Using NGS Survey Control Points
You can get National Geodetic Survey, NGS ground control points, as a shape file. You are limited to 1000 points per download, and you can select by quadrangle, county, or a bounding box.
New DEM . Open one DEM or multiple DEMs of the same area.
|Database. Open shape file with ground control points.|
|10 m USGS DEMs.
Map showing National Geodetic Survey ground control points. Note that many follow roads, and and not randomly or evenly distributed, although they do cover many of the major peaks
|You can use the Edit button on the
window, DEM options to add the elevations for the DEM or multiple
This database shows the locations of the bench marks, the elevation from NGS, and the elevation difference for each DEM (positive, the interpolated DEM elevation is higher than the bench mark).
|From the Plot button on the Table display window, Color code by DB field you can get the differences from the DEM for each of the bench marks. Note the small differences on the roads in flat terrain.|
The largest deviation in this region is -37 m, and this shows its location with the DEM reflectance view with overlaid contours ("overlay, contours" from the menu).
The interpolated DEM value is 1857 m.
|DRG, obtained by right clicking on the
map and loading Terraserver data.
The peak with the elevation control is shown as 6107 ft (or 1,861.41 m). Since this is almost exactly the value in the DEM (probably not surprising, since the DEM was produced from the contours on the paper map), either there is a big problem with the DEM and paper map, or the NGS database has a problem.
|SRTM DEM with the deviations from the control points.
Note the very close agreement along the roads.
A difference result, with two DEMs (33123379 is the SRTM). The SRTM is even lower than the USGS 10 m DEM, suggesting that the NGS control point is in error in the database.
|Contoured SRTM DEM. Note that the control point is not on the peak, suggesting a horizontal displacement error, and that the peak elevation (determined by the closed contour) closely matches that from the USGS 10 m data.|
Last revision 6/21/2006