top of page



Kp = 6                                                            Kp = 3


Northern Lights  

The Aurora Borealis (Northern Lights) and Aurora Australis (Southern Lights) are the result of electrons colliding with the upper reaches of Earth’s atmosphere. (Protons cause faint and diffuse aurora, usually not easily visible to the human eye.) The electrons are energized through acceleration processes in the downwind tail (night side) of the magnetosphere and at lower altitudes along auroral field lines. The accelerated electrons follow the magnetic field of Earth down to the Polar Regions where they collide with oxygen and nitrogen atoms and molecules in Earth’s upper atmosphere. In these collisions, the electrons transfer their energy to the atmosphere thus exciting the atoms and molecules to higher energy states. When they relax back down to lower energy states, they release their energy in the form of light. This is similar to how a neon light works. The aurora typically forms 80 to 500 km above Earth’s surface.

Earth’s magnetic field guides the electrons such that the aurora forms two ovals approximately centered at the magnetic poles. During major geomagnetic storms these ovals expand away from the poles such that aurora can be seen over most of the United States. Aurora comes in several different shapes. Often the auroral forms are made of many tall rays that look much like a curtain made of folds of cloth. During the evening, these rays can form arcs that stretch from horizon to horizon. Late in the evening, near midnight, the arcs often begin to twist and sway, just as if a wind were blowing on the curtains of light. At some point, the arcs may expand to fill the whole sky, moving rapidly and becoming very bright. This is the peak of what is called an auroral substorm.


The solar wind is be measured directly using a satellite positioned at the 1st Lagrange point. The solar wind  correlates closely to the sun spot activity. The sun rotates on a 27 +-3  day cycle. The sun is not solid so the rotation rate varies with local flow of the hot gas. 

Every 11 years or so, the Sun's magnetic field completely flips. This means that the Sun's north and south poles switch places. Then it takes about another 11 years for the Sun’s north and south poles to flip back again. The solar cycle affects activity on the surface of the Sun, such as sunspots which are caused by the Sun's magnetic fields. As the magnetic fields change, so does the amount of activity on the Sun's surface.  On the average there are 100 sun spots over a single side of the Sun at any one time. 

The appearance of an individual sunspot may last anywhere from a few days to a few months, though groups of sunspots and their associated active regions tend to last weeks or months. Sunspots expand and contract as they move across the surface of the Sun, with diameters ranging from 16 km (10 mi) to 160,000 km (100,000 mi).

The sun spot activity has a 11 year cycle, as can be seen the the 2 graphs.  imported as Kp2000 excel file.

The Northern Lights are tracked using the Kp Index. The Kp index is a three hour long quasi-logarithmic local index of the geomagnetic activity at the given location and time compared to a calm day curve. A magnetometer measures the maximum deviation of the horizontal component of the magnetic field at its location and reports this. The global Kp-index is then determined with an algorithm that puts the reported K-values of every station together. The Kp-index ranges from 0 to 9 where a value of 0 means that there is very little geomagnetic activity and a value of 9 means extreme geomagnetic storming.

There are large variations in Kp. By averaging out the earth rotation around sun using a 365 day rolling average, the underlying trends emerge. Kp rises with increasing sun spots, and then decays with a 2 year lag. The base Kp varies from 1 to 3. 

After removing the yearly cycle, a 27 day rolling average of the residual gives a monthly cycle of +- 1 Kp. This is the month to month variation in Kp, so the effect of rotation of the sun and the flipping of the suns magnetic field are eliminated. 

After removing yearly and monthly variations, the residual shows spikes in Kp of over 2.5 Kp that last for 1 day. 

An alternative approach is to find the maximum over a rolling 27 day window.  The spikes only last for a single months window, showing that the rotation of the sun does NOT appear to trigger sharp spikes in Kp in subsequent months. 


The Earth subtends an angle of 17 asecs when viewed from the sun (1/2 of the moon from earth). It also moves 1 degree every day roughly the same as the suns rotation.  A 3 hour window represents a 1/8 degree window, giving an scale to the outburst. The small angle subtended to a rapidly moving object supports the idea that spikes in aroura are caused by a "lighthouse beacon" directional outburst  from the sun that is around 10 -20 arcminutes in diameter.  Aligned burst events happen at rate of roughly 4 per year, using a threshold of 2.5 on the residual after yearly and monthly variation is removed.  Assuming 200 sun spots a month, gives 2,000 a year so 0.1% of the sunspots are aligned to earth. This is roughly the fractional angle of the bursts, suggesting that the 200 sun spots are randomly scattered across the sun. 


The data also shows some Kp spikes appearing 3-5 days BEFORE the max in solar flux as measured at the first Lagrange point (L1).  For stability, the spacecraft orbits the L1 point with a 100,000 miles diameter orbit. This is roughly  2 arc mins away from earth. This is the reason  that solar flux will lead or lag observed Kp by a few days. 

Forecasting the the aurora on different time scales can be done in different ways. The maximum in geomagnetic disturbance (Kp) lasts 1-2 nights. 

  • 15-45 Minutes: By measuring the solar wind and interplanetary magnetic field upstream of Earth it is possible to forecast the aurora quite accurately but only with a short lead time of 15-45 minutes. There is a location about 1.5 million km (1 million miles) from Earth towards the sun where gravitational forces and centripetal forces balance and satellites can remain almost stationary without expending fuel to stay in place. This location is called the 1st Lagrange orbital location (L1). This location is upstream of Earth in the solar wind. There are several satellites at this L1 location that provide measurements of the density, velocity, and magnetic field of the solar wind. These measurements provide a very accurate prediction of what the solar wind will be like a few tens of minutes later near Earth. These measurements provide a capability for accurate, short term aurora forecasts.  

