Comparison of graphical and statistical methods in determining the number of sampling units in vegetation studies of desert ecosystems of South Khorasan

Background and objectives: Determining the minimum number of samples and the minimum parameters to be measured that provide acceptable results is one of the most basic decisions that rangeland scientists need to make. Methods for determining the number of samples in rangeland vegetation measurements are different. graphical methods and statistical characteristics are the most common methods by measuring one or a number of dominant plant traits; The number of sampling units can be determined. Therefore, in the present study, the two methods were compared in the desert ecosystems with the predominance of Salsola richteri and Ammotamnus lehmannii in South Khorasan. Methodology: For the study, an area of ​​100 hectares of habitat was selected as key area, and to determine the number of sampling units, based on previous studies and vegetation distribution; 10 initial plots of 10×10 meters were randomly established inside the key area. Then by adding the number of plots and estimating the canopy cover percentage and the number of plant bases of Salsola richteri species at different times; The appropriate number of sampling units was determined by conventional graphical method in rangeland assessment studies and application of species accumulation curve (SAC). Also, the appropriate number of sampling units was determined by statistical methods including the first and second Cochran's sample size formula, Yamane's sample size formula, confidence interval, and coefficient of variation (CV). For this purpose, all statistical calculations were performed in the biotools package in R software environment. Results: Based on the application of graphical methods:  application of species accumulation curve (SAC) and conventional graphical method in rangeland assessment studies, the appropriate number of sampling units was identified 40 and 50, respectively. Based on the obtained results, using the first and second Cochran's sample size formula and the correction factor with an acceptable error rate of 10% for the study area, about 78 to 92 plots are required, respectively. If we continue the initial sampling to reach the power level of 100%, about 200 plots will be needed. Also, the number of plots obtained by range method (minimum: 53, maximum: 160 and optimal: 75) is closer to the Cochran's sample size formula. The weakest criterion is CV. Based on this method; the results showed that 10 plots have the lowest CV and the most appropriate number of plots in the study area. The most suitable number of plots in the study area is 75, because the optimal sampling design provides maximal information via minimal effort. Conclusion: The results showed that based on the application of graphical methods, more than 40 sampling units are suitable for measuring vegetation. Also, the lower cost of the species accumulation curve (SAC) method due to the use of fewer sampling units; The species accumulation curve (SAC) method is proposed to calculate the appropriate number of sampling units for the desert ecosystems of this region. The results of this study showed that there is no significant relationship between the coefficient of variation (CV) and the number of initial plots, ie increasing the number of plots does not necessarily reduce the standard deviation of the data, so it is suggested in studies If the number of initial plots are less than 40, do not use the graphical method or the relationship between the CV and the number of plots.