摘要：We used data for eight magnetospheric spacecraft providing magnetic observations in various magnetospheric domains during a six-day time period, including the June 2015 storm, and a five-day period including the March 2015 storm. For these time intervals, containing different solar wind regimes and different activity levels, we used three types of metrics to compare predictions of all existing types of magnetospheric magnetic field models, including empirical models (TA15, TS05, and T96), a data-adapted model (AM03), and a global magnetohydrodynamic model (Space Weather Modeling Framework/Block Adaptive Tree Solar Wind–Roe–Upwind Scheme coupled with Rice Convection Model). In total, the models are ranked in the order: AM03, TS05, TA15, Space Weather Modeling Framework, and T96. The regional scores may differ from the average ones (in particular, better scores are obtained at GEO, compared to the inner magnetosphere) and also the model effectiveness varies with activity conditions (for example, TA15 outperforms other models in quiet conditions, but TS05 leads during active times). Quite unexpectedly, although run at modest resolution, the community available global magnetohydrodynamic model closely approached the best scores of empirical statistical models during the moderate/high disturbed periods in the June event, suggesting that this kind of modeling may now compete with other type models. We also carried out the field line mapping from fixed points in the equatorial magnetosphere to the ionosphere, using different models to examine the storm time foot point excursions (which may be as large as ~10° CGLat during storm time sudden commencements) and their uncertainties, quantified by the difference between the model predictions.