  • Hours to Days: It is possible to predict geomagnetic activity and aurora a day or so in advance by detecting solar coronal holes on the sun and Coronal Mass Ejections (CME) near the sun.  as these coronal holes or CMES leave the sun, it is possible to predict their path to determine if they will impact Earth. This technique is more challenging and the accuracy of these medium range forecasts is not as good. It is often difficult to predict which direction the CME is going and how fast it is traveling. And knowledge of the strength and direction of the magnetic field within the CME introduces a big uncertainty in how strong the geomagnetic storm will be.. The primary sensor used for CME prediction is the Solar Coronagraph which measures the outflow of solar wind close to the sun. But the coronagraphs cannot measure the magnetic field within the CME. Measurements of coronal holes are made by solar imagers such as the SUVI instrument on the NOAA GOES satellite.

  • 27 Days: Active regions and Coronal Holes can last for many months and as the sun rotates, these active regions will repeatedly be directed towards Earth. They often produce similar levels of activity from one solar rotation the next. So if there was geomagnetic activity and aurora 27 days ago, there is a good chance that there will be aurora today… plus or minus a day or two.  The equator of the sun rotates at about 25 days. In a sixty degrees period, it rotates around about 30 days while at the pole it rotates around 36 days. One will notice that the rotation takes longer the farther it is from the equator. Unlike Earth, the sun is not solid. In fact, it is made up of gases and plasma. (

  • Years: Solar activity waxes and wanes on an approximate 11-year cycle. Thus, geomagnetic activity and aurora also follow an approximate 11 year cycle. During the 4-5 years near solar maximum, there are more solar active regions, larger solar flares, and larger and faster CMEs. However, there are some caveats to this 11 year forecast. Big solar cycles (with lots of sunspots) tend to produce more active regions and thus more auroral activity. The biggest solar flares and CME’s often occur near the end of the solar maximum period as the solar activity starts to wane. And even during solar minimum, the high speed solar wind streams from coronal holes can produce moderate geomagnetic activity (up to a Kp of 7) and thus, moderate auroral activity. Solar flux correlates with sun spot activity. 

  • "Middle latitudes" With reference to zones of geomagnetic activity, 20 degrees to 50 degrees geomagnetic latitude, Anchorage is 61 degrees north.  Other zones are equatorial, polar, and high latitude.

There is a general increase in solar activity associated with the flipping of the suns magnetic field. There is a random month to month variation. Some of the solar events emit a  narrow plume of electrons  roughly 10-20 arcmins in diameter, which occasionally hit the earth  (17 arcsecs in diameter, travelling at 1 degree a day roughly the same as earth around the sun).  During increased sun spot cycles, these events give rise to spectacular arouras with Kp >5 that typically last 1 day. These focused events mostly decay over the 27 days for the sun to rotate. Earth aligned events  occur at around 4 per year, 1-2 of them during dark skies. 

Wang-Sheeley-Arge (WSA)-Enlil is a large-scale, physics-based prediction model of the heliosphere, used by the Space Weather Prediction Center to provide 1-4 day advance warning of solar wind structures and Earth-directed coronal mass ejections (CMEs) that cause geomagnetic storms.

The modeling system consists of two main parts: 1) a semi-empirical near-Sun module that approximates the outflow at the base of the solar wind; and 2) a sophisticated 3‑D magnetohydrodynamic numerical model that simulates the resulting flow evolution out to Earth. The former module is driven by observations of the solar surface magnetic field, as taken over a solar rotation and composited into a synoptic map; this input is used to drive a parameterized near-Sun expansion of the solar corona, which is subsequently input into the second, interplanetary module to compute the quasi-steady (ambient) solar wind outflow. Finally, when an Earth-directed CME is detected, coronagraph images from NASA spacecraft are used to characterize the basic properties of the CME, including timing, location, direction, and speed. This input (the “cone” model) is injected into the pre-existing ambient conditions, and the subsequent transient evolution forms the basis for the prediction of the CME arrival time at Earth, its intensity, and its duration.

In the movie, the Sun is represented as a yellow dot, the Earth by a green dot, and the STEREO spacecraft by the red and blue dots.  The top row represents the WSA-Enlil predicted solar wind density and the bottom row the predicted solar wind velocity.  On the left is a pinwheel plot of the ecliptic plane, showing all of the solar wind structures that are likely to encounter Earth or which have recently encountered Earth, in what is effectively an 'overhead' view.  While the STEREO spacecraft are shown, this ecliptic slice does not normally pass through these satellites, though it is typically fairly close.  In the middle are meridional slices that go through the Earth, showing the solar wind structures that will encounter Earth from a 'side' view.  On the right, the predicted density and velocity values for the location of Earth and the two STEREO spacecraft are plotted. 

Early August, 4 plumes - in a 28 day cycle, one every 7 days on the average. Sudden plume forecast on 8/1/23 was a bust, 2 day forecast early by 1 day,  1 day forecast early by 6 hrs.

Long term forecast covers the next month. For best viewing the outburst needs to happen on a dark sky night while the moon is hidden.

That is the aroura lottery. 

2023-06-18 (1).png
Kp monthly.jpg
Screenshot (17).png

Actual max occurred July 14 - one day late  on a 1 month forecast.

Screenshot 2023-07-29 19.47.02.png

3 day forecast. UTC is close to GMT. Anchorage is UTC - 8, so 6am is UTC 14:00, 6pm is UTC = 18+8 = 2:00.   Sunset at 06:00 UTC 

Screenshot (43).png
Screenshot (45).png
bottom of page