Hierarchical variable probability sampling for carbon estimation

Abstract

This thesis implemented variable probability sampling into carbon inventory and tested the possibility of using sampling to correct LiDAR- assisted enhanced forest inventory for carbon estimation. At first, the efficiency of big BAF sampling was tested using simulation to develop sample designs and estimate costs. Then, list sampling with big BAF sample plots was implemented in a case study carried out at the 5th Canadian Division Support Base Gagetown. In this case study, double sampling with ratio estimation was used to calibrate LiDAR-derived estimates of total volume. Standing trees and dead woody debris carbon estimates were made. Our results showed that big BAF sampling is very efficient for carbon inventory and that a sampling to correct approach using list sampling can yield accurate estimates of aboveground carbon.