Real-time prediction of two geomagnetic storms using Solar Orbiter as a far upstream solar wind monitor

We present the first real-time predictions of coronal mass ejection (CME) magnetic structure and resulting geomagnetic impact at Earth for two events using far-upstream observations from Solar Orbiter during March 2024. While our approach assumes idealized conditions for CME propagation and scaling, in situ magnetic field data from upstream monitors still produced realistic predictions despite the large heliocentric distance between Solar Orbiter and L1 (0.53 and 0.60 au). Geomagnetic index predictions were made 15.3 and 4.3 hours before the CME shock arrival at L1, and 33.9 and 10.3 hours ahead of peak storm time; a large improvement over current L1-based nowcasting capabilities. We find that observationally constraining the simple drag-based models using the upstream in situ observations improved arrival time estimates for the two events in this study, although arrival time errors of several hours still remain. Our results show that good predictions of CME magnetic structure and geomagnetic indices with actionable lead-times can be made with far upstream spacecraft, even with longitudinal separations up to 10{\deg} from the Sun-Earth line, over heliocentric distance ranges where radial evolution effects dominate over longitudinal effects. Limitations include different expansion behaviors for individual CMEs and regions within. Future missions providing continuous data, including solar wind plasma parameters alongside magnetic field measurements, could account for preexisting disturbed conditions and improve geomagnetic prediction accuracy. Our findings demonstrate the substantial value of real-time upstream solar wind measurements for enhancing geomagnetic forecasting accuracy at Earth and provide critical validation for future dedicated upstream space weather missions.