Wang–Sheeley–Arge model

The Wang–Sheeley–Arge model (WSA model) is a semi-empirical model of the solar corona and ambient solar wind used to forecast solar wind speed and interplanetary magnetic field polarity. The model connects observations of magnetic fields in the photosphere to solar wind conditions near the Sun through potential field extrapolation and an empirical speed relationship. It serves as the coronal component for several heliospheric propagation models used in operational space weather forecasting, most notably the coupled WSA–ENLIL system operated by the U.S. National Oceanic and Atmospheric Administration's Space Weather Prediction Center.^[1][2][3]

The WSA model extends work by American solar physicists Y.-M. Wang and N. R. Sheeley Jr., who demonstrated that solar wind speed at 1 AU varies inversely with the magnetic flux-tube expansion between the photosphere and a spherical source surface in the corona.^[4] Their result built on the potential field source surface method of Martin D. Altschuler and Gordon Newkirk Jr. and on models that introduced a heliospheric current sheet, which provided a global coronal magnetic topology consistent with interplanetary observations.^[5][6]

American space physicist C. Nicholas Arge and colleagues generalized the Wang–Sheeley approach for operational forecasting. They introduced an additional predictor, the minimum angular distance of an open-field footpoint to the nearest coronal hole boundary, and combined it with the flux-tube expansion factor to better separate slow and fast streams.^[7][8]

In Sheeley's later historical account, he described the model's origins as beginning with "an attempt to understand why the solar wind speed was anti-correlated with the expansion factor of a coronal flux tube." At the time, he noted, "there was no single model that could relate coronal magnetic geometry to the observed distribution of solar wind speeds." The breakthrough came when "Yi-Ming Wang and I realized that the open-field regions producing the fast wind corresponded to areas of small flux-tube expansion," with the correlation emerging from potential-field extrapolations of photospheric magnetograms and interplanetary scintillation measurements. Arge's later contribution was to modify the Wang–Sheeley model so that the predicted solar wind speed matched spacecraft observations on a day-to-day basis, transforming what Sheeley characterized as beginning as "an empirical curiosity" into "a cornerstone of space-weather prediction" and "a procedure for deriving" solar wind speed from magnetogram-based coronal models developed to serve operational forecasting needs.^[1]

WSA begins with a current-free potential field extrapolation from the line-of-sight photospheric field to a spherical source surface, typically near 2.5 R_☉ solar radii. It then continues the solution into the outer corona with a Schatten current sheet representation of the heliospheric current sheet. This coupled PFSS–SCS approach maps open magnetic field regions and their connectivity to the source surface.^[5][6][9]

For each open field line, the model computes the flux-tube expansion factor, which compares the radial magnetic field strength between the photosphere and the source surface scaled by the square of the radii. It also evaluates the minimum angular distance to the nearest coronal-hole boundary at the footpoint. The WSA speed relation assigns a terminal solar wind speed near the outer coronal boundary as a function of these two quantities, with coefficients tuned to observations and to the chosen magnetic maps.^[4][7][8]

The WSA coronal solution is usually defined on an outer boundary between about 5 and 30 R_☉ solar radii. Forecast systems then propagate this solution through the inner heliosphere using either a kinematic upwind scheme, the Heliospheric Upwind eXtrapolation method, or a full magnetohydrodynamic model, to obtain wind time series at spacecraft and planets.^[10][11] The empirical WSA speed law has appeared in several equivalent forms in the literature that combine a decreasing function of expansion factor with an increasing function of coronal-hole boundary distance. Coefficients are re-estimated for different magnetogram sources, assimilation schemes, and heights of the source surface.^[7][8][12]

NOAA's Space Weather Prediction Center has used the coupled WSA–ENLIL+Cone system in operations since 2011 to provide 1 to 4 day warnings of solar wind structures and Earth-directed coronal mass ejections.^[2][3] NASA's Community Coordinated Modeling Center hosts WSA and WSA–ENLIL for runs-on-request and provides model documentation and versioning.^[13] In 2019 NOAA reported an operational upgrade of the WSA–ENLIL system to version 2.0.^[14] An open-data archive of operational outputs supports evaluation and downstream applications.^[15]

Independent studies have shown that WSA-based forecasts capture the large-scale wind structure, with typical root-mean-square speed errors near 100 km s−1 and high-speed stream arrival-time uncertainties near one day over multi-year samples.^[16] A community assessment characterized coupled WSA–ENLIL as "one of the workhorse models" for both research and forecasting, while noting that skill depends strongly on the magnetic boundary conditions and on the treatment of near-Sun acceleration.^[17]

WSA relies on global magnetic maps as its primary input, and the Air Force Data Assimilative Photospheric Flux Transport model produces ensemble maps by assimilating magnetograms with a localized ensemble Kalman filter, which are widely used to drive WSA and to quantify boundary uncertainties.^[18] Propagation schemes based on the Heliospheric Upwind eXtrapolation approach and its tunable variants are used to map WSA output to 1 AU with controlled complexity for ensemble forecasting.^[11]

Recent research applied machine learning techniques to improve WSA model performance by re-estimating model parameters or emulating the speed mapping while retaining the underlying physical predictors. These approaches demonstrated significant improvements over traditional fixed-coefficient formulations.^[19] Additionally, researchers investigated data-assimilative methods for calibrating the source-surface height and solar wind outflow parameters to optimize model accuracy.^[20]

Researchers have debated which predictor drives most of the WSA skill. British-American space physicist Pete Riley and colleagues concluded that expansion factor plays "only an indirect role" in determining the speed at 1 AU, while the distance to the coronal-hole boundary correlates more directly with observed speed.^[21] A 2025 analysis found that optimizing the source-surface height between roughly 1.8 and 3.1 R_☉ solar radii improved high-speed stream predictions for several rotations.^[22] Despite critiques, reviews routinely describe the coupled WSA–ENLIL system as a "workhorse model" for both research and operational forecasting communities, yet they also stress its limitations near the heliospheric current sheet and during active times.^[17]

• Coronal hole
• Enlil (heliospheric model)
• Heliophysics
• Solar wind

• NOAA SWPC: WSA–ENLIL Solar Wind Prediction
• NASA CCMC: WSA model page
• National Solar Observatory: ADAPT global magnetic maps