Jump to content


Best intersect methods?

- - - - -

  • Please log in to reply
1 reply to this topic




  • New Member
  • Pip
  • 1 posts
  • Sweden

This is what I have:

Plygon1 One polygon layer of squares crossed a country. This is the land area devided in quares.

Polygon2One polygon layer or water(lakes, rivers etc)

Polygon 3One polygon layer of a sediemnt


I will try to illustrate it for you in the picture. The blue in the picture is lakes and rivers. The black is sediment and the square are the green square polygons that a country is devided in.


The polygon with squares are devided in several small squares.

The polygon with waters and sediments cover sometimes the boundaries of the squares. Well they are widely scattered.


What I want is an result on a attribut table so I can calculate the sediment per land area and then make a thematic map of it. To be more specific. I want the landarea and the area with the sediment so I can calculate the difference between them.

I want that in a table every square should have the information of landarea and sediment area.


This is what I did

1. Step one. Erase all water from the land so I get the land area. I create a new column in the result with the name land_area with are equal to the shapearea.  [Polygon1]-[Polygon2]=RESULT1


2. Step 2. Erase all the sediment from the landarea. The result is called shape area. And then in the table calculate (land area-shape area)/landerea *100 top get the percantage.  [RESULT1-Polygon3t]=Result 2


3. in the RESULT 2 I now have Landarea, sedimentarea and the calculated percentage. So I make a themeatic map.


The result looks fine and each square seemes to be balanced to the accurate sediement spread. BUT the percentage is totaly wrong in some squares when i do a local check.




For me its logic when I have the an area of 10 and erase an erea of 5. The result should be 5. What I do now that the area of 5 is just a piece of a bigger polygon wich also are within another square. Maybe thats the problem? It dont understand whats wrong here.

Does anyone knows or does anyone have a better option here

Attached Thumbnails

  • Namnlöst-3.jpg

Edited by gridinter, 28 February 2013 - 08:38 AM.




  • Validated Member
  • PipPip
  • 27 posts
  • United Kingdom

Hi Gridinter,


This sounds as though it could be easier to do if you convert your data into raster format and then use something like zonal statistics to calculate values for each square.


How easy that is will depend on what software you're using - if you have ArcGIS and spatial analyst it should be fairly straightforward. Should also be do-able in qGIS which is open source. 

0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users