A different type of Burr distribution applying to hydro-meteorological data


Güçlü Y. S.

SPATIAL STATISTICS, vol.43, 2021 (Peer-Reviewed Journal) identifier identifier

  • Publication Type: Article / Article
  • Volume: 43
  • Publication Date: 2021
  • Doi Number: 10.1016/j.spasta.2021.100511
  • Journal Name: SPATIAL STATISTICS
  • Journal Indexes: Science Citation Index Expanded, Scopus, INSPEC
  • Keywords: Burr, Distribution function, Gamma, LogNormal, Probability, Modelling

Abstract

Probability (PDF) and cumulative distribution function (CDF) are used extensively to calculate probability features in engineering, natural sciences, hydrology, meteorology and many other fields. In this study, a different type of Burr distribution is suggested, which provides an alternative probability calculation. The type is fitted to the Weibull empirical equation results on the vertical axis versus measurements on the horizontal axis with coefficient of determination (R-2) almost one. The model generally produces better estimates according to mean absolute error and validation graph by comparing with the classical two parameter Log-Normal and Gamma distributions. Rainfall data measured at Ankara, Kocaeli, and Edirne stations are used to establish all models for 64 years available data in Turkey. It is shown also that proposed type of Burr distribution perfectly fits the outputs of the classical CDFs for each station data. Additionally, the PDF type is obtained by derivating the proposed CDF form. (C) 2021 Elsevier B.V. All rights reserved.