Authors: Satoshi Inoue (Center for Solar-Terrestrial Research, New Jersey Institute of Technology), Keiji Hayashi (George Mason University), Takahiro Miyoshi (Graduate School of Science, Hiroshima University

We conduct three different data-based magnetohydrodybamic (MHD) simulations for solar active region 12371 producing an M6.5 flare. First one is a data-constrained MHD simulation where the initial condition is given by a nonlinear force-free field (NLFFF) that is extrapolated from the photospheric magnetic field just before an M 6.5 flare and no time-varying photospheric magnetic field is imposed. Second one is also data constrained simulation but the initial condition is given by the magnetic field before the flare obtained from a data-driven simulation. In the data-driven simulation, the initial condition is given by the NLFFF 1 hour approximately before the flare and it is driven by the time-varying photospheric magnetic field. The last one is full data-driven simulation through before and after the flare. As a result, the highly twisted field lines are produced by the data-driven simulation before the flare that can lead the eruption in the second data-constrained simulation while the NLFFF is stable even at the flare occurring time in the first simulation. On the other hand, the full data-driven simulation also achieves the eruption but the sheared field lines are continuously formed even after the flare. The second data-constrained simulation reproduces well the phenomena associated with flares, including post-flare loops. However we found that the evolution of the bottom magnetic field is inconsistent with the evolution of the observed magnetic field. We will discuss these results and what is a better way for data-based MHD simulation.