Japanese surnames have diversity in their origins and regional distributions. Whereas their origins have been studied in philology and their regional distributions have been researched statistically, few studies have focused on the relationship between their origins and regional distributions. In this paper, we focus on surnames with Chinese characters indicating plants (namely, green surnames), and analyzed their statistical characteristics. First, we collected top-10,000 surnames from websites describing statistical information of surnames collected from telephone directories, and annotated green surnames based on a Chinese character dictionary. By analysis of the resulting 1,154 green surnames, we revealed the characteristic skewness of their regional distributions compared with not-green surnames. Then, we applied multi- level analysis to the data of green surnames including their frequency for each prefecture and their variables, and found the positive effect of vegetation related to the plant name in the green surnames on their frequency.
