Year of Award

2020

Document Type

Thesis

Degree Type

Master of Science (MS)

Degree Name

Geography (Cartography and GIS Option)

Department or School/College

Department of Geography

Committee Chair

Anna Klene

Commitee Members

David Affleck, Kevin McManigal

Keywords

remote sensing, forestry, montana, DNRC

Subject Categories

Forest Management | Geographic Information Sciences | Natural Resources and Conservation | Remote Sensing | Spatial Science

Abstract

Remote sensing can be utilized by land management organizations to save money and time. Mapping vegetation using either aerial photographs or satellite imagery and the applications for forest management are of particular interest to the Montana Department of Natural Resources. In 2018, the organization began a pilot program to test the incorporation of raster analysis of remotely sensed data into their inventory program and had limited success. This analysis identified two areas of improvement: the selection method of inventory plots and the imagery used for classification and metrics. This study found that selecting inventory plots using a generalized random tessellation stratified (GRTS) sampling design in spectral space would likely improve the representation of the population as the sample distributions for mean and standard deviation in all spectral bands were more concentrated about the population means. Analysis using Sentinel-2 based predictors produced results that were comparable to predictions built using predictors derived from high resolution National Agricultural Inventory Program (NAIP) imagery. The increased spectral/radiometric/temporal resolution of Sentinel-2 imagery may have compensated for its lower spatial resolution.

Share

COinS
 

© Copyright 2020 Ryan P. Rock