|
Title:
|
Using Multiple Resolution Digital Images and GIS to determine vegetation sampling locations at Whiskeytown National Recreational Area
|
|
Author:
|
McGovern, Ian Patrick
|
|
Abstract:
|
The Whiskeytown National Recreation Area located west of Redding, California has a very diverse landscape and associated plant diversity. High resolution (1-meter) multispectral imagery was to be used in devising a sampling plan for use in mapping the diverse vegetation. The high-resolution imagery did not perform well by traditional classification techniques, resulting in an unmanageable amount of spectral classes. Also this image had brightness banding along flight lines that could not be corrected. A Landsat ETM+ image provided the necessary results to map at an alliance level with a minimum mapping unit of 0.5 hectare. Due to the loss of resolution various methods were implemented to increase the usability of the imagery. A radiometric spectral merge was done with a 15-meter panchromatic image. Additionally the unit was split into six different physiographic zones based upon elevation and aspect. These different zones were classified separately by unsupervised ISODATA routines. A final supervised classification over the entire study area was performed following a separatability analysis of the six classified zones. Utilizing the GIS, spatial characteristics of the contiguous spectral classes were used to derive sampling locations based upon defined guidelines of area and geographic location. |
|
Description:
|
Thesis (M.S.)--Humboldt State University, Natural Resources: Planning and Interpretation, 2008 |
|
URI:
|
http://hdl.handle.net/2148/454
|
|
Date:
|
2008-12 |
