this paper, we propose a method for supporting the emergence of admissible scenario from KeyGraph with the ¡Ècondensed¡É visualized relations of concepts. We can count sub methods addressed in this study. First, based on a simplified ontology-based framework, we made a method for refining the structure of KeyGraph from a initial graph obtained for discursive writings. For realizing a further restructuring of KeyGraph, we also developed a technique for recomposing morphemes by applying rules of phrase compositions. Second, the hierarchical context classification enabled to focus the data for the user¡Çs interest. Third, by attaching scenic information i.e. pictures, to the refined KeyGraph, we realized admissible scenarios of earthquake signs. Here three participants of scenario creation meeting evaluated the admissibility on achieving consensus.