Date of Award

2010

Degree Type

Thesis

Department

Biological, Geological and Environmental Sciences

First Advisor

Clapham, Wentworth B.

Subject Headings

Vegetation mapping, Landsat satellites, Remote sensing, Colorado River Watershed (Colo.-Mexico), Lower Colorado River Riparian, LCRAS, Saltcedar, Vegetation Mapping

Abstract

In the Western United States, monitoring water usage is a complex task carried out by the U.S. Bureau of Reclamation (USBR). It may be argued that USBR's greatest challenge is equitably distributing the waters of the Colorado River, particularly the Lower Colorado River, where water rights have been established and contested several times. To help meet the demands of water management in the Lower Colorado River Basin, USBR estimates the amount of water lost from the basin each year via evapo-transpiration by riparian vegetation in the Lower Colorado River riparian zone. Key components of those estimates include maps of the vegetation itself, which provide a measure of the acreage covered by each dominant species. Previous mapping efforts have relied extensively on costly in-situ field measurements using the Anderson-Ohmart Classification scheme (which was developed for habitat evaluation, not species identification) and data-dense high resolution aerial photographs. This study employs low resolution Landsat imagery and simple classification and clustering algorithms to identify heterogeneous species assemblages in the Lower Colorado River as possible alternatives to Anderson-Ohmart and/or high resolution aerial photographs. Our results show that the method developed here is able to identify heterogeneous riparian species assemblages, but certain vegetative species can be mapped with greater accuracy than others. Pending an error assessment to be carried out in a future field season, we believe our method to be an inexpensive, relatively simple update to USBR's existing mapping procedures

COinS