Probabilistic solar wind forecasting using large ensembles of near-Sun conditions with a simple one-dimensional “upwind” scheme

Owens, M. J. ORCID: https://orcid.org/0000-0003-2061-2453 and Riley, P. (2017) Probabilistic solar wind forecasting using large ensembles of near-Sun conditions with a simple one-dimensional “upwind” scheme. Space Weather, 15 (11). pp. 1461-1474. ISSN 1542-7390 doi: 10.1002/2017SW001679

Abstract/Summary

Long lead-time space-weather forecasting requires accurate prediction of the near-Earth solar wind. The current state-of-the-art uses a coronal model to extrapolate the observed photospheric magnetic field to the upper corona, where it is related to solar wind speed through empirical relations. These near-Sun solar wind and magnetic field conditions provide the inner boundary condition to three-dimensional numerical magnetohydrodynamic (MHD) models of the heliosphere out to 1 AU. This physics-based approach can capture dynamic processes within the solar wind, which affect the resulting conditions in near-Earth space. However, this deterministic approach lacks a quantification of forecast uncertainty. Here, we describe a complementary method to exploit the near-Sun solar-wind information produced by coronal models and provide a quantitative estimate of forecast uncertainty. By sampling the near-Sun solar wind speed at a range of latitudes about the sub-Earth point, we produce a large ensemble (N = 576) of time series at the base of the Sun-Earth line. Propagating these conditions to Earth by a three-dimensional MHD model would be computationally prohibitive, thus a computationally-efficient one-dimensional “upwind” scheme is used. The variance in the resulting near-Earth solar wind speed ensemble is shown to provide an accurate measure of the forecast uncertainty. Applying this technique over 1996-2016, the upwind ensemble is found to provide a more “actionable” forecast than a single deterministic forecast; potential economic value is increased for all operational scenarios, but particularly when false alarms are important (i.e., where the cost of taking mitigating action is relatively large).

Item Type	Article
URI	https://reading-clone.eprints-hosting.org/id/eprint/73315
Identification Number/DOI	10.1002/2017SW001679
Refereed	Yes
Divisions	Science > School of Mathematical, Physical and Computational Sciences > Department of Meteorology
Publisher	American Geophysical Union
Download/View statistics	View download statistics for this item