Digital Analysis of Changes in Channel Morphology, Lower Virgin River, Arizona and Nevada

For a 30-mile stretch of the river, 8 separate reaches, each about 1 to 1.5 miles long, were being processed. Aerial photographs from 29 dates were optically scanned into digital images and registered to a standard coordinate system, using control points and image-processing techniques. Channel boundaries, sand bars, and the thalweg -- the deepest part of the channel -- were manually delineated onto prints of the registered images and then digitized into ArcInfo coverages. The coverages were verified for accuracy of locational coordinates and feature coding. These coverages were used as the base for further digital processing.

All digital processing was done using a combination of ArcInfo and FORTRAN programs developed for this study. The entire process was controlled through an interactive menu, designed for rapid sequential processing of a large number of image coverages for each reach. To begin processing, the reach orientation, or overall linear direction of the river within that reach, was delineated. The reach orientation was used in subsequent processing of each available image for that reach. Next, the area of the sand bars, the area of the channel, the thalweg sinuosity, and the maximum width of the channel meander belt, perpendicular to the reach orientation, were calculated for each image coverage. In addition, the reach was subdivided into a series of about 25 equal-width segments along the reach-orientation line. For each segment, lines were constructed that were perpendicular to the average orientation of the channel boundaries within that segment. (example 1) These lines were used to calculate the width of the channel at the center of each segment. The calculated data were stored for further geomorphic comparison and interpretation.

The channel-area features for any two selected images can be overlaid for planimetric analysis, showing changes from one date to another. (example 2) Also, several different types of stability maps can be produced using any of the delineated channel features. The selected channel features can be superimposed on a common base and the frequency of occurrence computed. (example 3) and (example 4) Regions with high frequency of occurrence represent greater channel stability. Stability maps are a quantitative as well as visual way of looking at channel stability.

By using digital methods, a large data set can be processed and analyzed quickly and easily. These results can then be used to document channel changes through time, assess channel stability, and help understand geomorphic processes.