降水・大気

[II-47] Statistical properties of precipitation's dependence by numerical simulation of simultaneous bivariate extreme value

Takuya Kunikado¹, 〇Ryota Imai¹, Toshikazu Kitano¹ (1.Nagoya Institute of Technology Graduate School)

Keywords:Numerical Simulation, Return Priod, Non-parametric, Dependence Fanction

Low frequency disaster cause enormous damage. Estimation of occurrence probabilities is required for countermeasures for disaster. Bivariate extreme value theory considering of dependence between different places is applied for estimation of those probabilities in case that a heavy rain occurs at different places in a watershed and cause a overflow at a merging section. For better data analysis, we generate a random number data and verify reproduction of estimates of precipitation data and obtain statistical properties by comparing estimation and given distribution.