Bias analysis of modified estimator of N-tree method for estimating stand density and crown cover area in Zagros Forests

Since N-tree sampling methods produce biased estimate, bias evaluation should be applied to N-tree estimator. The purpose of this study is to assess the N tree distance sampling method bias with modified estimator Kleinn and Vilcko in estimating number per hectare and canopy cover in Zagros forests. For this purpose, 6-10 trees from real forest map were tested with full inventories. For comparing the results of characteristics in different trees in N-tree method with full inventories, ANOVA was used. Evaluation of N tree distance sampling method was done by using accuracy criterion and standard error of mean. Bias was also evaluated via one-sample t-test and relative bias. Results showed that in estimating number per hectare and canopy cover, there is no significant difference between tree numbers with full inventory. Bias test for six and seven trees per hectare was significant, but it was not significant for eight, nine and 10 trees. Canopy cover for all tree numbers showed no significant difference. So, in order to achieve a maximum accuracy (minimum bias), at least eight trees within sample plot should be measured. Finally, this study implies that in forests with lower density and cluster pattern, bias change in N tree distance method does not follow a predictable decreasing pattern.