An Automated Procedure for Near-real-time Kp Estimates

The Kp index indicates geomagnetic disturbances in a simple manner. The index is derived from magnetic field data acquired at 11 ground stations distributed worldwide at subauroral latitudes (49°–62°), has values ranging from 00 to 90 in 28 steps, and is given at 3-hour intervals. Kp is widely used to study the dynamic relationship between the solar wind and the magnetosphere, to empirically specify the location of the plasmapause and other plasma regions/boundaries, and also as input to various models of the magnetosphere and ionosphere. Currently, the official Kp index is delivered with a delay of many days, so it is not useful for near-real-time monitoring of the state of the magnetosphere. We have developed an algorithm to derive an estimated Kp, denoted Kpest, using magnetometer data from nine ground stations that can transmit data in near real time. The algorithm is fully automated and includes a data-cleaning routine, a quiet-day-curve routine, and a routine to convert magnetic field deviations to Kpest. We have evaluated the performance of Kpest using archived magnetometer data and the official Kp. When data from all of the nine stations are available, the linear correlation coefficient between Kpest and Kp is 0.93. In addition, we find that a similarly high correlation between Kpest and Kp can occur when data from only one to three stations are used. We conclude that an automated procedure using data from a small number of ground stations can generate Kp estimates that are reliable in the context of near-real-time monitoring of space weather.