A Simple Meandering Model

During the past few days I created another simple physical meandering model that looks good (if nothing else). It is based on a sub-linear (exponent = 0.8) equation of the lateral force, again averaged (every 3 points) and shifted downstream (by 1 point) before change in position is calculated using rerody, lerody, lambda, and a scaling coefficient (1.0e8). Details can be seen in the Python code. The Topographic Steering Python code is also available for comparison.

Here is the new model run for 1000 iterations on a 10x2.5 m channel in a 1000x1000 m domain. Only the flow edges for the meandering nodes are shown in blue:

(Click for larger image)

In my 'spare' time I've also been working on new meshing routines that can triangulate an arbitrary region containing holes and multiple domains at arbitrary orientations, in anticipation of modeling the Missouri river as a macroscopic entity (potentially with islands), rather than as an ideal centerline/width/depth geometry. I don't know if the new meshing code will be required, but it might be good to have it ready just in case:

(Click for larger image)

(Click for larger image)

This completes the tasks for this month, which were:

1. To run some meandering tests with Child.
2. To become familiar with the Topographic Steering C++ code.
3. To create an external test environment for new meandering code outside of Child, but which can be called from Child.
4. To compare the TS model in C++ and Python.
5. To try some additional simple meanding models before moving on to JP.

Next month I will:

1. Work exclusively on the new JP meandering code.
2. Continue to enhance the graphics, meshing, and other ancillary code as necessary.