On the difference between real-time and research simulations with CTIPe

Understanding the thermosphere and ionosphere conditions is crucial for spacecraft operations and many applications using radio signal transmission (e.g. in communication and navigation). In this sense, physics based modelling plays an important role, since it can adequately reproduce the complex coupling mechanisms in the magnetosphere-ionosphere-thermosphere (MIT) system. The accuracy of the physics based model results does not only depend on the appropriate implementation of the physical processes, but also on the quality of the input data (forcing). In this study, we analyze the impact of input data uncertainties on the model results. We use the Coupled Thermosphere Ionosphere Plasmasphere electrodynamics model (CTIPe), which requires satellite based solar wind, interplanetary field and hemispheric power data from ACE and TIROS/NOAA missions. To identify the impact of the forcing uncertainties, two model runs are compared against each other. The first run uses the input data that were available in real-time (operational) and the second run uses the best estimate obtained in post-processing (research or historical run). The analysis is performed in a case study on the 20th November 2003 extreme geomagnetic storm, that caused significant perturbations in the MIT system. This paper validates the thermosphere and ionosphere response to this storm over Europe comparing both CTIPe model runs with measurements of Total Electron Content (TEC) and thermosphere neutral density. In general, CTIPe results show a good agreement with measurements. However, the deviations between the model and observations are larger in the ionosphere than in the thermosphere. The comparison of the two model runs reveals that the deviations between model results and measurements are larger for the operational run than the research run. It is evident for the storm analyzed here, that data gaps in the input data are impacting considerably the model performance. The consistency between simulation and measurements allows the interpretation of the physical mechanisms behind the ionosphere perturbations and the changes in neutral composition during this event. Joule heating in the Auroral region, generating meridional winds and large scale surges, is suggested to be the main driver of the positive ionospheric storm over central Europe. In the polar cap and Auroral region, convection processes dominate the thermosphere-ionosphere conditions. This study does not only illustrate the importance of working with a good estimate of the model forcing, but also indicates the necessity of using measurements and models, to get a better understanding of the most likely responsible processes for the observed storm effects.