MIONA: Machine Learning for Ionospheric GNSS and VLBI Data Assimilation

The project encompasses various tasks to enhance ionospheric models and forecasts. Initially, Vertical Total Electron Content (VTEC) will be derived from individual Global Navigation Satellite System (GNSS) stations using raw GPS observations, with plans to gradually include additional GNSS constellations and signals. A parallel effort involves estimating VTEC for Very Long Baseline Interferometry (VLBI) stations, addressing the differences between legacy VLBI and VLBI Global Observing System (VGOS) observations. Another significant aspect is aimed at enriching GNSS-derived Global Ionospheric Maps (GIMs) with VGOS ionospheric delays. This step will involve handling temporal and systematic differences. Machine learning techniques will be introduced to improve GIMs by integrating various ionospheric and space weather data sets. The final phase is dedicated to VTEC forecasting. It will utilize global GIM models and machine learning to predict VTEC up to 24 hours ahead, accounting for variations in geomagnetic conditions. Deep learning frameworks will be explored for more accurate predictions, and the model will be continually updated with new GIM solutions and external data to improve its forecasting capabilities.

Contacts:
Marcel Iten, ETH Zurich,
Arno Rüegg, ETH Zurich,
Benedikt Soja, ETH Zurich,