Type II radio bursts are slow-drifting emissions in dynamic radio spectra, first identified in the late 1940s, that arise from plasma oscillations excited by electrons accelerated at shock waves. When a shock disturbs the coronal plasma, electrons are accelerated and stream along magnetic field lines, exciting electrostatic oscillations at the local plasma frequency, which then convert into escaping electromagnetic radiation. Because the plasma frequency is proportional to the square root of the local electron density, the emission drifts from higher to lower frequencies as the shock travels outward through the corona’s tenuous, rapidly thinning atmosphere. Bursts observed in the metric range, roughly from tens to hundreds of megahertz, trace the low corona, where the plasma is densest and the shocks are young. It is precisely this region that matters most for space weather, because the early evolution of a CME and its shock determines whether the disturbance will evolve into a geoeffective interplanetary structure capable of compressing Earth’s magnetosphere days later.

The team assembled a sample of 23 coronal Type II bursts observed between 2007 and 2024 in the frequency range from about 150 MHz down to 20 MHz. The radio data were drawn from three complementary instruments: the Gauribidanur Low-Frequency Solar Spectrograph in India, the Culgoora radiospectrograph in Australia, and stations of the e-CALLISTO worldwide network of solar radio spectrometers. Each burst was matched to a CME observed in white light by the Large Angle and Spectroscopic Coronagraph aboard the Solar and Heliospheric Observatory, with supporting imagery from the Sun Earth Connection Coronal and Heliospheric Investigation package on the twin STEREO spacecraft and the Atmospheric Imaging Assembly on the Solar Dynamics Observatory. By cross-referencing the onset times of the radio emission with the projected positions and kinematics of the erupting structures, the authors were able to associate each burst with a specific CME and characterize both the radio signature and the erupting plasma.

The central result of the study is a clear anti-correlation between the frequency bandwidth of the Type II bursts and the angular width of the associated CMEs, quantified by a correlation coefficient of approximately minus 0.62. In practical terms, CMEs that span a large range of position angles in the coronagraph images tend to produce radio bursts whose emission is confined to a narrow band of frequencies, while narrower CMEs are associated with bursts spread across a wider frequency span. The inverse relationship is striking because angular width is often treated as a rough proxy for the lateral extent and energetic clout of an eruption. One might naively expect a broader shock front to sweep through a wider range of coronal conditions and therefore generate emission across a broader spectral window. The data say the opposite.

The authors interpret this counterintuitive result through the geometry of the shock and the density structure of the ambient corona. A Type II burst’s instantaneous bandwidth reflects the range of electron densities at which the shock is simultaneously exciting plasma emission. If a shock propagates through a relatively homogeneous medium, or if the emission is generated primarily in a localized region rather than along the entire shock front, the instantaneous bandwidth will be narrow. The anti-correlation with angular width suggests that wide CMEs, whose flanks spread out laterally over large position angles, tend to produce emission from restricted parts of the shock surface where the conditions for electron acceleration are most favorable. The study therefore points toward the flank regions of CME-driven shocks, rather than the nose or leading edge, as the likely sources of these metric Type II emissions.

This interpretation is reinforced by a second line of evidence in the paper: a careful comparison of heights. For each event, the team deduced the height of the Type II burst source at the onset time of the burst, derived from the emission frequency through coronal electron density models, and compared it with the estimated height of the associated CME or shock at the same moment, obtained by interpolating the CME’s measured position in coronagraph images. With the exception of a single event in the entire sample of 23, the height inferred from the radio emission was smaller than the height of the CME leading edge. In other words, the radio source consistently sat below the front of the erupting structure, a configuration that is difficult to reconcile with emission generated at the shock nose itself but entirely consistent with emission arising on the flanks of a laterally expanding shock, which lag behind the leading edge in height.

The lone exception in the sample is itself informative, because events in which the radio source coincides with or exceeds the CME height are the cases in which a nose or leading-edge origin is most plausible. The rarity of such cases in this sample of wide CMEs underscores the statistical conclusion. Previous imaging studies, including observations with the Low Frequency Array and the Murchison Widefield Array, have already shown that individual Type II bursts can have multiple, spatially separated sources and that different sections of a shock can accelerate electrons at different times. The present work adds a statistical dimension to that picture, suggesting that for broad, laterally extensive eruptions the flank geometry systematically dominates the radio output.

The technical foundation for the height estimates rests on well-established coronal density models that link emission frequency to plasma density, calibrated against eclipse and radio heliograph measurements, including multiplicative factors to account for density enhancements in streamers and active regions. The choice of density model introduces some uncertainty, but the systematic nature of the height comparison across 23 events, all analyzed consistently, gives the result robustness that a single event study cannot match. The frequency bandwidth measurements, meanwhile, were made directly from the dynamic spectra, capturing the span of frequencies over which the burst emission was present at a given stage of its evolution.

The implications extend beyond solar physics into the practical business of space-weather prediction. Interplanetary shocks driven by CMEs compress Earth’s magnetosphere on arrival, and the severity of the resulting geomagnetic disturbance depends on the shock’s speed, structure and history. Because Type II bursts are direct radio signatures of shock acceleration, they can in principle serve as early indicators of a shock’s strength and trajectory, and several research groups have explored using metric Type II observations to forecast shock arrival times at Earth. A relationship between burst bandwidth and CME angular width adds a potentially useful piece to that forecasting toolkit: a narrow-band metric burst following a wide CME would hint that the radio-emitting part of the shock is developing on its flanks, which may carry implications for where energetic particle acceleration is strongest and how the disturbance will couple to the heliospheric magnetic field.

The study also feeds into a long-running scientific debate about the origin of metric Type II bursts, which dates back more than half a century. Early proposals attributed the bursts to blast waves generated by solar flares, while decades of subsequent work have increasingly linked them to CME-driven shocks, with timing ambiguities persisting in individual events. Statistical studies like this one, which correlate a spectral property of the burst with a morphological property of the CME, sidestep some of the timing controversies and instead probe the physical relationship between the two phenomena. The result that wide CMEs produce narrow-band emission, plausibly from their flanks, offers a new observational constraint that any complete model of coronal shock emission must reproduce.

There remain open questions. The sample of 23 events, while respectable for a study of rare radio phenomena, is modest, and extending the analysis to a larger population of bursts, including those at decametric and interplanetary wavelengths, would test whether the anti-correlation persists farther from the Sun. Imaging spectroscopy with modern instruments capable of directly locating Type II sources, rather than inferring their heights from frequency, would provide the definitive test of the flank-origin hypothesis. The authors’ dataset, drawn from publicly available catalogs and spectrograph archives, is well suited for such follow-up, and the combination of ground-based radio spectroscopy with space-based white-light coronagraphy remains the workhorse technique for this field.

For now, the study delivers a clear and quotable takeaway: the width of a solar eruption and the width of its radio scream are inversely related, and the radio source usually hides below the CME’s leading edge rather than riding at its crest. As the current solar cycle continues to deliver energetic eruptions, each new wide CME accompanied by a narrow-band Type II burst will be another data point in a picture that is steadily coming into focus, one plasma-frequency drift at a time.