At Fukushima Daiichi Nuclear Power Plant unit 1 (1F1), it is expected that molten corium-concrete interaction (MCCI) occurred in a significant scale. large uncertainty in progression of the accident is inevitable due to lack of information (measurements) such as the timings of RPV depressurization and RPV failure (discharge of debris from RPV) and the RPV failure mode. Furthermore, significant uncertainties exist in the calculated results even with a set of well-defined boundary conditions among different severe accident analysis codes such as MAAP and MELCOR. Hence, this study aims influences of scenario uncertainties and modeling uncertainties in predicting the late phase of the 1F1 accident, where MCCI is expected to have occurred using MELCOR2.1. Two depressurization scenarios (early depressurization with steam relief valve failure and late depressurization with main steam line failure) and two RPV failure modes (penetration tube failure and RPV lower head rupture) on the subsequent MCCI are considered. Analyses with MELCOR indicated that different cases led to different debris compositions (metals and oxides) and amount on the basemat floor and led to differences in the radial ablation depths, while axial ablation depths showed negligible differences among different cases. However, in all cases, MCCI did not terminate 140 hours after the reactor scram. Sensitivity studies have been carried out with and without the presence of external water injection and parametric study with some of the parameters involved in MELCOR simulation relevant to MCCI. The results showed that the concrete ablation in the radial direction was suppressed by external water injection. However, the axial ablation was enhanced, indicating necessity to further investigate modeling of MCCI in MELCOR2.1.