UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Analysis of measurement errors associated with variable-radius plot forest sampling Omule, Stephen Agnew Yen’Emurwon

Abstract

Forest inventories form a basis of decision-making in most aspects of forest resources management. Implicitly, therefore, inventories must be conducted efficiently and with a minimum of error. This thesis analyses the errors of measurement during a variable-radius plot forest inventory cruise, with the objective of determining crew variation and bias in tree count, diameter, and total height measurements. Data for the study were collected at The University of British Columbia Research Forest, Maple Ridge, British Columbia, during the third year forestry students field school in mensuration. In each of the six lots of 10 pre-set sample plot centers, three to six 3~man student crews each established prism and relascope plots, took the tree count, and measured the diameters at breast height of all the "in" trees and the total heights of the first "in" trees. The crews took their measurements independently of each other, and had to complete the exercise within a period of 8 hours. The author's measurements and observations in the same plot centers formed the control to the study. The crew variation in tree count was calculated by the method of Analysis of Variance, and in diameter and height measurement by pooling the variation for each tree to obtain a weighted average variation per tree. Bias was evaluated by comparing the crew results with those of the control. The following results were obtained from the study: The average tree count was 9.5 trees in the basal area factor (BAF) - 6 prism plots and 7.4 trees in the BAF = 9 relascope plots. The coefficient of variation of observer tree count was 10.445 with the prism and 4.93% with the relascope. The percentage error in the determination of basal area per hectare from 2 5 prism plots was as much as ± 4.09%. and from 25 relascope plots ± 1.93% at the a = .05 probability level. About 37% of the tree counts in the relascope plots and 25% in the prism plots were measured without error. The maximum tree count error v/as ± 6 trees per plot. The average tree diameter was 52 .67 cm, and the measurer coefficient of variation was 8.16%. Only 6% of the tree diameter measurements were correct. Accuracy was lower at larger tree diameters. The average tree height was 31.38 m, and the measurer coefficient of variation of the tree height measurements was 21.86%. Crews measured tree heights with a significant bias. Only about 2% of the measurements were correct, with over 15% of the measurements being in error by ± 6.0 m or more. Height measurement was subject to larger and more sources of error. The results suggest that forest resouce managers should be more careful in using untrained crews in forest inventory work. They should establish rigorous field training programs, and outline and implement checkcruising guidelines.

Item Media

Item Citations and Data

Rights

For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use.