Search

In this study, Tourism Climate Index (TCI) was used for a network of 153 synoptic stations of Iran to determine climate comfort in each month. The monthly TCI distribution was determined according to one of the six classes of poor, optimal, summer peak, winter peak, dry season peak and dual peaks. Linear and nonlinear multivariable regression models were used to investigate spatial legitimacy between TCI, latitude, longitude and elevation of different stations for each month separately. For validation of the models, the statistics such as t-test, F-test, root mean square error, mean error, mean absolute relative error and determination of coefficient were used. The results showed that significance level of t- test probability is greater than 0.05 for latitude in April. Also, this variable is not significant at the 0.05 level in the F statistics and reduce the model’s determination of coefficient. So there is no need for the existence of this variable in three-dimensional model in this month. But all three variables could be involved in the model in January, February, March, May, November and December. To create a model with higher accuracy, nonlinear three-variable models are fitted when the index has been a regression equation just with two variables. The results showed that standard error of the estimated values of model parameters is high, except in February and December. The three-dimensional linear and nonlinear models for the estimation of TCI are presented for different months across Iran.