A goal of sustainable forest management using digital soil mapping (DSM) is to ensure that current and future generations have the best soil information so they can use forest resources wisely. This goal can be achieved using new technologies of generating digital soil maps and high-resolution light detection and ranging (LiDAR) data. Uncertainty in digital soil maps can be quantified using quantile regression (QR). The overall objective of this study is to generate several digital soil maps using different machine learning (ML) methods for forest management purposes and use a QR method to estimate their uncertainty. The study area is the Eagle Hill Forest (95 km2), located west of Kamloops, BC, Canada. Five soil properties were mapped and locations with soil erosion, displacement, and compaction and puddling hazards were displayed on maps and discussed. 90% prediction interval (PI) maps were produced and the performance of the QR method in uncertainty quantification of different ML models was illustrated by producing Prediction Interval Coverage Probability (PICP) plots.
Copyright is held by the author.
This thesis may be printed or downloaded for non-commercial research and scholarly purposes.
Supervisor or Senior Supervisor
Thesis advisor: Schmidt, Margaret
Member of collection