The Equinox Effect: Why Aurora Peaks in March and September

What 94 years of data actually show

The Ap index is the daily planetary measure of how disturbed Earth's magnetic field was, published by GFZ Potsdam and running unbroken since 1932. Averaging it by calendar month across those 94 complete years gives this:

March is the strongest month at 19.4% above the annual mean, September second at 17.5%. December is the weakest at 20.4% below it. Group them and the pattern is blunt: the four equinox-adjacent months (March, April, September, October) average an Ap of 15.5, the four solstice-adjacent months (June, July, December, January) average 11.5. That is 35% more geomagnetic activity around the equinoxes.

Is it real, or is it an artefact of averaging?

A 94-year average can hide a lot. A handful of freak years could produce that shape on their own, so the honest test is whether it repeats. We took each year separately and compared its equinox months against its solstice months.

It held in 88 of the 94 years. The six that went the other way were 1961, 1967, 1980, 2004, 2005 and 2013. A paired t-test across years gives t = 11.7 on 93 degrees of freedom, which is not a marginal result: it is the kind of number that leaves no room for coincidence. Whatever is causing this is switched on almost every year.

Nobody has settled why it happens

Here is the part that popular aurora writing almost always skips. The pattern has been known since Sabine noticed it in 1852. The cause is still argued over. Three mechanisms compete, and all three predict a peak near the equinoxes:

• Axial. Earth's heliographic latitude swings by ±7.25° over the year, peaking in early March and early September. Near those dates we sit over more active solar latitudes and catch a faster, denser solar wind.
• Equinoctial. The angle between the Earth-Sun line and Earth's magnetic dipole changes through the year, and that angle governs how efficiently the solar wind couples to the magnetosphere. Coupling is most efficient near the equinoxes.
• Russell-McPherron. The Sun's magnetic field reaches us wound into a spiral. Near the equinoxes, the sideways component of that field projects into a southward component in the coordinate frame that matters for reconnection, so an otherwise unremarkable solar wind behaves like a geoeffective one.

The three peak within about a month of each other, roughly 7 March to 7 April. That is the whole problem: a century of data still cannot cleanly separate predictions that close together. Russell-McPherron is the one popular articles name, usually with more confidence than the literature supports, and its own amplitude falls short of the observed variation by roughly a factor of four. Cliver and colleagues argued the equinoctial mechanism dominates; later work by Lockwood and others finds a mixture. It is genuinely open.

What it means for planning a trip

Less than the 35% figure suggests, because geomagnetic activity is only one of the things a night needs. The equinoxes give you more activity but shorter darkness than midwinter, and the two pull against each other. Which one wins depends on where you are standing.

That is why our month-by-month curves are calculated per location rather than published as one national rule. Each combines the measured activity above with that location's own darkness hours and ten years of its own cloud record. The answers diverge: Tromsø and the Greenland destinations peak in October, most of Canada in March, Svalbard not until February because it stays dark enough that late in the season. Compare the destinations in Norway, Canada, Sweden, Finland and Greenland.

The short version

Equinox months really do run about 35% more active, it really does repeat almost every year, and nobody can tell you with authority why. Treat it as a tilt in your favour when you are choosing between two workable months, not as a reason to override darkness, cloud or the moon. Those decide far more nights than the calendar does.