Evaluation of the sensitivity of statistical approaches for the analysis of forest inventories in the environmental inspection process

Abstract

Effective forest inventory auditing depends not only on field audits but also on the ability of the agency to correctly identify deviations in the data, making the choice of appropriate statistical approaches essential. This study assessed the sensibility of two approaches for forest inventory analysis, aiming to determine the required deviation in diameter that results in rejection by the environmental agency. Three types of sample inventories were used: native forest, linear projects, and forest plantation. For the audit, 10% of the plots were randomly selected, ensuring a minimum of 3 selected plots. The paired t-test was used to compare means at the tree level. The adherence test was conducted using the Kolmogorov-Smirnov (KS) test. To identify when the tests would start indicating significant differences between inventories, a gradually increasing deviation was added to the simulated checked data. No significant sensibility was found for sampling error considering the analyzed deviation. However, the t-test pointed significant differences for mean deviation greater or equal to 0.1 cm for diameter, while the Kolmogorov-Smirnov test shows significant differences starting at 0.5 cm for the same comparison.