N/APosted on - 04/01/2012
We were running few cases which are known with the new FRAGSTATS for having a better understanding for the nearest neighbor distances. According to the new documentation, edge-to-edge can be measured. It seems that the measured distance will be the centroid-to-centroid distance of the two possible nearest grid cells.
So to make it more clear, an example can be made. If there are two 30 x 30 m grid cells that are separated by one cell, the anticipated nearest neighbor distance will be 30 m. But the actual distance is reported around 60 m. To find out the problem, we did a sample test with multiple patch sizes. According to our test, we get to the following point that the deviation is because of simply to the distance from the centroid to the edge of the cell that is closed to the neighboring patch. We are not sure about it though,
Can anyone justify it?
Please verify our theory in FRAGSTATS
First, I would like to discuses the difference between edge-to-edge and centroid-to-centroid distance. Edge-to-edge distance is shortest distance between the edges of the two cells while centroid-to-centroid is the distance between the centroids of the two cells.
Next, what is the difference between edge and centroid? Edge is the surface around a the cell while centroid is the point at the center of the cell. To make it simple, here is an illustration that differentiates the two.
The green line around the two squares shows the edge.
The intersection of the two lines shows the centroid. This means that the centroid in your example is 15m away the edge.
So how it is computed?
Edge-to-Edge distance is the length of the grid between the cells while centroid-to-centroid is the length of the grid plus the distance of the two centroids on its edge .
This means that anticipated 30m distance is based on edge-to-edge measurement while the 60m distance is based on centroid-to – centroid measurement. This means that your conclusion which states that the deviation is the centroid to the edge of the cell is correct.