We created a database of over 100 simulations of geomagnetic storms from the time interval from 2010 to 2019, using the Geospace configuration of the Space Weather Modeling Framework. With this data set, we explore and quantify the errors that arise from the uncertainties of the input solar wind data set that comes from the OMNI database, using global magnetic indices such as SYM-H, AL and Cross Polar Cap Potential as measures of the model output that can be directly compared with observations. Our results show that the method of propagation of the solar wind data from the spacecraft location near L1 to the bow shock nose causes errors that are much smaller than those arising from the limitations in modeling the solar wind – magnetosphere – ionosphere coupling processes during strong solar wind driving. We devise a method to quantify these uncertainties, which may become useful in understanding the output of global simulations. These results will be useful as well for research investigations of magnetospheric processes as well as for forecasting space weather.
