River Simulations
August 2008

Page: 1 * 2 * 3 *** Research log, final report, and GRL paper

I am currently working on a year-long project for Oregon State University and the National Park Service to predict a 100 year migration corridor for the 59-mile segment of the Missouri National Recreational River from Gavins Point Dam to Kensler's Bend (approximately from Yankton, South Dakota, to Sioux City, Iowa). This prediction will be based on running multiple meandering simulations using the Johannesson-Parker 1989 model and other methods. I have written several programs in Python and Wx to digitize the banks of the river from USGS topographic maps and Google Earth satellite images, to calculate the river centerline and width, to digitize the valley walls and other simulation boundaries, to perform the meandering simulations, to maintain river points within the simulation boundaries, and to check for and remove loop cutoffs (switchbacks in the river which touch or cross each other). These programs will be used to run many slightly different simulations in order to perform a sensitivity analysis of river evolution with respect to input parameter values, and to accumulate 2d coverage statistics showing the location of the various simulated rivers over time. This data will be used in planning river management for the next century, and also to delimit the potential growth of housing and other structural/population development within the valley.

Here is the location of the study area:

Here is a mosaic made from several USGS topographic maps:

(Click for larger image)

(Click for larger image)

With a little work (e.g. separate x and y scaling), and some luck, the two mosaics can be made to superimpose well, even though the topographic maps were made at different dates from 1957-1996, and the satellite images were made much later (2006?). Although the position of the river has changed slightly, the endpoints and other man-made features are the same:

(Click for larger image)

Here is the elevation within the valley. It is essentially flat (within 30 m) over a distance of 100 km:

(Click for larger image)

Here is the valley in 3d:

(Click for larger image)

I've written a program that simulates a meandering river, using one of several different models. The program also accumulates the cutoffs which occur as the river loops back over itself. There are many examples of this kind of behavior in the literature, for all kinds of actual rivers:

(Click for larger image)

Here is an example of the evolution of a synthetic meandering river which starts from a straight line with slight random perturbations. The accumulated cutoffs are shown in green:

(Click for larger image)

I've written another program to digitize the coordinates of the river banks and valley walls, and to interpolate them to a resolution of 100 m using circular arcs:

The meandering routines require a centerline and width, so the program computes these based on the digitized bank coordinates:

Here is the entire digitized river and valley superimposed on a 1 km grid:

(Click for larger image)

Here are a few 'unconstrained' 100 year simulations of the river (i.e. the inlet has been allowed to move and the river has been allowed to exceed the boundary. All of the simulations shown below have greatly exaggerated motion of the river, in order to test the models and program. Less active, but more realistic, simulations will be shown on the next page):

(Click for larger image)

(Click for larger image)

(Click for larger image)

Here are a few 'constrained' simulations (i.e. the inlet has been fixed in place and the river cannot exceed the boundary. Again, these simulations show far more activity than is expected of the actual river over the same time period):

Here is a simulation of just the inlet of the river (at Gavins Point Dam):

(Click for larger image)

Continued on next page