High‐Latitude Conductance Modeling Under Solar EUV Forcing

Abstract

We develop a semi‐empirical model for high latitude EUV‐generated conductance that combines first‐principles results from GLOW with empirical calibration using strictly screened incoherent scatter radar‐satellite (DMSP) conjunctions (with both electron and ion precipitation energy fluxes less than 0.1 milliwatt per square meter). GLOW conductances are emulated with a compact neural network driven by solar elevation angle, F10.7, Ap, day of year, and geographic latitude. A uniform correction of 0.8 S is applied to both Pedersen and Hall conductances to bring the model into agreement with radar‐satellite conjunction measurements. Evaluation against a multi‐decade, six‐radar data set shows close agreement with observations across sites and time, and substantial reduction of the high‐latitude bias present in Moen and Brekke (1993, https://doi.org/10.1029/92GL02109 ) model (MB93). Further analysis of MB93 shows that the database used in that study does not fully remove precipitation effects. In comparison, our approach uses radar‐satellite conjunctions to fully separate EUV from precipitation events. We provide open‐source code to facilitate straightforward implementation and adaptation of the model in other studies.