To investigate the statistical sensitivity distributions of tropical cyclone (TC) forecasts over the Korean peninsula, total energy singular vectors (SVs) were calculated and evaluated for the TC forecasts during recent 10 years. SVs were calculated using the fifth-generation Pennsylvania State University-NCAR Mesoscale Model (MM5) and its tangent linear and adjoint models with a lanczos algorithm. Cases chosen were 25 TCs that affected the Korean peninsula among 230 TCs that were generated in the Northwest Pacific within the recent 10 years (2001-2010). A typhoon bogusing was applied to make the more realistic simulations. Sensitive regions were mainly near the midlatitude trough and TC centers, but they varied depending on the TC tracks and positions of the subtropical high. The cases were classified to the several groups by the clustering analysis based on finite mixture model. The elements of clustering were the TC's tracks and central pressures. While the groups that passed to the western and southern region of Northwest Pacific had the maximum sensitive region near the midlatitude trough, the groups passed to the more eastern and northern region of it had the maximum sensitive region near the front of TCs. Moreover each group had its own characteristic subtropical high pattern.