Using neutron monitor network data to improve the detection of space weather events

Neutron and muon monitor networks can be used as single multidirectional spectrograph, thereby improving the detection of Ground Level Enhancements (GLEs) and of signatures from approaching interplanetary disturbances (Usoskin et al., 1997; Mavromichalaki et al., 2004). The real-time analysis of such network data, however, is often hampered by a number of practical difficulties such as data gaps and noise. Here we show how multivariate statistical analysis methods known as blind source separation methods can help rapidly characterize network data while alleviating some of these problems.