Line intersect sampling (LIS) is a method used for quantifying post-harvest waste. It is often used by forest managers to quantify merchantable volume remaining on the cutover so that compensation may be exacted under stumpage contracts.
The theory has been extensively studied and will produce an accurate measure of harvest waste given the basic theoretical assumptions that: all logs are cylindrical, occur horizontally, are randomly orientated and randomly distributed. When these assumptions are violated, the method is not biased, although precision decreases substantially.
A computer simulation was completed to determine whether or not the LIS method is appropriate, given a clumped distribution of logs produced by processing at central sites in cutover before using a forwarder to extract to the landing. The software ArcGIS with the application ModelBuilder was used to produce the LIS Model for running LIS assessments.
It was determined through simulation that the conventional LIS method is not appropriate given these harvesting methods, as a level of bias was found in sampling determining that the LIS method underestimated true volume. T-tests confirmed the significance of this bias.
LIS volume estimates were not precise, with the range of estimates ranging from 0 m3/ha to double the true volume. An increase in sampling length by a third was found to increase precision by only a small amount. Therefore, it was determine that increased sampling is not worthwhile as the costs associated with it do not justify the small increase in precision.
Identifer | oai:union.ndltd.org:canterbury.ac.nz/oai:ir.canterbury.ac.nz:10092/10463 |
Date | January 2014 |
Creators | Tansey, Joshua |
Publisher | University of Canterbury. School of Forestry |
Source Sets | University of Canterbury |
Language | English |
Detected Language | English |
Type | Electronic thesis or dissertation, Text |
Rights | Copyright Joshua Tansey, http://library.canterbury.ac.nz/thesis/etheses_copyright.shtml |
Relation | NZCU |
Page generated in 0.002 seconds