The impact of magnetic field temporal sampling on modeled surface electric fields

Descriptions of the surface magnetic-to-electric field relationship, using magnetotelluric impedance tensors, have seen refinement in recent years. Correspondingly, models of geomagnetically induced currents flowing in power line conductors using input from more fundamental observables (e.g., the surface magnetic field or solar wind drivers) have become more sophisticated. Magnetometer data used for purposes of surface electric field modeling typically have cadences between 1 s and 1 min. Using 1-D and 3-D magnetotelluric impedance tensors, we quantify the effect of magnetometer cadence on modeled peak surface electric fields during the 22 June 2015, 17 March 2015, and 2 October 2013 geomagnetic storms. It is shown that 1-min magnetometer data sampling leads to median biases in the peak electric field of 10-20% magnitude attenuation and 2-17 degrees in direction (depending on the storm) across the United States using the Fernberg 1-D conductivity models and 11-26% (0.9-8 degrees) using the 3-D impedance tensors derived from the USArray magnetotelluric survey. The largest 10% of attenuations almost entirely occur near low-resistivity subsurface structures in the western and eastern United States, reaching a 50% median attenuation during the 17 March 2015 event. It is concluded that cadence, local geology, and the spectral content of the geomagnetic field all play a significant role in the bias.