Effect of sample plot size and shape on estimates of structural indices: A case study in mature silver birch (Betula pendula Roth) dominating stand in Järvselja

Abstract

Structural indices and characteristics have often been employed for assessing different aspects of forest stands. In order to assess the performance of such measures in forest sample plots of different size and shape, the current study is based on a data set for a relatively large (1 ha) silver birch dominated stand, and five simulated planar point patterns (a Poisson process, two clustered and two regular patterns) with similar intensity of points distributed on an area equal to the silver birch stand as reference stands. The analyses was based on repeated samplings of the stands, with randomly placed circular or square shaped plots of different sizes, with area ranging from 0.007 to 0.636 ha. Similar centre positions were considered for plots of different shapes, but the locations of the plot centres differed with size, so that plots of different sizes were able to cover the maximum possible area of reference stands. Some structural indices accounting for one or four nearest neighbours, and also the Point-pair correlation function was then quantified for each random sample plot. We used standard deviation and the differences of the mean estimates from the expected values to assess the sensitivity of the structural measures to the size and shape of sampling plots. The results indicated the significant effect of plot size and shape on structural indices. The correlation between plot size and estimates precision was positive, and small plots, regardless of their shapes, showed big variation among them while circular plots often provided more accurate estimates. Even though, as expected, expanding the plots increased the precision of the neighbourhood-based indices’ estimates (spatial measures), and also resulted in inaccurate estimates for some reference stands, depending on the arrangement of trees/points within that stand.