The goodness of fit of the weibull and rayleigh distributions to the distributions of observed wind speeds in a topographically diverse area
Stanton E. Tuller, Arthur C. Brett
Abstract
Stanton E. Tuller, Arthur C. Brett
Abstract
Abstract The goodness of fit of the Weibull and Rayleigh distributions to the observed wind speed distribution was investigated at seven coastal weather stations markedly influenced by local topography and possessing relatively long periods of record. Wind speeds will he Rayleigh or Weibull distributed if four conditions specifying the circular normal distribution for wind velocity are met. These conditions were not met at the seven stations because of the influences of local topography and storm systems. In all cases the null hypothesis that there is no difference between the Weibull or Rayleigh distributions and the observed distribution had to be rejected with the large sample sizes employed. The transformation of wind velocity involved in deriving the Weibull distribution helps make the distribution more circular normal and the Weibull distribution gave better fits than the Rayleigh distribution. The fits were generally best al those stations with the most circular normal wind velocity distributions and the lowest percentage of calms.
OpenAlex reports 29 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
Abstract The goodness of fit of the Weibull and Rayleigh distributions to the observed wind speed distribution was investigated at seven coastal weather stations markedly influenced by local topography and possessing relatively long periods of record. Wind speeds will he Rayleigh or Weibull distributed if four conditions specifying the circular normal distribution for wind velocity are met. These conditions were not met at the seven stations because of the influences of local topography and storm systems. In all cases the null hypothesis that there is no difference between the Weibull or Rayleigh distributions and the observed distribution had to be rejected with the large sample sizes employed. The transformation of wind velocity involved in deriving the Weibull distribution helps make the distribution more circular normal and the Weibull distribution gave better fits than the Rayleigh distribution. The fits were generally best al those stations with the most circular normal wind velocity distributions and the lowest percentage of calms.
Key concepts: Weibull distribution, Rayleigh distribution, Wind speed, Rayleigh scattering, Goodness of fit, Environmental science, Distribution (mathematics), Weibull modulus