Characteristics of Geomagnetic Storms: Comparison of Community Storm Databases and Storm Statistical Properties

While there are generally accepted guidelines for geomagnetic storm identification and categorization, the storm databases currently available do not necessarily use the same criteria for defining various storm characteristics. It is important to identify the various rules that different storm databases follow in order to allow for more accurate comparisons of storm properties; this is especially important for placing other geospace events in the context of storms, and knowing that these "places" within storms (phase distinctions) are defined similarly across databases. In this study, we focused on identifying how different databases define the temporal characteristics of storms, first comparing one specific storm characteristic across multiple databases, then using the most comprehensive database to compare multiple storm characteristics. The initial databases examined were the Kakioka Geomagnetic Event Catalogs, the International Service of Geomagnetic Indices (ISGI), the Space Weather Database Of Notifications, Knowledge, Information (DONKI), and World Data Center for Geomagnetism, Kyoto. We narrowed these down to Kakioka and ISGI as the most complete of the four, and used them to directly compare key storm database features: start time, storm classification, and phase length, as well as to determine their criteria for storm occurrence. Once these features were identified, the geomagnetic storm data over one solar cycle from the same databases were examined and the main phase durations of each storm were compared. This comparison revealed a slight discrepancy in the definitions of main phase start and end times. Next, using just the Kakioka database, the initial and main phase durations of storms over several years were examined in order to identify a pattern in these phase lengths, which could then be compared with timing within the solar cycle to reveal a trend in phase durations, where most storms seem to have similar durations of main and initial phases. Possible explanations for database discrepancies may be identified, as well as possible further applications for the revealed trends.