The feature/improve_coast_tracing_NO_GOOD version tried a different method of coastline tracing. This was tried partly to improve the robustness of coastline tracing, and partly in order to capture water bodies which are disconnected from the contiguous sea.
The approach used was:
- first a brute force (i.e. whole grid) search for all cells below SWL, then to mark these cells (this search can of course be sped up using OpenMP)
- then a raster coastline tracing of all water cells with at least one non-water cell as a neighbour
- then a vector tracing of the raster coastline.
The approach worked, but gave very intricate vector coastlines. So it was temporarily abandoned. However, in order to capture water bodies disconnected from the contiguous sea for UKGB, we may well need to revisit it.
The feature/improve_coast_tracing_NO_GOOD version tried a different method of coastline tracing. This was tried partly to improve the robustness of coastline tracing, and partly in order to capture water bodies which are disconnected from the contiguous sea.
The approach used was:
The approach worked, but gave very intricate vector coastlines. So it was temporarily abandoned. However, in order to capture water bodies disconnected from the contiguous sea for UKGB, we may well need to revisit it.