I am going to calculate least-cost paths toward a number of location, implementing Tobler's hiking function (i.e., considering slope as factor influencing speed of movement). I want to make sure that the paths will try to avoid some zones according to the land cover. The latter data is available as raster dataset.

Is it sound to reclassify the land cover types assigning a high value to the land-cover type I want the least-cost paths to avoid, and then multiplying the SLOPE by the reclassified LAND-COVER raster?

If it is ok, how should I proceed in case I want the paths to also avoid urban areas?

  • When using Least Cost Path Analysis you usually rank data on a scale from 1-5 or 1-9 for example by reclassifying. Just remember that a high value does not make it impossible to cross these cells. You can set these values to restricted. Each layer is classified based on the users knowledge of what they are modeling. So a 9 in your land cover should match the difficulty of 9 in your other data sets. You can also weight which layer/factor has more influence on the results. You use weighted overlay to sum the rasters not multiply. Jan 31 '17 at 10:40
  • Just be careful using restricted in your data, it will set the value for that cell to minus 1 from the lowest rank on the scale which is 0. In the past I have reclassified the zero from the output weighted overlay raster to something like 9999 making it virtually impossible for a path to cross these restricted cells. The ArcGIS documentation along with youtube tutorials should aid you in your process (youtube.com/results?search_query=least+cost+path+arcgis) Jan 31 '17 at 10:51
  • 1
    Thanks for replying. I am using Tobler's hiking function as a "tool" to predict movement speed across each cell as function of slope (and direction). I am using PATH DISTANCE followed by COST PATH tool. So, I am not reclassifying my data; I am inputting SLOPE raster as it is. That's why I was asking about "weighting" the SLOPE raster.
    – NewAtGis
    Jan 31 '17 at 11:04

An approach, not saying it is the best, is to exclude those areas when constructing the cost distance raster by setting them to NODATA.

When you build your cost distance raster using the Cost Distance tool you could set the cells in the cost raster to NODATA. These would correspond to your land uses that you want to avoid. You could use the SET NULL tool to achieve this. They would then act as a hard barrier and be excluded from all further analysis.

As an example of how to use SET NULL tool see screenshot below, the blue water courses will be converted to NODATA whilst all other cells remain their original value. The False raster does not have to be the same as the input raster.

Using Set Null tool The expression is selecting water and these will become NODATA, all other cells remain as their value.

  • Could you give me some more details as to how using the CON tool to set to NODATA the zones I wish to mask out. Thank you
    – NewAtGis
    Feb 3 '17 at 10:15
  • @NewAtGis It was the Set Null tool that you need to use, turns out CON cannot accept NODATA as a True value. Anyway the logic is identical just using a different tool that specifically create NODATA cell values. I have amended by reply.
    – Hornbydd
    Feb 3 '17 at 11:33

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.