Abstract

In this study, four types of circular probability distribution, namely circular uniform distribution, von Mises distribution, wrapped-normal distribution and wrapped-Cauchy distribution are considered in search of a  circular probability distribution provides the best fit for Malaysian wind direction data recorded at maximum speed. The data collected were classified as annual, northeast monsoon and southwest monsoon. Seven sites were identified for this study based on their geographic location, that is, whether they are situated on the east coast or west coast of Malaysia. The objective of the study was to see whether there are any differences in the circular probability distribution of the wind direction based on these two coasts. Graphical measures were used to gauge which distribution gives the best fitting. To further support the findings, two performance indicators or goodness of fit tests are used in the analysis, namely the mean circular distance and the chord length. Generally, it was found that the monsoon-based data fits well with the von Mises distribution as compared to the annual dataset.

Key words: Wind data, circular distance, chord length, circular probability distribution, performance indicator.