Precision and accuracy of sampling methods in thinned Pinus taeda L. stands

Abstract

Different sampling methods can be used in forest surveys. It is important to know the precision and accuracy of these sampling methods, and which one is the most appropriate in specific conditions of the forest population. The aim of this study was to compare estimates of a forest inventory performed by different sampling methods with forest census results. The sampling methods evaluated were the fixed-area method and the variable-area methods of Bitterlich, Prodan, and Strand. The data were obtained in a 15-year-old thinned stand of Pinus taeda L., located in the municipality of Teixeira Soares, southern Brazil, with a total area of 12.80 ha. Initially, the forest census was carried out, and subsequently, the sample units for each sampling method were distributed in the stand, with a common starting point. The variables used to compare the sample results with the census means were quadratic diameter, number of trees, basal area, and volume, per hectare. Precision and accuracy were evaluated by sampling error and whether the confidence intervals covered the population means, respectively. The fixed-area and Bitterlich methods stood out in precision for all variables analysed. The fixed-area, Bitterlich and Strand methods with proportion to height provided more accurate estimates. The Prodan method provided inaccurate and imprecise estimates for the variables under analysis, except for the quadratic diameter.

SAMPLING USING THE BITTERLICH METHOD IN Pinus taeda L. PLANTATIONS WITH DISTINCT MANAGEMENT SYSTEMS

“…For example, it is customary to use "appropriate" BAFs equal to 1 or 2 in Brazil, standardizing the count between 20 and 30 trees per point, being identified in studies involving different sampling methods in forest inventories carried out in forest plantations such as by Druszcz et al (2015) in Pinus taeda plantations with and without thinning, Miranda et al (2015) in commercial areas with Tectona grandis, Farias et al (2019) in a Pinus sp. stand with low tree density, and Miranda et al (2022) in thinned Pinus taeda plantations similar to the EXP3 of this study.…”

“…The estimates of the grouped variables of the Bitterlich sampling method were the same used by Druszcz et al (2015) and Miranda et al (2022), developed by Grosenbaugh (1958), and for average diameter by Nishizawa (1972), contained in Table 1. In which: "d" is the mean diameter at breast height at 1.30 meters from the ground (cm); K is basal area factor (m².ha); di is the diameter of tree "i"; gi is the cross-sectional area of tree "i" (m²); vi is the average individual volume (m³); n is the number of trees within the Bitterlich point; π is the constant PI.…”

“…Forestry products are essential for the well-being of society, commonly used in the form of assorted wood, reconstituted panels, charcoal, paper and cellulose, among other products. With the development of society, pressures on forests (native and planted) and their by-products have increased, requiring managers to develop the necessary management plans to guarantee the sustainable use of forest resources (FARIAS et al, 2002;FIORENTIN et al, 2015;MIRANDA et al, 2022).…”

“…Authors such as Grosenbaugh (1958), Pellico Netto andBrena (1997), Sanquetta et al (2014), Avery and Burkhart (2015), Farias et al (2002) and Miranda et al (2022) stated that there are two basic forms of forest inventories: through a census (or 100% inventory of the trees), but which is unfeasible in most cases due to the high costs and time involved in the field survey; or by sampling (from a part of the population that rarely exceeds 2% of the whole), which is a safer and cheaper way of estimating the parameters.…”

“…The chosen BAF will influence the number of individuals counted within the SU, and some authors, such as Pellico Netto and Brena (1997), Avery and Burkhart (2015), Machado and Figueiredo Filho (2014) and Sanquetta et al (2014), have suggested that the safe number to establish a sampling unit by this method is between 20 and 30 trees. Scientific works involving Bitterlich samplings in Pinus taeda L. stands by Druszcz et al (2015), Farias et al (2019) and Miranda et al (2022), and in Tectona grandis stands by Miranda et al (2015) are based on this range of 20 or 30 trees per SU.…”