摘要：We report a new solar extreme ultraviolet (EUV; 26–34 nm) proxy using solar radio fluxes at six different frequencies (410; 610; 1,415; 2,695; 4,995; and 8,800 MHz) from ground-based observations. The radio fluxes (2002–2008) are used to estimate the solar EUV flux under nonsolar flare conditions through an artificial neural network trained with coincidently observed solar EUV (26–34 nm) fluxes from the SOlar and Heliospheric Observatory/Solar EUV Monitor instrument. The radio fluxes at 610; 1,415; and 8,800 MHz are the three top contributors in the estimation of the solar EUV flux. The 1,415 MHz flux has the highest correlation coefficient (0.97) with the observed EUV flux. The estimated EUV flux retained the highest correlation coefficient (0.97) and had the lowest error in comparison with the observed EUV flux. The popular F10.7cm (2,800 MHz) is between 2,695 and 4,995 MHz, which have a minor contribution to the estimated EUV flux. On the other hand, F30cm (1,000 MHz) is between 610 and 1,415 MHz, the two major contributors to the estimated EUV flux. These features are consistent with a significantly better neutral density modeling using F30cm versus F10.7cm (Bruinsman, 2015, https://doi.org/10.1051/swsc/2015001). Similar results are obtained by applying the coefficients trained using a subset of the radio data (2002, 2004, and 2007) to a different subset of the data (2003, 2005, 2006, and 2008). These indicate that solar radio fluxes at multiple frequencies can be used to reliably retrieve solar EUV flux, one of the key parameters for space weather studies.
