Journal of Japan Society of Civil Engineers, Ser. G (Environmental Research)
Online ISSN : 2185-6648
ISSN-L : 2185-6648
Global Environment Engineering Research, Vol.29
ESTIMATION OF EXTREME CURRENT SPEED USING NORTH PACIFIC OCEAN MODEL AND REGIONAL OCEAN MODEL RESULTS
Takaki TSUBONODaisuke TSUMUNENaoto KIHARAMasamichi Ohba
Author information
JOURNAL FREE ACCESS

2021 Volume 77 Issue 5 Pages I_1-I_8

Details
Abstract

 We examined an approach to estimate the return level by interpolation and extrapolation of current velocities by combination of the JCOPE2M reanalysis and ROMS results. First, patterns for JCOPE2M were represented with Self-Organizing Maps (SOM), and then the statistics of the probability distribution of the scores of the principal components showing the vertical distribution of current velocity were calculated using the ROMS results for the pattern extracted period by the maximum likelihood method. Next, 27 years × 365 days of the velocities were produced by random numbers with the statistics of the pattern at each time of the period for the JCOPE2M reanalysis, and the annual maximum velocities for 27 years were set. Finally, the maximum likelihood method was used to calculate the statistic parameters for the extreme value distribution using the annual maximum velocities. We estimated the statistic parameters of the extreme value distribution for a total of 20 cases with various conditions: the SOM, probability density function, and extreme value distribution and evaluated the log likelihood substituting the 9 annual maximum velocities by ROMS into the extreme value distributions for all 20 cases. The log likelihoods showed that the combination of the 6 × 6 pattern of SOM, t-distribution for the probability distribution of the velocity, and Gumbel distribution for the extreme value distribution is the best approach for estimating the return level for velocities using the JCOPE2M and ROMS results.

Content from these authors
© 2021 Japan Society of Civil Engineers
Next article
feedback
Top