Open Collections

Brown, Madison Sophia

University of British Columbia

Non-stand replacing disturbances (NSRs) are events that do not result in complete removal of forest stands and generally occur at a low intensity over an extended period (e.g., insect infestation), or at spatially variable intensities over shorter periods (e.g., windthrow). The overall structural change associated with NSRs can impact both timber supply and ecosystem services, suggesting their detection and structural characterization is critical to inform forest inventory programs. The increased accessibility of high frequency revisit, medium resolution satellite imagery, has led to a subsequent increase in algorithms designed to detect sub-annual change across broad spatial scales. One of these algorithms, the Bayesian Estimator of Abrupt change, Seasonal change, and Trend (BEAST) has shown promise in detection of disturbance and has a documented relationship to change in Light Detection and Ranging (LiDAR) derived forest stand structural attributes. This thesis aims to evaluate the sensitivity of BEAST to detect and subsequently characterize NSRs across a range of severity levels and NSR types in a dry interior forest in Western Canada. First, satellite time-series data were processed for NSR sensitive indices and the BEAST algorithm was used to generate probability of change rasters and three time series variables capturing phenological variation (i.e., amplitude, slope, and trend). Second, to evaluate the detectability of NSRs, change probability rasters were compared to the occurrence, severity, and timing as mapped by historical survey polygons. In order to determine the BEAST’s ability to update forest inventories, probability distributions of major NSRs were compared between consecutive years of disturbances. Cumulatively, all levels of NSRs had higher and statistically significant (p < 0.05) mean probabilities compared with historically undisturbed areas. Third, to structurally characterize NSRs, three LiDAR derived forest attributes were modeled using the BEAST attributes amplitude, trend, slope (canopy cover, height and height variability). Then applied to a 5-year outbreak of aspen leafminer (Phyllocnistis populiella) and two-year budworm at (Choristoneura biennis) estimating an 11% decline in canopy cover. The thesis overall, allowed for structural changes attributed to a variety of NSRs to be tracked over time increasing opportunities for early detection, assessment, and mitigation strategies to be implemented.

Text

2025-07-10

Attribution-NoDerivatives 4.0 International

http://hdl.handle.net/2429/91546

Forestry

University of British Columbia

UBCV

http://creativecommons.org/licenses/by-nd/4.0/