Personal tools
Home » Members » ebert » Notes for Basin Generation
Document Actions

Notes for Basin Generation

by Colin Ebert last modified 01-03-2007 17:53

notes/tips/tricks for the basin generation, pour point generation, and basin naming for the srtm basins project

This is a brief rundown of the errors and workarounds for the marine threats projects that i worked on.





Basin Layers and Pour Points

North America:


    There is a single basin that would not cooperate properly and insisted on being an endorheic basin.  This small basin is located at -92.14 56.34, and has an area of 18.8 km2.  

    There were also several basins in Greenland that are the byproduct of the ice cap.  These basins were ignored due to their small size and location.
 
    There were 49 modified basins that required manual merging to complete.

South America:


    One abnormal basins found, basin id sa_05884, needed to be closed in after reprojection, large population inside polygon, so null area closed in and pour point moved to coast.

    There were 380 basins that required manual merging to complete.

Europe:


    There were no abnormal basins found, and only 3 endorheic basins in western europe, one in spain, and 2 in austria, the rest are products of the caspian sea.

    There were 2 basins that required manual merging to complete.

Africa:


    There were no abnormal basins found.

    There were 105 basins that required manual merging to complete, not including lake victoria and saline lakes like MakgadiKgadi salt pans.

Asia Land:


    There were no abnormal basins found

    There were 354 basins that required manual merging to complete, not including natural sinks such as the caspian sea  etc.....

Asia Islands:


    There were no abnormal basins found.

    There is only one endorheic basin in this layer and it did require merging by hand.  This is the crater in Mt. Tambora. 

Australia:


    Fixed one basin (FID 4384) (basin_id au_07365) that was internally draining along the coastline, the fix occurred after reprojection to mollwide. 

    There were 52 basins that required manual merging to complete. 

Pacific Islands:


    There were no abnormal basins found and no endorheic basins.



After Projection into Mollwide


Edge mapping:

    The following are the instructions for and problems with the edge matching of the following layers.


North and South America:

    Due to the small area between the two layers falling into Panama, the edge matching was done by hand using the editor in ArcMap using snapping.  South America was snapped to North America.
   
    There were a few errors with the edge matching process.  The main problem being that a few of the edge polygons were only mostly clipped and as a result of the projection and clipping combined turned into polygon lines smushed into the edge of the boundary.  There were 4 of these polygons that had to be removed by hand.  after this and the edge matching the result looks good. 

Europe and Asia Land:

    removed several extra points and small polygons from both layers that were missed in the clipping.

    calculated a new area field for the polygons to eliminate anything smaller that 1.  this should remove the small edge mismatched polygons that were smushed during the reprojection.  detail to follow:

  1. for Basin layer: add new field titled RP_area, type Double
  2. run Calculate Geometry on field to calculate new area in km2
  3. add new field titled area_diff, type Double
  4. run Field Calculator:  area - RP_area
  5. Select by Attributes: "RP_area" < 1 AND "area_diff" > 1
  6. Start editing asl_bas
  7. delete selected features
  8. save edits
  9. repeat for all basin layers, checking to make sure all polygons selected fall on edges or boundaries.  this process should eliminate the above steps found in North and South America.


Final Steps for Europe / Asia / Africa:

    The final edge matching for these layers was a bit of a pain.  After much trial and tribulation attempting to automate the process completely, we decided on a mixed approach to ensure that the process would work correctly.  The ends of the boundaries were matched by hand snapping polygons to each other.  both asia and africa were matched to europe for ease of processing.  After snapping the ends and each intersecting polygon border, the middle of every polygon from africa and asia was extened out.  the entire layers were the erased using the europe layer.  africa was then extened and erased again to match with asia (in the middle east).  this process ensured that there was even edges for all polygon borders. 

Built with Plone