Difference between GPS ellipsoid and sea level heights N

1
Difference between GPS ellipsoid and sea level
heights (N)
GPS heights need to be corrected by the geoidal
undulation N in order to integrate conventional
and GPS heights.
2
Geoid03 heights from the National Geodetic Survey
3
Counties (red), railroads (green) and NGS (geoid
control points for GEOID03 (black)
Area of interest
4
Gravity stations used in NGS Geoid models
5
Geoid03 in West Texas
6
Gravity data and geoid control used in Geoid03
7
Free-air gravity anomalies in west Texas
8
Free-air gravity anomalies of West Texas (used
for geoidal computations) (CI6 mgals)
Proposed road
NGS geoid control are black triangles.
Approximately 6 mgal gravity change results in
about 2cm geoid undulation. Note sparse data
along the western part of road. Geoid control are
used to force N surface through those known N
values so if control and/or gravity data are
sparse then N is less reliable. .
9
Elevation of gravity stations
10
Elevation of gravity stations (approximate
topography)(Contour interval 5 meters)
Note similarity to Free-air gravity anomalies.
11
Computing geoidal corrections

• Gravity data ( a version of the Free-air gravity
  anomaly) is used to compute the surface of the
  equipotential of gravity representing sea level,
  the geoid.
• GPS heights H are corrected by the geoid
  undulation correction N by h (geoid height)HN
• The surface N is then MODIFIED with a CORRECTOR
  SURFACE consisting of locations where both sea
  level (geoid) heights and GPS heights are known,
  which is N (considered control points).

12
Reliability of geoid corrections

• The reliability of the geoid computation is
  dependent on the distribution of the gravity
  datadata gaps are interpolated across
• If sparse gravity data-- less reliable.
• Another reliability factor is corrector surface.
• If widely spaced control, incorrect corrector
  surface.

13
Recommendations

• Gravity and control data in Big Bend area are
  most sparse and Free-air gravity anomalies and
  topography the most variable.
• Therefore original gravity values and corrector
  surface are less reliable.
• For 1 minute (lat/lon) grid prefer approximate 1
  minute data distribution.

14
Recommendations

• Coordinate with NGS
• At least--add regionally distributed gravity data
  in the west
• along railroad/road.
• Fill in data gaps at 1-5 km spacing along and on
  both sides of route.
• 10-20 stations/day20 days.
• Idea--Use university during the summer.
• Then provide data NGS to compute geoid or
  university could do it. Integrate with Geoid03.