Incoherent scatter radar measurements and modeling of high-latitude solar photoionization

A 9-year database of sunlit E region electron density altitude profiles (N-e(z)) measured by the Sondrestrom incoherent scatter radar (ISR) has been partitioned over a parameter space of 10.7 cm solar radio flux (F-10.7) and solar zenith angle (chi) to investigate long-term solar and thermospheric variability and to validate a contemporary EUV photoionization model. A two-stage filter, which rejects N-e(z) profiles with large Hall-to-Pedersen conductance ratio and incorporates an MLT-dependent correction factor for low-energy precipitation, is used to mitigate auroral contamination. Resultant filtered mean sunlit N-e(z) is compared with subauroral N-e measured for the same F-10.7 and chi conditions at the Millstone Hill ISR in order to confirm adequate high-energy auroral rejection. Mean Ne, as expected, increases with solar activity and decreases with large c. Radar model comparison indicates that across all parameter space and for altitudes from 105 to 180 km, the GLOW model estimates are within 5% of the ISR mean with the contribution from photoelectrons accounting for 30 to 50% of equilibrium ion density. Above 180 km, the GLOW model slightly overestimates the height of the F1 layer. Radar model comparison also reveals a low-altitude N-e enhancement for high solar activity at altitudes commensurate with 3 to 7 nm XUV and H Lyman-beta radiances. The variance of the ISR mean Ne is shown to be greatest at low F10.7 ( solar minimum). Simulated Ne variance envelopes, given by perturbing the GLOW model neutral atmosphere input by the measured A(p), F-10.7, and T-e extrema, are narrower than ISR derived geophysical variance envelopes at solar minimum. We find no evidence for solar cycle control of low-energy precipitation and thus we attribute the observed N-e variance at solar minimum to variability in solar EUV flux. In order to address estimation of N-e at altitudes where GLOW model photochemical equilibrium assumptions are invalid, we provide an empirical model for Sondrestrom quiet time photoionization N-e(z) as a function of F-10.7 and chi.