This question is related to the replies in this earlier thread (How to interpret GRASS v.kernel results?).

I am new to GIS mapping (arcGIS) and have been trying to work with the Kernel Density output. I have read on many threads that a good way to check Kernel Density output is to multiply the density values by the area of the cell and then sum those values over the grid. That sum should be equal to the number of points (in my case) in the original data.

To multiply the density values by the area of the cell: The only density values I see are when I 'Identify' a cell producing a single density value for a single cell and the density classification legend used to symbolize the raster in the Table of Contents. How do I extract the density values from each raster cell? (So I can multiply those values by the area).

Or should I be generating density values for each cell from the density classification legend presented in the Table of contents?

  • Can you clarify by 'density ranges' I thought you get a density value for each pixel in the table? Saying that the Kernel density (Raster/Value Attribute Table -VAT) will probably not be accessible as your dataset will be float. please post a snapshot of your attribute table just so that we understand what you meant by the range of the kernel density.
    – yanes
    Feb 5, 2016 at 19:17
  • @yanes, a KDE results in a floating point raster and as such, there will not be an attribute table. Feb 5, 2016 at 19:43
  • That is correct, that is why I am wondering where the density range table is from?
    – yanes
    Feb 5, 2016 at 19:45
  • 1
    I believe that the OP is referring to the legend in the TOC. Feb 5, 2016 at 19:48
  • oh-okay! thanks @JeffreyEvans, not used to that referred as a table. OP - please edit the question so that people know you are referring to the classification values used in symbolizing the raster in the Table of Contents (basically just for the purpose of symbolizing). Your data have distinct values for each pixel.
    – yanes
    Feb 5, 2016 at 20:00

2 Answers 2


I am not quite sure about the density table range. Since your data is float you won't have access to the VAT/rast. attr. table. One way, even though tedious, will be:

  1. Explore the minimum, maximum and average of the dataset to have some kind of idea of what decimal points you want to save. Then multiply your raster either by 100, 1000 or 10000, to make sure that you save as many decimal points as you choose.
  2. Once you multiplied your KDE raster with the desired factors use the INT function to change your dataset from float to integer.
  3. Check if your data now has an attribute table, if not, do not worry, use the Build a raster attribute table function and ArcGIS will automatically calculates a table for you.
  4. Now that you have a table add another column and divide the value column with the factor you used to integerize your raster (100,1000 or any other value used), let's name this as Orig_Value
  5. Add another column and name it for example KDE_Volume. Then use Field Calculator and calculate [Orig_Value]*[COUNT]*len * wid. here the length and width will be the same value for a square pixel therefore you can substitute it as len^2. where len = your cell resolution.

You confusion seems to stem from not understanding the nature of a raster and how analysis is applied to the data. ESRI has a good explanation of raster data. When you apply functions to rasters, unless a condition is defined, it is applied to every cell (value) in the raster.

The type of operator that you have in mind is often referred to as raster algebra and consists of applying mathmatical operators to the raster array. An example would be using the ArcGIS raster calculator to multiply the values of the raster by a specified value. It would simply be: "raster" * 900

Please do some research on raster data and your question will answer itself.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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