Year of Award


Document Type


Degree Type

Doctor of Philosophy (PhD)

Degree Name


Department or School/College

College of Forestry and Conservation

Committee Chair

LLoyd P. Queen

Commitee Members

Kelsey S. Milner, Melinda Moeur, Robert D. Pfister, Hans R. Zuuring


Agriculture, Forestry and Wildlife


University of Montana


The original design of this dissertation project was relatively simple and straightforward. It was intended to produce one single, dynamic, classification and mapping system for existing vegetation that could rely on commonly available inventory and remote sensing data. This classification and mapping system was intended to provide the analytical basis for resource planning and management. The problems encountered during the first phase of the original design transformed this project into an extensive analysis of the nature of these problems and a decade-long remote sensing applications development endeavor. What evolved from this applications development process is a portion of what has become a "system of systems" to inform and support natural resource management. This dissertation presents the progression of work that sequentially developed a suite of remote sensing applications designed to address different aspects of the problems encountered with the original project. These remote sensing applications feature different resource issues, and resource components and are presented in separate chapters. Chapter one provides an introduction and description of the project evolution and chapter six provides a summary of the work and concluding discussion. Chapters two through five describe remote sensing applications that represent related, yet independent studies that are presented essentially as previously published. Chapter two evaluates different approaches to classifying and mapping fire severity using multi-temporal Landsat TM data. The recommended method currently represents the analytical basis for fire severity data produced by the USDA Forest Service and the US Geological Survey. Chapter three also uses multi-temporal Landsat data and compares quantitative, remote-sensing-based change detection methods for forest management related canopy change. The recommended method has been widely applied for a variety of forest health and disaster response applications. Chapter four presents a method for multi-source and multi-classifier regional land cover mapping that is currently incorporated in the USDA Forest Service Existing Vegetation Classification and Mapping Technical Guide. Chapter five presents a study using nearest neighbor imputation methods to generate geospatial data surfaces for simulation modeling of vegetation through time and space. While these results have not yet been successful enough to support widespread adoption and implementation, it is possible that these general methods can be adapted to perform adequately for simulation modeling data needs.



© Copyright 2007 Charles Kenneth Brewer