Date of Award:


Document Type:


Degree Name:

Master of Landscape Architecture (MLA)


Landscape Architecture and Environmental Planning

Committee Chair(s)

Craig Johnson


Craig Johnson


Land-cover identification and mapping are an integral part of natural resource planning and management. Satellite imagery provides a way to obtain land cover information, particularly for large tracts of land such as those administered by federal and state agencies.

This study assesses the usefulness of the Brightness/Greenness Transformation of Landsat Thematic Mapper data for differentiating conifer forest types in northern Utah. Satellite data for the Logan Ranger District of the Wasatch-Cache National Forest were classified into 27 vegetation classes. Of these, nine were determined to be conifer classes and were used in subsequent analyses. Ten sites of each conifer class type were field checked and vegetation and physical site characteristics recorded.

The Brightness/Greenness Transformation was able to distinguish conifer areas from other vegetation types. High-density conifer classes were classified at 94 percent accuracy. Low-density conifer classes were classified correctly 65 percent of the time. The Brightness/Greenness Transformation alone met with limited success in distinguishing between conifer species. Each class showed great variability with respect to major overstory species. Analysis of variance indicated that none of the site factors measured consistently corresponded with the spectrally designated classes. While several factors differed significantly among classes, no factor was significantly different for all class-pair combinations.

Correlation analysis revealed that brightness, greenness, and wetness values related more to environmental values than to conifer species. Brightness was highly correlated with percent of exposed soil on the site. Greenness was highly correlated to the presence of deciduous and herbaceous vegetation. Wetness was highly correlated to total tree and conifer cover values.

Adding slope and aspect data to the Brightness/Greenness Transformation classes with the highest percentages of canopy cover did allow separation of lodgepole pine and Douglas fir. High percentage canopy cover sites on slopes less than 35 percent were classified as lodgepole pine with 89 percent accuracy. On slopes greater than or equal to 35 percent, Douglas fir was found with 79 percent accuracy.