Synthesising solar radio images from Atmospheric Imaging Assembly extreme-ultraviolet data

Abstract

During non-flaring times, the radio flux of the Sun at the wavelength of a few centimeters to several tens of centimeters mostly originates from the thermal bremsstrahlung emission, very similar to the EUV radiation. Owing to such a proximity, it is feasible to investigate the relationship between the EUV emission and radio emission in a quantitative way. In this paper, we reconstruct the radio images of the Sun through the differential emission measure obtained from the multi-wavelength EUV images of the Atmospheric Imaging Assembly on board Solar Dynamic Observatory. Through comparing the synthetic radio images at 6 GHz with those observed by Siberian Radioheliograph, we find that the predicted radio flux is qualitatively consistent with the observed value, confirming thermal origin of the coronal radio emission during non-flaring times. The results further show that the predicted radio flux is closer to the observations in the case of including the contribution of the plasma with temperatures above 3 MK than in the case of only involving the low temperature plasma as was usually done in the era of pre-SDO. We also discuss the applications of the method and uncertainties of the results.

Generating a radioheliograph image from SDO/AIA data with the machine learning method

“…Meanwhile, Alissandrakis et al (2019) modeled the sunspot-associated microwave emission based on the gyro-resonance emission mechanism with potential extrapolations of the photospheric magnetic field, in which the DEM was inverted from the EUV image of Atmospheric Imaging Assembly (AIA) on-board the Solar Dynamics Observatory (SDO) (Lemen et al 2011). More recently, Li et al (2020) find that the predicted radio flux is closer to the observations in the case that includes the contribution of plasma with temperatures above 3 MK than in the case of only involving low temperature plasma, and confirmed the thermal origin of the quiet corona radio emission. The predicted value of the DEM method depends on the physics model, including the derivation of n e , T e and magnetic field, and the emission mechanism.…”

“…Comparing with the DEM method, we can have higher precision of brightness temperature prediction with the machine learning method. The correlation of machine learning prediction is above 0.8 with an average value of 0.94 for the test cases, and the linear fit for the flux intensity of OBS and GEN is also much closer to y = x than the DEM prediction results (Zhang et al 2001;Li et al 2020). This may be partially due to the uncertainty of the DEM inversion and the emission mechanism assumed in the prediction.…”

Generate Radioheliograph Image from SDO/AIA Data with Machine Learning Method

“…On the other hand, Alissandrakis et al (2019) modeled the sunspot-associated microwave emission based on the gyro-resonance emission mechanism with potential extrapolations of the photospheric magnetic field, in which the DEM was inverted from the EUV image of Atmospheric Imaging Assembly (AIA) on-board the Solar Dynamics Observatory (SDO) (Lemen et al 2011). More recently, Li et al (2020) find that the predicted radio flux is closer to the observations in the case that includes the contribution of plasma with temperatures above 3 MK than in the case of only involving low temperature plasma, and confirmed the thermal origin of the quiet corona radio emission. The predicted value of the DEM method depends on the physics model, including the derivation of n e , T e and magnetic field, and the emission mechanism.…”