I've tweaked last week's trigonometric reaction diffusion model to get some interesting results.
The main change is that the reaction term is slightly different for each chemical species. For u, it's essentially cos(u)⋅v2 and for v it's essentially sin(v)⋅uv (the reaction term in the Gray Scott model is the same for both species: uv2).
Species values are now based between zero and one - I multiply by 2π inside their respective trigonometric functions (I also add 1 to the result and divide by two).
I've kept with the same nine point stencil for calculating the Laplacian but fiddled with the boundary conditions to make the initial state more conducive for some funky patterns.
This version of the model often displays global phase changes - where the entire simulation oscillates between being dominated by one species to another. But I've also been able to model travelling waves, (very) transient solitons and fluid like curves:
The Flash and Flex demo of Version 2.0 of my Trigonometric Reaction Diffusion model lives here and the source code lives here.