Document Type

Thesis

Degree Name

Master of Environmental Studies (MES)

Department

Geography & Environmental Studies

Faculty/School

Faculty of Arts

First Advisor

Barry Boots

Advisor Role

Thesis Supervisor

Abstract

Mountain pine beetle (Dendroctonus ponderosae Hopkins) is an endemic species in the forests of British Columbia that has become epidemic and reached infestation levels like never before. Different approaches have been taken in order to try and manage the forest and understand the processes affecting the behavior of mountain pine beetle. No single model has been entirely successful in unearthing the complexity of mountain pine beetle behavior. In this thesis, large spatial data sets of mountain pine beetle attacks, obtained from helicopter and ground surveys, and further adjusted for the incorporation of uncertainty, are studied using a spatial autocorrelation approach in a pattern-based analysis. The study of spatial patterns is carried out by simulating possible scenarios of the observed data set. Moran’s I is used to obtain an overall measure of spatial autocorrelation of the global pattern and Local Indicators of Spatial Autocorrelation, specifically Local Moran’s I, are used to identify local pockets of high levels of infestation (hot spots). Using a significance criterion, regions that have intense infestations are screened to retain those that are more pervasive, thus having a more robust set of results that can be more reliable. Different levels of significance can be used to allow for a more ‘liberal’ or ‘strict’ screening of results. Study of the sensitivity of the data model and detection approach is carried out by comparing the locations of hot spots obtained with different detection methods. A comparison between results derived from data sets containing only aerial data and those containing aerial and field data is useful to determine the impact and effectiveness of sending crews to groundtruth aerial surveys.

Convocation Year

2006

Convocation Season

Fall

Download

Share

COinS