Convection parameterization, model resolution, and output sampling frequency impacts on Joule heating in T-I models

Abstract

At polar and auroral latitudes, strong field-aligned currents flow along the Earth's open magnetic field lines. In the ionospheric dynamo region, field-perpendicular Pedersen currents close the circuit. Pedersen currents result in significant Joule heating of the thermosphere-ionosphere system, which is one of the main geomagnetic energy input mechanisms. High-latitude Joule heating has been shown to cause significant disturbances of the ionospheric electron density. In thermosphere-ionosphere (T-I) models, the high-latitude electrodynamics can be parameterized in several different ways, including purely empirical and data assimilation-based parameterizations. Other model settings like grid resolution and output sampling frequency also affect the Joule heating. We evaluate the Joule heating rates calculated from the TIE-GCM and WACCM-X models with those estimated from measurements by the EISCAT incoherent scatter radar in Tromsø, Norway (69.6°N, 19.2°E). We investigate model runs driven with the Heelis, Weimer, AMIE, and AMGeO convection parameterizations. T-I model runs that were performed with magnetospheric coupling to the GAMERA code are analyzed as well. The grid resolution of the investigated runs varies from ultra-high 0.25° to regular 1.25° and 2.5° resolutions. Output sampling frequencies of 5 minutes and 1 hour are compared. We show that all these settings affect the model Joule heating rates and how they compare to EISCAT measurements.