I have created a probability raster for the appearance of certain features, but want to turn this into a useful decision support tool. The thought is to have bounding boxes which enclose enough probability values to meet a threshold, such that it would require a much larger area to reach a 75% probability of find a feature when bounding low-probability pixels and a much smaller area to do so when bounding high-probability pixels. We have been experimenting with the Neighborhood tools in Spatial Analyst and wonder if there is a way to process these in reverse - instead of specifying a neighborhood area and generating statistics, to have the statistics dictate the size of the neighborhood. If anyone has any insight into this it would be much appreciated!
Best,
Doug