I have a detailed polygon shapefile with streets, structures/houses, and open areas as land cover data in the attributes table, and each land cover type has a corresponding manning 'n' value associated with it in another column in the table. The manning 'n' value is in decimal format.

I also have a 20'x20' grid shapefile from another program (FLO-2D) with specific grid numbers in the attribute table. I would like to overlay the grid on the land cover file to obtain a weighted manning 'n' value for each 20'x20' grid based on the percentage of land cover type that lies within each 20'x20' grid. There are approximately 400,000 grid cells in the shapefile. Both shapefiles have the same projection.

I'm using ArcMap 10.4 desktop, but do have the spatial analyst extension. The final product I need is the 20'x20' grid shapefile with a weighted average value of manning 'n' for each grid number in the attribute table.

I've tried the zonal statistics and the zonal statistics as table, but the results did not make sense. Is this possible with the 10.4 desktop version and only the spatial analyst?


1 Answer 1

  1. Intersect Grid with Land Cover
  2. Add a new double field called "AreaWeight" and calculate as Shape Area * Weight (in your case manning?)
  3. Summary statistics with Shape Area and AreaWeight as statistics field with SUM and Grid ID as case field
  4. Add a new double field called "Mean" to output summary table and calculate as SUM_AreaWeight/SUM_Shape_Area
  5. Add a join with Grid as target table, Sumstat table as Join Table and join on the common Grid ID field
  6. Copy Features to make the join permanent

You now have a Area-Manning-Weighted mean column.

enter image description here

  • Thanks - process makes sense! The intersect ended up empty tho, so i'll need to figure out what i'm doing wrong there.
    – EGM
    Commented Jun 15, 2018 at 16:25
  • @EGM check if the coordinate systems differ. If they do you could try reprojecting them so they are the same
    – Bera
    Commented Jun 15, 2018 at 16:27
  • that's what it was - I thought they were on the same system. thanks! Still working through the process - the intersect shapefile is huge
    – EGM
    Commented Jun 15, 2018 at 23:49
  • The intersect process is giving back multiple grids at the same location, which is making my intersect shapefile so big. querying 1 grid pieces gives back about 50 results. it appears to be assigning several grid ID's to a single grid as well. Is the order of the intersect important? I have been specifying the land cover first, and the grid second. I will try reversing that to see if it makes a difference.
    – EGM
    Commented Jun 16, 2018 at 2:53
  • further investigation: I have several rows of grids, and the first grid in the row is processed, and the next grid gets assigned it's original grid number as well as the previous grid number. The third grid gets assigned its original grid number, plus the previous two grid number...and the last grid gets assigned its original grid number and all previous grid numbers in that row. Is there an environment toggle that will keep this from happening?
    – EGM
    Commented Jun 16, 2018 at 4:23

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.