Impact of model resolution on predictions of ground magnetometer response to high-speed stream events

High-speed streams originating from coronal holes on the Sun are a significant source of driving modest geomagnetic storms. The currents flowing through the tightly coupled magnetosphere-ionosphere (MI) system can be quite large during these events. These changes in these currents can cause geomagnetic induced currents (GICs) that impact the operation of power grids. Networks of ground magnetometer stations are routinely used to monitor the evolution of these current systems. With the adoption of the first physics-based geospace model, space weather predictions are beginning to rely on these techniques for making forecasts. In this study we assess the impact of model resolution on the quality of the ground magnetometer predictions for a series of high-speed stream events. The numerical model used here is the Lyon-Fedder-Mobarry (LFM) global magnetohydrodynamic magnetosphere simulation. Recently we have developed a python based set of analysis tools that includes a Biot-Savart integrator for determining the magnetic field perturbations caused by the currents flowing in geospace. The tool can isolate the contributions for horizontal currents flowing in the ionosphere, field aligned currents flowing between the ionosphere and magnetosphere, and the currents flowing throughout the computational domain of the magnetosphere simulation. As part of an earlier study the LFM was used to simulate the interaction of geospace with two high-speed streams that where part of the Whole Heliosphere Interval (WHI). The WHI covers an entire solar rotation or Carrington Rotation, providing solar wind driving for 27 days. We repeat that simulation with two additional runs with the resolution doubled for each run. The combined set of the three WHI simulations provides robust data set over nearly an order of magnitude in computational resolution with broad range of driving conditions. Using the Biot-Savart integrator and magnetometer data we compute a variety of metrics ranging from simple RMS error to more advance spectral characteristics to quantitatively assess the impact of model resolution on LFM's ability to provide forecasts of ground magnetic field perturbations.