Effects of Small Scale Structure on Predictive Performance of the Space Weather Modeling Framework

Numerical models of the geospace environment, such as the operational configuration of the Space Weather Modeling Framework, require the solar wind and interplanetary magnetic field to be specified at the upstream boundary. For both science and operations, this boundary condition is usually taken from a solar wind monitor near L1, such as ACE or DSCOVR, and specified at 1-minute cadence or better. However, there are a number of circumstances in which this temporal resolution is not available, such as data gaps due to instrumental effects; historical events that do not have complete solar wind coverage; and synthetic events used for benchmarking and hypothesis testing, which are often specified as idealized drivers. In each of these cases, small-scale structure is missing from the upstream boundary condition driving the model. The extent to which the sub-hourly solar wind structure impacts the predictive performance of geospace models has not previously been examined. Using a set of 5 events that have been the basis of previous validation efforts, we simulate each event multiple times, progressively filtering and downsampling the driver time series to 1, 15, 30, and 60 minute cadences. We assess the change in model performance at high- and mid-latitude, as well as in local time. We quantify the reduction in predictive skill, and use a comprehensive validation suite to understand how the model performance differs for regions that are predominantly directly-driven versus those that are largely internally-driven.