As can be seen in these two map that show contaminated flower patches during two simulation runs with different seeds and hence landscapes, there is a distinct accumulation in the centre of the landscape:
I currently assume that this is a boundary effect and hence it should be possible to remove it by using periodic boundary conditions. We have those in the beginning but they became problematic when switching from ring to circle as the landscape shape as identifying opposite points on the circle distorts the geometry significantly.
Hence the plan here is to switch to Cartesian coordinates and use a square with sides of 1km length around the core zone which is a circle of 1km diameter. That should make it easy to consider the square as a torus in both latitudinal and longitudinal directions but still stay within the mapped buffer area of 2km diameter for which geo data is available.
