iterated for values of K=8 and then K=9. A value of 7 usually indicates that people in Tasmania will see an aurora, whereas a value of 9 indicates that an aurora may be visible from Sydney or Perth.
As stated above the SWS web site also runs a model showing where the auroral oval is expected to be at any given time, and thus how close it may be to various parts of New Zealand and Australia.
RESOURCES
Auroral webcams
Auroral Forecasts and Alerts
Auroral Information
Australian Space Academy