1

I have clipped buffer area around coordinate from one big raster (forest cover). Such created raster was loaded in R where PatchStat function (library SDMTools) gives fragmentation index.

After detailed look of such created raster I noticed that I have equal classification for non-forest area (0) and corners (also 0).

My question is whether such surplus of empty area make an influence on the value of fragmentation index. I think solution bellow may be helpful, but I do not know how to do it.

enter image description here

|improve this question
  • When you clip out the circular area of interest, make sure the exterior portions (shown in white) are represented by a distinctly different value than any of the actual data cells. I suspect--and perhaps you can confirm--that the white cells in the corners may have the same values as white cells within the circle: that would certainly cause problems with classification. – whuber♦ Apr 3 '14 at 14:46
  • Thank you @whuber. Yes, the values of white area within and outside the circle are both "0". So you recommend to clip it all again? A quite a lot of clipping (approx. 80). Is there some solution to my problem without clipping? For example some procedure to reclassify white areas in R ("overwrite all 0 vales to "XX" value within the range of "XY" from centre of the circle"?) – Ladislav Naďo Apr 3 '14 at 14:55
  • I believe that one way or another you will have to re-process your rasters to change how the corners are represented, so it would be more reliable to start over and do it right. This time, though, first go through the entire work flow for just one area, rather than all 80, to verify that it will succeed! – whuber♦ Apr 3 '14 at 15:22
1

Well! Solution was simple. Just take the circle shp., convert it into a raster (yellow circle) with equal resolution as master raster of forest cover. Than fuse this two rasters into one with raster calculator. Now I am able to distinguish between non-forest area within circle and empty areas in four corners.

If someone will have a similar problem (of classification) such solution is good.

However! Value of fragmentation index for forest area before and after applying solution is the same. So it was useless to my (non-existing) problem. I assume that for F-Index computation only forest area and perimeter are needed (and any other areas are irrelevant for the equation).

Many thanks to Imo Jakab for guidance and help!

enter image description here

|improve this answer

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.