[OSM-talk] Coastlines

David Groom reviews at pacific-rim.net
Thu Mar 23 14:14:14 GMT 2006

On 3/22/06, Jon Stockill <lists at stockill.net> wrote:
> There are several freely available datasets for coastlines:
> 1. VMAP0 - US department of defence vector map (contains far more than
> just coastlines, but can be rather "coarse".
> 2. GSHHS - Global Self-consistent Hierarchical High-resolution
> Shoreline. Tends to give more accurate coastlines than VMAP0 outside of
> the US (for some reason VMAP0 seems more accurate for the US). There are
> issues with inland water (such as the great lakes) though.
> 3. SWBD - Shuttle radar topography mission Water Body Data. Derived from
> the SRTM digital elevation model. Newer than the rest, potentially more
> accurate.
> For paths running along cliff tops you wouldn't want to trust your life
> to any of this data - but for the purposes of giving you a plausible
> landmass to draw maps on they're all useful.
> Jon

Unfortunately the SWBD shapefiles seem to have a number of potential flaws 
which may render them less useful than immediately might be thought.

Firstly their resolution is not that detailed, leading to a rather jagged 

Secondly they do not always follow the coast.

I have overlaid SRTM Water Body Data (SWBD), in ESRI Shapefile form over a 
MrSid format Landsat photo. The area concerned was the Solent, on the south 
coast of the United Kingdom.  Two extracts can be seen at: 
http://www.vectis-webdesign.com/misc/swbd.htm , the shapefile data is shown 
as a black line overlaid on the photo.

Where there is a large area of tidal mudflats these have been included 
within the outline of the coast, whereas they should not have been.

Also Ryde Pier has been included as part of the land, and in a rather 
distorted form.


More information about the talk mailing list