Accurate forecasting of the properties of coronal mass ejections (CMEs) as they approach Earth is now recognized as an important strategic objective for both NOAA and NASA. The time of arrival of such events is a key parameter, one that had been anticipated to be relatively straightforward to constrain. In this study, we analyze forecasts submitted to the Community Coordinated Modeling Center at NASA's Goddard Space Flight Center over the last 6 years to answer the following questions: (1) How well do these models forecast the arrival time of CME‐driven shocks? (2) What are the uncertainties associated with these forecasts? (3) Which model(s) perform best? (4) Have the models become more accurate during the past 6 years? We analyze all forecasts made by 32 models from 2013 through mid‐2018, and additionally focus on 28 events, all of which were forecasted by six models. We find that the models are generally able to predict CME‐shock arrival times—in an average sense—to within ±10 hr, but with standard deviations often exceeding 20 hr. The best performers, on the other hand, maintained a mean error (bias) of −1 hr, a mean absolute error of 13 hr, and a precision (standard deviation) of 15 hr. Finally, there is no evidence that the forecasts have become more accurate during this interval. We discuss the intrinsic simplifications of the various models analyzed, the limitations of this investigation, and suggest possible paths to improve these forecasts in the future.

Shock waves are ubiquitous in astrophysics and interplanetary space. In collisionless plasmas they transform directed flow energy into thermal energy and accelerate energetic particles. The energy repartition amongst particle populations is a multi-scale process related to the spatial and temporal structure of the electromagnetic fields within the shock layer. While major features of the large scale ion heating are known, the electron heating and smaller scale fields remain poorly understood and controversial. We determine for the first time the scale of the electron temperature gradient via unprecedented high time resolution electron distributions measured in situ by the Cluster spacecraft. We discover that half of the electron heating coincides with a narrow dispersive layer several electron inertial lengths (c/ωpe) thick. Consequently, the nonlinear steepening is limited by wave dispersion. The DC electric field associated with the electron pressure gradient must also vary over these small scales, strongly influencing the efficiency of shocks as cosmic ray accelerators.

[1] Geomagnetic activity has long been known to exhibit approximately 27 day periodicity, resulting from solar wind structures repeating each solar rotation. Thus a very simple near-Earth solar wind forecast is 27 day persistence, wherein the near-Earth solar wind conditions today are assumed to be identical to those 27 days previously. Effective use of such a persistence model as a forecast tool, however, requires the performance and uncertainty to be fully characterized. The first half of this study determines which solar wind parameters can be reliably forecast by persistence and how the forecast skill varies with the solar cycle. The second half of the study shows how persistence can provide a useful benchmark for more sophisticated forecast schemes, namely physics-based numerical models. Point-by-point assessment methods, such as correlation and mean-square error, find persistence skill comparable to numerical models during solar minimum, despite the 27 day lead time of persistence forecasts, versus 2-5 days for numerical schemes. At solar maximum, however, the dynamic nature of the corona means 27 day persistence is no longer a good approximation and skill scores suggest persistence is out-performed by numerical models for almost all solar wind parameters. But point-by-point assessment techniques are not always a reliable indicator of usefulness as a forecast tool. An event-based assessment method, which focusses key solar wind structures, finds persistence to be the most valuable forecast throughout the solar cycle. This reiterates the fact that the means of assessing the "best" forecast model must be specifically tailored to its intended use.

[1] Collisionless shock waves are a widespread phenomenon in both solar system and astrophysical contexts. The nature of energy dissipation at such shocks is of particular interest, especially at high Mach numbers. We use data taken by the Cassini spacecraft to investigate electron heating at Saturn's bow shock, one of the strongest collisionless shocks encountered by spacecraft to date. Measurements of the upstream solar wind ion parameters are scarce due to spacecraft pointing constraints and the absence of an upstream monitor. To address this, we use solar wind speed predictions from the Michigan Solar Wind Model. Since these model predictions are based on near-Earth solar wind measurements, we restrict our analysis to bow shock crossings made by Cassini within ±75 days of apparent opposition of Earth and Saturn. An analysis of the resulting set of 94 crossings made in 2005 and 2007 reveals a positive correlation between the electron temperature increase across the shock and the kinetic energy of an incident proton, where electron heating accounts for between ∼3% and ∼7% of this incident ram energy. This percentage decreases with increasing Alfvén Mach number, a trend that we confirm continues into the hitherto poorly explored high-Mach number regime, up to an Alfvén Mach number of ∼150. This work reveals that further studies of the Saturnian bow shock will bridge the gap between the more modest Mach numbers encountered in near-Earth space and more exotic astrophysical regimes where shock processes play central roles.

This study demonstrates two significant ways of improving persistence forecasts of the solar wind, which exploit the relatively unchanging nature of the ambient solar wind to provide 27 day forecasts, when using data from the Lagrangian L1 point. Such forecasts are useful as a prediction tool for the ambient wind, and for benchmarking of solar wind models. We show that solar wind persistence forecasts can be improved by removing transient solar wind features such as coronal mass ejections (CMEs). Using CME indicators to automatically identify CME‐contaminated periods in ACE data from 1998 to 2011, and replacing these with solar wind from a previous synodic rotation, persistence forecasts improve (relative to a baseline): skill scores for Bz, a crucial parameter for determining solar wind geoeffectiveness, improve by 7.7 percentage points when using a proton temperature‐based indicator with good operational potential. We also show that persistence forecasts can be improved by using measurements away from L1, to reduce the requirement on coronal stability for an entire synodic period, at the cost of reduced lead time. Using STEREO‐B data from 2007 to 2013 to create such a reduced lead time persistence forecast, we show that Bz skill scores improve by 17.1 percentage points relative to ACE. Finally, we report on implications for persistence forecasts from any future missions to the L5 Lagrangian point and on the successful operational implementation (in spring 2015) of the normal (ACE‐based) and reduced lead time (STEREO‐based) persistence forecasts in the Met Office's Space Weather Operations Centre, as well as plans for future improvements.