Modeling of daily maximum and minimum temperature changes in Bitlis province using Copula Method

Abstract

This paper aims to examine the relationship between daily maximum and minimum temperatures of Bitlis in Turkey between 2012-2017 years with Copula method. To present the relationship between the variables, we use copula families such as; Gumbel, Clayton, Frank, Joe, Gaussian and Survival Clayton copula. To explain dependence structures of the data set and to determine parameters of Gumbel, Clayton, Frank, Joe, Gaussian and Survival Clayton copula families, we calculate Kendall Tau and Spearman Rho values which are nonparametric. With the help of Kolmogorov Smirnov, Cramer Von Mises which are goodness of fit test, Maximum likelihood method, Akaike information Criteria ad Bayes information criteria, we find the suitable copula family for this data set. The results show that there is a strong dependence between daily maximum and minimum temperatures of Bitlis between 2012-2017 years.