Sudden Storm Commencement detection with SVM classifiers using ground magnetic data

When a cloud of plasma from a corona mass ejection hits the Earth's magnetosphere, a rapid perturbation of the magnetopause current systems occurs. It generates a signal that is detected worldwide by ground magnetic observatories in the form of a Sudden Storm Commencement (SSC). This signal has an amplitude of approximatively 20 nT but is similar to signals generated by other phenomena. Existing lists of SSCs had been set by human inspection of magnetic time series. We have implemented and tested a method to automatically detect SSC events using Support Vector Machines (SVM) classifiers within one-second data collected in the network of IPGP magnetic observatories.