Authors: Evangelia Samara (NASA/GSFC), Elena Provornikova (JHU/APL), Charles N. Arge (NASA/GSFC), Samantha Wallace (Embry-Riddle Aeronautical University), Andrew McCubbin (JHU/APL), Slava Merkin (JHU/APL)

Realistic modeling of the solar wind in the heliosphere is a complex task subject to many limitations related to the currently available solar observations and the capabilities of models per se. A major limitation, for example, is the lack of observations at the limbs, poles and back side of the Sun that does not allow the creation of trustworthy magnetograms which are the necessary drivers for coronal and heliospheric models. Other limitations arise from the models’ restricted capabilities to describe known physics because of numerical and computational complexity. Moreover, several ad hoc assumptions are frequently adopted to reconstruct the solar corona magnetic field and forecast the solar wind in the heliosphere. The extent at which these assumptions affect the final predictions is still not very well understood. In this work, we address the aforementioned limitations with state-of-the-art time dependent solar wind simulations which represent a critical step towards more realistic solar wind modeling. However, we ask the question, how far can we go with them and what else is needed to go forward? Our goal is not to (only) show nice solar wind predictions compared to observations at different spacecraft locations (PSP, Solar Orbiter, STEREO A, Earth), but to discuss why major discrepancies with observations occur and how/if we can overcome them. By using two models, the semi-empirical ADAPT-WSA coronal model and the MHD GAMERA inner heliosphere model, we explain solar wind predictions from their magnetic connectivity to the solar origins, and discuss how the emergence of active regions on the invisible side of the Sun can affect this connectivity and the overall forecast.