University of British Columbia
Classification of windthrow on outblock boundaries from Landsat 7 ETM
Abstract
dc:descriptionThe use of Landsat 7 ETM satellite data to detect windthrow on cutblock boundaries on North Vancouver Island was investigated in this study. Mature western hemlock (Tsuga heterophylla (Raf.) Sarg.), Pacific silver fir (Abies amdbilis (Dougl.) Forbes), and Sitka spruce (Picea sitchensis (Bong.) Carr) stands in that area are susceptible to windthrow due to the stand structure and topographic exposure. An endemic windthrow problem has arisen on cutblock boundaries, road edges, and riparian strips. Two approaches were used to predict the windthrow on the cutblock boundaries. The first used logistic regression analysis to model probability values for the pixels in the Landsat scene. The second used supervised classification to identify the windthrow pixels. The input data for both these methods included the six multispectral Landsat 7 ETM bands, and a series of forty vegetation indices that were calculated for the image data. The results indicate that using only the logistic regression models to identify windthrow pixels is insufficient. However, the supervised classification showed a 72.7% success rate in finding high severity windthrow pixels on cutblock edges. The highest accuracy in detection was for large areas of heavy windthrow.
Degree
thesis:*- Name thesis:degree_name
- Master of Science - MSc
- Level thesis:degree_level
- master's
- Discipline thesis:degree_discipline
- Forestry
- Grantor dc:publisher
- University of British Columbia
- Year dc:date
- 2003
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Ortlepp, Stephanie Maren
Rights
dc:rights- Statement dc:rights
-
- For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use.
- Language dc:language
- eng
Identifiers
dc:identifier.*- Handle dc:identifier
- http://hdl.handle.net/2429/15010