Channel Migration: A First Study
From Analyses of an Experimental Delta Basin and an Experimental Braided River with Varying Levels of Vegetation
Understanding the rates at which river channels migrate laterally is an important problem in sedimentology and geomorphology, with applications to paleoecology, stratigraphy, geoarchaeology, riparian ecology, watershed management, and understanding and preventing loss of life and property from natural hazards. Unfortunately, until recently, the proper computing facilities have not been sufficiently readily available to analyze this problem in a systematic and quantitative way for systems beyond simple meandering rivers. This work is a still-in-progress first foray into a simple method to analyze the migration rates of rivers under different environmental and sedimentological conditions.
My work analyzes timelapse images taken from two separate analog experiments for decorrelation in their patterns of water and land. The first experiment, the 2002 run of the Experimental Earthscape Facility (Jurassic Tank), contains data on an experimental delta fan with constant subsidence at a rate that increases in the downstream direction. This experiment also includes slow, rapid, and rapid-on-slow cycles of baselevel rise and fall. The second experiment is a simple braided channel that becomes vegetated. The vegetation forces the river into a more meandering and slower-migrating pattern than the initial braided system.
In each of these experiments, the water was filled with dye to make it apparent and separate from its substrate. By using the color difference between the dye and the sediment as well as the increasing concentration of the dye's color with the depth of the flow, it was possible to create a threshold that separated these images into a binary matrix of land and water. These white (wet) and black (dry) pixels were then compared in two main ways. To study migration, a series of images were compared to a single baseline image. To study an avusion timescale, each of these images was compared to the one before it. To This effect, the following equation was used to describe the change between images. Decorrelation was defined as the number of pixels that changed in a set of two images. This is the number of pixes that changed from wet to dry, plus the number of pixels that changed from dry to wet. This number is then scaled to the maximum number of pixels that could change. Since the number of wet pixels was always less than the number of dry pixels, the number of wet pixels in the picture with more wet pixels (of the two pictures being compared), times two, was the maximum amount by which a set of wet and dry pixels could decorrelate from itself Therefore, the decorrelation was scaled by this value. After this, the whole decorrelation term was subtracted from one, so the remaining solution was the correlation between the two river systems.