I have to apply uncertainty principle to computing flow accumulation using Monte Carlo method. I generate raster with random numbers which represents mistakes in creating DEM. Then I apply those rasters on DEM in iterations. For each DEM with added random numbers raster I compute flow accumulation. In each iteration it generates slightly different lines of flow accumulation. At the end of process I merge these lines into one raster. It looks like this: multiple flow lines
What I have to do is extract most probable flow line from this data. I use ArcGIS10 but it does not matter which software should be used. In ArcGIS I tried to use thin function or extract by attributes which gives me pretty good results. I also converted every line into polyline, merged them and applied buffer. Then I used thin function on buffer. This gives me also very good result.
I want to ask if you can help me with finding another solutions for extracting most probable line?