Flow Accumulation: 0 accumulation vs. invalid/nodata cells #263
Description
When computing flow accumulation with the accumulation method, the output seems to represent accumulation+1 for each cell. iiuc, this is because _sgrid._d8_accumulation_iter_numba guarantees that every valid cell gets at least one point of accumulation, even if it's endnode is itself. Nodata/invalid cells are set to 0 earlier in the callstack . It seems that min accumulation=1 helps to differentiate the nodata from valid cells in the output.
However, having min accumulation=1 results in accumulation rasters that are different than rasters from other tools like ArcGIS, QGIS, Whitebox Tools. For those tools, the min accumulation=0. To support interoperability between tools, I'd like to suggest some parameter (or tweak to the loop) that would allow for definition of a non-0 nodata, and a min flow accumulation of 0.
I'd be happy to push up a PR if you think it's a reasonable addition to the tool!