Using the accumulated
73 years worth of Kp data, we create a statistical model of geomagnetic storm
occurrence. We define a geomagnetic storm as the occurrence of a Kp value
above a given threshold, which is preceded by at least 2 days (16 readings) below
the threshold. This eliminates the temporal clumping of Kp, which distorts the
We find the wait times between storm occurrences can be described by an exponential
distribution. This implies the arrival of geomagnetic storms is a Poisson process.
We then determine the solar cycle variation of the occurrence rate through
maximum likelihood fitting. The range of variation in occurrence rate is found to
depend on the threshold chosen, with Kp ≥ 5 storms varying only a few percent,
while Kp > 8 storms vary by more than a factor of 4. Comparing the occurrence rate
of sunspots to that of storms of various thresholds shows a dependence on threshold
level. Large storms are seen to occur earlier on the solar cycle than small storms.