Approximation of the breast height diameter distribution of two-cohort stands by mixture models II Goodness-of-fit tests
The goals of this study are (1) to analyse the accuracy of the approximation of empirical distributions of diameter at breast height (dbh) using two-component mixtures of either the Weibull distribution or the gamma distribution in two−cohort stands, and (2) to discuss the procedure of choosing goodness−of−fit tests. The study plots were located in the Świętokrzyski National Park (central Poland) and in the Southern Appalachian Mountains (eastern USA). The results of the goodness-of-fit tests (chi-squared, Kolmogorov-Smirnov, Cramér-von Mises, and Anderson-Darling), normalised bias and normalised root mean square error, indicate that dbh empirical distributions of two-cohort stands are compatible with the mixture models investigated. The chi-squared test and the generalization of the Anderson-Darling test to discrete distributions should be used to assess whether empirical dbh data are consistent with a hypothesized ' distribution.