Abstract
This paper presents the reproducibility analysis of the distribution hydrological model using several rainfall data, such as XRAIN and ground observation rainfall data, applied to the upper basin of Kanda river in Tokyo. The highest reproducibility is observed by using high density deployed ground observation rainfall, among the simulation using other rainfall data including XRAIN with its spacial high resolution. From the simulation using XRAIN, direct use of XRAIN's spatial distribution data to the hydrological meshed model revealed higher reproducibility than the simulation using averaged XRAIN in the basin.