Uncertainty Estimates of Solar Wind Prediction Using HMI Photospheric Vector and Spatial Standard Deviation Synoptic Maps

Abstract

Current solar wind prediction is based on the Wang & Sheeley empirical relationship between the solar wind speed observed at 1 AU and the rate of magnetic flux tube expansion (FTE) between the photosphere and the inner corona, where FTE is computed by coronal models that take the photospheric flux density synoptic maps as their inner boundary conditions to extrapolate the photospheric magnetic fields to deduce the coronal and the heliospheric magnetic field configuration. Since these synoptic maps are among the most widely-used of all solar magnetic data products, the uncertainties in the model predictions that are caused by the uncertainties in the synoptic maps are worthy of study. However, such an estimate related to synoptic map construction was not available until Bertello et al. (Solar Physics, 289, 2014) obtained the spatial standard deviation synoptic maps; 98 Monte-Carlo realizations of the spatial standard deviation maps for each photospheric synoptic maps. In this paper, we present an estimate of uncertainties in the solar wind speed predicted at 1 AU by the CSSS model due to the uncertainties in the photospheric synoptic maps. We also present a comparison of the coronal hole locations predicted by the models with the STEREO/SECCHI EUV synoptic maps. In order to quantify the extent of the uncertainties involved, we compared the predicted speeds with the OMNI solar wind data during the same period (taking the solar wind transit time into account) and obtained the root mean square error between them. To illustrate the significance of the uncertainty estimate in the solar wind prediction, we carried out the analysis for three Carrington rotations, CR 2102, CR 2137 and CR 2160 at different phases of the solar cycle. The uncertainty estimate is critical information necessary for the current and future efforts of improving the solar wind prediction accuracies.