Iranian Journal of Applied Ecology

Mendeley  

Zotero  

Rostampour M. Assessment of Statistical Distribution and Data Transformation Methods for Vegetation Cover Percentage in the Rangelands of the Darmian-Sarbisheh Protected Area. Iranian Journal of Applied Ecology 2026; 14 (4) :71-89
URL: http://ijae.iut.ac.ir/article-1-1297-en.html

This study examined the statistical distribution of vegetation cover percentages across ten rangeland types within the Darmian-Sarbisheh Protected Area and evaluated the efficacy of various data transformation techniques. Vegetation cover was quantified in 600 plots of 4 m² each. Normality assessments employed skewness coefficients, graphical analyses, and formal tests-including Shapiro-Wilk, Anderson-Darling, and Lilliefors tests. Results revealed that nine of the ten vegetation types exhibited significant right-skewness and non-normal distributions, with only one type conforming to normality. Distribution fitting using the Akaike Information Criterion indicated that the normal distribution was inadequate for approximately 90% of the data, whereas log-normal and Weibull distributions provided better fits-accounting for 80% and 20% of the types, respectively. Among six transformation methods (square root, cube root, natural and base-10 logarithms, inverse, and Box-Cox), only logarithmic and Box-Cox transformations effectively normalized 50% of the datasets; the inverse transformation was entirely ineffective. Given the pervasive non-normality in semi-arid rangeland vegetation cover data and the limited success of transformations, the application of robust statistical techniques-such as Robust Principal Component Analysis-is recommended for ecological and rangeland management studies, ensuring more reliable analytical outcomes.