Hi everyone. I have a grid with many different clustered zones. I also have a point shape file with one point in each zone and the number of the of the points corresponds to the number of the zones. The point file also has a field indicating the number of nearest neighbors that I need to have in each zone. Does any one know I can flag those cells in each cluster that are nearest neighbors.

Bill Huber asked me for some context to the question I asked two days ago so let me provide that now. The zones are mini watersheds and the points are locations on either side of a concentrated water flow path. We used terrain analysis and a vegetative filter strip model to determine the buffer area required to reduce sediment loading by a certain percentage 75% in this case.

Here is a link showing a small example of what the analysis looks like. The selected points (cyan) require buffers of 10 to 18 grid cells. The different semitransparent colors represent the different watersheds for each of the points along the waterway.

Sorry if this is not clear.

best,

Tom

Bill Huber asked me for some context to the question I asked two days ago so let me provide that now. The zones are mini watersheds and the points are locations on either side of a concentrated water flow path. We used terrain analysis and a vegetative filter strip model to determine the buffer area required to reduce sediment loading by a certain percentage 75% in this case.

Here is a link showing a small example of what the analysis looks like. The selected points (cyan) require buffers of 10 to 18 grid cells. The different semitransparent colors represent the different watersheds for each of the points along the waterway.

Sorry if this is not clear.

best,

Tom

It occurred to me today that I don't necessarily want to ID those points in each mini-watershed that are closest to the points next to the stream. Rather, It would be good to select those points that have the highest contributing area or slope. The stream power index combines the contributing area and slope in a single index. So what i really want to do is a process where i take the point at the edge of the stream. Then I find the points that are within a grid unit of that point. Then determine which of these has the greatest stream power index value. Then that gets flagged as a buffer grid cell. Then find all of the grid points that are within one grid distance of these two grid points ..... and the process repeated till we have the quired number of buffer points.

I think this would work with your approach. I'm going to have to bang my head against the wall on this.

Bill, you have an incredible mind!

Best,

Tom