Decades of studies have suggested several criteria to detect Interplanetary coronal mass ejections (ICME) in time series from in-situ spacecraft measurements. Among them the most common are an enhanced and smoothly rotating magnetic field, a low proton temperature and a low plasma beta. However, these features are not all observed for each ICME due to their strong variability. Visual detection is time-consuming and biased by the observer interpretation leading to non exhaustive, subjective and thus hardly reproducible catalogs. Using convolutional neural networks on sliding windows and peak detection, we provide a fast, automatic and multi-scale detection of ICMEs. The method has been tested on the in-situ data from WIND between 1997 and 2015 and on the 657 ICMEs that were recorded during this period. The method offers an unambiguous visual proxy of ICMEs that gives an interpretation of the data similar to what an expert observer would give. We found at a maximum 197 of the 232 ICMEs of the 2010-2015 period (recall 84 +-4.5 % including 90% of the ICMEs present in the lists of Nieves-Chinchilla et al. (2015) and Chi et al. (2016). The minimal number of False Positives was 25 out of 158 predicted ICMEs (precision 84+-2.6%). Although less accurate, the method also works with one or several missing input parameters. The method has the advantage of improving its performance by just increasing the amount of input data. The generality of the method paves the way for automatic detection of many different event signatures in spacecraft in-situ measurements.

Taking the 32 storm sudden commencements (SSCs) listed by ISGI during 2002 as a starting point, we performed a multi-criterion analysis based on observations (propagation time, velocity comparisons, sense of the magnetic field rotation, radio waves) to associate them with solar sources. We identified their effects in the interplanetary medium, and looked at the response of the terrestrial ionized and neutral environment. We find that 28 SSCs can be related to 44 coronal mass ejections (CMEs), 15 with a unique CME and 13 with a series of multiple CMEs, among which 19 involved halo CMEs; 12 of the 19 fastest CMEs with speeds greater than 1000 km/s are halo CMEs. The probability for an SSC to occur is 75% if the CME is a halo CME. The complex interactions between two CMEs and the modification of their trajectories have been examined using joint white-light and multiple-wavelength radio observations. The solar-wind structures at L1 after the shocks leading the 32 SSCs are 12 magnetic clouds (MCs), 6 interplanetary CMEs (ICMEs) without an MC structure, 4 miscellaneous structures, which cannot unambiguously be classified as ICMEs, 5 corotating or stream interaction regions (CIRs/SIRs), and 4 isolated shock events; note than one CIR caused two SSCs. The 11 MCs listed in 3 or more MC catalogs covering the year 2002 are associated with SSCs. For the 3 most intense geomagnetic storms related to MCs, we note 2 sudden increases of the Dst, at the arrival of the sheath and the arrival of the MC itself. The most geoeffective events are MCs, since 92% of them trigger moderate or intense storms, followed by ICMEs (33%). At best, CIRs/SIRs only cause weak storms. We show that these geoeffective events (ICMEs or MCs) trigger or reinforce terrestrial radiowave activity in the magnetosphere, an enhanced convection in the ionosphere, and a stronger response in the thermosphere.