On the Statistics of the Wave Heights Observed at a Weather Station

Abstract

It is said that the statistics of wave heights follow their log． norlnal distribution ． The main part of the distribution ， however ， fits in very good manner ， the linearlty is ' destroyed in the high wave height Iegion ． The reason of this distortion can be thought as follows ； the observation lnethod ， number Qf data or sampling ， and meteorological condition and ！ or wave nature itself ． MQstly the wave statistics are obta ｛ ned from the ship ' s visual observations in the area settled in an ocean ， and the number of observations is not enough ． So the conseclltive 6 hourly records of Station T (29°N ， 135°E) for ユO years (1967 − 1976) are plotted， and are found that the data fit to log ． normal distribution very well ． Th the number of data was reduced to ユ ！ 2 through ユ 1ユ6 at rando 皿 ， and found that the linearllty became shaky in the high wave height region ． One reason Qf this distortion from the log − norma ； can be concluded by the result of data sampling ． In these cases ， extending the straight line of ／ 0 through 95 ％ ranges to the higher range may be reasonable ． Annual extreme values are also investigated ， and shown that the mean annual extreme value is nearly eq 岨 l to the wave height corresponding to the percentage of (」 V − 1) ！ 1V of all wave data ．