Abstract
Solar flares increase the electron number concentration in the daytime ionosphere, potentially affecting radiowave propagation over several frequency ranges. In this study, we use ionospheric observations to determine both peak magnitudes and time variations of solar flare X-rays without using the direct measurement from the flare. Ground-based observations of VLF transmitter phase perturbations are compared against measured X-ray flux levels during solar flares. Flare fluxes derived here from VLF phases on a west-east subionospheric path are compared with those from a previously analyzed north-south path. Using a wider selection of solar flares, including M-class flares for the first time, the best-fit equations and root mean square (RMS) errors are computed with improved standard deviation (SD) uncertainty estimates for the peak fluxes. Good agreement is found between peak long X-ray wavelength fluxes (XL, 0.1-0.8 nm) derived for M-class and X-class flares and those measured by the GOES satellites. Linear regression analysis on the two paths shows the uncertainties increase in inverse proportion to the path length. Investigations were made with a limited set of "operational" parameters that could be used to derive XL fluxes. No increases in RMS or SD uncertainty levels were introduced by the removal of satellite-based regression parameters such as the XL flux level measured before the flare onset. As such, these techniques support the idea of nowcasting M-class and X-class flares from entirely ground-based measurements.