Microwave Imaging Spectroscopy of Flares is Here

From RHESSI Wiki

Jump to: navigation, search

Number: 327
1st Author: Dale E. Gary
2nd Author: EOVSA and RHESSI Teams
Published: 16 July 2018
Next Nugget: The true dawn of multimessenger astronomy
Previous Nugget: Coronal nanoflares powered by footpoint reconnection
List all



The Expanded Owens Valley Solar Array (see our early EOVSA Nugget) is now fully operational, and obtained mutli-frequency microwave (MW) images and movies of the spectacular flares of September 2017 in the frequency range 2.5-18 GHz, with 1 s time resolution. A paper describing the first science results from EOVSA is now available (Ref. [1]), as summarized here and below.

The 2017 September 10 X8.2 Flare

The flare SOL2017-09-10, in NOAA AR 12673 reached GOES soft X-ray class X8.2. Figure 1 shows the EOVSA microwave (MW) and RHESSI hard X-ray (HXR) images at three times during the event, overlaid on AIA 193 Â EUV images. As usual for large limb events, the RHESSI thermal (~12 keV) sources are associated with the brightest EUV-emitting loops, while the nonthermal (~50 keV) coronal sources are higher, but still well centered above the growing EUV-emitting, cusp-shaped source. The MW sources, which generally come from much higher energy (~300 keV) electrons, are still higher and show some asymmetry in the emitting region, especially at later times when the MW source bifurcates and seems to avoid the densest part of the plasma sheet.

Figure 1: Comparison of AIA , RHESSI, and EOVSA images at the three times. AIA 193 Â images (in reverse grayscale of log intensity) are superposed with filled 50% contours of EOVSA microwave emission at 26 spectral windows, with hues shown in the color bar. RHESSI hard X-ray 30, 50, 70, and 90% contours are also superposed for two energy ranges. Left-hand panels have a 2×2 arcmin field of view (FOV), while right-hand panels have a larger 5×5 arcmin FOV. Panels a, b are at 15:54 UT early in the event, panels c, d are at 16:00 UT, near the peak of the radio and hard X-ray emission, and panels e, f are at 16:41 UT much later in the event.

Figure 2 shows height-time stackplots of AIA, EOVSA and RHESSI data taken along the dashed line in Figure 1d. This shows the steady rise in radio sources at exactly the same rate as the growth of the EUV loops, so that the radio source maintains its spatial relationship to the EUV-bright loops.

Figure 2: Height-time stackplots of AIA, EOVSA, and RHESSI data from a horizontal cut at vertical position y = −141′′ in Fig. 1d. The red vertical bars in each panel show the 50% contour height range of the RHESSI thermal sources at times t1 and t2, from Fig. 2a,c, while the blue bars show the corresponding height range of the RHESSI nonthermal sources. The time range covers 20 minutes, from 15:46-16:06 UT. (a) AIA 193 Â intensity, log-scaled. The black dashed curves schematically show the leading and trailing edges of the flux rope. (b) AIA 131 Â intensity, log-scaled. (c) EOVSA 5.42 GHz brightness temperature, linearly scaled. (d) EOVSA 13.42 GHz brightness temperature, linearly scaled. (e) The same AIA 193 Â intensity as in (a) overlaid with the 5.42 GHz brightness temperature contours at 30, 100, 300, 1000, and 3000 MK. (f) The AIA 131 Â intensity as in (b), overlaid with the 13.42 GHz brightness temperature contours at 100, 300, 1000, and 3000 MK.

MW Imaging Spectroscopy

The multi-frequency images at each moment in the flare form a spatial-spectral data-cube, which can be displayed as multi-frequency images, as illustrated in Figure 3a, at 28 frequencies (from 3.4 to 16.9 GHz) for the 15:54 UT time. The frequency content of these images can be visualized as in the “true-color” image in Fig. 3f, where the images at 28 frequencies are apportioned different red-green-blue weights according to their frequency. In addition to displaying images at different frequencies as in Figure 3a, we can display spectra at different spatial locations, as in Fig. 3b-e, which are line-of-sight spectra obtained at the 4 points labeled in Fig. 3f. The data points in Fig. 3b-e can be fit with gyrosynchrotron emission spectra from a multi-parameter homogeneous source, and the fit parameters (the magnetic field strength B and power-law index delta parameters are shown as text in each panel). Although these fits are preliminary, and shown only to illustrate the technique, it is clear that the changing spectral shape over the emission region reflects varying parameters vs. location in the source that can be exploited for imaging spectroscopy, including dynamic coronal magnetography (Ref. [2]).

Figure 3: (a) Individual images at 28 frequencies, from the location of the white box in the overview image in panel f. (b-e) Measured flux-density spectra (points with ±1σ error bars) in single pixels of the images in panel a, corresponding to locations 1-4 marked in panel f, and corresponding multi-parameter fits (red lines). (f) A “true-color” representation of the EOVSA data cube, combining images at the 28 frequencies shown in panel a.


This EOVSA first-science paper barely scratches the surface of the potential of the data for learning new details on the energy release, particle acceleration and transport in solar flares. Similar results as Fig. 3, both at other times in SOL2017-09-10 and in other flares (e.g. SOL2017-09-04 and SOL2017-09-09) are now being studied in detail, together with 3D modeling with GX Simulator. EOVSA has obtained such data on more than 100 flares in 2017 (very few so far in 2018 due to declining solar activity), as well as active regions. See the EOVSA web site for more details.


[1] "Microwave and Hard X-Ray Observations of the 2017 September 10 Solar Limb Flare"

[2] "Magnetography of Solar Flaring Loops with Microwave Imaging Spectropolarimetry"

Personal tools