From multi-decadal solar wind modelling to real-time forecasting, and on moving away from the ecliptic plane

I analyse the spatial distribution of solar wind sources and relate them to the properties of the interplanetary wind by means of an extended time series of data-driven 3D simulations that cover more than two solar activity cycles. Magnetic connectivity jumps are related with solar wind plasma signatures and with topological features of the global magnetic field. The occurrence frequency and amplitudes if such connectivity jumps vary with the epoch of the solar cycle and on the distance to the ecliptic plane.
The same solar wind model (Multi-VP) constitutes the core of the SWiFT-FORECAST service, based on model pipeline initially developed in the scope of the H2020 SafeSpace project, and later integrated on ESA's SWESNET and on the Virtual Space Weather Modelling Centre.  Ensemble forecasts are produced at a daily cadence and with a lead time of a few days.  I will address some of the main challenges related to the implementation and validation of these models and pipelines, as well as the pernicious issues that stem from the lack of observables between the two boundaries of the Sun­–Earth system, and from the dependence of "point" forecasts on the global properties of the solar atmosphere.