Characterizing the distribution of extreme geoelectric field events in Sweden

Abstract

Historically, Sweden has reported several impacts on transformers and transmission lines related to geomagnetically induced currents (GICs) that develop during strong space weather events. GICs are driven by the geoelectric field (E), and their intensity depends on various factors, including the lithology conductivity and the rate of change of the Earth’s magnetic field. The purpose of this study is to perform an extreme value (EV) analysis of the E magnitude at six different latitudes in Sweden and to express the maximum |E| that might be observed in 10, 50, and 100 years. We analyzed 10-s E data in Sweden, obtained from a 1-D model. This model incorporates 10-s geomagnetic measurements from the IMAGE network and the vertical Earth’s ground electrical conductivity in Sweden, extracted from a 3-D conductance map for the Fennoscandian region. Extreme E events tend to occur in clusters around geomagnetic disturbances (substorms and geomagnetic storms). Therefore, we applied two different methods to decluster the data. After declustering, Generalized Pareto (GP) distributions were fitted to the remaining extreme events that exceeded the 99.5th percentile. The EV analysis indicates that the shape parameter of the GP distribution depends on latitude. This implies that at higher geographic latitudes (64.52–68.02°N) the distribution decreases faster toward zero than at lower latitudes (58.26–62.25°N). As a result the expected maximum |E| in 100 years in central Sweden ranges between 4.0 and 8.5 V/km, while at higher latitudes, it ranges between 2.0 and 2.5 V/km, similar to the modeled geoelectric field values during the Halloween event in October 2003. In particular, around 60.50°N the distribution of extreme events exhibits the heaviest tail. When we additionally consider the effect of conductivity, the region of west Sweden around 60.50°N exhibits the largest expected maximum in 100 years with a value around 8.5 V/km. This is three times larger than the maximum modeled |E| at that latitude.